SERVICES

OIL, GAS & CHEMICALS

BLENDING OF PETROLEUM PRODUCTS
(INCLUDING SHIPBOARD BLENDING)
Oil Gas and Chemicals Services

Blending of Petroleum Products

Including Shipboard Blending

Fourth Edition

Page 1 of 83
Disclaimer This manual is designed and written to provide information with
regard to the subject matter involved. SGS makes no warranty,
express or implied that it is fit for any purpose whatsoever; or to the
absolute sufficiency of the material presented. It cannot be assumed
that every procedure is covered and that other or additional
procedures may not be required in a particular or exceptional
circumstance.

SGS assumes no responsibility for any inaccuracies in reproduction

or errors in interpretation of any authority. SGS reserves the right to
modify or amend this manual, without prior notification, but SGS
assumes no responsibility to update or issue corrections. Users are
urged to monitor underlying authorities for amendments, updates
and interpretations.

This manual is usually used in conjunction with a presentation.

Certain procedures contained in the manual may require or assume
expertise or experience. Caution is therefore advised and SGS
assumes no responsibility for any use by unauthorized or untrained
persons.

Copyright This manual has been compiled by SGS North America Inc. from
publicly available material, appropriate industry standards,
government publications and proprietary SGS data and
documentation. The SGS generated portions of this manual have
been produced as copyright material and SGS reserves all rights
under all applicable copyright laws both national and international.

No SGS generated portion of this manual may be reproduced, in

any form, without the express written permission of SGS North
America Inc. Copies of this manual are commercially available.
Additionally, SGS can provide blending consultancy services in
addition to customized blend training programs and manuals. Please
contact your nearest SGS office for additional information.

Use of any portion of this manual is subject to the conditions of the

disclaimers.

For further information, contact SGS North America Inc's. Oil, Gas
and Chemicals Services in Houston, Texas. Phone (281) 479-7170,
Fax (281) 479-2734.

Fourth edition published 2015

©SGS-NA 2015 All Rights Reserved

Page 2 of 83
Contents
Page

Chapter 1 - Hydrocarbon Structures .............................................................9

1.1 Paraffins
1.2 Olefins
1.3 Naphthenes
1.4 Aromatics
1.5 Non-hydrocarbon Elements

Chapter 2 – Refining Processes .....................................................................12

2.1 Straight-Run Process

2.2 Thermal Cracking Process
2.3 Catalytic Cracking Process
2.4 Vacuum Distillation Process
2.5 Visbreaking Process
2.6 Hydrocracking
2.7 Refining Yields

Chapter 3 – Fuel Oil Stability and Compatibility ...........................................20

3.1 Compatibility or Blending Reaction

3.2 Stability Spot Test
3.3 Problems Caused by Suspended Particles

Chapter 4 - Mixing Methods ............................................................................22

4.1 Agitation
4.2 Circulation
4.3 Density Blending
4.4 Heating
4.5 Inline Blending

Chapter 5 – Volumetric Shrinkage..................................................................24

5.1 API MPMS Chapter 12.3

5.2 Application

Chapter 6 – Principles of Blending.................................................................26

6.1 Linear Blending

6.2 Non Linear Blending
6.3 Developing a Blending Index
6.3.1 Data Collection
6.3.2 Developing Equation Forms
6.3.3 Computing Coefficients
6.3.4 Testing the Index
6.4 Interaction Blending Model
6.4.1 Using Interaction Coefficients
6.4.2 Obtaining Equations
6.5 Bench or Hand Blends

Page 3 of 83
Chapter 7 – Linear Blends ...............................................................................31
7.1 Density and API Gravity
7.2 Sediment and Water
7.3 Ash, Metals, Sulfur
7.4 Calculations

Chapter 8 – Non-Linear Blends.......................................................................34

8.1 Blending Indices or Blending Numbers

8.2 Viscosity
8.2.1 Viscosity Blending Indices
8.2.2 Example
8.2.3 Calculations
8.3 Pour Point, Freezing Point and Cloud Point
8.3.1 Pour Point
8.3.2 Freeze Point
8.3.3 Cloud Point
8.4 Flash Point
8.5 Vapor Pressure
8.6 Interactive Coefficients
8.7 Octane

Chapter 9 – Optimization .................................................................................43

Chapter 10 - Planning a Shipboard Blend .....................................................45

10.1 Key Roles

10.2 Blend Coordinator
10.3 SGS/Client Interface
10.4 SGS/Shore Terminal Interface
10.5 Blend Master/Vessel’s Chief Officer Interface

Chapter 11 - Load Plan Calculations ..............................................................49

11.1 Maximum Loadable Volume
11.2 Maximum Loadable Quantity

Chapter 12 – Loading Procedures ..................................................................54

12.1 Ship Tank versus Shore Tank Blending

12.2 Ship Considerations – Before Loading
12.3 Ship Considerations – During Loading
12.4 Ship Considerations – After Loading
12.5 Shore Considerations
12.6 Order of Loading
12.7 Mixing and Heating
12.8 Sampling

Chapter 13 - Special Operations .....................................................................60

13.1 Discharge Blending

13.1 Allowing for Changes
13.2.1 Changes during Loading
13.2.2 Changes at Discharge

Page 4 of 83
Chapter 14 - Limitations and Problems .........................................................63

14.1 Communication
14.2 Ship - Shore Volumetric Differences
14.2.1 Ships’ Capacity Tables
14.2.2 Other Factors
14.3 Imperfect Mixing
14.4 Variances between Tanks on Board
14.5 Draft, Stability and Vessel Stress Limitations

Chapter 15 – Manual Sampling .......................................................................67

15.1 Equipment
15.2 Types of Samples
15.2.1 All Levels and Running Samples
15.2.2 Spot Samples
15.2.3 Tap Samples
15.2.4 Manual Pipeline Sampling
15.2.5 Vapor Pressure Samples
15.3 Commodity Specific Guidelines

Chapter 16 – Automatic Sampling ..................................................................71

16.1 Automatic Sampling Concept

16.2 Key Elements
16.3 Types of Systems
16.4 Sampler Operations
16.5 Special Consideration for Marine Applications

Chapter 17 – Closed and Restricted Sampling Systems ..............................74

17.1 Vapor Control Valves [VCV]

17.2 Limitations and Problems
17.3 API MPMS Chapter 17.11 / EI HM 52
17.3.1 Scope and Introduction
17.3.2 Other Sections
17.3.3 Points to Note

Chapter 18 – Precision and Accuracy of Test Methods ...............................76

18.1 Accuracy
18.2 Precision
18.2.1 Repeatability
18.2.2 Reproducibility
18.3 Bias
18.4 Margins

Annex 1 – Hydrotreating .................................................................................78

A1.1 Process Description

A1.1 Process Description
A1.3 Summary

Page 5 of 83
Annex 2 - Straight Run Fuel Oil Indicators ....................................................80

A2.1 Overview
A2.2 Interpretation
A2.3 Limitations
A2.3.1 Distillation - ASTM D 1160
A2.3.2 Bromine Number - ASTM D 1159
A2.3.3 Flocculation test, Toluene or Xylene Equivalent
A2.3.4 Hot filtration test
A2.3.5 P Value
A2.3.6 - Asphaltenes, Density and Flashpoint

Annex 3 – Sediment by Hot Filtration ............................................................82

A3.1 Sediment by Extraction

A3.2 Sediment by Hot Filtration
A3.3 Ageing the Sample
A3.4 Test Methods
A3.5 Summary

Page 6 of 83
Figures
Page
Figure 1 – Paraffin Hydrocarbons ...............................................................................................................9
Figure 2 – Olefin Hydrocarbons ..................................................................................................................10
Figure 3 – Naphthene Hydrocarbons ..........................................................................................................10
Figure 4 – Aromatic Hydrocarbons .............................................................................................................11
Figure 5 – Typical Refining Yields ...............................................................................................................12
Figure 6 – Atmospheric Distillation ..............................................................................................................13
Figure 7 – Bubble Trays ..............................................................................................................................13
Figure 8 – Thermal Cracking.......................................................................................................................14
Figure 9 – Catalytic Cracking ......................................................................................................................15
Figure 10 – Vacuum Distillation ..................................................................................................................16
Figure 11 – Visbreaking ..............................................................................................................................17
Figure 12 – A Two Stage Hydrocracker ......................................................................................................18
Figure 13 – Refinery Yields .........................................................................................................................19
Figure 14 – Example of Rating 1 and Rating 5 Spot Tests .........................................................................20
Figure 15 – Jet Mixer Nozzle ......................................................................................................................22
Figure 16 – Jet Mixer Nozzle Arrangement .................................................................................................23
Figure 17 – Inline Blending System ............................................................................................................23
Figure 18 - Viscometer Tube.......................................................................................................................34
Figure 19 - Viscometer Bath .......................................................................................................................35
Figure 20 – Maxwell Viscosity Blending Index ............................................................................................35
Figure 21 - Pour Point Bath ........................................................................................................................37
Figure 22a - Flash Point Unit Diagram ........................................................................................................39
Figure 22b - Flash Point Unit Photo ............................................................................................................39
Figure 23 - Octane Engine ..........................................................................................................................41
Figure 24 – Shipboard Blend Operations Communications Flowchart ........................................................46
Figure 25 – Unmeasurable OBQ.................................................................................................................55
Figure 26 – Free Surface Effect on an OBO Carrier ...................................................................................66
Figure 27 - Sample Bottle Catcher and Beaker...........................................................................................67
Figure 28 – Spot Sample Locations ............................................................................................................68
Figure 29 – Probe Location in Pipeline .......................................................................................................71
Figure 30 – Probe Designs .........................................................................................................................71
Figure 31 – Automatic Sampler...................................................................................................................72
Figure 32 – Automatic Sampler with Fast Loop ..........................................................................................72
Figure 33 – Photo of an Automatic Sampler by Jiskoot ..............................................................................73
Figure 34 - Schematic of a typical Hydrodesulfurization unit in a petroleum refinery ..................................78

Tables
Table 1 - Volumetric Shrinkage ...................................................................................................................24
Table 2 - Example of a three component blend ..........................................................................................32
Table 3 - Sulfur blend calculation - Weight percent (correct) ......................................................................32
Table 4 - Sulfur blend calculation - Volume percent (incorrect)...................................................................32
Table 5 - Sediment and Water [S&W] blend calculation - Volume percent .................................................33
Table 6 - Component values for additional exercise - Linear Blending .......................................................33
Table 7 - Viscosity blend - Maxwell Index calculations ...............................................................................36
Table 8 - Viscosity blend - Refutas Index calculations ................................................................................37
Table 9 - Component values for additional exercise - Linear Blending .......................................................37
Table 10 - Component values for optimization ............................................................................................43
Table 11 - Sulfur blend optimization calculation check ...............................................................................43
Table 12 - Weight to observed volume calculation to ensure
weight of product fits volumetrically in vessel tanks ...................................................................50
Table 13 - How much of each component to load into each tank................................................................51
Table 14 – Vessel Load Plan for Component “A” ........................................................................................52
Table 15 – Vessel Load Plan for Component “B” ........................................................................................52
Table 16 – Vessel Load Plan for Component “C”........................................................................................53

Page 7 of 83
Preface
Blending of Petroleum Products, including Shipboard Blending, introduces the reader to the
science and art of blending; and, details the unique requirements of blending petroleum products
directly onto a ship or barge. It is used as an introduction for people new to the industry or as a
reference for those knowledgeable about some, but not all aspects of these processes. In order
to derive the maximum benefit from this book, it should be used in conjunction with its study
guide. Both the manual and study guide are usually used as course material for blending training
courses presented worldwide by SGS Oil, Gas and Chemicals Services.

Page 8 of 83
Chapter 1 - Hydrocarbon Structures
Crude Oil, the heavier petroleum constituents that occur naturally in liquid form in the ground,
contains many different hydrocarbons, each with its unique chemical structure and
characteristics. The refinery or processing plant does not deal with every one of these
compounds separately, however; it refines or processes groups of hydrocarbons with similar
chemical structures and physical traits. To understand how unrefined oil is converted into useful
products, it is necessary to be familiar with four types of hydrocarbons: paraffins, olefins,
naphthenes, and aromatics. All of these compounds are arrangements of carbon and hydrogen
atoms, and sometimes they include other elements such as oxygen, sulfur, and trace metals.

1.1 Paraffins

Natural gas, petroleum gas and a substantial portion of crude oil are composed of paraffin
hydrocarbons, some of which are shown in figure 1. The simplest is methane, the main ingredient
of natural gas. It has one carbon atom attached to four hydrogen atoms. The other hydrocarbons
in this group are chains of carbon atoms with attached hydrogen atoms. Examples are ethane
and propane, also components of natural gas. Each carbon atom can bond or attach to four other
atoms and each hydrogen atom can bond with one other atom. Paraffins are given names ending
in “ane”. Chemically they are very stable compounds. The paraffin group is large, varying in
number of carbon atoms from one to ninety or more. The number of carbon and hydrogen atoms
in each hydrocarbon can be
expressed by a formula; the
formula for ethane is C2H6,
indicating two carbon atoms
and six hydrogen atoms.
Hydrocarbons with very few
carbon atoms (C1 to C4) are
light in weight and are gases
under normal atmospheric
pressure. Hydrocarbons with
more carbon atoms (C5+) are
heavier and are either liquid or
solid under atmospheric
pressure. As the number of
carbon atoms increases, the
boiling point also
increases—more heat must
be applied to change the hydrocarbon from a liquid or solid to a gaseous state. Paraffins
with four or more carbon atoms can have more than one arrangement of atoms. All the carbon
atoms can form a single straight chain, or some of the carbon atoms can form branches attached
to the main chain. Butane, for example, has two forms: normal, or n-butane (straight chain), and
iso, or i-butane (branched chain). Isobutane is known as an isomer of normal butane because i-
butane has exactly the same formula but a different arrangement of atoms. This difference is
important because isobutane boils at a different temperature from normal butane and causes
different chemical reactions. The larger the paraffin molecule, the greater is the possible number
of isomers. Other paraffins include pentane (C5Hl2), hexane (C6H14), heptane (C7H16), octane
(C8H18), nonane (C9H20), and decane (C10H22). Because paraffins have the greatest possible
number of hydrogen atoms, they are called saturated hydrocarbons.

Page 9 of 83
1.2 Olefins

Like paraffins, olefins are chains of carbon atoms with attached hydrogen atoms (Fig. 2).
However, olefin chains do not have the greatest possible number of hydrogen atoms. If two
hydrogen atoms are missing, two carbon atoms in the chain will form a double bond to make up
for the deficiency. An example is
ethylene, an important petro-
chemical. Its carbon atoms are
capable of bonding to six hydrogen
atoms, but there are only four
hydrogen atoms, so the two carbon
atoms form a double bond. If four
hydrogen atoms are missing, the
carbon chain will have two double
bonds; an example is butadiene,
another petrochemical. Other
common olefins are propylene and
butylene. Like paraffins, olefins can
have isomers with branched chains
or with the double bonds at different
places in the chain. Olefins are
known as unsaturated hydrocarbons
because of their hydrogen deficiency. Because the carbon atoms of olefins are always seeking to
attach their “empty” bonds to other atoms, olefins are unstable and are easily used to make new
chemical compounds. They do not occur naturally in crude oil but are formed in the refinery by
the breakdown of larger hydrocarbon molecules. Olefins are very useful in creating certain
refinery products and petrochemicals.

1.3 Naphthenes

The carbon atoms of

naphthenes form rings rather
than chains, so naphthenes are
called ring compounds or
cycloparaffins. (Fig. 3)
Hydrocarbons in this group have
names that begin with the prefix
“cyclo” to indicate the ring
structure. An example is
cyclohexane, a hydrocarbon
often occurring in natural
gasoline. The carbon rings of
naphthenes are saturated with
hydrogen atoms.

Naphthenes are chemically

stable. They occur naturally in
crude oil and have properties
similar to the paraffins.

Page 10 of 83
1.4 Aromatics

Aromatic hydrocarbons are

compounds that contain a ring
of six aromatic carbon atoms
with alternating double and
single bonds and six attached
hydrogen atoms. (Fig. 4) This
type of structure is known as a
benzene ring. The most
important aromatics in refinery
production are the BTXs —
benzene, toluene, and xylene.
Aromatics occur naturally in
crude oil. They are also
created by refinery and
petrochemical plant processes.

1.5 Non-hydrocarbon Elements

2% to 10+% percent of a crude oil may be composed of other elements or compounds containing
oxygen, nitrogen, sulfur, and metals. Oxygen content can be as high as 2 percent and nitrogen
content as high as 0.8 percent. Sulfur content ranges from traces to more than 5 percent. If a
crude oil contains appreciable quantities of sulfur or sulfur compounds, it is called sour crude; if it
contains little or no sulfur, it is called sweet crude. Trace metals contained in crude oil include
sodium, magnesium, calcium, strontium, copper, silver, gold, aluminum, tin, lead, vanadium,
chromium, manganese, iron, cobalt, nickel, platinum, uranium, boron, silicon, and phosphorus. Of
these metals, nickel and vanadium occur in the greatest quantities. Product blending is one of the
most important steps in the transformation of crude oil into a finished petroleum product. It allows
the refiner to obtain the maximum amount of optimization in the allocation of the available
unfinished stocks as produced from the various refining processes and to utilize them in such a
way as to be able to meet product demand and specifications at the least cost. In other words
blending allows the refiner to make the most cost effect use of the raw material [crude oil]
available.

Page 11 of 83
Chapter 2 – Refining Processes
Refiners have much greater flexibility today than they had in the past. When crude oil was first
refined, the equipment was primitive and simple. The process consisted of an externally heated
drum of crude oil with a tube extended from it, delivering evaporated vapors to an area of lower
surrounding temperature; thus, condensing them into a liquid. As experience and knowledge
increased, equipment improved and the volume and number of products increased. New refining
methods and modern processes have become more numerous and very complex. In most cases,
the purpose behind these advances in design and technique of modern refining methods and
equipment is to produce more and more automotive fuel from the crude oil. In the US this is still
gasoline but in many parts of the world, especially Europe, automotive diesel is the predominant
fuel. Additionally, some advances have been made because of a need for specialty products,
particularly in the area of petrochemicals and jet fuel.

The amount of gasoline or diesel produced from simple distillation depends upon the feedstock
used and usually does not produce large amounts. Additionally, the current market demand for
the residual product that would result from
simple distillation is significantly lower than Figure 5 – Typical Refining Yields
the current market demand for automotive
fuel so modern refining processes optimize
the amount of high value product produced
from the crude oil in relation to the market
demand for these products. Refineries are
primarily interested in how much gasoline
and diesel they can produce, with fuel oil
production being of secondary concern.

Many petroleum products, such as

gasoline, diesel, jet fuel, kerosene,
lubricating oils, and solvents receive
additional treatment to remove many of
their impurities and unwanted properties,
but fuel oils do not. Fuel oils receive the
least attention of any of the refined products. This does not mean they are not properly refined,
but that they do not need to meet the rigid specifications of the other types of oils. When gasoline
is used in precision-built engines, and lubricating oil is used to fulfill a need for close tolerances
and great stability, their refining must be of the highest caliber; but, where the main purpose is
burning or combustion, the chemical structure is not as important. While the structure of fuel oils
will vary and, therefore, so will some of its combustion reactions, these variations are not usually
critical. One of the results of the flexibility that is given to the refiner by modern processing
technique is the considerable structural variation of residual fuel oils Modern refining methods
today basically produce two types of oil—straight-run and cracked. Formerly, all oils were straight
run, but today most oils are cracked.

2.1 Straight-Run Process

The first step in this refining process is heating the crude oil, thereby converting it into vapor.
These vapors are then condensed back into liquid form, but the refining equipment is so
constructed that these vapors are condensed at various levels in the fractionating tower. This is a
tall, cylindrical tower, varying in size, but approximating 1m to 2m in diameter and 25m to 35m in
height. Inside these towers are bubble trays which consist of baffles that permit the vapors to
pass through and upward, retaining the condensed liquids which, in turn, flow out of the tower into
the storage tanks. The lightest material condenses at the top, the heaviest, at the bottom. This is
a simplified explanation of the process but the principle is shown in figure 6. This figure shows

Page 12 of 83
 butane, light straight naphtha,
naphtha, kerosene, distillate, heavy
gas oil and residual fuel oil. These
are not the only ones that can be
obtained and the refiner has some
flexibility in regard to what is
produced. Crude oil is composed of
oils with various boiling ranges from
20°C to 550°C. The cut points
shown can vary and there is a
certain amount of overlap which the
refiner also uses to vary one yield
from another to obtain the maximum
benefit from the available material.

This simple crude-oil distillation is

Figure 6 – Atmospheric Distillation the fundamental separation process
on which all other refinery
processes, or steps, are based. In a few cases, fractions or oils produced by this method are
marketed without any additional processing; however, they are more likely to be used as
feedstock for a number of other refining processes, especially as starting material for the various
cracking processes. The straight-run method may also be referred to as topping or skimming.

After the initial straight-run distillation process, further

refining is accomplished using one of five of the more
important methods: thermal cracking, catalytic cracking,
vacuum distillation, visbreaking or hydro-cracking. The
process chosen is influenced by many factors which
may include: source of crude oil, refining techniques
available, demand for the different resultant products,
industrial activity, coal and gas prices, and the market
price between the crude oil and its refined products.

A high business activity can do more than anything else

for stabilizing the quality of the residual oils. As long as
the refinery can market its residual fuel oil, it does not
need to seek a way to reduce or eliminate the yield. Up
until World War II the demand for light and heavy
products was mostly in balance; however, after this the
demand for light product, mostly in the form of gasoline,
started to significantly outpace the demand for the
heavier product. This demand for light distillates forced
the refiner to get them from the heavy residual product,
and in order to do this they must be cracked. This
cracking produces both chemical and physical changes
in the heavy oils. One process or, as is more common
today, any combination of them can be used in the
production of products, with specifications varying Figure 7 – Bubble Trays
according to the refining procedures used.

2.2 Thermal Cracking Process

Thermal cracking uses heat and pressure to split or crack the molecules. As the
complicated hydrocarbon molecules that have not yet vaporized at 425°C are heated to higher
temperatures, the energy transfer at around 475°C is enough to split [or crack] the molecules

Page 13 of 83
into two or more smaller ones. Material for thermal cracking is the straight-run residue. In the
cracking of these straight-run oils, also known as reduced crude, the method is almost the same
as in the initial process. The difference is that a higher temperature and higher pressure are used.
It is from the need of high temperatures that the process is called thermal cracking. The thermal
cracking process can be divided into two types— liquid phase and vapor phase.
In the liquid phase process,
the cracking takes place
while the oil is in a liquid
state. In the vapor phase
process, the cracking takes
place while the oil is
completely vaporized. The
temperatures are a little
higher in the vapor phase
process than in the liquid
phase, but the pressures are
lower. The starting material is
first heated in the charge
heater, and from there it
goes to the reaction chamber
where, with the use of
increased temperatures and
Figure 8 – Thermal Cracking
pressures, the oil is broken
or cracked into lighter and
heavier materials. Chemically speaking, the molecule of straight-run oil is split and broken apart,
creating smaller or lighter molecules and larger or heavier ones. The light ones are formed
because a small piece of a large molecule is broken off, while the remaining portion unites with
another one of large size, forming a heavier molecule than the original. The cracked products
then go to the flash chamber where the materials are flashed into vapors. These vapors pass to
the fractionating tower where they are condensed, and flow to the storage tanks. As in the
straight-run method, different boiling oils are taken from the fractionating tower at various levels,
producing oils with different characteristics and specifications from those of the straight-run
process. The resulting petroleum products are called cracked oils. The products usually obtained
from the thermal cracking process are gasoline, diesel and residual fuel oil. Figure 8 illustrates
the thermal cracking process and its similarity to the straight-run method can be observed.

2.3 Catalytic Cracking Process

In the adolescent years of the petroleum industry, the proportion of the crude oil barrel that
consumers wanted in the form of gasoline increased faster than fuel oil. It became apparent to oil
men that the production of enough straight-run gasoline to satisfy the market would result in a glut
of heavy fuel oils. The economic symptoms of this problem were increasing prices of gasoline
and declining prices of the heavier cuts. To cope with this physical and economic problem, the
inventive process engineers developed a number of cracking techniques, the most popular of
which was cat cracking. In a cat cracker, straight-run heavy gas oils are subjected to heat and
pressure and are contacted with a catalyst to promote cracking.

A cat cracker uses a catalyst rather than heat to split the molecules. A catalyst is a
substance added to a chemical reaction which facilitates or causes the reaction; but when the
reaction is complete the catalyst comes out just like it went in. In other words, the catalyst does
not change chemically. It causes reactions between other chemicals. The boiling point for cat
feed is anywhere in the 340-600°C range. Heat is required to make the process work.
Temperatures in the cat cracker where the cracking takes place are about 480°C. The process is
designed to promote cracking in a specific way. The object is to convert the heavy cuts to
gasoline. Ideally, the entire product would be in the gasoline range, but the technology is not that

Page 14 of 83
precise. During the cracking process, several phenomena occur. As the large molecules crack,
the lack of sufficient hydrogen to saturate all the carbons causes some small amount of carbons
to form coke, which is virtually pure carbon atoms stuck together. As the large molecules break
up, a full range of smaller molecules from methane on up are formed. Finally, the large
molecules, consisting of several aromatic or naphthenic rings plus long side chains, are likely to
crack where the side chains are attached. The resulting molecules, though lower in carbon count,
are more dense or heavier; that is, their density is higher. They also tend to have higher boiling
temperatures. The irony is that these molecules form a product which is heavier than the feed.
Products of cat cracking are the full range of hydrocarbons, from methane through to the residue
range, plus coke.

There are three parts to the cat cracking hardware:

 the reaction section,

Figure 9 – Catalytic Cracking
 the regenerator; and,
 the fractionator

The guts of the cat cracker

are its reaction section.
The cat feed is pumped
through a heater mixed
with catalyst being
pumped into a line called
a riser that travels to the
bottom of a large vessel
which looks like a water
tank, called a reaction
chamber. By the time the
feed reaches the vessel,
the cracking process is
well underway, so the residence time in the vessel is only seconds. In fact, in the more modern
cat crackers, all cracking takes place in the riser. The reactor is only used as a
catalyst/hydrocarbon separator. That’s done with a cyclone, a mechanical device using
centrifugal motion. The catalyst is one of two types:

 Beads; or
 Particles

The beads are 3 mm (⅛ inch) to 6 mm (¼ inch) in diameter; however, they are pretty much out of
vogue now. The particles are much smaller and look like baby powder. The particles have two
unusual characteristics. If a container of them is shaken or tilted, the powder reacts just like a
fluid. Hence, the process is sometimes called fluid cat cracking. The second characteristic is
common to both the beads and the particles but is not apparent to the naked eye. Under a
microscope it can be seen that each has a large number of pores and, as a consequence, a
tremendous surface area. Because the influence of the catalyst depends on contact with the cat
feed, this tremendous surface area is important to the process. In the catalytic cracking process
the same reaction occurs as in thermal cracking, but it is obtained by a different method. Starting
material for catalytic cracking is often straight-run residual, but thermally cracked products,
especially the No. 2 distillate, are also used at various times.

It is also very important for the catalyst separation process to remove as much of the catalyst as
possible, especially the very small particles that are sometimes referred to as “cat-fines”. A
modern FCC catalyst has four major components: crystalline zeolite, matrix, binder, and filler.
Zeolite is the primary active component and can range from about 15% to 50% weight percent of
the catalyst. The zeolite used in FCC catalysts is composed of silica and alumina tetrahedra with

Page 15 of 83
each tetrahedron having either aluminum or a silicon atom at the center and four oxygen atoms at
the corners. It is a molecular sieve with a distinctive lattice structure that allows only a certain size
range of hydrocarbon molecules to enter the lattice. In general, the zeolite does not allow
molecules larger than 8 to 10 nm (i.e., 80 to 90 angstroms) to enter the lattice. If the catalyst is
not fully removed from the residual material it will likely have unacceptably high levels of
aluminum and silicon, which could put the residual fuel oil off-specification. High levels of
aluminum and silicon are a result of the refining process and do not come from the crude oil as is
the case of heavy metals such as vanadium.

2.4 Vacuum Distillation Process

Suppose you had the hot dog concessions in both the Astrodome in Houston and Mile High
Stadium in Denver. On average, to boil water to cook hot dogs you’d heat the water to 100°C in
Houston but only 98°C in Denver? This temperature gradient represents the effect of different
atmospheric pressures caused by change in altitude. When people refer to the air in the
mountains being thinner, they really mean less dense, which is because the pressure is less. The
relationship between pressure and boiling temperature is simple. The process of heating helps
the molecules to absorb enough energy to escape from the liquid form. The rate at which they
escape depends on the rate heat is delivered and the pressure of the air above them. The lower
the pressure, the less energy has to be transmitted, and the lower the temperature at which vapor
will begin forming in the liquid, i.e., boiling.

The point is—the lower the

pressure, the lower the
boiling temperature.
Applying the pressure-boiling
temperature relationship to the
crude oil cracking problem, the
straight run residue will crack if
the temperature is too high
and it still requires separating
into more cuts. The solution is
to do the fractionation at
reduced pressure. Straight run
residue is pumped directly
from the distilling column to
the flasher. The straight run
residue is taken into a squat,
large diameter column where
the pressure has been Figure 10 – Vacuum Distillation
lowered. The pressure in a
vacuum flasher is about 0.3 to
0.4 of an atmosphere (4.5 to 5.5 psi). At that reduced pressure, the lighter fractions of the residue
will begin to boil (flash). Flashing is a cooling process. To counteract the cooling process,
superheated steam (steam under pressure with temperatures above 400°C) is also introduced to
the vessel. Heat is transferred from the steam to the residue, maintaining the temperature and the
flashing. The other function of the steam is vacuum control. Pressure in the vessel is kept very
low by a vacuum pump at the top of the vessel. Several streams can be drawn off the flasher.
The remaining residue is called vacuum bottoms. The distillate produced is frequently used as
charge stock for the catalytic cracking process. When reduced crude oil is used as the charge,
the vacuum bottoms will be higher in density and much higher in viscosity, than the original
feedstock. These vacuum bottoms can be used to make asphalt or blended with a cutting oil or
distillate, to make No. 6 fuel oil. The bottoms can also be used as charge stock for the
visbreaking process when it is necessary to decrease the high viscosity. This reduces the

Page 16 of 83
percentage of expensive cutting stocks needed for obtaining a lower viscosity to make a suitable
residual fuel oil.

2.5 Visbreaking Process

The process name of "visbreaker" refers to the way in which the process reduces (i.e., breaks)
the viscosity of the residual oil. It is sometimes referred to a poor-man’s thermal cracker and
takes flasher bottoms and thermally cracks part of it. It permits reflashing of the residue to reduce
the volume. Cutter stock is added to the visbreaker residue to make it loose enough to be
marketable as residual fuel. The benefit is that the cutter stock added back is less than the
cracked product freed up by visbreaking, so there is a net reduction of residual fuel being
produced. The cutter stock can be cat cracked heavy gas oil, cycle oil, or even a cut of the
flashed distillate from the visbreaker.

Visbreaking is essentially thermal cracking but differs in intensity. The hardware is simpler and
cheaper to operate; however, only about 20-30% of the residue is destroyed. There are two types
of visbreaking processes – coil visbreaking and soaker visbreaking.

The term coil (or furnace) visbreaking is applied to units where the cracking process occurs in the
furnace tubes (or "coils"). Material exiting the furnace is quenched to halt the cracking reactions:
frequently this is achieved by heat exchange with the feedstock material being fed to the furnace,
which in turn is a good energy efficiency step, but sometimes a stream of cold oil (usually gas oil)
is used to the same effect. The gas oil is recovered and re-used. The extent of the cracking
reaction is controlled by regulation of the speed of flow of the oil through the furnace tubes. The
quenched oil then passes to a fractionator where the products of the cracking (gas, LPG,
gasoline, gas oil and tar) are separated and recovered.

In soaker visbreaking, the bulk of the

cracking reaction occurs not in the furnace
but in a drum located after the furnace called
the soaker. Here the oil is held at an
elevated temperature for a pre-determined
period of time to allow cracking to occur
before being quenched. The oil then passes
to a fractionator. In soaker visbreaking, lower
temperatures are used than in coil
visbreaking. The comparatively long duration
of the cracking reaction is used instead.

Visbreaker tar can be further refined by Figure 11 - Visbreaking

feeding it to a vacuum fractionator. Here
additional heavy gas oil may be recovered
and routed either to catalytic cracking, hydrocracking or thermal cracking units on the refinery.
The vacuum-flashed tar (sometimes referred to as pitch) is then routed to fuel oil blending.

From the standpoint of yield, there is little or nothing to choose between the two approaches.
However, each offers significant advantages in particular situations. The cracking reaction forms
petroleum coke as a byproduct. In coil visbreaking, this deposits in the tubes of the furnace and
will eventually lead to fouling or blocking of the tubes. The same will occur in the drum of a soaker
visbreaker, though the lower temperatures used in the soaker drum lead to fouling at a much
slower rate. Coil visbreakers therefore require frequent de-coking. This is quite labor intensive,
but can be developed into a routine where tubes are de-coked sequentially without the need to
shutdown the visbreaking operation. Soaker drums require far less frequent attention but their

Page 17 of 83
being taken out of service normally requires a complete halt to the operation. Which is the more
disruptive activity will vary from refinery to refinery.

The lower temperatures used in the soaker approach mean that these units use less fuel. In
cases where a refinery buys fuel to support process operations, any savings in fuel consumption
could be extremely valuable. In such cases, soaker visbreaking may be advantageous. Because
of the slightly lower temperatures and pressures, the cracking is less severe. If too-extensive
cracking occurred, a large amount of low-grade gasoline would be produced, whereas this
process is, generally, only employed to obtain large volumes of distillate as charge stock for the
catalytic cracking process. When vacuum bottoms are used as charge, the residual produced, is
lower in density and lower in viscosity, than the original feed. The viscosity, while still high, needs
considerably less cutting oils to make No. 6 fuel oil. The average consumer will not encounter
many of the individual components just as they are produced, for the oils are frequently altered by
blending to meet the specifications of the various grades of fuel oils.

2.6 Hydrocracking

Previously we have discussed a number of different types of refining process such as

atmospheric distillation, thermal cracking, catalytic cracking, vacuum distillation and visbreaking.
The increasing market demand for middle distillates and lower sulfur transportation fuels has
resulted in the industry adding conversion capacity through a refining process called
“hydrocracking”. Hydrocracking is a very flexible process option that can be used on almost all
refinery streams. It produces gasoline components from light or heavy gas oils, cracked gas oils
or straight run gas oils from the crude distillation unit. It can also produce jet fuel and diesel fuel
from heavy gas oils.

The Concept

While hydrocrackers are very expensive, they work from a relatively simple process, which is
catalytic cracking in the presence of hydrogen. They also offer a lot of flexibility to the refiner
because varying the combination
of catalyst and hydrogen allows
for a wide range of feedstocks to
be cracked.

There are a number of different

designs of hydrocrackers, which
come in one, two or three stages.
The example that will be
described is a two stage version.
Feed, such as cracked heavy gas
oil is mixed with hydrogen vapor,
heated to 290ºC (550ºF) to 400ºC
(750ºF), pressurized to 8275 kPa
(1,200 psi) to 13,800 kPa (2,000
psi) and fed to the first stage
reactor where it passes through
the catalyst. At this point around
40% to 50% of the feed is
cracked to gasoline fractions. The Figure 12 – A Two
high pressure is needed because Stage Hydrocracker
the feedstocks used are usually
from other cracking processes
and are not easy molecules to
separate (crack).

Page 18 of 83
The catalysts used in this process are typically sulfur compounds with cobalt and molybdenum or
nickel with alumina, which in the world of petroleum refining are neither exotic nor expensive.
Most systems used fixed catalyst beds although about 20% of them use fluidized beds with
moving catalyst, similar to that of a traditional fluid cat-cracker.

The material leaving the first stage reactor is cooled sufficiently to liquefy all but the hydrogen and
methane, which are sent to the hydrogen separator where they are split for recycling back into the
process. The liquid is charged to a fractionator that cuts out the light streams such as gasoline,
kerosene/jet and light gas oil. The bottoms stream is then sent to the second stage reactor.
The process in the second stage reactor is essentially the same as the first (mix the feed with a
hydrogen stream and pass through a catalyst); however, the conditions are more severe utilizing
significantly higher pressure and temperature. The material from the second stage then has the
hydrogen removed from it and is usually sent to same fractionator as the material from the first
stage.

The high pressure used in hydrocracker reactors is one of the primary reasons why they are so
expensive. For example, the reactor walls can require 15 cm (6 inch) thick specialty steel
construction. Additionally, elaborate quench systems are built into most hydrocrackers because
the process is exothermic and the possibility of runaway cracking, which would dangerously
accelerate the cracking rate giving off more and more heat as hydrogenation takes place, is a
constant concern. Refiners use hydrocrackers in an integrated refinery as a pivot point to move
from maximum diesel and distillate fuels in winter to maximum gasoline and jet fuel in the
summer. They produce no “bottom of the barrel” residues such as residual fuel oil, pitch or coke.
Currently there are about a dozen different types of hydrocrackers in use; however, they are all
expensive to build and operate, which is their only real drawback.

2.7 Refining Yields

The refining processes when used individually or in combination give the refiner considerable
flexibility in how crude oil is processed, what products will be produced and in what quantities.
There are many possibilities for choosing the most cost effective process or processes which
demonstrates the craft or skill of the refiner. The more complex the refinery, the more likely the
finished product will be blended and the greater the importance of blending to the overall process.

Figure 13 – Refinery Yields

Page 19 of 83
Chapter 3 – Fuel Oil Stability and Compatibility

The subject of stability in petroleum products is complicated, technical and often over
emphasized. It is complicated because the meaning of stability is broad. A simple definition of
stability is “the resistance of an oil to breakdown”. Stability is mostly a concern with fuel oils.
Fuel Oils are composed of many hydrocarbons of which four types are present in the largest
volume. They are, in order of increasing instability:
 paraffinic
 naphthenic
 aromatic
 olefinic
The stability of these hydrocarbons is governed by their resistance to oxidation, or ability to
pickup oxygen, and their reaction to heat and high temperatures. Heavy fuel oils are subject to
great amounts of physical stress from heat, pressure and handling, in addition to contamination
from water and tank bottoms. Sometimes, during the cracking process the oil will be cracked too
much, producing free carbon and other insolubles. These particles often remain in suspension
and unnoticed until they are blended.

3.1 Compatibility or Blending Reaction

Compatibility is similar in that it is the tendency of fuels to produce deposits when mixed. This
organic precipitate or sediment is a chemical compound called asphaltene. Various chemical
compounds in the residual oil form a protective film around the asphaltenes, maintaining a stable
equilibrium. When blended with distillates, this film is broken causing the asphaltene particles to
drop out of solution. These particles are originally small in size; but eventually coagulate into
larger particles and settle at the bottom of the tank. The type of distillate, its composition, amount
of blending, and the amount of asphaltenes present in the residual, all determine the blending
reaction and percentage of sediment produced. Blending small amounts of distillate does not
precipitate large amounts of sludge. This is due to the maintained equilibrium, but as the
percentage of distillate is increased, the balance is changed. Simply stated, the more distillate
used as a diluent, the more waxy precipitation is formed. Until all asphaltenes are precipitated
from the residual oil, any additional amount of distillate added to the blend will precipitate more
asphaltenes. Thermally cracked distillates are better blending oils than are catalytically cracked
oils or straight-run oils. This is because they have a greater percentage of aromatic hydrocarbons
which help maintain the proper equilibrium in the residual after blending. Also, aromatic oils are
better solvents for the heavy type chemical compounds found in residuals.

3.2 Stability Spot Test

The usual specifications for a residual oil do not reveal its stability or solubility in distillate oils.
ASTM D-4740 is the standard test method for stability and compatibility of residual fuels by spot
test. In the stability procedure, a drop of the preheated homogenized sample is put on a test
paper and placed in an oven at 100°C. After 1 hour the test paper is removed from the oven and
the resultant spot examined and rated. A stable blend that does not precipitate heavy particles
results in a
Figure 14 – Example of Rating 1 and Rating 5 Spot Tests uniform brown
stain or blot with
no inner ring, and
is rated 1. A blend
that precipitates
heavy compounds
or insoluble
particles results in

Page 20 of 83
a very dark solid or nearly solid area in the center. The center area will be much darker than the
background. This is rated 5. There are also ratings of 2, 3 and 4. In the compatibility procedure,
a blend composed of equal volumes of the sample fuel oil and the blend stock are tested in the
same way as described for the stability procedure.

Sediment by Hot Filtration is another common test that is used to assess the stability of fuel.
Additional information on this can be found in Annex 3.

3.3 Problems Caused by Suspended Particles

Due to these solid particles floating in the oil, there is actually a colloidal solution present. That is,
solid material suspended in a liquid, and thus the oil is not a true, complete liquid. This condition
affects combustion to the greatest degree in smaller types of equipment and burners. In actuality,
the unit is trying to burn two types of material at the same time, the liquid oil and the solid,
insoluble substances. This results in two combustions requiring different ratios of air and different
methods of burning. Sometimes an oil may be unstable yet this is not noticed, principally because
of not knowing where to look for some of the results. The following items are due to instability of
oil:

 Excessive amount of waxy deposits in storage tanks. These deposits are found in most
tanks, but they should not accumulate rapidly.
 Large carbon formation in preheaters, due to coking of some excessively heavy
carbonaceous materials.
 Large formation of soot and black smoke caused by the incomplete combustion of the heavy,
precipitated, insoluble compounds.
 Thick waxy, gum-like substances found on screens and strainers.

Page 21 of 83
Chapter 4 - Mixing Methods
Blending is the processes of mixing various components such that they become a totally new
product. While this new product will be a function of the components that went into its
manufacture the underlying question is usually, at what point do these individual components
become this new product? The answer to this question is invariably tied to what type of mixing
was employed and how effective was it.

The methods employed in blending and mixing are as varied as are the results obtained. The
method used in any one situation is usually dictated by the equipment or physical plant that is
available. There are five primary methods currently in use. Some methods lend themselves to
certain products for obvious reasons; but, in the main it is not possible to state with conviction that
one method is superior to the other. The methods are agitation, circulation, density blending,
heating, and in-line blending.

4.1 Agitation

Agitation is a physical mixing process that is usually accomplished by the use of variable pitch
propeller mixers that are built into the tank. As these are specially built or retro fitted they are
limited in their availability but are never-the-less more common, as they are very effective. The
speed and effectiveness of this type of mixing mostly depends upon:

 The number of propellers available (typically one to three)

 The size of the propellers and the speed at which they rotate

Not surprisingly, the greater the number of propellers, the greater their size and the faster they
rotate, the more effective and faster the mixing. These mixing systems are usually designed to
mix a tank in the range of 6 to 12 hours.

4.2 Circulation

Circulation is accomplished by pumping the product from the bottom of the tank, through a
circulating pump and returning it to the tank, often at the top, although some of these circulation
systems are very sophisticated. The type used mostly for gasoline blending is called a “jet mixer”.
The jet part is a variation on the “venturi” effect, which creates a low pressure at the point where
the circulating
liquid leaves the
jet nozzle and
creates a second
circulation
process as a
result. The
Figure 15 – Jet Mixer Nozzle effectiveness of
these mixing
systems depends
greatly on their design; however they tend to have many “jets” at various angles as shown in the
photograph below.

Another type of circulation method is to pump air or nitrogen into the bottom of the tank. The
surface-tension around the bubbles created cause very small amounts of material from the
bottom of the tank to float to the top. As the bottom material is moved to the top of the tank it is
displaced with material above it creating a circulation. This method can be effective but has some
specific limitations. Nitrogen is the safest method but is very expensive. Air is more economical
but potentially much more dangerous. The use of either nitrogen or air can release unacceptable

Page 22 of 83
levels of volatile organic compounds [VOC] emissions and has a tendency to drive off the light
ends. Additionally they both significantly increase the potential for static electric
discharges. Because of the environmental impact, this method is becoming less common.

Figure 16 – Jet Mixer Nozzle Arrangement

4.3 Density Blending

Density Blending, also referred to as “Gravity Blending”, along with heating, is the most common
mixing method used for onboard blending. In most cases, the heaviest [densest] oil is pumped
into a tank first, and then progressively lighter [less dense] oil is pumped into the bottom of the
tank. The lighter oil will move up through the heavier oil mixing as it moves. When blending
multiple components this way, each component must be pumped in order of its density, beginning
with heaviest material and finishing with the lightest. This method is limited in its effectiveness but
is nevertheless widely used as it can be used on all products.

4.4 Heating

Heating, may not sound like a mixing method and is obviously not applicable to light products and
distillates; however, heat is absolutely necessary to obtain a good blend with residual fuel oil. As
the material in immediate contact with the tank’s heating coils becomes warmer it becomes less
dense and moves toward the upper part of the liquid and is displaced by cooler liquid from above
it; which, in turn, heats up and repeats the process, creating the circulation within the tank thus
giving the desired mixing effect. Residual fuel oil temperatures typically range from 50°C [122°F]
to 65°C [150°F] and the higher the temperature, the lower the viscosity, which further makes
blending easier. It should be noted, however, that vessels with coated tanks [which most do
today] will not usually heat or carry a cargo above 65°C [150°F] due to the possibility of coating
breakdown. Where possible, heating should be used in conjunction with other methods. The
down side of this process is that it is slow and typically takes days and even weeks to fully mix a
tank. Also, there is a cost associated with heating the cargo.

4.5 Inline Blending

Figure 17 – Inline Blending System
Inline Blending is usually very effective but
requires specialized and expensive equipment.
This method typically uses a constant rate pump
and variable rate pumps all of which are
metered. A flow computer controls the variable
rate pumps to insure that the correct amount of
blend stock is added. All pumps feed into a
common line which usually has a static in-line
mixer at the point where they join together.
Although initially developed for small batches
such as truck loading racks (gantries) many
systems are now large enough to accommodate
ship loading.

Page 23 of 83
Chapter 5 – Volumetric Shrinkage

Refining is the process of taking crude oil and separating it into its various commercial products
or fractions such as naphtha, gasoline, jet fuel, diesel, etc. This separation of the fractions can be
considered as the opposite of mixing. If you had a refinery that did not have any operational
losses and you weighed the crude oil going into the charge and then summed the weight of all the
products coming from the run down, the weight in should equal the weight out. However, if you
made the same comparison using volumetric measurements, you would find that the volume out
was greater than the volume in; typically, in the range of 5 – 10%.

The simple explanation for this is that larger molecules tend to have their atoms packed closer
together than the smaller ones do. In the refining process when the molecules are separated, the
lighter ones expand at a proportionally greater rate than do the heavier ones resulting in a larger
volume of finished product for a corresponding volume of feedstock. Of course there is no such
thing as a free lunch and should the products be mixed back together, the reverse of the process
takes place – hence the shrinkage. Therefore, whenever a light hydrocarbon mixes with a heavier
hydrocarbon a phenomenon known as volumetric shrinkage takes place. Actually, this process
happens any time two hydrocarbons are mixed; however, it is only measurably noticeable when
there is a large difference in the densities of the two materials

For high density or full boiling components the changes are negligible. However, when low
density hydrocarbons are blended with heavier stocks, volumetric losses can be significant; and,
has the most effect when a very light material such as condensate is used to blend with crude oil;
therefore, the larger the difference in densities of the materials to be mixed, the greater the loss.
This is generally not a problem with shipboard blending but it does come into play on occasion
when a very light cutter stock is used to bring down the viscosity. It also has a measurable effect
in gasoline blending when butane is added.

Let us assume that during the cracking process a Cetane [C16H34] molecule is split into the
following three smaller molecules; Octane [C8H18], Hexene [C6H12] and Ethylene [C2H4]. If we
were able to crack a cubic meter (1000 litres) of cetane completely into these three components,
the chemistry says that there would be 50% Octane, 38% Hexene and 12% Ethylene. These are
percentage by weight, not volume. The density of cetane at 15°C weighs 863.1 kg/m3, while
octane, hexene and ethylene are 707.5, 671.4 and 371.8 respectively. The table 1 below shows
how this breakdown affects the volume to weight ratio

Table 1 - Volumetric Shrinkage

Weight Percent Mass of Fraction Density Volume of Fraction
Product 3
of Fraction kg kg/m Litres
Octane 50% 431.6 707.5 610.0
Hexene 38% 328.0 671.4 488.5
Ethylene 12% 103.6 371.8 278.6
Combination 100% 863.1 1377.0
Cetane 100% 863.1 863.1 1000.0

5.1 API MPMS Chapter 12.3

There is only one commonly available document on volumetric shrinkage and that is API MPMS
Chapter 12.3 – Volumetric Shrinkage Resulting from Mixing Light Hydrocarbons with Crude Oils.
These tables are entered with the API Gravity or density difference and the percentage that the
lighter component represents in the mix.

Page 24 of 83
Example: Blend 90,000 Bbls of Heavy Crude Oil, API 10.0 with 10,000 Bbls of condensate, API
60.0
Step 1: Find the percentage of the light
component [C]:
Excerpt from Table 3 ‐ API MPMS Ch 12.3

C = 10,000 x 100 = 10% Light‐Product Concentration

100,000 API
 Gravity  Percent Shrinkage
Step 2: Find the API Gravity difference [G]. Difference 9 10 11
48 0.1198 0.1319 0.1438
G = 60.0 – 10.0 = 50.0 49 0.1256 0.1383 0.1507
Step 3: Go to table three to determine the 50 0.1315 0.1448 0.1578
shrinkage value for 10 percent light component 51 0.1376 0.1515 0.1651
and a 50.0 API Gravity difference. 52 0.1438 0.01584 0.1426
Shrinkage factor = 0.1448
Step 4: Calculate the physical shrinkage.

100,000 x 0.1448 = 144.8 Bbls.

100

5.2 Application

Caution must always be exercised when using API MPMS Chapter 12.3. There has only ever
been one comprehensive study done, which was in the 1970’s. The document that first came out
was Recommended Practice 2609C [RP2609C]. It later became API MPMS Chapter 12.3 when
they included a metric version of the tables but they still used the original study to derive the data.
There are dozens, if not hundreds of new crude oils that have come onto the market since that
study was conducted and no one really knows how accurate the data is in today’s world, although
it remains the only comprehensive study available. Additionally, the tables in API MPMS Chapter
12.3 were only derived from a study of mixing mostly unrefined light hydrocarbons with crude oil.
Therefore the applicability of the tables in API MPMS Chapter 12.3 to refined products is
unknown and cannot be recommended.

Generally, volumetric shrinkage is not a problem with shipboard blending but it does come into
play on occasion when a very light cutter stock is used to bring down the viscosity. It also has a
measurable effect in gasoline blending when butane is added. However, beyond its impact on
quantity determination it is a useful topic to introduce the subject of weight percent versus volume
percent, which is a key issue in blending calculations.

Note: API MPMS Chapter 12.3 is a copyrighted document of the American Petroleum Industry and can be purchased
from their web site at http://www.api.org

Page 25 of 83
Chapter 6 – Principles of Blending

6.1 Linear Blending

When two or more petroleum products are mixed [blended] a number of changes or reactions
take place. These reactions are physical and chemical. As stated in chapter two, one of the skills
of the refiner is to choose those blend components from the various downstream processes that
will produce a material with the desired specifications for the least cost. To make these decisions,
a blender must be able make reasonable estimations concerning the individual specifications of
the blended product.

The units used to report the many analytical specifications vary considerably. This is one of the
controlling factors in determining the final result when mixing two blend components. Results are
given in weight percent, volume percent, temperature, pressure, time, pass/fail or simply as a
number on an arbitrary scale.

To calculate the result of a blend, the principle task is to put the data into a form which can be
blended linearly. A property that will blend linearly is one where the final blend quality can be
expressed as the average quality of the individual blend components. In other words one of the
following equations must be satisfied:-

Pb = ∑Wi Pi or Pb = ∑Vi Pi

Where:
Pb = Property of the total blend
Pi = Property of Component i
Wi = Weight Fraction of Component i
Vi = Volume Fraction of Component i

Those properties whose analysis is presented in weight percent or volume percent are linear and
as such fit directly into the above equations. Examples are; relative density, sulfur, sediment and
water, ash, metals, etc.

6.2 Non Linear Blending

Many other petroleum properties, such as viscosity, do not blend linearly on any basis. To be able
to calculate the results of mixing these properties, the analytical values must be converted into
alternate values, which can then behave in a linear fashion. These alternate values are referred
to as blending numbers, blending indices or blending factors and must satisfy the following
equations:

Ib = ∑Wi Ii or ∑Vi Ii

Where:
Ib = Blending Index for Total blend
Ii = Blending Index for Component i

It should be noted that properties can blend linearly on a molal bases in addition to weight and
volume. From a practical side, molal information is not usually available to the shipboard blender
[it is frequently not available to the refiner either] and will not be reviewed in this seminar. For
example, vapor pressure can be blended on a molal basis with good accuracy. However, when

Page 26 of 83
blending calculations are based on volume flows, the vapor pressures must be converted to
volume blending indices as vapor pressures do not blend linearly on a volumetric basis.

6.3 Developing a Blending Index

There are four tasks necessary when developing a blending index.

 Collecting the Blend Data

 Deriving the Blending Function
 Adjusting Equation Coefficient
 Testing the Blend Function

6.3.1 Data Collection

Data Collection is one of the most important tasks in the development of blending indices and
there are two basic requirements. The data must be accurate and well distributed. The
difference between the calculated and the observed result will depend upon how closely the
products used to develop the indices, resemble that of the products that are actually being
blended.

The blending data should not only include all potential blending formulations, but blends that do
not fall into the normal range of blends. By including blends at the extremes, one is able to select
and test a system which will provide sensible results up to and beyond the limits of current
blending practice. Extrapolating results beyond the data population almost always increases the
uncertainty.

6.3.2 Developing Equation Forms

Selecting an appropriate equation form to represent the blending function requires a knowledge
of the property being modeled. In making this selection there are three options:

 A theoretical model
 An empirical model
 Combination of theoretical and empirical

If the analytical components are related to other qualities, then a theoretical model based on
physical laws is used to formulate a blending index system. For example: flash point is related to
the vapor pressure or boiling point of its blend components.

Many of the product properties used in the petroleum industry are empirical in nature and
therefore do not correspond to any simple physical or chemical property. For instance, the sulfur
content of a product is clearly related to the quantity of sulfur present in the product; whereas, the
octane number has no such simple interpretation. Conversely, the basic chemical and physical
properties of a product affect the results of empirical tests. Using this knowledge, one can
develop equations to describe the blending performance of the blend components by using the
techniques of multiple regression analysis and non linear optimization.

6.3.3 Computing Coefficients

Once the equation form has been determined, one must compute the coefficients which best
represent the data. The coefficients for a theoretical blending function are derived from the theory
itself, whereas the coefficients for the empirical system must be determined by statistical analysis.
Both sets of coefficients must be adjusted to fit the available blend data.

Page 27 of 83
6.3.4 Testing the Index

Finally, one must test the blending system against independent data. These tests consist of
calculating the blend property using the index and comparing the calculated value with the
observed value. The results are statistically analyzed. The mean and standard deviation of the
results (calculated-observed values) are the most common factors used in analyzing the reliability
of the blending function. These indicate if there is an overall bias in the blending index and give
an indication of an error. The standard deviation may include four types of error:

 Error in the blending index system

 Error in the observed property of the blend
 Error in making up the blend
 Error in the properties of the blend components

It is necessary to remove any error attributable to the last three sources from the standard
deviation to determine the error of the blending index system. Careful laboratory procedures
coupled with duplication of test blends will minimize any error resulting from poor blending
techniques and inconsistent or inaccurate techniques in determining the product property values.
The use of blends consisting of three or more components with approximately equal contributions
also helps to minimize errors associated with determining the property values.

6.4 Interaction Blending Model

This approach works well with octane, Rvp, and atmospheric distillation blending, although it can
be used for any non-linear blending. In the case of distillation, equations are more accurate in
terms of % evaporated at specified temperatures, rather than in terms of temperature at which
10%, for example, is evaporated

The interaction blending model proposed by Morris is one of the most widely used schemes for
blending studies.

6.4.1 Using Interaction Coefficients

The key element in interaction equations is the interaction coefficient and these can be developed
by testing each component and the 50/50 blend of the two components. Some pairs of
components have positive interactions and some have negative interactions. When interaction
coefficients between individual components are known, interactions between mixtures can be
calculated. As a practical matter, interaction coefficients for pairs of gasolines can be expected to
be quite small.

For blends of two components, the interaction blending equation is as follows:

Q = a1 X1 + a2X2 + b1,2X1 X2
Where:

Q = blend property or quality

a1 = property of component 1
a2 = property of component 2
X1 = Volume fraction component 1
X2 = volume fraction component 2
b1,2 = interaction coefficient for components 1 and 2

Page 28 of 83
For three or more components, there is an additional linear term for each component and an
interaction component for each pair of components. For example, the equation for a blend with
three components would be as follows:

Q = a1 X1 + a2X2 + a3 X3 + bl,2X1 X2+ bl,3X1X3 + b2,3X2X3

The linear coefficients can be changed as often as required to reflect changes in the quality of the
individual components. However, the interaction coefficients need not be changed unless the
component characteristics change significantly.

To undertake a basic interaction blending study, test all of the blend components plus 50:50
blends of all possible pairs of components. The total number of tests required for a blending study
equals:

n components + n (n -1)/2 blends

An interaction equation to describe the blending behavior of five components would have 5 linear
terms and 10 interaction terms. Experience has shown that interaction equations can accurately
predict the quality of any multi-component blend.

6.4.2 Obtaining Equations

Since most blending studies are conducted with linear programming models, linear blending
values must be developed from the blend data. This can be accomplished by calculating the first
derivative of the interaction model for a specific blend composition.

Interaction equations are developed by carrying out interaction blending studies. To obtain good
results, considerable technical supervision is required. The blending task is not trivial; blend
compositions must be accurate. The real cost of a blending study can be expected to be in
excess of $100,000.

Blending, like most things, is getting more sophisticated and increased sophistication is usually
accomplished by greater complexity. When working with traditional products such as fuel oils, the
current blending models, most of which have been around for thirty years or more, still provide a
useful and viable tool to predict final product quality. However, many petroleum products,
especially “boutique” fuels such as reformulated gasolines have numerous components resulting
in very large permutation combinations that can only be realistically calculated with the use of
computers.

In addition to the complexity of the calculations, these programs require a lot of input data and
many of them are self correcting, based upon the continuous streams of data that they may
possibly have access to. This data can be historical or even “real time” from in-line analyzers. The
development of such data is usually the proprietary information of the organizations that develop
the programs and process the data into useable forms. Therefore it is anticipated that less and
less information regarding blending models will be entering the public domain.

6.5 Bench or Hand Blends

It is critically important that everyone involved in a blending operation understand that

calculated blend results are theoretical and that it is unrealistic to expect the final result to be
identical to the calculation. The purpose of blend calculations is to give the blender a “best
estimate” starting point for the compilation of a bench blend, also known as a hand blend.

Page 29 of 83
A bench blend takes actual component samples and mixes them in the exact ratio of the
proposed blend. This will give a physical sample of the proposed blend. It is necessary to run
actual laboratory analysis to confirm or refute the results of the blend calculation. It is not unusual
for the bench blend results to vary from the calculated results. Usually the differences are not
large, but most blends are adjusted based upon the results of the bench blend. This requires the
bench blend be remade and a new analysis run on the new bench blend. Sometimes a client will
require only critical specifications to be rerun. There are a number of factors that contribute to
differences between the calculated blend results and the actual blend results:

 Precision of the individual component analytical results.

 Precision of the bench blend analytical results.
 Possibility of a Chemical Reaction.
 Precision of the Blending Indices

Those specifications whose analysis is given in percent weight or percent volume will typically
produce bench blend results that are close to the calculated results as these are only affected by
the precision of the test methods. Those specifications that rely on blending indices are prone to
differences between observed and calculated results.

However, the important element to realize is that the calculated blend input produces results of
theoretical output. When blending, never rely on calculated, theoretical data for your finished
product. Always check calculated results with actual analytical tests that were conducted on an
actual bench blend of the final mixture.

Page 30 of 83
Chapter 7 – Linear Blends
Linear blending is the simplest of the blending techniques that we will discuss as it is the sum of
average ratios (volumetric or weighted) of the individual component values. As previously stated
blend specifications given in volume or weight percent, will blend on a linear basis. In other words
the percentage of the specification in the blend component will be directly proportional to the
percentage of that component in the final blend. However, some specifications blend on a
volumetric basis, while others blend on a weight [mass] basis. It is important not to confuse
these two. Additionally, a set of product specifications will typically contain analyses of both
types, in addition to those that will require non-linear blending; therefore it will usually be
necessary to calculate the components of the blend on both a volumetric and a weight basis.

As you can see from the example in table 2, the weight percentage and the volumetric
percentage differ and this number is significant enough to affect your calculations. You will find
that this situation becomes more pronounced, the greater the spread in density. The calculation
process is the same regardless of which percentage basis is used. The important aspect to
remember is to use the appropriate percentage basis for the type of analysis being calculated and
do not confuse the processes.

7.1 Density and API Gravity

Quite frequently, the density of the blend can be calculated by dividing the total volume into the
total weight. However, if these figures are not readily available, calculate the final density using
the volume percent bias. API Gravity, unlike density, does not blend linearly on volume. This is
due to the relationship
of API Gravity to Relative Density. This relationship is:

API Gravity = 141.5 - 131.5

R. D.
Where: R.D. = Relative Density

To calculate API Gravity, first convert it to Relative Density or Density

7.2 Sediment and Water

Sediment and water by the centrifuge method gives a result in volume percent. Water by
distillation gives a result in volume percent; however, sediment by extraction initially gives a result
in weight percent. It is necessary to convert these to the same units. Traditionally, the sediment
by extraction result is converted to volume percent and added to the water by distillation number
to give a total sediment and water result in volume percent.

7.3 Ash, Metals, Sulfur

Ash and sulfur are reported in weight percent. Metals are typically reported in parts per million,
weight percent. It is not necessary to convert the ppm into an actual percent, ppm values work
equally as well and are more familiar than percentages. Some of the metals analysis typically
requested in fuel oils is vanadium, sodium, magnesium, aluminum, silicon, etc. This is not a
complete list, by any means. Regardless of the test, if the analytical result is given in weight
percent then the blend calculations are figured on weight percent.

7.4 Calculations

The following table [2] is an example of a three component blend (A, B & C) that we will use
numerous times throughout this manual. Typically, the blender will start with the volume (m3 @

Page 31 of 83
15°C) and a density at 15°C. Multiplying the volume by the density will give the weight of each
component.

To calculate the volume percent of component A, divide its volume (1,000 m3) by the total
volume of the final blend (5,000 m3), which when expressed as a percentage is 20%. The same
calculation is then performed for the remaining two components.
To calculate the weight percent of component A, divide its weight (980 metric tons) by the total
weight of the final blend (4,305 metric tons), which when expressed as a percentage is 22.8%.
The same calculation is then performed for the remaining two components.

Table 2 - Example of a three component blend

Blend Component Data Provided Blend Component Data Calculated
Component Volume Density @ 15°C Weight
3 Volume % Weight %
Identifier m @ 15°C kg/m 3 Metric Tons
Comp. A 1,000.0 980.0 980.0 20.0 22.8
Comp. B 2,500.0 865.0 2,162.5 50.0 50.2
Comp. C 1,500.0 775.0 1,162.5 30.0 27.0
Totals  5,000.0 861.0 4,305.0 100.0 100.0

Example 1.

In the blend shown in Table 2, sulfur values are assigned for components A, B and C , of 1.5%wt,
0.7%wt and 3.5%wt., respectively. What would be the sulfur content of the final blend?

The property of the total blend is equal to the summation of the component fraction [volume or
weight] multiplied by the property of that component.

Pb = ∑Wi Pi or Pb = ∑Vi Pi

Table 3 - Sulfur blend calculation - Weight percent (correct)

Blend Component Data Provided Blend Component Data Calculated
Component Sulfur Values Weight percent
Calculation Increment
Identifier of components of components
Comp. A 1.50 22.8 1.50 x 22.8% = 0.342
Comp. B 0.70 50.2 0.70 x 50.2% = 0.351
Comp. C 3.50 27.0 3.50% x 27% = 0.945
Totals  100.0 1.638

The mathematics for this calculation is not difficult. Obtain an incremental value for each
component by multiplying the component value by its percentage in the final blend. Add the
incremental value of each component to attain at the final blend value. In the example in Table 3,
the result is: 1.64% Wt. [Rounded]

Table 4 - Sulfur blend calculation - Volume percent (incorrect)

Blend Component Data Provided Blend Component Data Calculated
Component Sulfur Values Weight percent
Calculation Increment
Identifier of components of components
Comp. A 1.50 20.0 1.50 x 20.0% = 0.30
Comp. B 0.70 50.0 0.70 x 50.0% = 0.35
Comp. C 3.50 30.0 3.50% x 30.0% = 1.05
Totals  100.0 1.70

Page 32 of 83
What would happen to our calculation above if we used the volume percent basis, by mistake? As
you can see from table 4, the result obtained by using the [incorrect] volume fraction is 1.70%

Example 2. Using the same blend as shown in table 2, S&W values are assigned for components
A, B & C of 0.25%vol, 1.10%vol and 0.76%vol., respectively. Calculate the S&W content of the
final blend. As shown in table 5, the volumetric fraction is used because S&W is reported in
volume percent. Otherwise, the process for performing the calculation is the same as for the
previous sulfur examples. The result is 0.83% [Rounded]

Table 5 - Sediment and Water [S&W] blend calculation - Volume percent

Blend Component Data Provided Blend Component Data Calculated
Component S&W Values of Volume percent
Calculation Increment
Identifier components of components
Comp. A 0.25 20.0 0.25 x 20.0% = 0.050
Comp. B 1.10 50.0 1.10 x 50.0% = 0.550
Comp. C 0.76 30.0 0.76% x 30.0% = 0.228
Totals  100.0 0.828

Note:
When calculating the incremental value it is normal to round off to one or two decimal places
more than is used on the actual result value. However, the final result is rounded back to the
customary number of decimal places. Do not lull yourself into a false sense of precision because
the calculator can round to an infinite number of decimal places. The final result cannot have any
greater precision [and will usually have less] that that of the input data.

Additional Exercise

A client wants to make the following four component blend as defined in Table 6. Calculate the
theoretical final blend values for:

Table 6 - Component values for additional exercise - Linear Blending

Component Volume Density @ 15°C Sulfur S&W Vanadium
Identifier m 3 @ 15°C kg/m 3 Wt. % Vol.% ppm
Comp. A 1,250.0 960.0 2.80 0.35 300
Comp. B 2,000.0 865.0 1.65 0.20 50
Comp. C 1,500.0 825.0 1.10 1.27 25
Comp. D 1,750.0 780.0 0.35 0.68 156

 Density
 Sulfur
 S&W
 Vanadium

The answers to this exercise are found in your exercise study guide, which accompanies this
book.

Page 33 of 83
Chapter 8 – Non-Linear Blends

8.1 Blending Indices or Blending Numbers

When dealing with analytical values that do not blend on a linear basis it is necessary to use
blending indices (or blending numbers) to convert the analytical value into a number which will
blend on a linear basis. Once you have converted the analytical value into an index value, the
calculation process is the same as previously reviewed in Chapter seven on Linear Blends.

Ib = ∑ Wi Ii or ∑ Vi Ii

There is an additional step needed in order to arrive at the analytical value. The result obtained
from the equation above is the blending index value for the final blend, not the analytical value. It
is therefore necessary to convert this blending index back to an analytical value. This is achieved
by working backwards in the same table used to obtain the blending indices for the first part of the
equation. It is important, when working with blending indices, to determine if the index in question
was developed for volumetric blending or weight [mass] blending. Because many of these indices
were developed in the U.S.A. there is a tendency for them to be for volumetric blending; however,
this should not be taken for granted. Also, you cannot tell this simply by looking at the index
values. The index must state whether it is volumetric or weight [mass] basis as this is something
that only the compilers of the index know for certain.

To help put non-linear blending into better context we will explain the calculation process along
with and in conjunction with a review of some of the test methods whose analytical values do not
blend on a linear basis. Specifically:

 Viscosity
 Pour Point, Freezing Point and Cloud Point
 Flash Point
 Vapor Pressure

8.2 Viscosity

Viscosity can be defined as the resistance of the liquid to

flow under gravity. Establishing test methods that
produced analytical values for viscosity were some of the
earliest analysis conducted on petroleum and as a result
of this a number of viscosity scales were developed,
such as Engler, Redwood, Redwood Admiralty, Saybolt
Furol, Saybolt Universal, etc., to mention just a few. The
problem with having so many different viscosity scales is
that it is very difficult to compare values that use different
scales as these were all empirical in nature and had no
underpinning commonality to allow for ease of
comparison. The trend, today, is for the kinematic
procedure to supersede these older empirical
procedures, although some of them are still occasionally
used. Kinematic viscosity is a function of the absolute
viscosity; specifically, it is the ratio of the absolute
viscosity to the relative density of the oil at the
temperature at which the viscosity is measured.

Note: Absolute viscosity is the force required to move a Figure 18

plane surface of one square centimeter over Viscometer
Tube

Page 34 of 83
 Figure 19 another plane surface at the rate of one
Viscometer centimeter per second, when the two
Bath surfaces are separated by a layer of
liquid one centimeter in thickness. The
unit of absolute viscosity is the poise;
however, absolute viscosities are often
reported in centipoises (one poise
equals 100 centipoises).

ASTM D 445 - Standard Test Method

for Kinematic Viscosity of
Transparent and Opaque Liquids

Kinematic viscosity is determined by

measuring the time for a fixed volume of
liquid to flow under gravity through a
calibrated glass capillary viscometer at a closely controlled temperature. The kinematic viscosity
is the product of the measured flow time and the calibration constant of the viscometer. The
calibration constant is obtained by timing the flow of a viscosity oil standard through the
viscometer at a precise temperature. The unit of
kinematic viscosity is millimeters squared per Figure 20
second [mm2/s], formerly known as the centistoke Maxwell Viscosity Blending Index
[cSt] expressing kinematic viscosity.
Viscosity Blending Viscosity Blending
mm 2/s Number mm2/s Number
Note: The dynamic viscosity can be obtained by
multiplying the kinematic viscosity by the 100 25.90
density of the liquid. 0.2 120.00 200 23.10
0.3 106.50 300 21.60
The repeatability for viscosity analysis by ASTM D- 0.4 97.00 400 20.65
445 for residual fuel oils at 50°C is 1.5% of their 0.5 90.00 500 19.90
mean for results; and at 80°C and 100°C is 1.3% of 0.6 84.50 600 19.30
their mean results plus 8 mm2/s. The reproducibility 0.7 80.50 700 18.80
for residual fuel oils at 50°C is 7.4% of their mean 0.8 77.00 800 18.40
for results and at 80°C and 100°C is 4% of their 0.9 74.00 900 18.05
mean results plus 8 mm2/s.
1 71.50 1000 17.75
8.2.1 Viscosity Blending Indices 2 59.50 2000 15.80
3 52.00 3000 14.80
There are a number of viscosity blending indices 4 48.50 4000 14.10
available and doubtless many more covered by 5 46.00 5000 13.55
intellectual property rights, which are not available. 6 44.00 6000 13.10
In the U.S., two blending indices in common use are
7 42.50 7000 12.75
the Maxwell and the Chevron Viscosity Blending
8 41.30 8000 12.50
Function Two. Both of these index systems are for
volumetric blending. The Refutas system is a 9 40.25 9000 12.20
blending index for weight [mass] fractions and is 10 39.55 10000 11.95
widely used in Europe where products are bought 20 34.45
and sold on a weight basis. In all of these indices 30 32.10
the viscosity value to enter the table with, must
40 30.45
be in mm2/s at the same temperature.
50 29.25
Maxwell blending numbers are presented in the 60 28.30
table on page 16 and are the ones in most common 70 27.50
usage by SGS. This preference is by tradition rather 80 26.90
than a particular study which has shown the Maxwell 90 26.40

Page 35 of 83
to be superior to the Chevron or other indices. The Maxwell index has served well over the years,
however, as previously stated, no index is all things to all blends.

The viscosity index in your exercise study guide is the Refutas viscosity blending numbers index
for weight [mass] blending. In our exercise on viscosity blending we are going to take our original
blend from table 2 and assigning the components viscosity values, we will calculated the final
blend viscosity on a volumetric basis using the Maxwell index and on a weight [mass] basis using
the Refutas index. For ease of reference Table 2 is repeated on the following page.

8.2.2 Example

In the following blend, the viscosity values for components A,B & C are 3 mm2/s @ 50°C, 70 cSt
@ 50°C and 600 mm2/s @ 50°C, respectively. Calculate the viscosity of the final blend using both
the Maxwell and Refutas viscosity indices. The calculations and differences are shown below.

Table 2 - Example of a three component blend

Blend Component Data Provided Blend Component Data Calculated
Component Volume Density @ 15°C Weight
Volume % Weight %
Identifier m 3 @ 15°C kg/m 3 Metric Tons
Comp. A 1,000.0 980.0 980.0 20.0 22.8
Comp. B 2,500.0 865.0 2,162.5 50.0 50.2
Comp. C 1,500.0 775.0 1,162.5 30.0 27.0
Totals  5,000.0 861.0 4,305.0 100.0 100.0

8.2.3 Calculations

This is the same blend that was used in the linear blending examples in the previous chapter.
You will notice from the example in table 7 that the calculation gives the index value of the blend
not the actual result. It is therefore necessary to go back to the index and working in reverse,
determine the actual blend value that equates to the index value.

Table 7 - Viscosity blend - Maxwell Index calculations

Blend Component Data Provided Blend Component Data Calculated
Component Viscosity Volume percent Index
Calculation Increment
Identifier mm 2/s @ 50°C of components Number
Comp. A 3.0 20.0 52.0 52.0 x 20.0% = 10.40
Comp. B 70.0 50.0 27.5 27.5 x 50.0% = 13.75
Comp. C 600.0 30.0 19.3 19.3 x 30.0% = 5.79
Totals  100.0 29.94

Visc. Index #
44.3 40.0 30.45
50.0 29.25

By interpolation, an index number of 29.94 represents a viscosity of 44.3 mm2/s. The same
situation applies to the example in table 8 which uses the Refutas index. Here the index value for
the final blend is 29.79

Page 36 of 83
Table 8 - Viscosity blend - Refutas Index calculations
Blend Component Data Provided Blend Component Data Calculated
Component Viscosity Weight percent Index
Calculation Increment
Identifier mm 2/s @ 50°C of components Number
Comp. A 3.0 22.8 15.17 15.17 x 22.8% = 3.46
Comp. B 70.0 50.2 32.04 32.04 x 50.2% = 16.08
Comp. C 600.0 27.0 37.95 37.95 x 27.0% = 10.25
Totals  100.0 29.79

Visc. Index #
37.6 37.0 29.72 29.79
38.0 29.83
Using the same interpolation process, but on the different index, we determine that the final blend
value is 37.6 mm2/s. These two examples clearly illustrate that different indices produce different
theoretical values, further emphasizing the need for analyzing the final bench blend. The
difference between the two methods is 6.7 mm2/s. This is quite large and is not typical. The
difference between the two methods in the additional exercise is 1.2 mm2/s. This variation in
difference is indicative that the blend index equations are probably derived from different models.

It is a good idea to keep a record of actual versus theoretical blend results as this can prove
invaluable when determining which index to use.

Additional Exercise

Using the information in table 9 below, calculate the viscosity of the final blend using the Maxwell
and Refutas
Table 9 - Component values for additional exercise - Linear Blending indices. The
Component Volume Density @ 15°C Viscosity worksheets and
3 3 2 answers to this
Identifier m @ 15°C kg/m mm /s @ 50°C
exercise can be
Comp. A 1,250.0 960.0 850.0 found in your
Comp. B 2,000.0 865.0 0.8 exercise study
guide, which
Comp. C 1,500.0 825.0 75.0
accompanies
Comp. D 1,750.0 780.0 700.0 this book.

8.3 Pour Point, Freezing Point and Cloud Point

8.3.1 Pour Point

The pour point [ASTM D-97] of oil is the

lowest temperature at which the sample will
pour or chill, at specified conditions, without
disturbance. From the practical point of view
it is the lowest temperature at which the oil
is capable of flowing under very low forces,
such as gravity. The material tested is
placed in a stoppered jar which has a
thermometer running through the stopper
into the test material. The test sample is
heated to 9°C above its anticipated pour
point and is then placed in the cooling Figure 21
medium of the test apparatus. It is checked Pour Point
every 3°C until the test specimen shows no Bath

Page 37 of 83
sign of movement when held in the horizontal position for 5 seconds. 3°C is then added to the
temperature at which zero flow was noted; and, this is recorded as the pour point of the material.
The repeatability of this method is 3°C and the reproducibility is 6°C. A diagram of pour point
apparatus is shown above in figure 21

8.3.2 Freeze Point

The freezing point [ASTM D-2386] of an aviation fuel is the lowest temperature at which the fuel
remains free of solid hydrocarbon crystals that can restrict the flow of fuel through filters if present
in the fuel system of the aircraft. Another way of describing this is that it is the fuel temperature at
which solid hydrocarbon crystals, formed on cooling, disappear when the temperature of the fuel
is allowed to rise under specified conditions of test. This is an important test for aviation fuels;
however, these are not usually blended, although “re-batching” of jet fuel is becoming more
common.

8.3.3 Cloud Point

The Cloud Point [ASTM D-2500] is the temperature of a liquid specimen when the smallest
observable cluster of hydrocarbon crystals first occurs upon cooling under prescribed conditions.
In other words, the temperature at which petroleum oils begin to form wax crystals and become
cloudy in appearance as they are cooled towards their pour points.

As with viscosity blending, there are a number of different blending indices used for calculating
blend results of these analyses. The NOVA/Huskey and Reid-Allen indices for pour point, Hu-
Burns for all three and the Maurin method for freeze point. The NOVA/Huskey is most appropriate
for products blended at a specific facility with a relatively constant source of components. The
Reid-Allen Method is more suitable to blend components that do not come from the same source;
however, it requires knowledge of the 50% distillation temperature. The Hu- Burns method
developed a way of overcoming the need for the 50% distillation temperature by the introduction
of a constant in the equation. This resulted in a series of indices for varying constants and is not
an unusual situation for the development of blending indices. The concept is that, a refinery or
terminal would determine which constant value worked best for them and use it for their blend
calculations. Hu and Burns determined that a constant value of 0.05 worked best for cloud
point on an industry wide basis; and, a constant value of 0.08 gave the best industry wide
results for pour point. The Maurin freeze point method is a weight [mass] index of which there
appears to be no data available on the accuracy of the system.
Note: Where the constant value is shown on an index, this is for identification purposes only. It is
not necessary to factor the constant into the index numbers as this has already been
done. Some of these blending indices can be located in your exercise study guide.

While these three specifications are commonly included in final blend requirements, it is unusual
to blend to meet these particular specifications. Of the three pour point is the most common and
is also the most likely to be affected by a chemical reaction within the mixture. It is not unheard of
for the observed pour point of the bench blend to be outside of the minimum and maximum pour
points of the components of the blend. The Hu-Burns index numbers are for volumetric blending
and are used in exactly the same way that the Maxwell blending numbers are used in the
viscosity calculations.

8.4 Flash Point

The flash point [ASTM D-93] Pensky-Martens Closed Cup [PMCC] of a petroleum product is the
minimum temperature at which it gives off sufficient flammable vapor for an ignition source to
ignite it, under specified conditions [open cup, closed cup]. As with pour point there are a number
of indices to choose from. The Wickey-Chittenden method reported that 71% of the deviations
were within reproducibility of observed flash. Hu-Burns, as with their pour point indices,

Page 38 of 83
developed a variable series with a constant. They determined that a constant of -0.06 gave the
best industry wide results. This index can be found in the exercise study guide. The Thiele
method developed a flash point blending index for lubricating oils. Henri Maurin also developed a
weight [mass] based flash point blending index but again, the accuracy of the method was not
reported. The Hu-Burns index numbers are for volumetric blending and are used in exactly the
same way that the Maxwell blending numbers were used in the viscosity calculations. However,
estimating flash points of blended products usually results in a calculated flash point higher than
the observed flash point of the bench blend. From personal experience we have found that flash
point blending does not always produce satisfactory results. Extreme caution should be exercised
and results always backed up by observed flash points on actual blends.

Figure 22b
Figure 22a Flash Point
Flash Point Unit Photo
Unit Diagram

8.5 Vapor Pressure

Although many test methods are developed for determining vapor pressure, the one most
commonly used for petroleum products is the Reid Method [ASTM D-323]. This standard test
method is used to determine the vapor pressure of volatile, non-viscous petroleum products. The
Reid vapor pressure (Rvp) test is widely used as a criterion for blending gasolines and other
petroleum products. The Reid method employs a metal cylinder or bomb, having two chambers
(one for liquid and one for air) which are designed for coupling together. The bomb is fitted with a
pressure gage and is constructed to withstand high pressures. The test product is placed in the
liquid chamber, and the bomb is agitated and immersed in a water bath at 100°F. This process is
repeated and pressure readings taken at intervals. These readings are averaged to obtain the
uncorrected pressure which is converted to Reid vapor pressure (Rvp) by use of an ASTM table.

The Rvp is an absolute pressure exerted by a mixture as determined at 37.8°C [100°F] and a
vapor-to liquid volume ratio of 4; expressed as kilopascals [kPa] pounds-force per square inch

Page 39 of 83
(psi). Although the Rvp provides a convenient approximation of the absolute vapor pressure of a
partly vaporized sample at 100°F, it is not equivalent to the true vapor pressure. The Reid vapor
pressure differs from the true vapor pressure of the sample due to some slight vaporization of the
sample and the presence of water vapor and air in a confined space. In general, true vapor
pressures are higher than Rvp’s. ASTM D 323 is not applicable to liquefied petroleum gases or
fuels containing oxygenated compounds other than methyl t-butyl ether (MTBE). For
determination of the vapor pressure of liquefied petroleum gases, refer to Test Method D 1267 or
Test Method D 6897. For determination of the vapor pressure of gasoline-oxygenate blends, refer
to Test Method D 4953.

Vapor pressures will blend linearly on a molal basis according to Raoult’s Law and there are
several methods available. However, none of them lend themselves to use in volumetric linear
programming models. A blending index method to meet Rvp specifications for gasolines and jet
fuels is based upon the following equation:-

RVPI = RVP1.25

Where:
RVPI = Reid vapor pressure blending index
RVP = Reid vapor pressure, psia

A copy of these Reid Vapor Pressure Blending Index Numbers for Gasoline and Jet Fuel are
found in the exercise study guide. They are used in the same manner as other previously
described volumetric blending indices.

Additional Exercise

An additional nonlinear blending exercise for calculation of:

 Pour point
 Flash point
 Reid vapor pressure
is presented in your exercise study guide.

8.6 Interactive Coefficients

For blending situations where there are many component possibilities such as with gasoline
blending or where the analytical result has multiple values such as distillation, the use of blending
indices may not be the most practical approach and in cases such as these the interaction
blending proposed by Morris is one of the more widely used models for blending studies.

This approach works well with octane and atmospheric distillation blending, although it can be
used for any non-linear blending application. In the case of distillation, equations are more
accurate in terms of percent evaporated at specified temperatures, rather than in terms of
temperature at which 10%, for example, is evaporated. For our purposes we will use octane as
an example.

8.7 Octane

The power obtained from a conventional reciprocating internal combustion gasoline engine is
limited by abnormal combustion known as knock. Knock is defined as non-homogeneous pre-
ignition ahead of the spark. In other words, when the fuel-air mixture in the cylinder of a spark-
ignition engine burns spontaneously in localized areas instead of progressing from the spark, this
explosive decomposition produces a characteristic noise or knock. The accompanying pressure
surge cannot be converted into mechanical work by the piston, which must move in a fixed

Page 40 of 83
rotation driven by the crankshaft, and is lost as heat to the engine cooling system and exhaust
gas. In addition to loss of power and fuel economy, knocking can result in engine damage and
may even burn a hole in the piston, in extreme cases. For these reasons, the antiknock quality of
gasolines is one of their most important properties; and, the antiknock quality of motor gasolines
is normally expressed in terms of octane numbers.
In the various test methods used to measure antiknock quality, the test fuel is compared to blends
of two pure hydrocarbons, normal heptane and isooctane (2,2, trimethylpentane). Normal
heptane is quite low in its resistance to knock and has been arbitrarily assigned an octane value
of zero; isooctane has good antiknock properties and has, therefore, been assigned a value of
100. The octane number of a fuel is defined as the volume percentage of isooctane in a blend of
n-heptane which matches the test fuel in knock intensity under standardized conditions. In other
words if a grade of gasoline is said to be 90 Octane this means that when run through a CFR
“Knock” or octane engine, it match a blend by volume of 90% Iso-octane and 10% n-heptane.

There are two recognized laboratory

Figure 23 engine test methods for evaluating
Octane Engine antiknock quality of motor fuels,
namely the Research Method
(ASTM D 2699) and the Motor
Method (ASTM D 2700). Both
methods use the same basic test
engine but differ in the operating
conditions of the engine. The
research octane number (RON),
which is also referred to as the F-1
octane number, is representative of
the fuel performance during city
driving when the speed is generally
low and acceleration frequent. On
the other hand, the motor octane
number (MON), or F-2 octane
number, indicates the relative fuel
performance at high speeds.
Generally, the RON is higher than
the MON, and the difference
between the two ratings is known as
the “sensitivity” or “spread” of the
fuel. The octane number of a
gasoline blend will not necessarily
equal the volumetric average of the
octane numbers of the blend
components. As the differences in
antiknock rating, hydrocarbon type,
and other properties increase
among the blend components, the
deviations from ideal volumetric blending usually become more pronounced. The deviations
between observed and calculated antiknock ratings can be as high as several octane numbers
when blending common fuel components. To enhance the reliability of gasoline blending
calculations, a number of blending methods have been proposed.

Most modern blending systems are based on the interactive blending approach proposed by
Morris. In this system, the refiner must test the octane numbers of the individual components and
50:50 blends of all combinations of the potential blend components. For unleaded blends of three
components the interaction blending equation is as follows:

Page 41 of 83
ONb = AX1 + BX2 + CX3 + ica,bX1 X2+ ica,cX1X3 + icbcX2X3
ONb= octane number of blend
A = octane number of component 1
B = octane number of component 2
C = octane number of component 3
X1 = volume fraction of component 1 in blend
X2 = volume fraction of component 2 in blend
X3 = volume fraction of component 3 in blend
ica,b = interaction coefficient for components A and B
ica,c = interaction coefficient for components A and C
icb,c = interaction coefficient for components B and C

From a practical point of Component % Vol MON

view, it is unlikely that we
[SGS] would be in a A Light Straight Run 50.0 80.0
position to develop
interaction coefficients.
B Alkylate 30.0 90.0
Some interaction C Reformate 20.0 96.0
coefficients that were
developed by Morris are
available; however, there
value for industry wide
application is unknown
and should be considered
highly doubtful, unless
verified by observed
results. The same
equations work for both
Research octane number
[RON] and motor octane
numbers [MON];
however, you cannot mix
them up. The following is
an example of an
interaction coefficient
calculation for a three
component gasoline
blend.

MON = AX1 + BX2 + CX3 + ica,bX1 X2+ ica,cX1X3 + icbcX2X3

MON = (80 x 50%) + (90 x 30%) + (96 x 20%) Linear blend calculation
+ (8.8 x 50% x 30%) + (9.1 x 50% x 20%) + (-2.8 x 30% x 20%)
MON = 40.0 + 27.0 + 19.2 + 1.32 + 0.91 – 0.17
= 86.2 + 2.06
= 88.26
= 88.3 rounded

Page 42 of 83
Chapter 9 – Optimization

The previous chapters dealt with calculating the theoretical final result of mixing “Components A
through n”; however, this is not very real-world. A person does not arrive at work one day and
decide to mix two or three tanks together to see what they get. Typically, they will have material
at a certain specification (Component “A”) that they need to change so as to meet a new
specification, typically a sales specification, and they need to know how much of Component “B”
they will have to blend with Component “A” to meet that sales specification.

Optimization is the process whereby you calculate how much of “Component A” must be mixed
with “Component B” to reach a desired specification. To some extent this is the reverse of our
previous exercises because the final result is already known and the blender usually has a base
stock [typically the least expensive] into which he or she is adding a blend stock to bring the final
mix to the desired specification. Typically it is the quantity of blend stock that is being calculated.
Optimization calculations use all of the previously discussed techniques; however, it also requires
the use of an algebraic function to solve for one unknown.

Table 10 - component Base Blend Blend In this chapter we will

values for optimization Stock Stock #1 Stock #2 work some optimization
problems using the base
Quantity, m/t 10,000 5,000 3,000 and blend stocks shown in
Density @ 15°C 0.9204 0.8840 0.8632 Table 10. Exercise 1. how
Sulfur, wt. % 1.21 0.22 0.13 much of blend stock #1
2
417.0 34.6 41.8 must be added to the
Viscocity, mm @50°C
base stock to bring the
sulfur of the mix down to 1%?

In all optimization calculations it is necessary to develop a balanced equation and then solve the
unknown. In the case of a linear blend value such as sulfur, this balanced equation is:

Final Quantity x Final Sulfur = (Base Quantity x Base Sulfur) + (Blend Quantity x Blend Sulfur)

Our unknown in this case is the quantity of blend stock # 1 that is needed, which we will assign a
value of “y”; therefore the final blend quantity will be (10,000 + y) m/t. Placing these values into
our equation we get:

[10,000 +y] 1.0% = [10,000 x 1.21%] + [y x 0.22%]

Using standard algebraic rules we can reduce this formula as follows:

10,000 + y = 12,100 + 0.22y
10,000 + 0.78y = 12,100
0.78y = 2,100
y = 2,100 ÷ 0.78
y = 2,692.3 metric tons

It is usually a good idea at this time to check your blend calculation results to ensure that they will
have the desired effect.

Table 11 - Sulfur blend optimization calculation check

Blend Component Data Provided Blend Component Data Calculated
Component Sulfur Values Weight of Weight percent
Calculation Increment
Identifier of components Components of components
Base Stock 1.21 10,000.0 78.8 1.21 x 78.8% = 0.953
Blend Stock 0.22 2,692.3 21.2 0.22 x 21.2% = 0.047
Totals  12,692.3 100.0 1.000

Page 43 of 83
This check provides much more than just confirmation of the calculation. It is highly unlikely that
as a blender you would need exactly the contents of both tanks; in this example 12,692.3 m/t.
However, from table 11 you can determine that a blend of 78.8% of the base stock with 21.2% of
Blend Stock #1 will produce a blend with a 1% sulfur and knowing these rations allows the
blender to make whatever quantity is needed up to a maximum of 12,692.3 m/t.

In Exercise 2 we use a non-linear blend value (viscosity) for our calculation: How much of
blend stock #2 must be added to the base stock to give the final mix a viscosity of 380 mm2/s @
50°C? Use the Maxwell Index for your calculations.

This calculation is very similar to that of the sulfur exercise; however, there three additional steps.
It is first necessary to convert the viscosity analytical values to blending index values and
because the Maxwell Index is a volumetric based index, we must convert the metric ton quantities
to cubic meters at 15°C and then convert the answer from cubic meters at 15°C back to metric
tons..

Our unknown is the volume of blend stock # 2 that is needed, which we will call “Z”. 10,000 m/t
with a density 0.9204 = 10,865 m3; therefore the final blend volume will be (10,865 + Z) m3

Our next action is to convert the viscosity analytical values to index values using the Maxwell
Blending Index

Viscosity Index #
Base Stock 417 mm2/s 20.52
Blend Stock # 2 41.8 mm2/s 30.23
Final Blend 380 mm2/s 20.84

As in our previous example we must derive a balanced equation only in this case we use index
values rather than analytical values to derive the following equation:

Final Vol. x Final Index Value = (Base Vol. x Base Index Value)+(Blend Vol. x Blend Index Value)

Placing these values into our equation we get:

(10,865 + Z) 20.84 = (10,865 x 20.52) + (z x 30.23)

Using standard algebraic rules we can reduce this formula as follows:

226,426.6 + 20.84z = 222,949.8 + 30.23z
226,426.6 = 222,949.8 + 9.39z
3,476.8 = 9.39z
Z = 3476.8 / 9.39
Z = 370.3

Final Blend Quantity = 370.3 m3 or 319.6 m/t [370.3 x density]

Optimization is the most complex of the exercises studied thus far, and lends itself to
computerization. The exercises we have reviewed are the simplest forms of optimization since we
were only dealing with a two component blend and we only had one critical specification. In real
life, optimization is more involved. As you increase the number of blend components and critical
specifications the numbers of variables increase exponentially as does the number of
calculations. Complex blend optimization models are extremely intricate and need a computer
with very sophisticated and usually customized software to be practical in terms of speed and
accuracy.

Page 44 of 83
Chapter 10 - Planning a Shipboard Blend
10.1 Key Roles

Unlike single product loadings or multiple product loadings that are kept segregated, component
blending demands a very high level of teamwork and continuous monitoring from all involved. A
typical blending team consists of:

 BLEND COORDINATOR, usually a specialized operations supervisor in the respective office.

 BLEND MASTER, a qualified senior field inspector, preferably with Merchant Marine Deck
Officer experience. This is not an absolute requirement provided the individual in question is
sufficiently knowledgeable of shipboard procedures be capable of gaining the respect, and
more importantly the cooperation, of the vessel’s Chief Officer
 SHORE INSPECTOR
 LABORATORY CHEMIST OR TECHNICIAN must be fully conversant with the cargo being
loaded and with the principles of blending the product in question. It is not unusual, especially
in large laboratories, for
 Different technician/chemists to work on the various analytical test specifications required. A
senior chemist/technician is responsible for coordinating the work of the individuals who are
performing the tests, to review all results for logic; and, to liaise with the Blend Coordinator to
insure results are conducted in a timely manner. These, together with the vessel’s Chief
Officer and the client representative, form the nucleus of those responsible for directing the
operation and maintaining the necessary level of communication throughout the loading.

It is not always necessary for different individuals to carry out these functions and depending
upon the size and complexity of the blend it may be possible for one individual to perform multiple
functions; however, it is important that the functions themselves are carried out. Whenever
possible, personnel assigned to these tasks should not be changed during the blending
operation. Maintaining continuity of personnel directly involved in blending operations plays an
important role in reducing the risk of errors and omissions. Risks may otherwise be introduced
with unfamiliarity of the blend when in progress.

Blending in general and shipboard blending in particular is as much art as science. As seen in the
following flow chart [figure 14] on page 24, much simultaneous action and organization goes into
the execution of a shipboard blend.

10.2 Blend Coordinator

The SGS Blend Coordinator is a senior operations supervisor or manager, who is fully familiar
with all of the aspects of shipboard blending. The Blend Coordinator is the primary link between
SGS and the Client. Additionally, he or she must liaise with the Chemists/Technicians, Blend
Master, Inspector and the shore facilities. The Blend Coordinator monitors the entire operation
from start to finish. Prior to the vessel’s arrival and during the preliminary data gathering,
sampling and testing of the various components from the designated terminals, shore tanks and
pipelines, the Blend Coordinator remains the key person in maintaining communication with the
client, and in helping to develop the product to the required specifications.

The client will inform the Blend Coordinator of the initial blend components and the final blend
specifications, including quantity. This is initially verified by means of the SGS Blend Program.
The sample source, together with related values, e.g. target volumes, densities, viscosity, sulfur,
etc. are introduced into the blend program. The program will display the blended percentage and
volume of each component plus the final calculated blend specifications. The preliminary
information gathered at this stage may be further developed in the program by adjusting the
physical volumes of individual components or by completely changing components to maximize

Page 45 of 83
the most cost effective use of the components available to the customer for meeting the final
blend specifications. The Blend Coordinator then makes arrangements to have all components
sampled and brought back to the laboratory for testing. When the specifications of the
components are known, a bench blend will be composited and the results compared with the final
blend specifications. This process lasts from a couple of days to a couple of weeks. Changes in
the initial planning stages are the norm rather than the exception. As the price and availability of
the individual components vary, so does their participation in the finished product.

Figure 24 – Shipboard Blend Operations Communications Flowchart

All of this activity usually takes place prior to the vessel’s arrival. In some cases there may be
blend components not immediately available for sampling. The vessel may have appreciable
quantities of OBQ, the pipeline may be laid in a different material from the shore tank, the product
to be loaded may not be available. As useful as blend programs are in this situation, it is never
prudent to rely on calculated blend data. It is preferable to put the blend together with as many
actual components as are available and keep the calculated one to a minimum. However, at
some point prior to final loading, a bench blend from actual components should be composited for
verification purposes and possible adjustments if necessary.

10.3 SGS/Client Interface

It is important that the following steps be satisfied and information gathered and discussed before
beginning the job.

Page 46 of 83
Complete review of nomination requirements, including:
 Intended number of blend stock components, together with relevant identification details.
 Quantity to be delivered.
 Location to be delivered from.
 Method of delivery, i.e. pipeline or barge.
 Typical specifications if known.
 Final product designation.
o As this is a blend the finished product will be new material and there is a certain
amount of latitude in what it can be called; however, we cannot call it something that
we know it is not. It is usually best to stick with generic descriptions.
 Targeted QUANTITY and SPECIFICATION of finished product.
 Approximate loading temperature and temperature to be maintained during voyage.
 Expected status of shore tanks after delivery, e.g. will any shore tanks be stripped? If so, this
may result in the terminal delivering less product than planned, which will then require a
revision of the total blend.
 Expected loading rate during each receipt.
 Draft restrictions at each terminal.
 Verification of charterer’s instructions for component distribution and, if relevant, individual
shore tank distribution.
 Verification of charterer’s instructions concerning the pre-loading vessel tank inspection
including the level of cleanliness required.

10.4 SGS/Shore Terminal Interface

Shore tank measurement and sampling will be conducted by the SGS inspector using recognized
industry standards. The terminal personnel may not be aware that the vessel is blending and that
the material to be loaded will only represent a certain percentage of the finished product. A
meeting should be scheduled with the appropriate terminal supervisory personnel during which
the requirements for loading will be explained in detail.

The shore line content must be verified and, if necessary, flushed and reset with a compatible
product and to insure full condition before loading. If only one shore line is being used, and more
than one product is to be loaded from the terminal, arrangements should be made to load the
heaviest product first and the lightest product last. This configuration will allow for the most
effective commingling of products on board while effectively flushing the lines with the lighter
product at completion. As the product in the shore line is being delivered to the vessel as part of
the blend cargo, the chemical and physical properties must be compatible with the blend
specification. The line content must be sampled in sufficient quantity to allow for testing as:

 A component.
 Integration into the laboratory hand blend.
 For retention.

The shore line should be sampled from as many accessible points as is practically possible. The
shore terminal is responsible for calculating the stop gauges and confirming the volumes to be
loaded together with estimated product temperature. To assist the terminal and crosscheck
information received from the customer the SGS shore Inspector will present to the terminal the
Blend worksheet from the Blend Coordinator and verify the shore-tank stop gauges and volumes.
The Shore Inspector will also relay this information to the Blend Master on board the vessel.

An effective means of communication must be established between the shore terminal and the
vessel. The terminal should recognize its responsibility to inform the vessel of any intended or
unscheduled stops and provide a final countdown before completion of loading, i.e. 2500 bbls to

Page 47 of 83
go, 1500 bbls to go, 1000 bbls to go, etc. This form of communication will allow the vessel
personnel/ SGS Blend Master to fine-tune volumes on board if necessary. The expected loading
(pumping) rate and volumes must be agreed between terminal and vessel before the start of any
movement. Any deviation from intended blend worksheet and load plan, by either ship or shore
should trigger a notice of apparent discrepancy/protest upon consultation with the Blend
Coordinator and SGS’ client.

10.5 Blend Master/Vessel’s Chief Officer Interface

In addition to the key meeting requirements as outlined in such standards as the APl’s Manual of
Petroleum Measurement Standards Chapter 17.1 or the equivalent Energy Institute or ISO
standards, the charterer’s instructions must be reviewed with the Chief Officer. Under normal
circumstances, the Charterer will contact the vessel independently and obtain from the Master or
Chief Officer available tonnage based on the target density. In consultation with the Chief Officer,
a load plan must be prepared. This will take into consideration all product information known at
the time together with relevant vessel information supplied by the Chief Officer. It is important that
both parties participate in this exercise and that both fully understand the plan and agree to its
objectives. We stress the word IMPORTANT because under normal circumstances our client
does not require “constant attendance” by our Blend Master for cost/economic reasons.
Therefore, both parties must be in full agreement with the plan before the SGS Blend Master
departs the vessel following the start of loading. The Chief Officer must sign and acknowledge
this plan. Also, a copy of the load plan must be made available to the Blend Coordinator, for
review. The Blend Coordinator may wish to input suggestions and use the information to maintain
an overview of shipboard operations.

The vessel’s Chief Officer should be made aware that our calculations will take into consideration
any significant amount of arrival O.B.Q., excluding segregated slop, although this will not be
evident from the vessel load plan prepared at each respective terminal. Using an estimated cargo
temperature, stop gauges must be calculated by the ship, verified by the Blend Master and
entered on the load plan sheet. Normally, cargo will be shut down by the delivering terminal at the
completion of loading, per shore stop gauges. It is necessary for the Blend Master to monitor the
vessel ullages during the topping-off operation, at each terminal, to insure that no contamination
occurs if for any reason the terminal continues to deliver product in excess of their own stop
gauges. In case of under or over delivery by the terminal, the client and Blend Coordinator must
be advised so that the appropriate corrective actions can be instigated [i.e. resume loading or
adjust blend calculations, load plan and subsequent component correction). A notice of apparent
discrepancy/protest may be issued to the shore terminal upon consultation with the Blend
Coordinator and client.

The vessel’s Chief Officer must be informed of the sequence in which the components will be
received from each terminal, respective shore tank identifications and component specifications.
Unforeseen circumstances may dictate changes in any of these factors and replanning will be
necessary. Each load plan sheet must show the calculated draft and, for comparison, the
maximum allowable draft for the respective terminal. Not all Chief Officers are well-versed in the
art of on-board blending. Therefore, it is important that the Blend Master take an active role in the
information-gathering process in preparation for loading, this is not to suggest, however, that he
should become involved with or interfere with the Chief Officer’s own duties. Regardless of the
complexity of any blend operation, the Blend Master’s task remains within the bounds of normal
SGS operating procedures as advisors/consultants and not operators of the ship. Opening or
closing valves, setting of stop gauges, starting or stopping pumps, switching tanks, etc. are the
sole responsibility of the ship.

Page 48 of 83
Chapter 11 - Load Plan Calculations
The actual component ratio for the final blend and the order of loading for each component will
usually have been decided prior to the vessel arriving at the first loading berth. While this is a
significant step in the process, there remains the important step of getting the right amounts of
each component into the right tanks. Most ships are paid on the amount of freight they carry. The
more freight they carry, the more they are paid. If a ship does not load its maximum cargo, the
charterer will have to pay “dead freight”. In other words the charterer will be penalized for the
difference between what the ship loaded and what it could have loaded. Therefore, it is in
everyone’s interest to insure that the ship loads as much cargo as possible. There are two
primary limiting factors when determining how much a vessel can load.

11.1 Maximum Loadable Volume

This is a simple concept. The maximum loadable volume is the total amount of cargo tank space
available on the vessel. Ships do not load their tanks above the 98% capacity point except in
unusual circumstances. In practical terms it is 98% of the total cargo tank space that is available.
Therefore, this 98% total cargo tank capacity is the maximum Total Observed Volume [TOV]
that the vessel can load. However, petroleum products are bought and sold on a weight [mass] or
volume at standard temperature basis. Therefore the amount of cargo that the vessel can load is
going to also depend upon the density and temperature of the cargo.

11.2 Maximum Loadable Quantity

The second factor limiting how much a vessel can load is its maximum tonnage or deadweight.
This is usually based upon a maximum draft restriction, although there are other factors that
come into play. They are:

 Maximum draft due to loadline restrictions

 Maximum draft due to berth or harbor restrictions
 Vessel stress [bending moments, sheer, etc.]
 Trim
 Stability

All vessels have loadline restrictions which limit the maximum draft they can load to. This limit
varies slightly depending upon the location of the vessel and the time of year. The vessel cannot
exceed this draft, other than for its permissible fresh water allowance. The maximum draft due to
berth or harbor limitations is usually information that the charterer determines in advance. It is
always a good idea to verify these items.

The vessel cannot load a cargo in such a manner that it will put unacceptable levels of stress on
the hull. The cargo must be fairly evenly distributed throughout the vessel to avoid excessive
bending moments. Also, it cannot have many empty tanks next to full tanks as this causes
unacceptable sheering forces. On most ships there is a specialized computer called a “loadicator”
or some other similar name which calculates these factors. They can be calculated manually from
the vessel’s hydrostatic tables but this is quite tedious. Trim is not a major factor but, does come
into play on occasion. On many ships, if all tanks were loaded evenly, [i.e. they all had the same
ullage] it would be trimmed one or two meters by the head. This is unacceptable as it affects the
vessels handling and fuel consumption. Additionally, excessive trim increases the vessels
maximum draft, which might not necessarily effect how much it can load [the loadline is always on
the amidships draft marks] however, it could cause problems if there are berth or harbor
restrictions. Most ships calculate their load plans so as to leave the loading port slightly by the
stern to allow them to arrive at the discharge port on even keel. Stability is only a problem when

Page 49 of 83
dealing with OBO [Ore/Bulk/Oil] multipurpose carriers. This will be discussed in greater detail in
the chapter on loading procedures.

The first thing that the Blend Master must get from the Chief Officer is the maximum loadable
quantity [weight] per tank that will allow the vessel to sail at its desired maximum draft with
acceptable trim. The Blend Master in conjunction with the Chief Officer must then calculate what
the anticipated Total Observed Volume will be for each tank, based on the estimated final
density and temperature; and, insure that this amount does not exceed the 98% capacity of the
tank. If any of the TOV amounts do exceed the 98% tank capacities, then the load plan must be
recalculated. Once the individual tank quantities are agreed upon, the Blend Master and Chief
Officer can calculate the quantities of each component to be loaded on a tank by tank basis as
per the blend percentages. This is limited to weight and standard volume, with the exception of
the first component to be loaded. It is now possible to calculate the vessel tank stop gauges
[ullages] for the first component. To calculate the vessel tank stop gauges [ullages] for all
components is possible, but not practical. The ullages appertaining to successive components
are based on actual, not computed readings; and, the estimated loading temperature is best
obtained from the shore tank immediately prior to product transfer.

Table 2 - Example of a three component blend

Blend Component Data Provided Blend Component Data Calculated
Component Volume Density @ 15°C Weight
Volume % Weight %
Identifier m 3 @ 15°C kg/m 3 Metric Tons
Comp. A 1,000.0 980.0 980.0 20.0 22.8
Comp. B 2,500.0 865.0 2,162.5 50.0 50.2
Comp. C 1,500.0 775.0 1,162.5 30.0 27.0
Totals  5,000.0 861.0 4,305.0 100.0 100.0

Using the quantities from table 2 [copied here for ease of reference], the blend is to be loaded
onto the M/V Lollipop (a good ship) into tanks IC, 2C, 3C & 4C. The Chief Officer advises us that
the quantities to be loaded into these tanks are 500 m/t, 1,000 m/t, 1305 m/t and 1,500 m/t
respectively. In conjunction with the vessel's Chief Officer, we determine that this particular
loading plan will work.

Table 12 - Weight to observed volume calculation to ensure weight of product fits volumetrically in vessel tanks
3 GOV
Tank Metric Density GSV M Est. VCF 98% Go /
Number tons to @ 15°C @ 15°C Temp. Table M3 Tank No Go
1C 500 861 581 58°C 0.9646 602 750 G
2C 1,000 861 1,161 58°C 0.9646 1,204 1,225 G
3C 1,305 861 1,516 58°C 0.9646 1,572 1,740 G
4C 1,500 861 1,742 58°C 0.9646 1,806 1,840 G
TOTAL 4,305 861 5,000 58°C 0.9646 5,184 5,555 G

The next step is to calculate how much of each component will go into each tank, based upon the
blend percentage. There are two ways this can be achieved, arriving at the same answer. You
can use the weight percentage to arrive at weights for each component fraction. However, as you
will be working, in the main, with gross standard volume [GS.V.} it is usually easier to do this with
volume percentage.

The advantage of using tables 12 and 13 is they not only provide a uniform tabulated method of
presenting the information, they also provide a means of checking your answers. This is the very
important initial planning stage where it is critical to double check all figures, as a mistake here
will have consequences throughout the entire blend.

Page 50 of 83
Table 13 - How much of each component to load into each tank
Tank Number  1C 2C 3C 4C
Final GSV, Bbls@15°C 581 1161 1516 1742 Totals 
Comp. A 20%, Vol 116.2 232.2 303.2 348.4 1000.0
Comp. B 50%, Vol 290.5 580.5 758.0 871.0 2500.0
Comp. C 30%, Vol 174.3 348.3 454.8 522.6 1500.0

The next calculation in the process is to determine the individual vessel tank stop gauges
[ullages] for component A, which is the first to be loaded. This is achieved with the use of a
Vessel Load Plan. The first three columns list the tanks to be loaded, their 98% capacity and
their calculated volume after loading. The next column lists the GSV on board before loading. As
this is the first port, this will be the OBQ quantity. The next column lists the GSV to load as
calculated in the previous table. The next column is the calculated GSV on board after loading
component A. [This is the OBQ plus the GSV to load]. The next two columns list the estimated
density and temperature after loading component A. The next column lists the volume correction
factor [VCF] calculated from the density and temperature of the previous columns. The following
column is the calculated gross observed volume [GOV] quantity for component A after loading.
This is obtained by dividing the calculated GSV on board after loading by the volume correction
factor. With the GOV, it is possible to go the vessel’s tank capacity tables to calculate the stop
ullages. In order to arrive at the observed stop ullage it is necessary to calculate the trim after
loading to be able to calculate the appropriate trim correction.

The last three columns allow the user to put the actual ullages after loading plus the actual GOV
volume and compare this with the calculated GOV volumes. It is normal for there to be minor
discrepancies; however, large differences must be investigated and it may be necessary to
modify the blend. Of course this is done only after consultation with the Blend Coordinator
and Client. The same form is used for each subsequent component. The GSV on board quantity
is taken from actual ullages and not from calculated ones. With respect to the estimated density
and temperatures after loading, it will be necessary to take into consideration the density and
temperature of those components already on board. Topping off a blend always requires extra
attention. It is important to insure that each tank has sufficient room for the last component,
keeping in mind that the Chief Officer may need to do some trim adjustments in the final segment
of the loading. The Blend Master should work as closely as possible with the Chief Officer at this
stage to anticipate and to insure that final adjustments do not compromise the integrity of the
blend.

The following three pages are the complete Vessel Load Plans for the M/T Lollipop having loaded
our initial blend. The first load plan is for the first component to be loaded and this will be
completed prior to the start of loading. The second load plan is not completed until after the first
component has been loaded and measured as there are usually slight variations between the
calculated stop gauges and the actual stop gauges and the blend master will always want to start
a plan with the most accurate information available. The primary reason for these slight variations
is due to trim. Ship tanks are usually loaded two to four at a time so the trim of the vessel will
differ slightly as different tanks complete cargo operations. The same applies for the third and any
subsequent components. The effects of trim reduce as the vessels gets closer to its fully loaded
condition.

Page 51 of 83
 File 123456
Vessel M/T Lollipop
Port Somewhere
Vessel Load Plan - Page 1 of 3 Date xx/xx/xxxx
Cargo Residual Fuel Oil Blend
Shore Volum e Units:  Bbls  M3  Other Inspector:
Delivery Density @ 15°C Observed VCF To Deliver Actual Delivery
Tanks Temperature Tbl _54B_ GSV GOV GSV GOV Term inal / Witness:
Comp. A 980.0 66.0 0.96512 1000.00 1036.14 1003.92 1040.20
Vessel's Staff:

Estim ated Forw ard Draft 3.20

Estm ated After Draft 9.75
Totals 1000.00 1036.14 1003.92 1040.20 Estim ated Trim 6.55 by the stern
Vessel Volum e Units:  Bbls  M3  Other
Tanks to GSV on Board GSV GSV Dens. after Estimated VCF GOV Stop Trim Trim Corrected
Load Before Load to Load After Loading Loading Temp. Tbl _54B_ After Loading Ullage Corr. Stop Ullage

1C 3.40 116.20 119.60 980.0 63.0 0.96653 123.74 16.75 0.46 16.29
2C 0.00 232.20 232.20 980.0 63.0 0.96653 240.24 15.68 0.12 15.56
3C 6.30 303.20 309.50 980.0 63.0 0.96653 320.22 15.02 -0.03 15.05
4C 3.50 348.40 351.90 980.0 63.0 0.96653 364.09 15.31 0.21 15.10

Totals 13.20 1000.00 1013.20 1048.29

Table 14 – Vessel Load Plan for Component “A”

File 123456
Vessel M/T Lollipop
Port Somewhere
Vessel Load Plan - Page 2 of 3 Date xx/xx/xxxx
Cargo Residual Fuel Oil Blend
Shore Volum e Units:  Bbls  M3  Other Inspector:
Delivery Density @ 15°C Observed VCF To Deliver Actual Delivery
Tanks Temperature Tbl _54B_ GSV GOV GSV GOV Term inal / Witness:

Comp. B 865.0 60.0 0.96309 2500.00 2595.81 2498.50 2594.25

Vessel's Staff:

Estim ated Forw ard Draft 6.85

Estm ated After Draft 10.10
Totals 2500.00 2595.81 2498.50 2594.25 Estim ated Trim 3.25 by the stern
Vessel Volum e Units:  Bbls  M3  Other
Tanks to GSV on Board GSV GSV Dens. after Estimated VCF GOV Stop Trim Trim Corrected
Load Before Load to Load After Loading Loading Temp. Tbl _54B_ After Loading Ullage Corr. Stop Ullage

1C 153.60 290.50 444.10 898.0 61.0 0.96407 460.65 10.01 0.03 9.98
2C 237.60 580.50 818.10 898.0 61.0 0.96407 848.59 8.96 0.02 8.94
3C 307.30 758.00 1065.30 898.0 61.0 0.96407 1105.00 9.37 -0.03 9.40
4C 387.30 871.00 1258.30 898.0 61.0 0.96407 1305.20 9.52 0.02 9.50

Totals 1085.80 2500.00 3585.80 3719.44

Table 15 – Vessel Load Plan for Component “B” (Loaded After Component “A”)

Page 52 of 83
 File 123456
Vessel M/T Lollipop
Port Somewhere
Vessel Load Plan - Page 3 of 3 Date xx/xx/xxxx
Cargo Residual Fuel Oil Blend
Shore Volum e Units:  Bbls  M3  Other Inspector:
Delivery Density @ 15°C Observed VCF To Deliver Actual Delivery
Tanks Temperature Tbl _54B_ GSV GOV GSV GOV Term inal / Witness:

Comp. c 775.0 50.0 0.96094 1500.00 1560.97 1503.40 1564.51

Vessel's Staff:

Estim ated Forw ard Draft 10.50

Estm ated After Draft 11.00
Totals 1500.00 1560.97 1503.40 1564.51 Estim ated Trim 0.50 by the stern
Vessel Volum e Units:  Bbls  M3  Other
Tanks to GSV on Board GSV GSV Dens. after Estimated VCF GOV Stop Trim Trim Corrected
Load Before Load to Load After Loading Loading Temp. Tbl _54B_ After Loading Ullage Corr. Stop Ullage

1C 445.60 174.30 619.90 861.0 58.0 0.96453 642.70 4.22 0.01 4.21
2C 817.00 348.30 1165.30 861.0 58.0 0.96453 1208.15 1.31 0.00 1.31
3C 1051.50 454.80 1506.30 861.0 58.0 0.96453 1561.69 3.80 0.00 3.80
4C 1249.40 522.60 1772.00 861.0 58.0 0.96453 1837.16 0.92 0.00 0.92

Totals 3563.50 1500.00 5063.50 5249.71

Table 16 – Vessel Load Plan for Component “C” (Loaded After Components “A” and “B”)

Page 53 of 83
Chapter 12 – Loading Procedures
12.1 Ship Tank versus Shore Tank Blending

Blending is seldom undertaken by the user. Most blending is conducted by the refiner [producer]
or the supplier. Most receivers require a product that is already blended and have a tendency to
object to receiving their product in component pieces. However, with shipboard blending, it is not
unheard of for the component parts to be placed on a vessel and kept segregated during the
voyage; then, blended during discharge. This situation is quite different from shipboard blending
during a loading and is usually limited to discharges into shippers own tanks. This topic will be
discussed in greater detail in chapter 15.

Most blending is performed in shore tanks and the greatest degree of accuracy is achieved when
entire shipments are pre-mixed in shore tanks prior to transfer. However, on board blending is a
recognized means by which a cargo may be prepared to specification in the vessels tanks, by
volumetrically blending individual components. Each component must be identified by its unique
physical and chemical properties; which, when blended with other identifiable components will
result in a product of saleable specifications.

On board blending offers certain conveniences to suppliers, particularly to trading organizations

that have limited [or no] terminal facilities. Some of the principle reasons to support shipboard
blending are:

 Limited storage space.

 Freedom to choose suppliers.
 Limited inventory.
 Considerable physical distance between product components.

In comparison to other methods of blending, i.e. batch blending in a storage tank or in-line
metered blend components; on board blending is a complex alternative. Some of the downside
aspects of this type of blending are:

 Greater number of variables and unknowns

 Errors are more difficult to fix
 Physical mixing is limited

12.2 Ship Considerations – Before Loading

The vessel tank inspection on arrival is conducted in an identical manner to any other tank
inspection for the product in question. It is becoming more frequent for pre-loading tank
inspections to be limited to “deck-level” surveys. There are a number of reasons for this but most
are concerned with time and/or regulations.

 If the vessel is going to back load, the time to clean tanks and the cost of disposing the slops
can be problematic.
 If the cargo is to be inerted [which most are, even fuel oils], it will typically take 24 to 48 hours
to inert all of the tanks, depending upon the size of the vessel’s inert gas generators.

It is important that everyone involved understands the limitations of this type of survey and that
we do not give the impression that a “deck-level” survey is as effective as going into the tanks
and visually inspecting them. If there is any likelihood that remnants from the previous cargo
could place the cargo to be loaded off-specification then first-foot samples should be taken and
analyzed for critical specifications such as flash point, color, etc. It must be kept in mind that even

Page 54 of 83
 though the tanks are measured, a significant
amount of unmeasurable cargo can remain
aft of the gauging points and in cargo lines.
This is an area where the Blend Master and
Unmeasurable OBQ Blend Coordinator must work together as
any significant OBQ volumes must be taken
into consideration as part of the finished
product; therefore, samples must be drawn
for testing. This scenario does not occur
frequently but must be confirmed by
Figure 25 – Unmeasurable OBQ sounding of all tanks.

Slop volumes must be sampled and measured separately. The slop weight may influence the
total quantity of cargo which can be safely loaded in relation to draft restrictions at load or
discharge port. Load-on-top slop is normally not permitted; and, similar to permanent ballast
tanks and designated void spaces must be checked for inadvertent loading or presence/ absence
of oil before and after loading at each terminal. If the vessel maintains segregated slops, a
sample should be taken prior to loading the first component. This is for reference purposes in
case of a cargo contamination. It is quite common to back load a fuel oil blend cargo behind a
crude oil discharge. In this case the major concern will be that of flash point. The surveyor must
pay particular attention to the vessel’s lines, ensuring that they are all dropped back into a tank
before gauging for OBQ.

Clingage can become a factor on very high pour material such as Low Sulfur Waxy Residue
[L.S.W.R]. This does not usually present itself as a quality problem but can affect the final vessel
quantity. This will result in larger than expected vessel over shore figures. You may find that your
client will want to base the Bill of Lading on the vessel quantity with the vessel experience factor
[VEF] applied, rather than the shore tank figure. This is a commercial issue between buyer and
seller. However, it is important in these instances to clearly indicate what was done and how any
figures were arrived at. At the completion of the vessel tank inspection, copies of our reported
findings are presented to the Chief Officer and signed by him. Only SGS approved field forms are
to be used. We do not normally sign papers presented by ship or terminal.

12.3 Ship Considerations – During Loading

Once loading has started and sufficient product has entered the cargo tanks, an attempt is be
made to determine cargo temperatures as soon as they are likely to have stabilized. The cargo in
the tanks will offer a more realistic assessment of the loading temperature, thereby allowing
adjustments to be made accordingly to stop gauges. It must be recognized, however, that any
gauging / sampling activity may only be conducted with prevailing safety regulations fully
complied with, especially those relating to static electricity.

Toward completion of loading, the Chief Officer should reduce the loading rate to a level which
will allow him to top off the tanks up to the predetermined stop gauges without unnecessary risk
of overfilling the tanks involved. Overfilling of any single component loaded after the first
component can create irreversible consequences that would invalidate the entire blend plan.
Should this occur, the Blend Coordinator must immediately be advised to discuss with the
customer how to adjust the blend based on a new (theoretical) calculated blend confirmed by
bench blend testing, a new load plan and shore stop gauges. Notices of apparent discrepancies/
protests may be issued at this time to the ship and/or shore terminal.

12.4 Ship Considerations – After Loading

After completion of loading at EACH terminal or of EACH component, the cargo tanks must be
gauged to determine:

Page 55 of 83
 The total quantity received
 The total quantity now on board

It is recommended to check void spaces, slop tanks and permanent ballast tanks, especially if
appreciable ship/shore discrepancies are noticed. If the client’s instructions call for segregated
parcels, they must be calculated independently to insure the product received is distributed
according to the load plan. Always obtain the Chief Officer’s signature on the SGS load plan. It
should be noted that the VESSEL LOAD PLAN, does not replace the vessel ullage report.

After completion of loading of ALL components, the total cargo on board must be gauged and
sampled and all void spaces, slop tanks and permanent ballast checked. Sampling of the vessel,
especially with the limitations of closed and restricted systems, can be time consuming and of
limited effectiveness, especially if the cargo is stratified. This will be discussed later in the chapter
on sampling; however, a sampling plan should be agreed with the client such that the samples
obtained meet the client’s requirements.

12.5 Shore Considerations

While the shore operation tends to lend itself more closely to a typical loading than the vessel
operation does, there are still a number of areas where the shore inspector must be particularly
vigilant and coordinate with both the Blend Coordinator and the Blend Master.

The shore pipeline is frequently the area of largest concern because its contents represent an
unknown element both from a quality and quantity aspect. The larger the line, the larger the
amount of uncertainty. Where possible the shore line should be displaced, prior to loading, with
material from the delivering shore tank. This not only insures the quality but helps to insure that
the line is full. It should be noted that displacement is not an absolute guarantee that all air or
vapor has been removed from a line. Especially if the line has many changes of direction in it and
the displacement was done at a very slow pumping rate. If the shore pipeline is not displaced, it is
imperative for a sample to be taken so that its effect on the blend can be verified. It will also be
necessary to know the capacity of the line to adjudicate its impact.

It is quite possible for different components of the blend to be in different tanks at the same
terminal. The important element to insure if this situation occurs is that each tank must be
pumped separately. As far as the blend is concerned they are separate operations. In this
situation it will often be desirable to displace the line contents of the first component with that of
the second component. This can pose some interesting shore line accounting. It will be
necessary to reduce the stop gauge of the first component by the amount of the shore line. It is
also very important that the vessel knows that when the second component starts up, a volume of
the first component [equal to that of the capacity of the shore line] will be used to finish off the
loading of the first component. It will also be necessary to increase the stop gauge of the second
component by the amount of the line capacity, unless the shore line is displaced into the vessel,
after loading, as a separate operation.

It is always helpful if the shore inspector who is monitoring the loading is also the same one who
did the pre-sampling for the bench blend. This will put him or her in a much better position to
verify that the product in the tank has not changed since it was sampled. If there is more product
in the tank immediately prior to loading than there was when the tank was sampled, it will be
necessary to re-sample the tank to insure that the quality has not changed. If there is less product
in the tank, than when it was pre-sampled, it is still necessary to take a sample and at least verify
one or two critical specifications, to insure that stratification did not affect the remaining product. It
is always a good practice to sample a tank whenever it is gauged and to gauge a tank whenever
it is sampled and this is especially the case for shipboard blending. When a tank is presampled
for the bench blend, the blend coordinator should know if the full amount of the component is in

Page 56 of 83
place at the time of sampling. If it is not, then enquiries can be made to determine where the
balance will be coming from and how it might impact the final blend. It is also useful information
for the blend coordinator to know if the shore tank in question contains a significant amount of
product more than the component amount. Again product stratification is the concern and the
Blend Coordinator may recommend levels [upper, middle, lower] testing. It is important, in this
situation, to sample the tank after loading for possible testing or retain in the event of a dispute.

12.6 Order of Loading

For shipboard blending, gravity mixing is one of the principle means of getting a homogeneous
final product. Blending by gravity consists of loading the cargo with the highest density first and
then in descending order of density with the component having the lowest density being loaded
last. Density is usually reported at a standard temperature. Density is also directly proportional to
temperature which will affect the observed density. A product at a cooler temperature can have
a higher observed density than another product at a higher temperature, even if its density at
standard temperature is lower than the other product’s density at standard temperature.

In some situations the client may decide to load the vessel outside of the gravity loading
progression. This tends to happen under the following circumstances:

 Different components at the same terminal, which require the vessel to make multiple
dockings. A possible alternative to this situation is to blend the multiple components into one
component prior to the arrival of the vessel. This will not affect the overall blend but will
change the load plan.
 If an intermediate component is at another port, a possible alternative would be to use
another vessel or barge to bring this component to the mother vessel and perform a vessel to
vessel transfer.
 The berth for the component in question may not have sufficient water to accommodate the
vessel at that stage of loading. Again, the use of a barge or other such lightering vessel is an
alternative when there are draft restrictions.
.
12.7 Mixing and Heating

A primary unknown in any shipboard blend is that point at which the individual components
become a new product that takes on the specifications of the blend itself. As discussed in
previous chapters, components that have widely different densities or temperatures do not mix as
readily as those whose densities and temperatures more closely resemble one another. In a
typical blend some components will mix well and other components will mix poorly. Additionally,
once the second component is loaded, you are dealing with an intermediate blend that is
probably not completely blended and changes with each additional component.

Not just one mixing process taking place in a vessel’s tanks during the loading and the voyage
but, a number of mixing processes. Whether one is more significant to the mixing process than
the other will depend upon the circumstances of the blend and the voyage. These are:

 Effects of Gravity Mixing and Heat, if applicable

 Turbulence from the Loading Operation and from the Vessel Movement

In any mixing situation, a circulation or movement of the component pieces through one another
will produce the greatest amount of mixing. Thus, the effects of turbulence, especially from
movement of the vessel will have a much lesser effect than heating. Heating creates a natural
circulation in the product. The material in immediate contact with the heating coils becomes
warmer; therefore, less dense than the remainder of the tank. This less dense material will rise

Page 57 of 83
through the product and is displaced by cooler more dense material which is heated and the
process starts all over again.

It will take time for this process to work its way to the material at the top of the tank. Also, it is
possible to have a density inversion if a very light component is the top layer. Even though the
product beneath is at a higher temperature, it is possible for the observed density to be higher.
This is one of the reasons why the lightest component should never be loaded first. If this cannot
be avoided, an alternative is to load it into a tank or tanks that are not designated for that
particular cargo and then transfer it internally once enough other material has been loaded to
insure some mixing. This significantly increases the complexity of the load but means the
difference between having a mixed or a stratified cargo at the discharge port.

12.8 Sampling

For testing purposes, unless a problem is suspected or known to exist with either the product or
the storage tank, samples from a full tank will comprise upper, middle and lower levels drawn in
accordance with ASTM D-4057, API MPMS Chapter 8, Section 1 or equivalent. However,
stratification is the number one enemy in producing a good blend. If stratification is suspected,
sampling should be adapted to meet the circumstances so as to determine the extent of
separation. To ascertain whether a tank is stratified the upper, middle and lower samples are
usually analyzed for density and possibly other key tests such as viscosity, sulfur, etc.

If it is determined that the product is stratified, remixing should be started. The terminal will
typically conduct this operation using one of several methods, the more common being:

 In-tank blenders or mixers

 Having the product circulated outside of the tank through the external pipeline system and
returned to the tank.

If re-sampling indicates the mixing is ineffective in breaking strata, then the components are
considered incompatible and therefore inappropriate for normal blending purposes.

In a tank of non-stratified compatible product, typical sampling in preparation for the bench blend
will include upper, middle and lower levels of all tanks included in the blend. In preparing the
various level samples for testing, only those drawn from levels related to the product to be
delivered will be included in the batch designated for the proportional laboratory bench blend. If
our calculations show that only product from the lower third of a particular tank will be delivered
then only the sample representing that level will be included in the bench blend. Those not
included in the bench blend will be retained. These may ultimately be required if additional
product is requested and/or if the delivered volume from the respective shore tank is more than
was anticipated in the first instance. This exercise will proceed, only after full consultation with the
client. The laboratory will test individual samples from each source (shore tanks, pipelines) for
consistency with the various suppliers’ own specification. Providing these agree, the laboratory
will refer to the computer-generated (theoretical) calculated blend and prepare a proportional
hand blend of the individual components. Testing is conducted to determine if there are any
significant variations with the specifications of the final product blend; and, the computer
generated theoretical blend in order that adjustments can be made before the vessel arrives at
the first loading terminal. All test results are conveyed to our customer for approval. If cleared by
our client, the components in question will be earmarked for loading.

In most cases, test results of the final proportional hand blend sample prepared in the laboratory
are used as the official custody transfer quality. Depending upon purchase/sales contract
conditions, we may be requested to provide after load vessel composite samples for official
custody transfer quality testing purposes; however, we advise against this whenever possible. By
its nature, shipboard blending does not allow for a good homogeneous mix of product

Page 58 of 83
immediately after loading; and, the variables prevalent in the overall makeup of a cargo on board
a vessel consisting of up to ten or more components, do not lend themselves to obtaining a
representative sampling immediately after loading.

Page 59 of 83
Chapter 13 - Special Operations
13.1 Discharge Blending

There are two types of discharge blending. The first involves a blended cargo which is either not
completely mixed or is off specification. The second is when the individual components are
loaded as segregated parcels to be blended either in the receiving shore tank or as the product is
being discharged.
Regardless of the type of discharge blend, it is critically important to know the exact specifications
of the product on board, on a tank by tank basis. This will involve a considerable amount of
sampling of the vessel on arrival. The exact nature of the sampling will depend upon how the
product was loaded. At a minimum, it is necessary to take upper middle and lower samples. It
may be necessary to take additional levels if stratification is determined. It is also a good idea to
take temperature profiles of the vessel’s tanks as this can give an indication of possible
stratification. This is needed when working out vessel arrival quantities; and, with electronic
thermometers it does not take a lot of time.

Most terminals are uncomfortable with the idea of blending from the vessel to the terminal, even
though they frequently have better facilities for mixing product than the vessel does. On discharge
blending the Blend Master must work more closely with the terminal than he or she would on a
blend loading. A discharge blend operation is usually trying to achieve one of three results:

 Blend material already in a tank.

 Insure that individual tanks receive the same product, from a vessel that has a stratified
cargo.
 Insure that individual tanks receive a specified blend from individual components on a vessel.

Shore tanks that contain petroleum products are usually a lot smaller than those that contain
crude oil. Therefore, a typical vessel discharge operation will involve multiple shore tanks. The
difficulty in discharge blending is to insure that all tanks receive the same or a specified blend.
This will often require the vessel to discharge entire vessel tanks into specified shore tanks and
may cause operational problems on the vessel that include:

 High levels of stress, especially sheer

 Problems with trim and list
 Reduced pumping performance
 Problems with cargo stripping

The Blend Master needs to be innovative, persuasive and diplomatic to get the vessel to perform
in such a manner. This will be best for the blend but not necessarily easiest for the vessel. Most
vessels are limited in the number of empty tanks they can have next to full tanks. It is normally
the case in these situations to stop the discharge after the first shore tank is completed and
transfer product into the vessel tanks that were emptied. This will reduce the sheer and bending
forces on the vessel.

One of the advantages of discharge blending is that mixing can be quite effective using the
vessel’s pumps and discharging through the vessel manifold. Most tankers have three or four
main cargo pumps and with a cooperative Chief Officer it is possible to use these like an in-line
blender. The rate on each pump is adjusted based upon discharge rate as determined from
individual tank ullages. In many cases the remote ullage readings from the cargo control room are
adequate for this purpose. This works best if the components or tanks are limited to two or three.
Some bunker barges use this system to great advantage allowing them to deliver varying grades
of intermediate fuel on demand.

Page 60 of 83
Many of the problems associated with loading a blended cargo are the same as those of
discharge blending:

 The contents of the shore pipe line must be taken into consideration and displaced where
possible.
 The blend calculations are the same.
 It will be necessary to discharge the heaviest component first. Terminals prefer to fill one
shore tank completely and then move to the next one. Although this makes more work for the
vessel it provides a greater element of control. All of the vessel's tanks are controlled within a
relatively small area [either the cargo control room or the main deck]. Switching valves in
some terminals can require three or four people; and, it may be necessary to drive from one
tank to the other.
 Excessive amounts of ROB may affect the final discharged product, although this is unusual.

Discharge blending is generally associated with a problem loading; however, there are certain
situations where discharge blending can be very cost effective, especially at the buyers own
terminal.

13.2 Allowing For Changes

This is the most difficult aspect of a shipboard blending operation. Once the second cargo
component is loaded, the only changes that can be easily made are those that do not require
material to be pumped off.

The importance of careful pre-planning and meticulous attention to detail when setting up a blend
cannot be over emphasized. Once a blend progresses beyond its first component you have a
partially mixed product that can be extremely difficult, if not impossible, to get a representative
sample of. This is a problem that gets worse with the addition of each component.

13.2.1 Changes during Loading

For example, what would happen if a terminal inadvertently started pumping from the wrong tank
[one that had not been sampled or tested] at an intermediate component stage. Once the material
enters the vessel’s tank, some mixing takes place. This means that it cannot be pumped off
because it is not the same product that was loaded. If you did pump a comparable volume back
off the ship you would almost certainly leave some of it on the vessel and discharge some of the
previously loaded components. However, as it does not mix completely at this point, you cannot
easily sample the product to get a base stock from which to calculate a new blend. All situations
are dictated by the circumstances of the case. It is necessary to advise the client immediately and
issue the necessary protests. Unfortunately, no amount of letters of protest will fix the cargo. The
first point to determine is the quality of the material that went onboard. It may be possible to
sample what is left in the shore tank or there may be line samples available. You might get lucky
and find out that the material that was loaded by mistake was of a better quality than the intended
component. Murphy’s Maritime Law very rarely works this way. Depending upon where in the
loading operation the mistake was discovered, can be a big factor. If it was early, and only a few
tanks were involved, it might make sense to take multi level samples from the tanks in question
and to use that as a base from which to calculate a new blend for those tanks.

Vertical stratification is not usually a problem on vessel tanks but in a situation such as described
above, where the product was loaded slowly into large tanks, it is prudent to get level samples
from more than one location in the tank or tanks. Changes will require new bench blends to
confirm calculated blend figures and new load plans. Changes may impact the order of rotation
for loading the remaining components. The optimum loading order may not be achievable at this
late

Page 61 of 83
13.2.2 Changes at Discharge

This would typically be covered under the previous heading of discharge blending, however a
situation could arise where it was commercially advantageous to change a blend at the discharge
port. This would involve one of the following scenarios:

 Adding a component or components to the original blend

 Discharging part of the original blend and adding additional components to make a new blend
 Discharging one parcel completely, spreading the remaining parcel throughout the vessel and
adding additional components to make a new blend

The last situation is the one that occurs most frequently. Again, stratification is the primary
concern but provided the material is adequately mixed, the re-blend of the parcel remaining on
board is no different from any other blend loading. Mixing is a concern, especially if the composite
density of the remaining components is greater than the density of the product left on board. This
could create a significant stratification problem. Even without any density problems, mixing is still
a concern if there is a short voyage to the final discharge port. Applying additional heat will help;
but, whether or not the product is fully mixed by the time it reaches its destination, is often a
function of the time it takes to get there. The second scenario is unusual but not unique. It works
well in theory and the calculations and back-loading are usually straight forward. However you
tend to run into the same basic problems, which are:-

 How representative is the sample of the cargo remaining on board? This product will be used
as the base for all future blends, both calculated and bench blends.
 How well will the final product mix and will it be fully mixed by the time it gets to where it is
going?

The first scenario is very unusual. It is unlikely that there would be enough room on board to add
additional cargo. You could probably rely quite heavily on the load port specification for your final
blend. There would still be the mixing problem, which would be significant if the component has a
high density or differs considerably in temperature and or density.

Page 62 of 83
Chapter 14 - Limitations and Problems
14.1 Communication

Clear, concise communication between all parties is a fundamental requirement if the margin for
human error is to be minimized during a complicated blending operation. SGS takes measures to
assist in meeting these parameters by:

 Insuring that involved company personnel are in attendance during key meetings.
 Continuity in personnel allows the team to remain familiar with the job at hand. It keeps an
overview of the operation until successful completion and insures that important details are
not overlooked.

14.2 Ship - Shore Volumetric Differences

It is well known that shore tank and marine vessel tank capacity tables, in most cases, are not
produced to the same level of precision, which results in volumetric differences between ship and
shore transferred quantities.

According to studies carried out many years ago, they concluded that the average volume, as
taken from a vessel’s capacity table has an absolute accuracy of ±1%. Today, this number is
likely on the high side and from our experience we would put the figure closer to ±0.5%. No one
can say with certainty what the actual number is but land based tank strapping techniques
produce accuracy of about ±0.1% so no matter how you slice it the variation in precision is
significant and the resulting volumetric differences can be quite large.

This difference is significant to shipboard blending because the total volume of a component is
usually shipped on shore tank measurements while the incremental separation into each ship’s
tank is based upon ship measurement, which has a lower level of precision. This tends to have
the most impact on the first component and may require some adjustment in stowage if this
variation is significant.

It should be noted that it is possible to physically measure ships’ tanks and produce capacity
tables with comparable accuracy to those of shore tanks; and, a few ships do have such tables.
Most gas carriers and a few chemical tankers have capacity tables produced from physical
measurements but the majority of the rest of the tanker fleet do not. The differences in table
accuracy come from the way in which ships’ capacity tables are typically generated.

14.2.1 Ships’ Capacity Tables

There are two primary aspects that are specific to the generation of a vessel’s capacity tables that
impact their accuracy:-

 Most ships capacity tables are calculated from the vessel’s drawings as opposed to actual
measurement. Therefore, the tables are only as accurate as the closeness of the vessel to its
drawings. This leads to variations in the accuracy of individual tank’s total volumes as
compared to one another. For example; if the bulkhead between two tanks is a few inches
off, one tank will be larger and one tank will be smaller than the drawings specify.
 The other major contributor to this problem is the way in which the deadwood in the tank is
accounted for. Deadwood is the internal frames, webs, stiffeners, etc. that supports the shell
plating of the tank; and, it is usually accounted for in the capacity tables as a straight line
linear deduction. However, on most vessels 70% of the deadwood is located in the lower
25% of the tank. This results in the accuracy of the capacity tables varying with the liquid
level in the tank. Typically, volumes in the lower levels of the tank will tend to be overstated or

Page 63 of 83
 more overstated than volumes at the upper levels of the tank. This can be particularly
problematical for shipboard blending where blend components are “layered” into the tank.

14.2.2 Other Factors

While ships’ capacity tables and the way in which they are generated impact ship to shore
variation the most, there are a number of other factors that need to be taken into consideration.
Some of these are:-

 Vessel’s capacity tables frequently lack important information, such as:-

o Reference gauge heights.
o Zero gauge volume, which can add considerable uncertainty to OBQ/ROB volumes.
o Position of the gauge port with respect to the centerline of the vessel and the forward and
after tank bulkheads.
 Trim and list correction tables have their own precision limits. The horizontal plane of the tank
must be square or rectangular for maximum accuracy.

Accuracy of these tables is inversely proportional to the amount of trim or list. Excessive amounts
of trim or list will have an impact on gauge accuracy due to the deflection of the tape away from
the vertical.

 Trim and list corrections do not work collectively, unless there is a special combination
table. When both situations exist, make every effort to have the vessel remove one or both.
List is usually the easiest for a vessel to remove.

Operational and maintenance issues can also affect the accuracy of the capacity tables.

 Permanent and temporary tank deformation can alter a tanks reference height and/or volume.
This can be caused by effects such as Hogging, Sagging, Racking or Twisting. Additionally,
grounding and collisions can cause tank deformation, which is common with barges.
 Repairs and modifications, including new seals on the expansion truck hatch lid and
retrofitted standpipes.
 Capacity tables must be applicable to the gauge point. Many tanks have more than one
location from which to gauge. Therefore, it is imperative to take measurements from the
gauge port that was used to calculate the tank's capacity table. This is a common problem on
ships that were retrofitted for closed measurement. It is important to insure the “zero-ullage”
of the closed measurement device matches the “zero ullage” of the capacity tables.

With the advent of double-hull tankers, some of these problems are improving, especially those
attributable to deadwood because more of the deadwood is now contained in non-cargo spaces
such as cofferdams and double bottoms. Additionally, there is a growing trend, although
admittedly a slow one, to move away from calculating ships’ capacity tables from their drawings
and to physically measure the tanks. However, even if all ships’ capacity tables were generated
from physical measurement there are still some measurement issues that we will never be able to
escape from.

Very small amounts of trim and list will always impact accuracy if they are too small to correct for.
In this situation, the bigger the tank the bigger the impact; and, the trend today is for ships’ to
have a small number of large tanks rather than a large number of small tanks, as was more
prevalent in the past. Also, the further away the gauging point is from the geometric center of the
tank, the greater the impact.

Additionally, ships are required to empty their tanks at the conclusion of each voyage so there will
always be issues relating to the measurement of ROB, and clingage.

Page 64 of 83
14.3 Imperfect Mixing

Much has already been discussed regarding this topic. Imperfect mixing is the “Achilles Heel” of
any blending operation, whether ashore or on board a ship. The methods employed in blending
will determine the extent of a good blend and mixture, which will affect the stability of that blend.
Other factors affecting blends are the viscosity and gravity of the oils and the percentages used.
Oils of high viscosity make blending more difficult, as will a wide spread in density among the
blending oils. The more light oil used in a blend, the more extensive the mixing methods should
be. It is easier for a heavy oil to absorb a little light oil than it is for a lot of light oil to absorb a little
heavy oil. The reason is, when a little light oil is added to heavy oil, it does not appreciably cool
the heated heavy oil. When adding a little heated heavy oil to a large volume of cool light oil, the
heavy oil is cooled by the light oil. This increases the viscosity of the heavy oil causing difficulty in
mixing the two oils. Oils of similar densities will mix better and faster than those with a wide
difference in density.

The stability of blended oils will be better when the blend components are close in density as
there are fewer tendencies to separate or stratify; as in the example of low density oil rising to the
top and high density oil sinking to the bottom. It should be clearly understood by all parties that
the blend is unlikely to be completely mixed immediately after loading. Should this condition
continue throughout the voyage and the vessel arrive at the discharge port with a stratified
product, the situation can be problematic. Solutions to this situation are limited and will be
governed by the circumstances; such as, the type of vessel, the cargo, the discharge facility, etc.

Imperfect mixing will impact the effectiveness of upper, middle and lower spot samples or running
samples to produce a representative sample. The first action is to sample the vessel in a way that
insures the resulting composite is representative of the cargo itself. This will determine whether or
not the cargo as a whole is within specification. Once this is determined, the Blend Master for the
discharge can weigh the available options for mixing.

 If the receiver is cooperative or has an immediate need for the product, the most cost
effective approach is to blend the product off the vessel, as described in Chapter 13 on
Discharge Blending.
 If discharge blending is not an option, it may be possible to mix the product using the vessels
equipment.

Most vessels can circulate product within a tank; but, are limited by the number of pumps they
have. This is frequently the best option but depends upon the construction of the vessel. The
vessel's pumps can move large volumes of oil [which is good]; but, tend to have the suction and
fill close to one another or on the same horizontal plane, [which is not good]. If a tank can be
emptied or almost emptied it may be possible to circulate product from one tank to another.

14.4 Variances between Tanks on Board

One potential consequence of the type of loading referred to is you can end up with tanks on
board the vessel that are not uniform from a specification point of view. Even though the overall
vessel composite is within the blend specification, some tanks are in specification and some are
out of specification. This is a significant problem. In this situation, it is necessary to blend product
from individual tanks together, not just mix the contents of the tanks. You may stand a better
chance of persuading the receiver to discharge blend into the shore tanks as the variation
between the different tanks is not likely to be large and the product should blend well.

Blending different tanks on board in a situation like this will definitely sort out the men from the
boys. This requires the vessel to pump from tank A into tank B, while at the same time pump tank
B into tank A; at the same speed. The potential for overflowing a tank is high and most vessels
will not acquiesce to this request, unless some cargo is lightered. Even so, it will be a difficult and

Page 65 of 83
lengthy operation. Variances between tanks on board do not usually happen unless there is an
unusual loading situation or a component is overloaded into one or two tanks.

14.5 Draft, Stability and Vessel Stress Limitations

There is little that the Blend Master or Blend Coordinator can do about draft limitations, except to
be aware of them and to anticipate potential problems that draft limitations might impose. The
Blend Master should be familiar with, or make it his or her business to be familiar with the tidal
situations at the port in question. Working the tides is very important. It may be necessary to do
as much as possible of the ship and shore inspection prior to arrival so that the vessel has the
maximum amount of time alongside with a favorable tide. When draft is a problem at loading, it
may be necessary to load the last one or two components at an anchorage, from a barge or
lightering vessel. This does not usually present a problem; although the OBQ on the barge must
be taken into consideration if it is significant. It is not advisable to have more than one component
on a lightering barge unless they are segregated and can be loaded separately.

Stability of the vessel is a significant factor on OBO [Ore/Bulk/Oil] multipurpose carriers. OBOs
have an advantage over conventional tankers as they have significantly fewer tanks which makes
the operation easier. A typical OBO has seven to nine very large cargo tanks with two small slop
tanks aft by the pump room. Some carry oil, as well as ballast, in their upper hopper wing tanks
but this adds to the stability problem rather than taking away from it. Because an OBO’s main
cargo tanks are so large, the loss of stability due to the free surface effect of a slack tank,
severely restricts the amount of slack tanks they can sail with. Most can sail with a maximum of
two or three slack tanks. They may move around a protected harbor with more than their
mandatory slack tank maximum, but they cannot put to sea in that condition. The shape of the
upper portion of the main cargo tanks is such that when full, the free surface effect is significantly
reduced.

This becomes
problematic for the
Blend Master [and the
Chief Officer]. The
very act of shipboard
blending requires all
tanks to be slack until
the final component is
Figure 26 – Free Surface Effect on an OBO Carrier loaded. If an OBO
loads a blend cargo at
more than one port, it
will be necessary to
load the components
from the first port into
the least amount of
full tanks. When it
arrives at the next
port, the cargo on
board will need to be transferred into the remaining cargo tanks prior to loading at this port. If
more than two ports are involved, this becomes a complicated operation. It requires many internal
transfers of blends within blends. Incomplete mixing is the biggest problem; and, if internal mixing
is available from the vessel, it is desirable. Limitations due to stress on the vessel must be
anticipated and planned for when possible. Problems due to bending moments and sheer forces
are likely to have a greater impact on OBO vessels. When proceeding from one port to another
with a part cargo on board, it may be necessary to ballast a cargo tank. This will add an additional
element and the Blend Master must monitor the possibility of contamination of the cargo with sea
water.

Page 66 of 83
Chapter 15 – Manual Sampling

The concept of sampling is to take a small portion of a product that is representative of a much
larger portion. Usually, the amount of sample taken is only a tiny fraction of a percentage of the
entire cargo. Laboratories typically have millions of dollars invested in equipment, some of which
can measure down to parts per billion. However, all of this investment in equipment is only as
good as the sample that is taken.

Samples of crude oil and petroleum products are taken to determine physical properties (such as
density, S&W and chemical components, such as sulfur and benzene), and compliance with
quality specifications such as those required in sales contracts. Samples may also be taken to
insure that the commodity is not contaminated during transfer. Sampling technique is important
because oil and chemical products may not be homogeneous and may contain impurities.
Sampling methods are designed to capture a sample of the material in a tank; or, flowing through
a pipeline that is representative of the whole. There are many conditions which may cause a
sample to be non-representative; such as stratification resulting from temperature, density or
other differences in product composition. If these differences are extreme it may be impossible to
obtain a representative sample by manual methods.

15.1 Equipment

A variety of equipment is used to perform manual

sampling, which include:

 Natural Fiber Cord or Chain

 Amber , Clear Glass, Plastic Bottles or Metal
Cans (Sometimes Lined)
 Sample collectors such as a Cage and Bottle,
Beakers, Bottom Samplers or Flow-through
Samplers
 Caps, Restricted Flow Caps, Split Corks, or
Stoppers

Sampling equipment must be properly maintained to

obtain good samples. Sampling equipment must first
of all be clean. Some analytical tests measure
impurities to ppm (parts per million) and even ppb
(parts per billion). Therefore, a very small amount of
residue in a sample container can result in an
erroneous test result. This is the principle drawback of Figure 27
the beaker sampler. It is difficult to insure that all Sample Bottle
residues from the previous sample are removed. In Catcher and
addition, sample cords must not contain residue from Beaker
incompatible material.

Certain sample containers should not be used with

some products. A product requiring extreme care is
jet fuel. ASTM has written one entire standard (ASTM
D 4306) to describe the proper sampling containers to
be used, depending upon the tests to be performed.
In general, epoxy-lined steel cans are preferred for all
testing except metals. The can must be rinsed three
times with the product being sampled before use.

Page 67 of 83
High density polyethylene containers are preferred for trace metal and lubricity analysis.
Aluminum containers must not be used. Also, equipment containing brass or copper must not
come into contact with the sample or the cargo.

Gasoline and naphtha samples are not taken in clear glass bottles or unlined cans. Cool the
sample containers before filling, by immersing them in the product. This also rinses the bottle with
the product to help insure contaminants are removed.

15.2 Types of Samples

Manual samples usually fit into one of two broad categories. A running or all-levels sample, which
is designed to obtain a single sample that is representative of the entire tank; or, spot samples,
which are samples that are taken at a specific place or spot in the tank (or flowing stream). Spot
samples by definition are not considered representative of a tank and are usually combined with
other spot samples, such as upper, middle and lower, so as to make a composite sample that is
considered representative of the tank’s contents.

15.2.1 All Levels and Running Samples

An all levels sample is taken by lowering a weighted, stoppered bottle to the draw-off level of the
tank, but above the level of any free water, pulling out the stopper with a sharp jerk, and raising
the bottle at a uniform speed so it emerges from the liquid about 75-80% full. For light products or
deep tanks a restricted opening may be necessary to avoid filling the bottle. No product should be
poured from the bottle before closing. If the bottle is more than 80% full the sample should be
discarded, and the sampling repeated.

 It can be problematical to take all level samples reliably as it is difficult to remove the stopper
at the bottom of the tank. This may cause the sampler to insert the cork too loosely, so that it
falls out of the bottle before reaching the bottom, an event that may remain undetected and
ruin the sample. Another problem can occur if the sampler fails to begin raising the bottle as
soon as the stopper is out. The sample will then contain too much product from the lower part
of the tank.

A running sample is obtained by lowering an unstoppered bottle or beaker to the draw-off level of
the tank, but above the level of any free water, and raising it again to the surface at a uniform
speed so the bottle is about 75-80% full when withdrawn from the liquid. The sample must be
removed from the liquid 75-80% full, or discarded. Flow restrictors are used if the bottle cannot be
moved through the liquid rapidly enough to achieve this result.

15.2.2 Spot Samples Figure 28 – Spot Sample Locations

Spot samples are those

samples taken from a specific
location in a tank or from a
pipeline. Spot samples are
taken by lowering a stoppered
bottle to a designated point in
the tank, removing the stopper
with a sharp jerk, and allowing
the bottle to fill completely
before withdrawing it. After
removing the bottle from the
tank, product is poured from
the bottle so it is 75-80% full
before closing. Spot samples

Page 68 of 83
can be taken from any point in the tank; however, API MPMS Chapter 8.1 defines certain specific
points where spot samples are routinely taken. These include:

TOP 15 cm (6 inches) below the upper surface of the liquid

UPPER Middle of the upper third of the liquid
MIDDLE Middle of the liquid
LOWER Middle of the lower third of the liquid
BOTTOM Bottom surface of the tank, sometimes called the Dead Bottom
3”, 6", etc. Sample taken a specified distance from the bottom of the tank, usually using a
bottom sampler (sometimes referred to as a Bacon-bomb) with an extension rod

15.2.3 Tap Samples

Some tanks have sample taps for sampling. When possible it is preferable to take spot samples
from the top of the tank, since the taps are usually fixed and therefore do not sample from the
middle of each third of the product. Tap samples must always be clearly labeled with the height of
the tap and the height of the liquid in the tank. Tap samples are not usually used for custody
transfer purposes, and are taken only when specifically requested.

15.2.4 Manual Pipeline Sampling

Manual pipeline sampling is not recommended for custody transfer because automatic pipeline
sampling is superior to the manual method. Most manual pipeline samples are spot samples that
are taken for quality control purposes, during a movement. When manual pipeline sampling
becomes necessary, the samples require labeling with the exact time the sample is taken, and a
pumping log of the transfer obtained to verify the pipeline activity at the time of sampling.

15.2.5 Vapor Pressure Samples

Samples that are to be tested for volatility (vapor pressure, Rvp) require special handling. The
most common type of sample is a running sample; however, spot samples can be taken.

 Under no circumstances is it allowable to decant a vapor pressure sample from one container
to another. The sample must be taken in the container that it will be transported to the
laboratory in. Vapor pressure samples are usually taken in glass bottles using a cage
sampler or a lead weight and a cord
 Spot samples will be full when pulled from the liquid. Quickly pour out approximately 20% of
the liquid, cap the bottle and invert it
 It is common practice to place inverted vapor pressure samples in ice, which is carried in a
cooler
 It is possible to use metal cans, provided that the sample is taken in the can and the cap
makes a vapor tight seal
 Plastic bottles are not permitted
 Once tested the sample cannot be retested so if retain samples are required additional
untested samples must be taken

15.3 Commodity Specific Guidelines

Up until recently, international standards refrained from giving specific sampling instructions for
the various commodities and in the main that is still the case; however, the current version of API
MPMS Chapter 8.1 / ASTM D 4057 - Standard Practice for Manual Sampling of Petroleum and
Petroleum Products – 4th Edition, October 2013 has certain specific requirements for both crude
oil and products. Many of these specific requirements are contained in Sections 9 and 10 of the
standard.

Page 69 of 83
In many ways Section 9 – Sampling Requirements, Considerations and Procedures is the “meat”
of the standard and includes such topics as:

 Requirements for Mixing and Compositing Samples

 Non-homogeneous (Stratified) Product (Section 9.2.8 is especially important)
 Marine Cargoes, Sampling Methods and Procedures (these are mostly the same as in
previous versions)
 Sampling Using Closed and Restricted Equipment
 Manual Pipeline Sampling
 Sampling Package Lots (Cans, Drums, Barrels, or Boxes)
 Solid and Semi-Solid Sampling
 Sampling into Fixed Volume Sample Cylinders
 Sampling into Single Floating Piston Cylinders (FPCs),

There is a huge amount of information in this section, so it is important to take the time to read it
and fully understand both what it contains and what it implies.

Section 10 – Special Instructions for Specific Products provides specific additional equipment,
container, sampling and sample handling guidance for the referenced products. The principal
section is that for crude oil and some examples (BUT NOT ALL) of the types of specific
requirements are:

 10.2.4 – Plastic containers are prohibited

 10.2.5.2 – S&W determination should be performed on a running or all-levels sample. DO
NOT perform this test on a composite of Upper, Middle or Lower spot samples.
 10.3.1 – if samples are to be tested for vapor pressure, H2S, or any other tests in which
retention of light ends is critical, subsamples for these test methods should be taken first
before any other sample handling procedures are performed.

There is also a section on refined Products. While this part covers many different types of
products and specific test methods relating to these products, Section 10.7 is very significant as it
essentially prohibits the use of metal cans; specifically, “10.7 Sample Containers—Use only
brown glass or wrapped clear glass bottles as containers, since it is difficult to make certain that
cans are free of contaminants, such as rust and soldering flux.”

Again, you must take the time to read it and fully understand both what this section contains and,
more importantly, what it implies.

Another important document in this area is the Energy Institute’s HM 92 - Model sampling
plans for crude oil and liquid hydrocarbon products. This document provides model sampling
plans for the manual sampling of liquid crude oil and petroleum products relating to shore tank,
road tanker, railcar, drum container sampling, and vessel loading or discharge operations.

These model plans will not be mandated unless referenced by contract or order confirmation and
are designed to complement existing sampling standards by providing guidance on the type,
location and volume of sample which should be drawn in order to provide sufficient sample for
retention and also to enable a typical series of laboratory tests to be performed for the
determination of product quality. For loading and discharge operations, the plans again provide
recommendations regarding sufficient samples for retention and also for analysis and quality
assurance relating to custody transfer. Due to the ease of reference it is anticipated that they will
become widely used and accepted.

Page 70 of 83
Chapter 16 – Automatic Sampling
16.1 Automatic Sampling Concept

In the main, automatic sampling is limited to discharges of crude oil. The reason is economic. An
automatic sampler, when functioning properly, is more likely to give you a representative sample
than manual methods, especially with a non-homogeneous mixture such as the free water that is
found in crude oil. As this free water is deducted from the outturn of the cargo, a receiver of crude
oil will want to use the most effective method possible to insure the amount of water in the sample
is representative of the amount of water in the cargo. There is no technical reason for limiting
automatic sampling to crude oil. Automatic sampling works as well in petroleum products as it
does in crude oil. They are often found as part of an automatic processing or blending system.

When inspecting a cargo movement that is using an automatic sampling system a number of
procedures must be followed to insure the unit is working properly. However, before we get into
specifics, we will review the main elements and features of these automatic systems. The review
will not go into the depth of understanding that a measurement technician needs to know, rather it
will cover the main elements that are common to all or most of these types of samplers and some
key critical aspects specific to these elements.

16.2 Key Elements

A typical automatic sampling system consists of the following elements.

 Stream conditioning, upstream

of the sampling location. The term
stream conditioning refers to ways
in which free or entrained
sediment and water is mixed to
form a homogeneous mixture.
This material must be uniformly
dispersed at the sample point and
is achieved by way of an adequate
flow velocity through the piping
system or by use of static or Figure 29 – Probe Location in Pipeline
mechanical [jet] mixers upstream
of the sample probe.
 A device to physically extract a grab from the flowing stream. This device is often referred to
as a probe. Some typical probe designs are shown in figure 18.
 A flow measurement device. If meters are available they will act as the flow measurement
device. However, there are alternate measurement devices such as a pitot tube or an
ultrasonic probe. If the flow
rate during the total delivery
varies by less than ±10%
from the average flow rate,
a representative sample
may be obtained by the
time proportional control of
grabs. This method is rarely
used for marine vessel
transfers. It is limited to
interstate and intra-state
Figure 30 – Probe Designs pipeline movements.

Page 71 of 83
 A means to control the total volume [number of grabs] of sample extracted. Grabs must be
extracted and collected in a flow proportional manner that provides a representative sample
of the entire parcel volume. The minimum number of barrels per grab can be calculated by
the formula, 0.079548 d2. Where: d = diameter of the pipeline in inches. This formula equates
to one grab for every 80 lineal feet of product moved through the pipe. The number of grabs
must also be such that enough sample is collected for the amount of analysis required.
 A sample receiver to collect and store the grabs. The sample must be maintained in the
sample receiver without altering the sample’s composition. Venting of vapors during filling
and storage must be minimized.

16.3 Types of Systems

There are two types of automatic sampling

systems used, both of which produce
representative samples if properly designed and
operated. One system locates the extracting
device directly in the main line, whereas the other
system locates the extracting device in a sample
loop, referred to as a “fast loop”. In a loop-type
system a probe is located in the main pipeline and
directs a portion of the fluid flow into the sample
loop. The controller which operates the sample
extractor receives a signal from the measurement
device. For sample loop installations a flow
indicator must also be installed in the sample loop
to insure that the flow through the loop is similar to
Figure 31 – Automatic Sampler the flow through the pipeline. Some loops have a
“low-flow” alarm.

Another important aspect of automatic sampler design is the location of the sample probe. While
it is acceptable to place a sampling probe in any part of a pipeline that provides adequate stream
conditioning, the location of choice is to place it in the downstream section of a vertical piping
loop. This gives the benefit of the
additional mixing from two 90° elbows.
Figure 17 illustrates a typical vertical loop
pipe configuration.

The previously mentioned systems are

fixed installations. There is another type of
automatic sampler system which is a
portable unit. These systems are used
mostly on marine vessel transfers. The
flow sensor and the grab probe are located
in a spool piece which fits onto the ship’s
manifold between it and the cargo hose. A
controller reads the flow and activates the Figure 32 – Automatic Sampler with Fast
probe accordingly which sends the Loop
appropriate number of grabs to a sealed
sample container.

16.4 Sampler Operations

One of the problems with automatic sampling systems is that they have many moving parts,
require regular maintenance and need to be closely monitored before, during and after the cargo
operation.

Page 72 of 83
 Before the operation it is important to insure the extractor probe, piping and sample receiver
are clean and contain no traces of the previous cargo. This necessitates witnessing the
purging of the system with a cleaning solvent. Also, it is necessary to insure that the valves
on the sample receiver are properly set and exit valves are sealed.
 During the operation it is important to insure that sampling was continuous [i.e. the sampler
did not fail or malfunction at any time]. This is accomplished by checking it at intervals during
the operation and reviewing plant operations logs.
 After the operation you should insure that all seals are intact. It is necessary to witness the
mixing of the sample, if done on sight, and the distribution of the sample.
Figure 33 – Photo of an Automatic Sampler by Jiskoot

16.5 Special Consideration for Marine Applications

Automatic samplers work best when the flow rate is consistent and fast. These conditions do not
usually exist on the start up or finish of a marine movement. It is recommended that a marine
vessel commence discharge using tanks with the least [preferably none] amount of free water in
them. Once the flow rate levels off, the tanks with free water in them can be pumped.

Page 73 of 83
Chapter 17 – Closed and Restricted Sampling Systems
The object of any type of sampling is to obtain a small portion of a much larger lot or shipment
that is representative of the lot or shipment. While it is generally recognized that automatic (flow
proportional in-line) sampling is superior to manual sampling in respect to obtaining a
representative sample, (and this is even referenced in API MPMS Chapters 8 and 17.2), such
equipment is not widely available, except for crude oil.

Therefore, at the majority of locations, samples are drawn manually from vessels and shore
tanks. However, as more and more environmental and safety regulations are introduced the
ability to draw samples through open tank access ports is being prevented, resulting in samples
being drawn through the use of closed or restricted systems.

17.1 Vapor Control Valves [VCV]

A vapor control valve is usually a quarter-turn valve and standpipe arrangement that allows entry
of gauging and sampling equipment into the tank while it [the tank] is under inert gas pressure.
This is achieved by the attachment of specialized equipment to the top of the standpipe, which
then allows the vapor control valve to be opened and equipment entry into the tank, without
undue loss of the tank’s inert gas pressure. The specialized equipment required for drawing
samples through these vapor locks can restrict the types of sample that may be drawn; and, may
not allow samples to be drawn in accordance with current international manual sampling
standards.

Development of equipment is ongoing and some already exists which though effective are subject
to operational constraints.

17.2 Limitations and Problems

Some of the problems encountered when working with this equipment are already having an
impact on operations, particularly where the cargo is non-homogeneous, of variable quality and/or
viscous, such as residual fuels. Some of the more commonly experienced problems are:

 Significant additional time is required for sampling, particularly of ships’ tanks, to allow
repeated fitting, operation and removal of the equipment.
 The diameter of most vapor lock systems dictates that samples can only be of a limited
volume, normally less than 400 ml.
 The sample containers are normally constructed of stainless steel so it is not possible to see
how full they are. This limits the ability to draw All Level or Running samples in accordance
with the standards.
 The design of the equipment often allows only zone or spot sampling. As most systems have
only one sample container this must be used repeatedly, leading to potential contamination
during the sampling process.
 The cleanliness of the vapor lock system is usually unknown and it is not practicable to clean
while in use.
 Though the majority of closed/restricted equipment originates from two manufacturers the
absence of an international standard regarding the type of vapor lock fitting has resulted in a
great number of different types and sizes in operation on vessels, sometimes preventing the
use of the inspection companies own equipment.
 Where vessel equipment has to be used it may not be clean or in a serviceable condition.
 Almost all closed and restricted sampling systems require the sample to be decanted in the
field, which prohibits the use of such samples for certain types of analysis, most notable of
which are vapor pressure, hydrogen sulfide and mercury.

Page 74 of 83
17.3 API MPMS Chapter 17.11 / EI HM 52

API MPMS Chapter 17.11 / EI HM 52 - Measurement and Sampling of Cargoes on Board

Tank Vessels Using Closed and Restricted Equipment

The first edition of this standard was published in May 2009 and was a new standard, when
issued, although much of the content is also in API Chapter MPMS 17.2. The purpose of this
standard was to update the information in Chapter 17.2 and to place measurement and sampling
by use of closed and restricted systems into its own standard.

17.3.1 Scope and Introduction

The scope of this document is quite short and states, “This document provides guidance on the
use, maintenance and calibration of restricted and closed measurement and sampling equipment
for ship operators and cargo inspectors. It also provides guidance to shipbuilders and ship-
owners on preferred size and positioning for gauging and sampling fixtures on vessels”.

The introduction to this standard is VERY IMPORTANT. Specifically, it states, “It should be
noted that when taking vessel samples it may not be possible to obtain representative
samples of non-homogenous cargos due to the inherent nature of the material and also
the sampling restrictions”

17.3.2 Other Sections

The other principal sections of this standard are:

Section 5 – Types of Equipment Section 6 – Standpipes

Section 7 – Procedures Section 8 – Vapor Control Valve
Section 9 – Gauge Points Section 10 – Vessel Calibration

17.3.3 Points to Note

The following are some of the highlights of this standard that are additional or different from the
information contained in Chapter 17.2.

Section 7.3.2.4 – Stratification and Nature of Cargo. This repeats and further expands the
caution from the introduction. Specifically, it goes on to state in paragraph 3, “In such cases, the
taking of additional samples (e.g. multiple spot or comparative shore tank samples, etc.)
and use of alternative sampling methods should be considered. If there is disparity in
related analysis results, it is recommended that all associated results obtained be
reviewed and considered, particularly when those results are to be used for custody
transfer purposes. This section is important because it gives us a level of authority to compare
additional results and samples if the vessel samples are out of specification.

Section 8 – Vapor Control Valves (Recommended Size and Type). This section starts out by
stating that the size and location of the vapor control valve is critical. It further goes on to state,
“The recommended size of vapor control valves is 100 mm (4 in.) nominal
diameter…..Vapor control valves smaller than 100 mm (4 in.) in diameter are suitable for
gauging but can severely limit the type of sampling equipment that can be used and,
ultimately, the quality of the sample obtained.

It further goes on to state that the use of 48mm (2 in.) vapor control valves often requires multiple
samples to be taken and that time can become a significant consideration in such instances.

Page 75 of 83
Chapter 18 – Precision and Accuracy of Test Methods

There is a tendency to use the terms “precision” and “accuracy” interchangeably; however, they
are not the same.

18.1 Accuracy

Accuracy is concerned with correctness. If a measurement process produces the correct results,
it is accurate, and the measured value is also accurate. Strictly speaking, only a counting process
is exact and hence accurate. A measured value is an estimate and may be “close enough” to the
correct value to be considered “errorless.” Therefore, accuracy is the degree of agreement of a
measured value with the true or expected value of the quantity of concern. In other words, if
shooting at a circular target, accuracy is how close you get to the center of the target or “bull’s-
eye”.

18.2 Precision

Precision, on the other hand, is the degree of mutual agreement characteristic of independent
measurements as the result of repeated application of the process under specified conditions. In
other words, precision is how often you can hit the center of the target or the bull’s-eye.

Repetitive measurements in the same measurement system such as a laboratory test method will
differ from one another and the scatter of the values, when plotted is a measure of the precision;
specifically, the narrower the scatter, the higher or greater the precision. When discussing
precision in relation to analytical testing, this is defined in terms of repeatability and
reproducibility.

18.2.1 Repeatability

Repeatability is defined as the precision of a test method expressed in terms of the agreement
attainable between measurements made by a single operator using the same apparatus and
techniques. It is further defined as the standard deviation of results obtained by the same
operator using the same instrument in successive measurements on the same sample.
Repeatability conditions are the conditions under which test results are obtained. That is the
same test method in the same laboratory by the same operator with the same equipment in
the shortest practical period of time using test units or test specimens taken at random from a
single quantity of material that is as nearly homogeneous as possible.

 Repeatability limit (r)--the value below which the absolute difference between two individual
test results obtained under repeatability conditions may be expected to occur with a
probability of approximately 0.95 (95 %) or 19 times out of 20.

NOTE I The "same operator, same equipment” requirement means that for a particular step in the
measurement process the same combination of operator and equipment is used for every test
result. Thus one operator may prepare the test specimens, a second measures the dimensions
and a third measures the mass in a test method for determining density.
NOTE 2 “in the shortest practical period of time” means the test results are obtained in a time period not
less than in normal testing and not so long as to permit significant change in test material.

18.2.2 Reproducibility

 Reproducibility is defined as the precision of a test method expressed in terms of agreement

expected between measurements made in different laboratories using similar apparatus and
the same procedure. It is further defined as the standard deviation of results obtained by

Page 76 of 83
 different operators using the same or different instruments in different laboratories on the
same sample. Reproducibility conditions are the conditions under which test results are
obtained. That is, with the same test method on identical material in different laboratories.
 Reproducibility limit--The value below which the absolute difference between two test
results obtained under reproducibility conditions may be expected to occur with a probability
of approximately 0.95 (95 %) or 19 times out of 20.

NOTE I Identical material means either the same test units or test specimens are tested by all the
laboratories or test specimens are taken at random from a single quantity of material that is as
nearly homogeneous as possible.

18.3 Bias

In a stable measurement process, a large number of individual values will tend to converge
toward a limiting mean, which should be the true value; however, if it is not then the process is
said to be biased.

Bias is a systematic error inherent in a method or caused by some artifact or idiosyncrasy of the
measurement system. Temperature effects and extraction inefficiencies are examples of the first
kind. Blanks, contamination, mechanical losses, and calibration errors are examples of the latter
kinds. Bias may be both positive and negative, and several kinds can exist concurrently so that
net bias is all that can be evaluated, except under special conditions.

18.4 Margins

A margin is that amount of the precision limit [usually reproducibility] a supplier or manufacturer is
willing to give away to insure that specifications are maintained. Trade is more global than in the
past. As a result, markets are more open but the technical margins in commercial transactions
are much tighter. How different companies approach this depends upon their trading philosophies
and what they deem critical to their business reputation. For those who believe that having a
reputation for always meeting a sales specification then the giveaway will be close to the limits of
precision. For those Companies that believe the lowest price always wins out in the long run, then
their giveaway will be as small a percentage of the precision limits that they can financially justify.
There is nothing wrong with either approach. A Company picks one or the other, or something in
between, depending upon how they run their business. If the precision of a test method is not
taken into consideration and blends are regularly planned to the limit of their specifications, you
should [in theory] only expect to have your finished product in specification 50% of the time, over
a period of time. If the full limit of reproducibility is used as the test or blend margin then [in
theory] the finished product should be in specification 19 times out 20.

In the current business environment, the scales are tipped more toward the latter philosophy than
the former. This keeps the pressure on to find ways of reducing the limits of precision and
therefore tightening the market. This is not only limited to repeatability and reproducibility of test
methods. How good is the sampling? How reliable is the quantitative element? Sampling and bulk
quantity determination methods do not have stated precision levels such as repeatability and
reproducibility. However, they do tend to have certain acknowledged levels of acceptability, even
if these levels are unofficial and variable.

Page 77 of 83
Annex 1 – Hydrotreating
Environmental regulation continually requires finished petroleum products to have fewer and
fewer impurities, especially sulfur. Crude oil contains a wide range of organic sulfur compounds,
including thiols, thiophenes, organic sulfides and disulfides, and many others. These sulfur
compounds are not removed as part of the normal refining processes and therefore find their way
into the finished product and must be removed. One of the key processes for removing these
sulfur compounds is called hydrotreating or more correctly, hydrodesulfurization. In this process
hydrogen is introduced into a mixture to increase the overall hydrogen count. The petroleum then
undergoes what is called hydrogenolysis, which is when the carbon and sulfur bond contained
within the petroleum is split, and a hydrogen atom is linked to both the carbon and the sulfur
atoms. Through this process, the sulfur can be removed from the petroleum until it reaches an
acceptable level.

The industrial hydrodesulfurization processes include facilities for the capture and removal of the
resulting hydrogen sulfide (H2S) gas. In petroleum refineries, the hydrogen sulfide gas is then
subsequently converted into byproduct elemental sulfur or sulfuric acid (H2SO4). In fact, the vast
majority of the sulfur produced worldwide is byproduct sulfur from refineries and other
hydrocarbon processing plants.

A1.1 Process Description

In an industrial hydrodesulfurization unit, such as in a refinery, the hydrodesulfurization reaction

takes place in a fixed-bed reactor at elevated temperatures ranging from 300°C to 400°C and
elevated pressures ranging from 30 to 130 atmospheres of absolute pressure, typically in the
presence of a catalyst consisting of an alumina base impregnated with cobalt and molybdenum
(usually called a CoMo catalyst). Occasionally, a combination of nickel and molybdenum (called
NiMo) is used, in addition to the CoMo catalyst, for specific difficult-to-treat feed stocks, such as
those containing a high level of chemically bound nitrogen.

Figure 34 - Schematic diagram of a typical Hydrodesulfurization unit in a petroleum refinery

The liquid feed (at the bottom left in the diagram) is pumped up to the required elevated pressure
and is joined by a stream of hydrogen-rich recycle gas. The resulting liquid-gas mixture is

Page 78 of 83
preheated by flowing through a heat exchanger. The preheated feed then flows through a fired
heater where the feed mixture is totally vaporized and heated to the required elevated
temperature before entering the reactor and flowing through a fixed-bed of catalyst where the
hydrodesulfurization reaction takes place.

The hot reaction products are partially cooled by flowing through the heat exchanger where the
reactor feed was preheated and then flows through a water-cooled heat exchanger before it flows
through the pressure controller (PC) and undergoes a pressure reduction down to about 3 to 5
atmospheres. The resulting mixture of liquid and gas enters the gas separator vessel at about
35°C and 3 to 5 atmospheres of absolute pressure.

Most of the hydrogen-rich gas from the gas separator vessel is recycle gas, which is routed
through an amine contactor for removal of the reaction product H2S that it contains. The H2S-free
hydrogen-rich gas is then recycled back for reuse in the reactor section. Any excess gas from the
gas separator vessel joins the sour gas from the stripping of the reaction product liquid.

The liquid from the gas separator vessel is routed through a reboiled stripper distillation tower.
The bottoms product from the stripper is the final desulfurized liquid product from
hydrodesulfurization unit. The hydrogen sulfide removed and recovered by the amine gas treating
unit is subsequently converted to elemental sulfur in a Claus process unit or to sulfuric acid in a
wet sulfuric acid process.

A1.2 Other Hydrotreating Processes

Hydrotreating processes differ depending upon the feedstock available and catalysts used. For
example, hydrotreating can be used to improve the burning characteristics of distillates such as
kerosene. Lube-oil hydrotreating uses catalytic treatment of the oil with hydrogen to improve
product quality. The objectives in mild lube hydrotreating include saturation of olefins and
improvements in color, odor, and acid nature of the oil.

Hydrotreating also can be employed to improve the quality of pyrolysis gasoline (pygas), a by-
product from the manufacture of ethylene. Traditionally, the outlet for pygas has been motor
gasoline blending, a suitable route in view of its high octane number. However, only small
portions can be blended untreated owing to the unacceptable odor, color, and gum-forming
tendencies of this material. The quality of pygas, which is high in diolefin content, can be
satisfactorily improved by hydrotreating, by the conversion of diolefins into mono-olefins which
provides an acceptable product for motor gas blending.

A1.3 Summary

Hydrodesulfurization or hydrotreating is a catalytic chemical process widely used to remove sulfur

from natural gas and from refined petroleum products such as gasoline, jet fuel, kerosene, diesel
fuel, and fuel oils. The purpose of removing the sulfur is to reduce the sulfur dioxide (SO2)
emissions that result from using those fuels in automotive vehicles, aircraft, railroad locomotives,
ships, gas or oil burning power plants, residential and industrial furnaces, and other forms of fuel
combustion.

Another important reason for removing sulfur from the naphtha streams within a petroleum
refinery is that sulfur, even in extremely low concentrations, poisons the noble metal catalysts
(platinum and rhenium) in the catalytic reforming units that are subsequently used to upgrade the
octane rating of the naphtha streams.

While the most common use of hydrotreating is the removal of sulfur, it can also be used to
remove other impurities such as, nitrogen, oxy-compounds, chloro-compounds, aromatics, waxes
and metals.

Page 79 of 83
Annex 2 - Straight Run Fuel Oil Indicators
A2.1 Overview

Straight run fuel oil (also known as “atmospheric residue from crude oil distillation” or long
residue) is usually considered more valuable than “cracked” fuel oil because it can be used as a
feedstock for further processing. In particular these fuels are subject to vacuum distillation and
conversion by means of catalytic cracking, visbreaking, etc. The determination of whether or not a
material is “straight run” can be problematical because no international standard is available for
the analytical testing of fuels on straight run criteria. However, the following criteria are generally
accepted by the industry for the determination of straight run material.

TEST CRITERIA TEST METHOD SPECIFICATION

Distillation ASTM D 1160 smooth curve
Bromine nr. on fraction IBP - 360oC ASTM D 1159 max 5 g/100 g
Flocculation test SMS 305 passes (25)
Toluene equivalent Exxon 79-004 max 25 - 30
or
Xylene equivalent BP method 230 max 25
Shell hot filtration test SMS 2696
existent max 0.15
accelerated max 0.15
potential (accelerated - potential) max 0.03
Asphaltenes IP 143 to report (usually low)
Density ASTM D 1298 to report (usually low)
Flashpoint ASTM D 93 to report (usually low)

A2.2 Interpretation

Based upon the above test results’ falling within the stated specifications we would normally
state, that in our opinion such a sample meets the usual accepted criteria for straight run product.
However, given that fuel oil is a mixture, it is only the aggregate characteristics that have been
determined and it is possible that the sample could contain some non straight run components,
within the overall matrix.” i.e., a blend of straight run material with some cracked material might
still fall within the criteria.

A2.3 Limitations

While the above referenced tests have proved effective in analyzing for straight run fuels, it is not
possible to categorically state, from analysis that a material is straight run and we must always
qualify our reports or certificates of analysis to this effect. An example of the suggested wording
is:
"The above tests were carried out on the ... (hand blend, composite, other - state) ... sample,
representing the cargo of .... Based upon the reported results falling within the stated
specifications, in our opinion the sample meets the usually accepted criteria by the industry for
“straight run” fuel oil. However, given that fuel oil is a mixture, it is only the aggregate
characteristics that have been determined and it is possible that the sample could contain some
non straight run components, within the overall matrix."

The following information explains the significance of the various quoted tests.

A2.3.1 Distillation - ASTM D 1160

Page 80 of 83
If the product does not contain cracked material, it can be distilled up to a boiler bottom
temperature of 400°C, a smooth curve and no observed cracking at the stipulated/indicated test
vacuum. In contrary, a product that contains cracked material will show cracking during the
distillation and a boiler bottom temperature of 400°C cannot be achieved, the distillation has to be
discontinued and no smooth curve will be obtained. D 1160 “smooth curve and no observed
cracking at the stipulated/indicated test vacuum”.

A2.3.2 Bromine Number - ASTM D 1159

This test is performed on the distillate fraction IBP - 360°C, which is obtained by test method
ASTM D 1160. The result of this test is a measure for olefins present in this fraction. Cracked
material contains olefins of the higher Carbon numbers. Additionally, if all the indicators point to a
‘straight run fuel’, except the bromine number result is found to be above 5 g/100 g, try re-running
the distillation at a lower pressure than the advised 10 mm Hg and test the IBP-360 fraction for
bromine number again. We have experienced that when distilling at 10 mm Hg with bottom
temperatures between 300 - 400°C, cracking may occur and the result of the bromine number will
be affected by this cracking
.
A2.3.3 Flocculation test, Toluene or Xylene Equivalent

These tests are a measure for the n-Heptane insolubles. The principle of these different tests is
more or less similar. The main difference is the treatment of sample and type of filtration paper
used. Samples are dissolved in blends of different ratios of the respective aromatic component
and n-Heptane. Asphaltenes, higher olefins (gums), Naphthalenes, etc. are soluble in the
Aromatic component, however insoluble in n-Heptane. After heating of the different solutions a
drop is placed on a filter paper and the stain (precipitate) is compared with standards.

A2.3.4 Hot filtration test

A measure for the n-Heptane insolubles, compared to the former tests which give a rating, this
filtration test gives an absolute figure for the n-Heptane insolubles plus the normal sediments.
The results obtained from the accelerated test will give an idea of the behavior of the product
during storage (ageing of product). The principle of this test is that the product reacts with Cetane
under controlled conditions, if cracked material is present the product will become unstable and
the content of n-Heptane insolubles will increase. See Annex 3 for additional information.

A2.3.5 P Value

This test gives same as the tests mentioned under 3 and 4 an idea of the stability of the product.
The more cracked material, the less stable the product becomes. The product reacts with
different concentrations of Cetane under controlled conditions. A drop is transferred to a
microscope slide and examined under a microscope for insolubles.

A2.3.6 - Asphaltenes, Density and Flashpoint

If no cracked material is present the results will usually be low.

Other companies sometimes use other methods. One often used is the Peptization Value.

Page 81 of 83
Annex 3 – Sediment by Hot Filtration
Most of us, when we think of sediment in crude oil or fuel oil think of inorganic sediment such as
sand or other similar grit-like material that has no hydrocarbon content of any kind. While this is
correct as far as it goes, there is another type of sediment that is found in residual fuel oils that
we can refer to as organic sediment and this type of sediment is different in just about every way
from inorganic sediment and it is extremely important not to confuse the two because the
significance of these tests (i.e., what they are being tested for) are very different.

A3.1 Sediment by Extraction

Inorganic sediment is determined by the Sediment by Extraction method ASTM D 473. This is a
very common test for crude oils and fuel oils and is usually used in conjunction with water by
distillation or water by Karl Fischer to determine sediment and water [S&W] content. This test
method covers the determination of sediment in crude oils and fuel oils by extraction with toluene.
A representative oil sample, contained in a refractory thimble, is extracted with hot toluene until
the residue reaches constant mass. The mass of residue, calculated as a percentage, is reported
as sediment by extraction.

In other words, the only material that will remain in the thimble to be measured as sediment is
material that is not soluble in toluene.

A3.2 Sediment by Hot Filtration

This method is more properly called Determination of Total Sediment in Residual Fuels,
although it is frequently referred to by its more descriptive name of Sediment by Hot Filtration. A
weighed quantity (10 g) of the oil sample is filtered through the prescribed apparatus at 100°C.
After solvent washing and drying the total sediment on the filter medium is weighed.

This method is run primarily to give an indication of the stability of the fuel oil, which typically
manifests itself by the precipitation of asphaltene resulting from a breakdown of the protective
coating around the asphaltene causing it to combine into large particles that precipitate (fall) out
and settle to the bottom of the storage vessel. This precipitation of asphaltenes from a residual
fuel oil during storage and handling can cause severe difficulties which, in extreme cases, can
render the fuel unfit for use. Once out of solution, it is extremely difficult to repeptize the
asphaltenes back into their original state.

Note: Asphaltene is soluble in toluene so these two methods have no correlation to one another.
While it is possible to filter inorganic sediment by hot filtration it is not possible to filter
asphaltene based sediment using sediment by extraction.

A3.3 Ageing the Sample

While we may be asked to run the Total Sediment (Hot Filtration) on a sample as taken, we are
more likely to be asked to run this test after the sample has been “aged” and the method
describes two ways in which the sample can be aged.

A. Thermal aging, referred to as Total Sediment – Potential [TSP]. Thermal ageing is a well
established technique for determining whether sediment will precipitate from residual fuel oils
during storage and handling. Aging is achieved by placing the sample in the air well of an oil
bath, maintained at 100°C ±0.5°C, for 24 hours ±15 min.
B. Chemical aging, referred to as Total Sediment – Accelerated [TSA]. Chemical ageing
looks at the balance between the required aromaticity of the asphaltenes and the available

Page 82 of 83
 aromaticity of the oil phase, and tests whether a specified amount of a normal alkane disturbs
this balance to the extent that asphaltene precipitation occurs. Aging is achieved by the
controlled addition of 2.5 ml ± 0.02 ml of hexadecane to the sample and placing it in the well
of the oil bath at 100 ◦C ± 0.5°C for 60 min ± 2 min.

A3.4 Test Methods

ASTM, IP and ISO all have test methods for The Determination of Total Sediment in Residual
Fuels and all three include instructions for aging the samples should Total Sediment – Potential
or Total Sediment – Accelerated be required. The three test methods in question are:

 ASTM D 4870
 IP 390
 ISO 10307

IP 390 and ISO 10307 considered technically identical; whereas ASTM D 4870 is considered
technically equivalent as there are some differences in the calculation procedures.

A3.5 Summary

Not all sediment is created equally. Sediment by Extraction and the Determination of Total
Sediment in Residual Fuels (whether aged or not) are very different test methods that look for
very different aspects within the material. It is critically important not to confuse one with the other
or to suggest or recommend one as an alternative to the other.

Page 83 of 83