Engineering design for temperature, flow and pH control

in the brewery process

Hector Rodriguez Rangel
Jose Francisco Uribe Vazquez

May 31, 2019

Abstract

One of the most important economic activities in Mexico is the beer production and its export. In
the brewery process the equipments, instrumentation and a good control strategy for each variable like
temperature, flow and pH determines the flavour and quality of the product. In this paper is going to be
explained the engineering design for the variable measurement, the tuning for each control loop but also
the reasons for making those choices and getting a better product.

1 Introduction to germinate. In this phase the enzymes (amylase)

are formed which are needed for the starch separa-
One of the most important economic activities in tion. The germinating process is then interrupted
our country has been the beer production and its at the optimal stage by the drying (kilning). The
export. Mexico is the fourth into the global produc- green malt is dried carefully at 80o celsius and tastes
tion and the first one in export. In fact, the beer is slightly sweet when dried. The malt sugar later
considered one of the principal economic engines in serves to feed the yeast cultures, which then form
Mexico and it is something to be proud about. 1 the alcohol. 1
The beer is more than just water, hops, malt and
yeast. In the beer making process various ingredi-
ents are mixed, processed and sometimes the struc-
ture of the raw materials is altered. The brewing
process is made up of ten production steps from the
fresh barley to the finished beer. When the prod-
uct is bottled is ready for its consume. Some of the
problems inside the brewery process are the mea-
surement and control of temperature, flow and pH
in some sections of the general process.
Inside the beer production, the pH is a huge chal-
lenge because this variable determines the bitterness Figure 1: Brewing process flow diagram
of the product. Also the temperature play a role int
the way that the pH can change. A pH less than 4.2 2. Milling: The finished malt is milled, similar
can result in an acid product and that is why the to the making of flour, so that it dissolves better in
importance of giving a solution for better control water. Malt mills produce various grades of crushed
and measurement of the variables. malt: husks, groats, meal, semolina and powder
(from the coarsest to the finest).
2 The brewing process 3. Mashing: The milled malt is mixed with
water in the mash tun – mashing-in. The starch in
1. Malting: Firstly the fresh barley gets soaked the grist dissolves and sugar, protein and tannin are
in water and put in the so-called germinating boxes released. The so-called malt extract is produced by

1
this mashing process. is used for guessing another variable. In the selec-
4. Lautering: The mash is filtered in the lauter tion of the sensor for this paper, the main criteria is
tun as the husks sink and the wort is separated from the type of industry: the food industry. This indus-
the solid substances (spent grains). The wort is then try requires of instruments that satisfy the national
further used in the brewing process, whereas the standards as the world ones. 2
spent grains are generally used as cattle fodder. In that way, the selected sensor is a Resistance
5. Wort boiling: In the brewing pan or wort Temperature Detector (RTD); that its electrical re-
kettle the wort is boiled and the hops are added. sistance is variable according to the temperature.
The taste of the beer depends on the type and The model is the iTHERM TrustSens TM371.
amount of hops used: the more hops the bitterer Specially designed for use in hygienic and aseptic
the beer. As the water evaporates the wort is con- applications in the food & beverage and life sciences
centrated to the original wort, the malt enzymes industries is one of them features.
are deactivated and tannin and protein elements are The field of application is given by some parame-
separated out to form the so-called trub. ters like:
6. Wort clarification: Here the wort is fed into a) Measuring range: –40 to 160 o C (–40 to 320
the whirlpool and starts to rotate. Hop particles o F)
that have not dissolved and protein – the so-called b) Pressure range up to 50 bar (725 psi)
trub – form a cone in the middle of the container, c) Protection class IP67/68 or IP69K
and the clear wort can be tapped off at the side. d) Communication: Analog output 4 to 20 mA,
Then the clear wort is cooled to a temperature of HART protocol.
between 10 and 20o Celsius in the wort cooler. Some other characteristic or application are:
7. Fermentation: This takes place in a fermen- self-calibrating, metric style, compact temperature
tation tank where special brewing yeast is added. probe, hygienic design and hygienic process connec-
The yeast turns the malt sugar into alcohol and tion.
carbon dioxide. As soon as the malt sugar has fer-
mented the yeast sinks and is collected. Depending
on the type of yeast and the wort preparation used
a top- or bottom-fermented beer is produced.
8. Storage: Young beer is stored anything from
three weeks to three months in a storage tank at
between 1 and 2o celsius. A secondary fermentation
takes place; remaining yeast particles and protein-
tannin sink to the bottom. The beer becomes clear
and acquires its characteristic colour.
9. Filtration: At this last stage any substances
still remaining after fermentation and secondary fer-
mentation, such as yeast particles, hop resin and
Figure 2: The iTHERM TrustSens TM371 Self-
protein are extracted and the beer gets its final clear
calibrating temperature sensor
colour.
10. Filling: Bottled beer and cans are filled in
the filling station. The beer is filled using counter
pressure to prevent any carbon dioxide from escap- 3.1 Temperature control
ing. Barrels are filled directly in the brewery.
In real life, most of the industrial processes are non-
linear and can be expressed as an approximation
3 Temperature such a first order plus time delay (FOPTD) or sec-
ond order plus time delay (SOPTD). The form of
The temperature measurement is one of the most FOPTD can be expressed as follows: 3
important parts of a process in the industry. Almost
all physical phenomena is affected by it and that is Kp e−T Ds
one of the reasons for the which one the temperature G(s) = (1)
Ts + 1

2
 The process transfer function is represented as:

5e−1s
G(s) = (7)
90s2 + 33s + 1

Which is in the form of SOPTD.

3.1.1 PID Controller

The block diagram of a closed loop control based on

PID setup of heat exchanger system is shown in the
figure below which has a classical PID controller.

Figure 3: Temeperature process

SOPTD can be expressed as follows:

Kp e−T Ds
G(s) = (2)
(T1 s + 1)(T2 s + 1)

Where Kp is the gain of the process, TD is the Figure 4: Block diagram of a closed loop control
time delay and T is the time constant of FOPTD based on PID
system and T1 and T2 are the time constant of
SOPTD. A classical PID controller can be expressed as
Determining the transfer function model for heat
exchanger we have the exchanger response to steam 1 4u
G(s) = Kp + Ki + (8)
o
flow gain is 50 C/kg · sec − 1, whose time con- s + K d 4t
stant is 30 seconds, exchanger response to variation
of process fluid flow gain 1 o C/kg ·sec−1, exchanger Where Kp is the proportional gain, Ki is the in-
response to variation process temperature gain 3 tegral gain and Kd is the derivative gain.
o C/o C, capacity of control valve 1.6 kg/sec, time

constant for control valve is 3 sec and time constant

for sensor is 10 sec.
Thus, the transfer function model for the heat
exchanger system is:

50e−1s
G(s) = (3)
30s + 1 Figure 5: SIMULINK diagram
Transfer function model for valve is:
0.13
G(s) = (4)
3s + 1 3.1.2 Simulation Results
Transfer function model for the sensor is: For the classical PID controller we have that the
0.16 time response is acceptable, we have 40 seconds time
G(s) = (5) settling without overshooting which is good for the
10s + 1
process because we do not want to have a change
Transfer function model for disturbance is: in the temperature. The proportional gain used is
1 Kp =2, integral gain Ki =0.0265 and derivative gain
G(s) = (6) Kd =9, these gains produce this response:
10s + 1

3
 4.1 PID Controller
In this simulation 4 , we have two main constraints:
- The height of the liquid in tank cannot be more
than 55 cm.
- Flow range can be in the range of 20 to 200
cm3/sec.
Thus, taking in consideration these constraints we
decided to use a flow of 97.2 cm3 /sec whit a height
Figure 6: Simulation results for temperature control of tank of 49.1 cm.
The transfer function model is written as
1.9
G(s) = (15)
334s + 1
Retaking the disturbance model of temperature
control, we have
1
G(s) = (16)
10s + 1
Figure 7: schematic diagram of a single tank system
The simulation structure is shown below
4 Flow
The schematic diagram of a single tank system is
shown in the figure below.
Mass-Balance equation is given as:
Qi − Q = AdH/dt (9)
Figure 8: SIMULINK diagram
Where Qi is the inlet flow rate (cm3 /sec), Q is
the outlet flow rate (cm3 /sec), A is the area of the
tank (cm2 ) and H is the height of the liquid of the
4.1.1 Simulation Results
tank (cm). Valve equation can be written as
√ As we can see, there is an appropriate response of
Q=K H (10)
the process. There is no overshooting which can let
Apply Taylor’s theorem to linearize this equation improve the response of the level in the tank and
about steady state height Hs the control of the process. Although the simulation
dH √ H − Hs looks great, this cannot represent the response in
A = Qi − K( H + 0.5 √ ) (11)
dt H real life, mainly because many issues are not con-
At Steady State sidered.
dHs √
A = Qi s − K H (12)
dt
On subtracting equation
√
dH H
A = qi − 0.5K √ (13)
dt Hs
Where qi = Qi − Qi s, h = H − Hs
Taking Laplace transform we have
H(s) R Figure 9: Simulation results flow control
=√ (14)
Qi (s) Ts + 1
√ √ The gains to produce this response are: pro-
2 Hs 2A Hs
Where R = and T = portional gain Kp =50, integral gain Ki =0.15 and
K K
derivative gain Kd =0.2.g

4
5 pH
For more information, search in ResearchGate the
paper: Simulation of the Soil pH Control System
Using Fuzzy Logic Method. In this paper, the re-
searches show how can be implemented a fuzzy logic
control in a pH process 5 .

6 CONCLUSIONS
The engineering design is the most important part
of the industry. In the brewery process there are a
lot of parts in the process that, in fact, there are
not controlled. This problem cause troubles in the
quality product because the pH, temperature and
light are the main factors that determine the beer
flavour.
In the other hand, the design and tunning of
the controller is based on empirical knowledge, this
means the process works but in a non-ideal way
mainly for the complex of the process, they don not
have the appropriate knowledge or the poor engi-
neering.
In conclusion, we have seen that if the process is
well known there is no restriction for doing all things
better: tunning, control design and instrumentation
could be improved.

References
[1] https://www.braeuamberg.at/en/brewing-
process/

[2] Controller Design for Temperature Control of

Heat Exchanger System: SImulation Studies

[3] DESIGN OF INTERNAL MODEL CONTROL

AND INTERNAL MODEL FEED-FORWARD
CONTROL FOR LIQUID LEVEL SYSTEM.
Rohit Singh.

[4] Simulation of the Spoil pH Control System Us-

ing Fuzzy Logic Method. M.A. Abu, E.M.M.
Nasir, and C.R. Bala.