Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2, pp.

96108 (2007)

CFD INTEGRATED DESIGN OF SCREW COMPRESSORS

Ahmed Kovacevic*, Nikola Stosic, Elvedin Mujic and Ian K. Smith
Centre for Positive Displacement Compressor Technology, City University, London, SEMS,
Northampton Square, EC1V 0HB, London, United Kingdom
*E-Mail: a.kovacevic@city.ac.uk (Corresponding Author)
ABSTRACT: Positive displacement screw machines are used in variety of applications such as compressors,
expanders, blowers, vacuum pumps, liquid and multiphase pumps. To improve their appearance, efficiency and
robustness they are designed with the aid of analytical tools, based often on one-dimensional flow models solved by
numerical methods that are confirmed by experiment. Continuing demand for further improvements has led to the need
for improved assessment of fluid flow losses in the inlet and outlet openings and how these are affected by the shape of
the ports, the deformation of machine components due to the effects of pressure and temperature gradients and their
effect on performance, the behaviour of multiphase flows and many other effects. These require more advanced
analytical procedures, based on three dimensional numerical flow analysis and fluid-structure interaction. The way to
estimate these phenomena is to use CFD analysis and to integrate the results with three dimensional CAD systems. As
computers become cheaper and faster and advances are made in numerical methods, such techniques are becoming
available for everyday use by design engineers. This paper describes how CFD is merged with other design software by
means of an integral management system to obtain interactive control of the entire design process of screw compressors.
The methods described are of considerable scope and can be applied, not only to screw compressors but also to any
other type of twin rotor rotary machines with parallel axes, such as gear pumps, multiphase pumps, vacuum pumps and
roots blowers.
Keywords:

Screw Compressor, CFD, Computational Continuum Mechanics, CAD, Design Integration

They may operate as oil flooded, with other fluids

injected during the compression or expansion
process, or without any form of internal lubrication.
Their designs may vary, depending on the choice of
rotor proportions and profile, the number of lobes in
each rotor and in the shape and size of their ports
and openings. To obtain the best combination of
these for a given application requires a set of well
defined criteria governed by an optimisation
procedure.
Such a procedure should include not only, the
generation and optimisation of the rotor profile and
clearance distribution, but also the housing ports,
bearings, seals and the lubrication system if full
advantage is to be derived from their potential and
maximum performance gains are to be achieved.

1. INTRODUCTION
Twin screw machines of the rotary positive
displacement type are capable of efficient operation
with variety of fluids over a wide range of operating
pressures and flow rates. Compared to other types
of positive displacement machine, their design is
simple, with the two rotors comprising the only
moving parts. Thus, they can rotate at relatively
high speeds and are therefore both reliable and
compact. Consequently, a large percentage of all
positive
displacement
compressors
now
manufactured and in use are of this type.
Screw machines can be used both as compressors
and expanders of working fluids which may be
gases, dry vapours or multi-phase mixtures with
phase changes taking place within the machine.

Received: 6 Dec. 2006; Revised: 31 Jan. 2007; Accepted:15 Feb. 2007

96

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

Fig. 1

Typical examples of oil injected and dry screw compressors.

Early designs of screw compressors were based on

the assumption of an ideal gas in a leak proof
working chamber going through a compression
process which could reasonably be approximated in
terms of pressure-volume changes by the choice of
a suitable value of the exponent n in the
relationship pVn = Constant. The advent of digital
computing made it possible to model the
compression process more accurately and, as time
passed, ever more detailed models of the internal
flow processes were developed based on the
assumption of one-dimensional flow.

equations which describe the instantaneous flow of

fluid, work and heat transfer through the system.
These could be solved numerically to estimate
pressure-volume changes through the suction,
compression and delivery stages and hence
determine the nett torque, power input, fluid flow
and isentropic and volumetric efficiencies in a
compressor. In addition, the assumption that any oil
passed through the machine as a uniformly
distributed spray with an attributed mean droplet
diameter, enabled account to be taken of oil
injection. Such models were then refined by
comparing predictions, derived from them, with
experimentally derived performance data.
A typical result of such modelling was the
generation of a suite of computer programs called
SCORPATH for the performance prediction and
optimisation of screw compressors. More details of
this are given in the recently published book on
Screw compressor modelling and performance
analysis by Stosic et al. (2005). Similar work was
also carried out by other authors such as Fleming
and Tang (1998) or Sauls (1998). Despite their
speed and relatively accurate results, these works
lacked the possibility for analysis of some
important phenomena that affect compressor
performance and reliability.

2.1

2.2

This requires a complete understanding of the fluid

flow inside the machine and at the inlet and outlet
openings, the influence of the shape of compressor
ports on the performance, the effects of high
pressure and temperature gradients on the
deformation of compressor elements and the reverse
effects of these on the performance. Three
dimensional CAD and CFD tools highly integrated
with traditional computational tools for design and
manufacture are thus being developed to evaluate
these effects.
2. NUMERICAL ANALYSIS OF SCREW
COMPRESSOR PROCESSES

One dimensional models

The assumption of dimensionless non-steady bulk

fluid flow and steady one dimensional leakage flow
through the working chamber, together with
suitable flow coefficients through the passages, and
an equation of state for the working fluid, made it
possible to develop a set of non-linear differential

Three dimensional models

Screw compressor performance may be estimated

more precisely by the use of Three-dimensional
Computational Fluid Dynamics (CFD) or
Computational Continuum Mechanics (CCM).
Although the number of papers published in this
area has increased recently, only a few deal with the
97

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

coupling, given by Peri (1990) and later discussed

and improved by Demirdi et al. (1992) and
Demirdi and Muzaferija (1995) allowed
application of the method simultaneously to fluid
flow and solid body stress analysis.
The usual practice for the analysis of solid body
deformation or fluid-structure interaction is to
couple Finite Volume (FV) code with Finite
Element (FE) solvers using a specially designed
interface. Most CFD and FE vendors use that
procedure to take advantages of both FV and FE.
Although well established, this procedure in many
situations may not be entirely suitable. One
important example is the simultaneous conjugate
heat transfer in both a solid and a fluid. This is the
case of the interaction of fluid flow and solid
deformation in a screw machine. However, this may
be estimated by use of Computational Continuum
Mechanics (CCM). In this case, the small
deformation of solid parts is caused by the large
pressure and temperature gradients, which are
generated as a consequence of the fluid flow.
Although these deformations are relatively small,
they are of the similar order of magnitude as the
compressor clearances, and may thus significantly
change the leakage flow within the machine.
There are many CFD and FE codes currently
available. However, the authors most commonly
employ COMET, developed by Star CD, mainly
because that was the first commercial code on the
market that enabled simultaneous fluid flow and
solid structure calculations. Today many CFD
codes allow similar calculations and are integrated
with CAD systems. Most of these today use a
modern polyhedral topology of computational
control volumes to map the geometry, as discussed
in Peric (2004). Polyhedral cells allow both easier
grid generation and more accurate solution in a
mesh with a lower number of cells. This improves
the computational speed.

application of computational fluid dynamics to

screw compressors. These began with the work of
Stosic et al. in 1996. That paper gives basic
principles of three dimensional numerical
modelling applied to positive displacement screw
machines. However, it was not successful due to the
relatively poor grid generation applied. Kovaevi,
Stoi and Smith published a number of papers
between 1999 and 2005. These papers introduced
3D numerical analysis to the screw compressor
world. In later years, the authors published a series
of papers related to both, grid generation in screw
compressors and 3D numerical performance
estimation (Kovacevic et al. 2003 and 2005). These
include fluid solid interaction in screw machines
(Kovacevic et al. 2004). A recent book on CFD in
screw machines by Kovacevic et al. (2006) give a
comprehensive overview of the methods and tools
used.
A number of commercial CFD software packages
are currently available which may be able to cope
with the complex flow through screw machines and
be integrated with CAD systems. However
developed these codes are, there are still limitations
in their use for some applications. For the analysis
of screw machines, a moving, stretching and sliding
mesh has to be produced to map the working
chamber. The grid generators contained in these
packages are still not capable of coping with these
requirements.
2.3

Computational continuum mechanics

Computational fluid dynamics (CFD) covers a

broad area, which attracted the interest of many
investigators at the beginning of the computer era.
It is based on the numerical simulation of the
conservation laws of mass, momentum and energy,
derived for a given quantity of matter or control
mass. Three main groups of methods have been
developed through the years as described by
Ferziger and Peri (1995). These are the finite
difference, finite element and finite volume
methods. The finite volume method has been used
for this application and is briefly described.
Of particular interest for the analysis of screw
compressors is the calculation of unsteady flow
with moving boundaries. The space conservation
law introduced earlier and further discussed by
Demirdi and Peri (1988 and 1990) allowed
calculation in complex geometries and various grid
topologies. Detailed analysis on pressure-velocity

2.4

Grid generation

Grid generation problems are mainly connected

with computational fluid dynamics but the
applicability of the concepts used in making
numerical grids is not in any way limited to this
area.
Applying the principles of analytical grid
generation through transfinite interpolation with
adaptive meshing, the authors have developed an
automatic numerical mapping method for an
98

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

CAD systems also relate to the selection and/or

calculation of mechanical properties of compressor
elements. Some are selected from standard
databases such as bearings, bolts and screws, keys
and key-ways, dowels etc., while others are
designed and calculated for deflections and
strength. The majority of CAD software allows
such calculation through integrated CAE
components. However, these generally require
substantial action by the user in order to set up and
relate loads and restraints, calculated outside the
CAD system, with geometry created within the
CAD system, as reported in Kovacevic (2004).

arbitrary screw compressor geometry, as explained

in Kovacevic (2005). This was later used for the
analysis of the processes in screw compressors. On
that basis, the authors have developed an interface
program called SCORG (Screw COmpressor Rotor
Geometry Grid generator), which also enables a
grid, generated by the program, to be directly
transferred to a commercial CFD or CCM code
through its own pre-processor.
The interface employs a novel procedure to
discretise rotor profiles and to adapt boundary
points for each particular application so that the
numerical mesh generated for the rotor domains is
directly transferable to a Computational Fluid
Dynamics or Computational Continuum Mechanics
code. This is required to overcome problems
associated with moving, stretching and sliding rotor
domains and to allow robust calculations in
domains with significantly different ranges of
geometry features (Kovacevic, 2006). Stationary
parts of the compressor domain can be generated
either by use of a commercial grid generator or in
SCORG and in any case they are directly integrated
in the CFD or CCM code.

4. INTEGRATION OF SCREW
COMPRESSOR DESIGN SOFTWARE
A prerequisite for success in the highly competitive
market of screw machines is ability to design,
analyse and produce machines quickly. These
activities need to be automated to be used by design
engineers in industry.
However, a great obstacle to obtaining fast screw
compressor design by industrial engineers was
difficulty in simultaneous implementing both
geometric and non-geometric information quickly.
Thus, what was required was a concurrent design
process to perform an iterative procedure in the
design phases so that the exchange of information
between software elements could be performed
several times. This had to be done through design
integration package that increases interoperability.
The DISCO (Design Integration for Screw
Compressors) software was developed to integrate
tools for the design and manufacture of screw
machine components in a user friendly environment
suitable for industrial use. It manages both
geometric and non geometric information transfer
between the software components used.
These are related to the heat and fluid flow,
optimisation parameters, boundary conditions and
operational parameters that are organised so that the
function, behaviour and structure integration is
embedded in the code. The interface basically
consists of five modules named SCORPATH,
SCORG, SCOCAD, SCOCFD and SCONOISE.
The organisational scheme of DISCO is given in
Fig. 2. More information can be found in Kovacevic
(2005).

3. APPLICATION OF CAD SYSTEMS IN

THE DESIGN OF SCREW COMPRESSORS
Both preliminary and detailed mechanical design in
traditional CAD systems is limited to the
representation of geometric data and other types of
information
relating
to
geometry.
The
implementation of the function, behaviour and
structure of a specific machine cannot not be
accounted for in them.
The engineering design community has been
developing new classes of tools to support
knowledge-based design, product data management
(PDM) and concurrent engineering. When
contrasted with traditional CAD tools, these new
systems are improved but are still mainly database
related. Also, although these systems can represent
some kinds of non-geometric information such as
design process details, bills of materials, etc, the
representation of the design object itself is still
generally limited to geometric considerations.
Evidence of this may be found by reviewing the
projects described on the IAI web page. Due to
these limitations, further interoperability is required
to enable full interaction between any CAD system
and performance estimation software.

99

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

their efficiency being high. The task was given to

find the best design of a family of two-stage air
compressors with shaft input powers between 22
and 312 kW at 815 bars discharge pressure. The
design of an efficient second stage for one of these
units is described here.

Fig. 3 Rotor profiles optimised for a two-stage oil

flooded screw compressor.

Fig. 2

Altogether 19 variables were used for this

multivariable optimisation, 9 for each compressor
stage plus the interstage pressure. Both 1st and 2nd
stage rotors are presented in Fig. 3, where it can be
seen that there is a distinctive difference between
them. The first stage rotors are slender to achieve
high displacement, while the second stage rotors are
stronger to survive the high pressure loads.
The computing time required to calculate the
performance from the input data using this program
is measured in tens of seconds while the whole
optimisation process required less than one hour of
computer time on an ordinary PC.
Diagrams generated directly from DISCO,
presented in Fig. 4, show the compressor
performance and rotor bearing loads, obtained from
the thermodynamic model as the basis for the
calculation and selection of the compressor
elements.
The CAD model of the compressor is shown in
Fig. 5. This 3D solid model of the assembled
compressor is generated directly in MDT 7 by use
of a parameter database generated by SCOCAD.
Since a fully parametric approach is applied during
the entire integration with the CAD system the
changes that had to be made are implemented
directly and easily. Manufacturing drawings are
generated from the 3D model. A part of the report
on the calculation and selection of machine
elements performed based on the forces calculated
in SCORPATH, is presented in the same figure.

Organisational scheme of DISCO.

Procedures explained here have been extensively

used both for research purposes and in everyday
industrial design of screw machines. The examples
of applied academic research are shown in Chapter
5 where it is clearly outlined that the integrated
approach allowed fast modification of design
parameters in order to analyse processes and
phenomena within these machines efficiently. This
approach allowed significant reduction in time for
optimisation and mechanical design of screw
machines which allows industrial user to perform
entire optimisation loop within hours rather then
days and get basic CAD models ready for
manufacturing within days rather then months.
5. EXAMPLES OF USE
5.1

Two stage oil injected compressor

Oil injected screw compressors have been used for

air compression for a long time. Today, more than
ever, their successful marketing is dependent on

100

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

Fig. 4

Fig. 5

Performance calculation results.

CAD Model of the compressor and the part of the report on housing thickness calculation.

and discharge ports is displayed in the left hand side

of the same figure. The rotor domains are generated
by SCORG, while the ports and other stationary
mesh parts are generated by the commercial CADCFD interface Star-Works of CD-Adapco. The
integration was performed fully automatically.

As the next step, the numerical grid for CCM

calculations is generated by SCORG. The fluid and
solid parts of the machine each contain about
700,000 numerical cells. A cross section through
the fluid part of the mesh generated by SCORG is
shown on the right hand side of Fig. 6. The
complete numerical mesh which includes suction

Fig. 6

Numerical mesh for CCM calculation.

101

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

Fig. 7

Distribution of pressure and velocity in the cross sectional view of the machine.

female rotors respectively, and their centre distance

is 90 mm. The rotor length to diameter ratio is 1.65.
The numerical mesh for the test case in this study
comprises 513,617 cells of which 162,283 describe
the solid part of the rotors, 189,144 other cells map
the fluid parts between the rotors while the rest
specify the suction and discharge ports and oil
openings. A cross section through the mesh for the
rotors and their fluid paths is presented in Fig. 9.
The calculations were carried out on an office PC.
A converged solution was achieved with 120 time
steps in approximately 30 hours of computing time.
In the case of the oil injected air compressor, the
results of the 3D calculations were compared with
measurements obtained from an experimental
compressor of identical dimensions. The pressure
fluctuations within the machine were measured with
piezoresitive transducers, positioned in the male
rotor side of the housing to cover as much as
possible of the whole process. Measured results
were obtained for suction conditions of 1 bar abs
pressure and 20oC temperature with discharge
pressures of 6, 7, 8 and 9 bar. Good agreement was
obtained both for the integral parameters and the
instantaneous values, shown in the right diagram of
Fig. 10. The complete report on the experimental
verification of this numerical procedure is given in
Kovacevic et al. (2003). The new experiments
which involve advanced measuring techniques such
as Laser Doppler Velocimetry (LDV) are being
undertaken in order to compare instantaneous
velocities across the computational details. Detailed
report of the achieved comparison is given in Nouri
et al. (2006).

With this number of cells, the numerical solution of

fluid flow and rotor deformation was obtained on a
PC by use of the COMET CCM solver in less than
30 hours. Here, it is important to emphasise the fact
that the 3D numerical simulation of fluid flow and
structure analysis is used as a check rather than as
an optimisation tool because of the time required to
obtain results for all the required 240 time steps. It
is therefore performed only once at the end of the
design process, before the prototyping of the
machine.
The pressure and velocity distribution field obtained
from the CCM calculation are presented on the left
of Fig. 7. The deformation of the male rotor is
presented on the right of the same figure. The rotor
distortions, which do not exceed 6 micrometers in
this case, are magnified 20,000 times in order to be
visible. The pressure distribution on the rotor
surface is also given in the same figure. It is clear
that the highest deformation is in the critical area of
the machine but since it is not substantial, the
leakage flow is practically unchanged and the
performance is hardly affected by it.
5.2

Fluid solid interaction

Fluid and solid interaction is presented here for

three common applications of screw compressors,
each with the same geometry, as shown in Fig. 8.
These are: an oil-injected air compressor of
moderate pressure ratio, a dry air compressor, of
low pressure ratio, and a high pressure oil flooded
compressor. In all cases, the rotors are of N type
with a 5/6 lobe configuration. The rotor outer
diameters are 128 and 101 mm for the male and

102

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

Fig. 8

Fig. 9

Cross section of rotors (left) and compressor (right).

Numerical mesh for rotors (left) and their fluid parts (right).

Fig. 10 Rotor deformations (left), P- diagram - comparison of CFD and measurements(right).

results visible, the numerical grid deformations are

enlarged 200 times.
In the oil free air compressor, due to the lack of
cooling oil, the air temperature rise is such that at 3
bar discharge pressure, the exit temperature has an

The rotor deformations caused mainly by the

pressure rise in the working chamber are presented
in the left diagram of Fig. 10. These make the
clearance gap between rotors larger. To make the

103

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

average value of 180oC. The deformation in this

case is presented in left diagram of Fig. 11. The
fluid temperature in the immediate vicinity of the
solid boundary changes rapidly, as shown in the
right diagram of the same figure. However, the
temperature of the rotor pair is lower due to the
continuous averaging oscillations of pressure and
temperature in the surrounding fluid. This is shown
in the right diagram of Fig. 11, where the
temperature distribution is given in cross section for
both the fluid flow and the rotor body. The
deformation, presented in the figure, is increased
1,500 times in order to make it visible. It has a
maximum value of 50m.
The high pressure oil injected application was taken
as a CO2 refrigeration compressor with suction
conditions of 30 bar and 0oC and discharge
conditions of 90 bar and 40oC. In this case, the large
pressure difference was the main cause of the rotor
deflection with the highest deformation in excess of

15m, as shown in Fig. 12. The deformation pattern

of the rotors is similar to the low pressure case but
with slight enlargement at discharge.
The change of clearances caused by rotor
deformation affects the compressor integral
parameters, as shown in the right diagram of
Fig. 12. The reduced rotor clearance due to the
temperature dilatation increases both the
compressor flow and the power input. However, the
effect on the flow is greater than on the power and
hence the specific power decreases, or more
conventionally, the efficiency is increased.
In other cases, rotor deflections caused by the
pressure make the clearances larger. For a moderate
pressure rise, the effect on delivery and power
consumption is negligible. In the CO2 refrigeration
application, the rotors deform more and the
decrease in delivery and rise in specific power
becomes more pronounced.

Fig. 11 Rotor displacement vectors and temperature distribution for an oil free compressor.

Fig. 12 Deformations of a high pressure oil injected compressor.

104

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

5.3

were meshed by ANSYS CFX and ICEM tools into

a tetrahedral mesh. These two domains were
connected through transient interfaces (GGI).
The pressure increase in the trapped volume is
shown in the left of Fig. 14. The iso-surface of the
oil vapour volume fraction in the right figure
represents the cavitation bubbles. Their movement
in the near wall region causes the erosion induced
by cavitation. The author concluded that cavitation
is in phase with the period of the tooth meshing.
Damage due to cavitation occurs likewise in the two
axial gaps in terms of gap erosion as well as on the
rotors on the trailing side of the flank.

Cavitation in a gear pump

The assembly, the functionality and the numerical

mesh of a gear pump are presented in Fig. 13.
During operation, damage due to cavitation and
erosion occurs at the rotor shafts and in the gaps.
The work presented here is the property of CFX
Berlin. Steinman (2006) outlined that the main
challenges in this computation were the relatively
complex geometry, the relative moving and
deforming grids and transient interfaces and
cavitation.
The hexahedral numerical mesh of moving parts
was generated by SCORG while the stationary parts

Fig. 13 Geometry and numerical mesh of the gear pump.

Fig. 14 Pressure distribution in the gear pump (left) and the occurrence of cavitation (right)

105

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

oscillations as a function of the shape of the port

and the cross sectional area of the connecting
flange. These predictions agree reasonably well
with the measured results. However, this model
does not take into account the shape of the
discharge chamber which may play important role
in the whole process.
Further steps are therefore undertaken to overcome
this problem. Firstly, pressure fluctuations in the
discharge port are calculated in the commercial
CFD code. Hexahedral meshes are used to map
moving rotor domains while a polyhedral mesh is
generated from CAD-CFD interfaces for stationary
domains. The left hand of Fig. 16 shows the
pressure oscillations obtained on dry compressor
with a full 3D model of the entire compressor. This
set of calculations required 240 time steps each of
which took no more then 10 minutes of a normal
PC time to obtain a converged solution. The results
obtained by this model agree well with
measurements (Mujic et al. 2005).

5.4 Prediction of noise generation sources in

screw machines
Noise identification and analysis in screw
compressors have been an important issue in the
majority of operations that involve such kinds of
rotating machines for a long time. Here some
details of ongoing research and development of
methods to reduce noise emission from screw
compressors is presented.
Pressure fluctuations in the discharge port affect not
only aero acoustics in that domain but also a
mechanically generated noise due to rotor rattling.
Preliminary measurements of gas pulsations in an
industrial compressor with different sizes of
discharge port show the change in nature of the
pressure oscillations in the port in Fig. 15. It is
believed that adequate porting can decrease the
level of noise and increase the performance of the
machine. A one dimensional thermodynamic
method was derived which can estimate pressure

Fig. 15 Measurements of pressure oscillations in the discharge port of an industrial compressor.

Fig. 16 CFD calculation of blended 3D and 1D models (left) and results from the genuine 3D model.
106

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

Despite the accurate results, the speed of the

calculation is still limiting factor for the application
of this model in industry. It seems that 1D
procedure which calculates pressure in the
compression chamber when integrated with 3D
CFD model in the novel commercial CFD code
StarCCM+ might be a suitable option. Since the 1D
chamber model runs as the user subroutine in CFD
software it allows easy mapping of boundary
conditions and fast overall calculations. The 3D
CAD model is produced in DISCO for SolidWorks
together with the macro for importing geometry and
running the case. A polyhedral numerical mesh is in
this case automatically generated for StarCCM+ by
use of a commercial CAD-CFD interface developed
by the CFD code vendor. The results of the
numerical methods which blend 3D CFD analysis
with 1D calculation of the compressor
thermodynamics is presented on the left of Fig. 16.
The model could be generated in matters of minutes
and it allows calculation of the entire compressor
cycle in only a couple of hours on an ordinary PC.

ACKNOWLEDGEMENTS
The authors wish to thank Dr A. Steinman of CFX
Berlin for kind permission to publish part of his
research on and analysis of a gear pump.
REFERENCES
1. Demirdi I and Muzaferija S (1995).
Numerical Method for Coupled Fluid Flow,
Heat Transfer and Stress Analysis Using
Unstructured Moving Mesh with Cells of
Arbitrary Topology. Comp. Methods Appl.
Mech. Eng. 125:235255.
2. Ferziger JH and Peri M (1996). Computational
Methods for Fluid Dynamics. Springer, Berlin.
3. Kovacevic A, Stosic N and Smith IK (2002).
Solid-Fluid Interaction in Screw Compressors.
XVI International Compressor Engineering
Conference at Purdue, July 2002.
4. Kovaevi A, Stoi N and Smith IK (2003).
Three Dimensional Numerical Analysis of
Screw Compressor Performance. Journal of
Computational Methods in Sciences and
Engineering 3(2):259284.
5. Kovaevi A, Stoi N and Smith IK (2004). A
Numerical Study of Fluid-Solid Interaction in
Screw Compressors. International Journal of
Computer
Applications
in
Technology
21(4):48158.
6. Kovaevi A (2005). Boundary Adaptation in
Grid Generation for CFD Analysis of Screw
Compressors. International Journal for
Numerical Methods in Engineering 63: xxx.
7. Kovaevi A, Stoi N and Smith IK (2006).
Screw Compressors Three Dimensional
Computational Fluid Dynamics and Solid Fluid
Interaction. Springer Berlin Heidelberg New
York.
8. Nouri JM, Guerrato D, Stoi N and Kovaevi
A
(2006).
Cycle
Resolved
Velocity
Measurements Within a Screw Compressor.
18th International Compressor Engineering
Conference at Purdue, Lafayette, Indiana, USA.

6. CONCLUSIONS
The screw compressor is a mature product. Despite
its now established role in industry, efforts continue
to make advances in every aspect of its design,
manufacture and mode of operation. Although
improvements so gained are most likely to be
evolutionary, there is still scope for revolutionary
methods or procedures to achieve a better product.
The new generation of methods and tools for the
research and development of these machines give
potential for the screw compressor to continue in
use for a long time. Moreover, these tools can be
used for other types of machine with a similar
configuration to screw compressors and expanders
such as gear pumps, multiphase pumps and vacuum
pumps.
With the advent of computing technology,
improvements in numerical methods and the
interdisciplinary merging of analytical methods, an
engineer is today in a position to design positive
displacement machines faster and more accurately.
Design Integration for Screw Compressors
(DISCO) is tool which helps by the parametric
merger of available CAD, CFD and other design
tools.

107

Engineering Applications of Computational Fluid Mechanics Vol. 1, No. 2 (2007)

9. Mujic E, Kovaevi A, Stoi N and Smith IK

(2005). Analysis and Measurement of
Discharge Port Influence upon Screw
Compressor
Noise.
9th
International
Research/Expert Conference Trends in the
Development of Machinery and Associated
Technology, TMT 2005, Antalya, September
2005.
10. Peric M (2004). Flow simulation using control
volumes of arbitrary polyhedral shape,
ERCOFTAC Bulletin 62 (September).
11. Steinman A (2006). Numerical Simulation of
Fluid Flow in Screw Machines with Moving
Mesh
Techniques
in
ANSYS
CFX.
Schraubenmaschinentagung 2006, Dortmund
VDI Berichte.
12. Stoi N, Smith IK and Zagorac S (1996). CFD
Studies of Flow in Screw and Scroll
Compressors, XIII Int. Conf on Compressor
Engineering at Purdue.
13. Stosic N, Smith IK and Kovacevic A (2002).
Optimization of Screw Compressor Design.
XVI International Compressor Engineering
Conference at Purdue, July 2002.
14. Stoi N, Smith IK and Kovaevi A. (2005).
Screw Compressors Mathematical Modelling
and Performance Calculation. Springer Berlin
Heidelberg New York.

108