MS5100 RESEARCH METHODOLOGY

ASSIGNMENT 2
Research Question

Firas Zuhair Altair (13618050)

M Ghullam Reza A (23621304)
Abdurrahman Birry (23621305)

Lecturer: Dr. Arie Wibowo, S.Si., M.Si.

DEPARTMENT OF AEROSPACE ENGINEERING

FACULTY OF MECHANICAL AND AEROSPACE ENGINEERING
BANDUNG INSTITUTE OF TECHNOLOGY
2022
Paper 1:
Title : Stall/surge dynamics of a multi-stage air compressor in response to a load
transient of a hybrid solid oxide fuel cell-gas turbine system
Author : Mohammad Ali Azizi, Jacob Brouwer
Journal : Journal of Power Sources, Volume 365, 15 October 2017
Year : 2017
Impact factor : (10 citation indexes)

Resume of the paper:

Compressor stall/surge is one the major challenges that comes from the usage of hybrid solid
oxide fuel cell-gas turbine (SOFC-GT). This paper investigated a solution that can reduce the
significant effect of this problem to the hybrid SOFC-GT system performance during sudden power
demand perturbations. A 1.7 MW multi-stage compressor model, which is similar to the other
industrial compressor available on the market (Kawasaki gas turbine GPB17), was developed in order
to analyse compressor stall/surge for this 4 MW hybrid SOFC-GT, applied in locomotive system.
CFD analysis was used since it could provide the flow distribution in the compressor impellers during
dynamic hybrid system operation and also simulate instabilities near the stall/surge line. At the
boundary condition set up, a transient pressure that resulted from a load perturbation was applied in
the CFD model of the compressor. The simulation result showed that the multi-stage radial design of
the compressor increased the ability of the compressor to maintain air flow rate during transient step-
load changes. It means that the compressor keeps providing positive rate of air flow during the mild
stall/surge event due to these changes. In the other words, the multi-stage compressor was able to
handle pressure perturbation without going to deep stall/surge, making it one of the best suitable
options for hybrid SOFC-GT application.

What is Done in This Research:

• Exploring about dynamic operation of a hybrid solid oxide fuel cell gas turbine system,
• Evaluating stall/surge dynamic phenomenon due to pressure perturbation,
• Analysing the multi-stage radial compressor in response to the pressure dynamics,
• Using CFD analysis for simulating a multi-stage compressor,
• Maintaining positive air flow without entering deep surge conditions.

Future Works:
Evaluation of the usage of deadtime block in the open loop back up system that can even prevent
the start of surge/stall phenomenon.
Paper 2:
Title : Design of a Centrifugal Compressor for Micro Gas Turbine: Investigation
of Scaling and Tip Clearance Effects
Author : Dario Barsi, Alberto Bottino, Andrea Perrone, Luca Ratto, Pietro Zunino
Journal : Open Journal of Fluid Dynamics, Volume 9, Issue 1 (March 2019)
Year : 2019
Impact factor : 0.96

Resume of the paper:

This paper investigated the performance comparison between a centrifugal compressor for micro gas
turbine (MGT) and the scaled one by using CFD investigations. The scaled one means the MGT
underwent a modification from 100 kW electrical output to 250 kW with the boundary condition of
the maintained compressor pressure ratio and the same geometry. The investigation was divided into
two parts. The first part focused on comparing the performance of the original and the scaled MGT
while the second one deeply studied about the tip gap effect in the scaled MGT configuration. The
result showed that scaling up the machine leads to increase the efficiency, up to 1.4% from the original
design, although still keeping the technological limit in impeller manufacture. The second part of the
investigation was variating the gap height in five values from the ideal case (zero gap), with 0.2 mm
increment and maximum value of 1 mm. This further investigation showed that the tip gap height
variation highly affects the machine efficiency, up to 10% decrease from the ideal case.

What is Done in This Research:

• Scaling up a centrifugal compressor for micro gas turbine (MGT) to increase the electrical
output from 100 kW to 250 kW,
• Maintaining the compressor pressure ratio and the geometry while scaling up the machine,
• Studying about the tip gap height variation effect on the machine performance,
• Generating the computational mesh using Autogrid and doing numerical simulation through
CFD (Computational Fluid Dynamic) analysis.
Future Works:
• Usage of simple corelations of thermodynamic, aerodynamic and geometrical characteristics
for determining the global effect of tip gap variation,
• The analysis of blade angle and blade span variation effect in response to the compressor
efficiency.
Paper 3:
Title : Assessment of Turbulence Model Predictions for A Centrifugal Compressor
Simulation
Author : Lee Gibson, Lee Galloway, Sung in Kim, Stephen Spence
Journal : Journal of the Global Power and Propulsion Society
Year : 2017
Impact factor : (20 citation indexes)

Resume of the paper:

The research conducted in this paper is to compare the simulation results of a centrifugal
compressor with a diffuser vane using a different turbulence model to the experimental results. Global
parameters such as pressure ratio and efficiency can be predicted accurately if the right turbulence
model is chosen. This is due to the precise turbulence model, providing numerically stable with short
simulation and accurate flow details for centrifugal compressors. Computational Fluid Dynamics
(CFD) was chosen as the numerical method used to simulate the steady-state centrifugal compressor.
To compare the results of three eddy-viscosity models and two Reynolds stress type models. The
turbulence models investigated in this study were:
1. Spalart-Allmaras (SA)
2. Shear Stress Transport (SST)
3. A modification to the SST model denoted the SST-curvature correction (SST-CC)
4. Reynolds stress model of Speziale, Sarkar and Gatski (RSM-SSG)
5. The turbulence frequency formulated Reynolds stress model (RSM-ω).
From this comparison of turbulence models, it was found that the five turbulence models gave
similar results to the experiment (below 2% difference) for global parameters with three operating
conditions. However, this turbulence model has its own advantages and disadvantages in predicting
flow in a centrifugal compressor. For the SA model in near surge condition, it gives a poor flow
prediction when compared to the experiment. Meanwhile, SST-CC provides better accuracy for
curved rotating flow on the impeller and RSM-ω predicts wake and separation flow in the diffuser
well. In addition, the SST model provides stable, robust, and time-efficient results for obtaining global
and local flow feature parameters.

What is Done in This Research:

In this study, the authors have compared the effect of the turbulence model for the numerical
simulation of ANSYS CFX to the experimental results. To compare the effect of this turbulence
model, it will be reviewed in terms of global parameters, flow details at the exit of the impeller and
diffuser, blade loss factor, and in terms of computational time. The important results of this
comparison can be seen as follows.
1. The SST-CC turbulence model gives the most accurate results for the pressure ratio
parameters and isentropic efficiency at all test points where the difference is 0.28% and 1.05%
compared to the experimental results. However, for each model it has given an accurate result
for the second test point (M) about 1%.
2. For the distribution of circumferential velocity and vorticity this will predict the secondary
flow in the centrifugal compressor. From the circumferential velocity distribution, it is found
that SST-CC gives the most similar results to the experimental results and SST-CC gives
accurate results also from the vorticity distribution so that this model can provide the highest
accuracy on the pressure ratio variable.
3. In predicting the occurrence of blockage and loss, the turbulence model is compared using the
cp and Yp parameters from the experiment in the diffuser section. From the simulation results,
the SST turbulence model gives the most accurate results where the difference in cp and Yp
values is 0 and 0.3, respectively, with experimental results.
4. The SST turbulence model is the most efficient model in terms of time and number of
iterations for the three test points where this model requires 5.27 hours wall clock time and
674 iterations to reach the convergence point.

Future Works:
In this paper, the simulation of a centrifugal compressor is limited to using steady-state
conditions for each operating condition, which in predicting flow under conditions near surge is less
accurate than using full-stage transient conditions, so the authors need to perform full-stage transient
simulations in surge conditions. In addition, the authors also need to conduct further research using
different software, such as ANSYS Fluent to analyse centrifugal compressors so that later a
comparison of software can be obtained that gives results close to the experiment and in a shorter
time.
Paper 4:
Title : Aerodynamic analysis of the small-scaled centrifugal compressor for
micro-turbojet engine applications
Author : A.A. Sebelev, A.S. Tikhonov, V.A. Aleksensky, A.A. Shengals, O.I. Klyavin
Journal : Journal of Physic Conference Series, IOPSCIENCE
Year : 2017
Impact factor : (2 citation indexes)

Resume of the paper:

The research that was conducted in this paper is an aerodynamic analysis of a small-scaled
transonic centrifugal compressor for micro turbojet engine (TJE) applications using a CFD model.
The experimental data collected for the gaged JetCat P200-RX micro-TJE were used to validate the
CFD model. The loss balance diagram was obtained by estimating the loss coefficients of the
impeller, varied diffuser, and deswirler for four design points corresponding to 70%, 80%, 90%, and
100% rotation speeds. The flow angle spanwise variation demonstrated an intensive flow separation
zone at the top 25% of the vaned diffuser span due to the pressure shock appearance. It was shown
that the main source of the losses in the investigated compressor is the deswirler due to non-optimal
flow tuning conditions. The diffuser loss coefficient was estimated as 0.18 at the compressor full
load.

What is Done in This Research:

• Build a 3D model of the compressor and representation of its “fluid” model,
• Numerical approach used to represent the compressor “fluid” model,
• Using The Ansys CFX commercial solver to perform 3D steady-state Reynolds Averaged
Navier Stokes (RANS) simulation,
• Validation of the numerical model conducted based on experimental data collected for the
gaged JetCat P200RX micro-TJE in terms of the outlet temperature and compressor total-to-
total pressure ratio,
• Performed analysis of total outlet temperature analysis showing good agreement between
numerical and experimental data,
• Performed analysis of the loss balance diagram showing that total-to-total impeller efficiency
decreases from 0.864 at 70% rotation speed to 0.818 at 100% rotation speed,
• Performed analysis of the flow angle span variation at the vaned diffuser inlet and at the
deswirler inlet,
 • Shown that the main source of losses for the compressors of interest is the diffuser which
combines the vaned diffuser and deswirler.

Future Works:
In this paper, the authors have obtained the loss coefficient of the impeller, diffuser, and
deswirler for four design points corresponding to the percentage of rational speeds. in addition, the
authors also need to vary the mass flow and analyse its effect on the loss coefficients of the
impeller, diffuser, and deswirler.
Paper 5:
Title : Optimization of a Centrifugal Compressor Using the Design of Experiment
Technique
Author : Mohammad Mojaddam, Keith R. Pullen
Journal : Applied Sciences MDPI
Year : 2019
Impact factor : (24 citation indexes)

Resume of the paper:

The research carried out in this paper is to optimize the impeller using the design of experiment
technique where the technique uses the response surface method as a surrogate model and the Box-
Behnken methodology as a method to identify the most optimal point. The parameters considered in
this study are meridional geometry, rotor blade angle distribution, and start location of the main blades
and splitters. The shape of the diffuser geometry will be entered as meridional geometry and will be
optimized using the following four stages
1. Stage 1: Parameter optimization of meridional geometry of centrifugal compressor
2. Stage 2: Optimization of blade angle on impeller along hub to shroud
3. Stage 3: Position optimization of leading and trailing edge impellers for main blades and
splitter blades
4. Stage 4: The Box-Behnken technique is used to select the most influential parameter from the
previous stage.
From the optimized centrifugal compressor model, it is known that there is an increase as 3%
for efficiency and 11% for pressure ratio compared to the initial impeller design point.

What is Done in This Research:

In this research, the authors have optimized using a design experimental technique on a centrifugal
compressor with an increase in the pressure ratio and isentropic efficiency at each stage. The
important results of the optimization for each of these stages can be seen as follows.
1. In Stage I there are 5 main parameters in designing the meridional geometry, namely M1 to
M5 where the curvature of the meridional geometry gives the most optimum giving an increase
of 2.5% and 7.24% for efficiency and pressure ratio, respectively.
2. In stage 2, it is known that there are 4 main parameters that affect the distribution of the blade
angle, namely B1 to B4 where parameters B1 and B2 are factors that give a 4% effect on
changes in the pressure ratio and efficiency of the centrifugal compressor.
 3. At stage 3, it can be seen that there are 4 main parameters that affect the inlet geometry and
splitter Blade, namely I1 to I4 where the optimum point gives the highest increase for the
pressure ratio of 0.34% and 3% for efficiency.
4. In stage 4, 6 main parameters will be obtained using the Box-Behnken method which will
then form a second-order polynomial function according to the response surface. These six
parameters will provide the optimum point with an increase of 3% for efficiency and 11% for
pressure ratio.

Future Works:
In this paper, the authors optimize the meridional geometry, blade angle distribution, and
blade inlet geometry and splitter so that furthermore the authors can optimize the effect of tip gap on
the impeller, fillet on the impeller blade and scaling of the centrifugal compressor size. In addition,
the authors need to optimize the geometry of the diffuser to reduce the occurrence of loss and
separation in the flow. In addition, the authors can use other surrogate models such as Artificial Neural
Networks to optimize centrifugal compressors.
Determining Research Question
Globally, each of these papers talked about centrifugal compressor which underwent
numerical analysis through Computational Fluid Dynamic (CFD) method. Paper 1 explained about
phenomenon around surge/stall region by using mass flow as the parameter constraint. Tip gap effect
and scaled compressor comparison were discussed in paper 2, which can be used later as important
considerations in designing centrifugal compressor. Paper 3 provided information regarding which
turbulent model can give as closest as possible numerical simulation model compared to the one used
in experiment, while having the lowest computational cost. Aerodynamic phenomenon which led to
the loss at impeller, diffuser and deswirler on some variations of compressor rotational speed was the
aim of paper 4 study. Paper 5 focused on compressor optimization using design of experiment
technique, optimizing pressure ratio and isentropic efficiency of the predefined design of a
compressor. Paper 1 and paper 2 were related for their study about mass flow and surge/stall
prevention, also in determining the optimum value of tip gap clearance. There were both in paper 3
and paper 4, a discussion about loss due to the usage of diffuser which is simulated numerically using
CFD method (ANSYS CFX), considering turbulent model variation and compressor rotational speed.
Based on the explanation above, paper 2, 3 and 4 can be used for reference in determining
parameter constraints for the optimization process of centrifugal compressor. This optimization aims
for obtaining the most optimum pressure ratio and isentropic efficiency, by using Design of
Experiment technique just like paper 5 discussed. Off-design analysis also needs to be performed in
order to get working range of this compressor.

Research Question:
“How do the change of mass flow and compressor rotational speed affect pressure ratio and
isentropic efficiency of the off-design of centrifugal compressor in the micro turbojet engine
application by using CFD analysis?”
 REFERENCE

Azizi, M. A., & Brouwer, J. (2017). Stall/surge dynamics of a multi-stage air compressor in
response to a load transient of a hybrid solid oxide fuel cell-gas turbine system. Journal of
Power Sources, 408-418. doi:https://doi.org/10.1016/j.jpowsour.2017.09.010

Barsi, D., Bottino, A., Perrone, A., Ratto, L., & Zunino, P. (2019). Design of a Centrifugal
Compressor for Micro Gas Turbine: Investigation of Scaling and Tip Clearance Effects.
Open Journal of Fluid Dynamics, 49-62. doi:https://doi.org/10.4236/ojfd.2019.91003

Gibson, L., Galloway, L., Kim, S. i., & Spence, S. (2017). Assessment of turbulence model
predictions for a centrifugal compressor simulation. Journal of the Global Power and
Propulsion Society., 142-156. doi:https://doi.org/10.22261/JGPPS.2II890

Mojaddam, M., & Pullen, K. R. (2019). Optimization of a Centrifugal Compressor Using the
Design of Experiment Technique. Journal of Applied Science, 1-19.
doi:https://doi.org/10.3390/app9020291

Sebelev, A., Aleksensky, V., Tikhonov, A., & Shengals, A. (2021). Aerodynamic analysis of the
small-scaled centrifugal compressor for micro-turbojet engine applications. Journal of
Physics Conference Series, 12-17. doi:http://dx.doi.org/10.1088/1742-6596/1891/1/012017