See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/338100625

Canopy design to minimize restrictions and maximize heat transfer

Conference Paper in AIP Conference Proceedings · December 2019

DOI: 10.1063/1.5141249

CITATIONS READS

3 6,392

3 authors, including:

Veena Parthan Nagaraja S R

Cummins Inc. Dayanada Sagar University
1 PUBLICATION 3 CITATIONS 48 PUBLICATIONS 211 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Veena Parthan on 21 December 2019.

The user has requested enhancement of the downloaded file.

Canopy design to minimize restrictions and
maximize heat transfer
Cite as: AIP Conference Proceedings 2200, 020079 (2019); https://doi.org/10.1063/1.5141249
Published Online: 20 December 2019

Veena Vattaparambil Parthan, P. I. Shyamkumar, and S. R. Nagaraja

AIP Conference Proceedings 2200, 020079 (2019); https://doi.org/10.1063/1.5141249 2200, 020079

© 2019 Author(s).
 Canopy Design to Minimize Restrictions and Maximize
Heat Transfer
1
Veena Vattaparambil Parthan, 2P I Shyamkumar and 1 a)S R Nagaraja
1
Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham,
Bengaluru, India.
2
Cummins Engine Business Unit, Cummins India office campus, Survey No. 21, Balewadi, Pune, India
a)
Corresponding Author: [email protected]

Abstract. A Canopy or an Enclosure is a crucial part of a genset which protects the generator from external environmental
factors and maximizes its efficiency. Most of the genset enclosures are designed to meet the noise requirements set by legal
authorities but it is also important that such designs can expel the heat generated by different parts of the generator such as
engine, alternator, and radiator. This report focusses on the design of a Canopy for QSK 60 G4 Cummins Genset to minimize
the restrictions and simultaneously maximize the heat transfer across the system. This method of optimizing the two objectives
i.e. pressure drop and heat transfer are carried out by parametrizing the louver angles and length of louvers for two different
models on ANSYS. Genetic Aggregation in RSO (Response Surface optimization) is used to constrain the objective functions
to the limiting range which is set as per the experimental tests carried out for detecting the pressure drop and LAT (Latent
Ambient Temperature).

Keywords: Genset, Multi-objective Optimization, ANSYS Design Xplorer, Canopy

INTRODUCTION

Today’s world survives on electricity for all basic needs and an outage in power could be fatal in different sectors.
This brings the necessity of generators which can be the sole provider of power during events of emergency. A
generator requires a source of fuel to function and diesel is widely used as the fuel because of its efficiency and
lower prices. Nowadays, generators using gas as the source of fuel i.e. Biogas, Natural gas and LPG are being
manufactured. Commercially, a diesel generator is referred to as a ‘Genset’; particularly when a diesel engine is
used in combination with an alternator. With the current pace at which technology is advancing and with changing
regulatory laws, it has become a necessity to change the product design to meet the customer requirements and to
keep up with the competitive market place. It is possible to take charge of these changes by utilizing computational
fluid dynamics for designs that involve fluid flow. The analysis of these models helps in reducing costs associated
with the production of prototypes which adds as an advantage. These models can be optimized as per the
constraints and or new regulatory laws in an easier and less expensive way using these computational tools.
Prashant P. Pandav, Surendra Barhatte & Nitin Gokhale [1] carries out thermal analysis of an existing genset. The
paper was focussed on reducing footprint occupied by the system. The exhaust system including exhaust pipe and
manifold, head block were analysed using CFD simulation to determine the cooling performance. D. S. Mole, P.
S. Yadav, M. D. Kandalkar & N. V. Karanth [2] focusses on the optimization of an enclosure for a high-speed
genset (3000 rpm) to meet MoEF limits of noise levels. Noise Source Identification (NSI) was carried out to
identify the major sources of noise from the genset. A hybrid approach comprising DOE (design of experiments)
and Boundary element method were used in the enclosure design. Gholap and Khan [3] used multi-objective
optimization for a household refrigerator that works on irreversible Carnot cycle. In this optimization, the
objective functions were the minimization of energy use and the cost associated with the bill of materials. These
two objective functions are inversely proportional to each other imposing a problem with optimization. Hence a
Pareto Optimal model was adopted to obtain an optimum solution using response surfaces. Shabliy and Cherniaev
[4] optimized the design of gas turbine compressor blades by utilizing finite element analysis and computational
fluid dynamics in ANSYS through a single path of interaction between the fluid and the blade structure. The
adiabatic efficiency and minimum equivalent stress were the objective functions which were optimized by varying
the structure of the compressor blades. This optimization was carried out on ANSYS using Genetic Algorithm in
Response surface optimization.
1st International Conference on Manufacturing, Material Science and Engineering (ICMMSE-2019)
AIP Conf. Proc. 2200, 020079-1–020079-8; https://doi.org/10.1063/1.5141249
Published by AIP Publishing. 978-0-7354-1951-3/$30.00

020079-1
 METHODOLOGY
Experimental Set-up
A radiator mounted fan arrangement for 2000kVA QSK 60 G4 genset was set-up to study the main reason which
was leading to misalignment issues of the fan and crank pulley. Further, this set-up was tested to evaluate the
performance of the radiator in an enclosed condition. The data that was obtained from this test helped in
determining the restriction across the canopy and the LAT (Limiting Ambient Temperature) for the radiator. The
test was carried out by at100% and 110% load conditions.
A U-tube manometer which +/- 8 inches long was placed in the duct near the louvered canopy inlet and outlet
such that while varying the opening of the duct the restriction was found to be between 0.7 inches to 1 inch of
‫ܪ‬ଶ ܱ. The velocity of the air at the two ends of the canopy was measured using an anemometer. Thermocouples
and temperature meters where located at different locations to measure the temperature across different sections
of the system. The table below showcases the data collected for the canopy.
TABLE 1: Experimental Test data.

Load Canopy Air Inlet Canopy Air Inlet Canopy Air Outlet Canopy Air Outlet
% Temperature I (°C) Temperature II (°C) Temperature I (°C) Temperature II (°C)
100 28.6 28.5 69 58.4
110 29.95 29.9 71.42 64.22

From the experimental test, the Latent Ambient Temperature (temperature between the inlet of the canopy till the
low-temperature after-cooler) is supposed to be in the range of 5°C - 7°C. The rise in temperature across the
canopy is observed to be 40°C and is nominal to have values between 40°C - 50°C. The restriction across the
canopy due to the components of the genset is expected to be between 0.7-1 inch of H2O. This value is obtained
using the U-tube manometer through adjustment of the duct opening.
From the measured data of the volumetric flow rate, the canopy inlet air flow is found to be 65734 cubic feet per
min and the canopy outlet air flow is noted as 83675 cubic feet per min. The mass of incoming air flow is not
equal to the mass of outgoing air flow which violates the law of conservation of mass. This discrepancy in the
values of inlet and outlet air flow could be due to errors in the measurement of airflow data.
Proposed Models
Model I

FIGURE 1: Simplified Geometry of Canopy with Genset of Model I.

FIGURE 1 represents Model I of Canopy with three louvers which are window-shaped flaps. The Canopy has an
inlet louver and two exit louvers which are at an angle of 45 degrees. The internal components of the Cummins
QSK60 G4 genset CREO model are simplified for ease in meshing and analysis. The pipes and connectors that
offer less resistance to the flow of air are eliminated from the Space-Claim model of genset. Further, two
compartments on either side of the louver areas are created to accommodate the fluid flow region. The louvers
and the genset are subtracted from the fluid region. The components of genset are replaced in the subtracted
regions. The fluid and solid regions are made to share the topography with the share topology module in the
Design Modeller. The three louver angles are parametrized in Design Modeller and not in Space Claim as the
latter does not keep a history of actions that are carried out.

020079-2
 FIGURE 2: Fluid Geometry of Model I.

FIGURE 2 shows the fluid region of the canopy along with the fluid sections on either side i.e. the inlet and the
outlet.

Model II

FIGURE 3: Simplified Geometry of Canopy with Genset Model II.

FIGURE 3 represents Model II of Canopy with opening areas on four sides of the structure. The Canopy has two
inlet openings and two exit openings with louvers which are at an angle of 45 degrees.

FIGURE 4: Fluid Geometry of Model II.

FIGURE .4 shows the fluid region of the canopy along with the fluid sections on all four sides i.e. the two inlets
and the two outlets.

Boundary Conditions and Set-Up

The boundary conditions to be applied for the scaled-down models are tabulated below.

TABLE 2: Boundary Conditions for Canopy Design.

Inlet Mass Flow Rate 0.052kg/s

Inlet Temperature 320C
Outlet Temperature 680C
Total heat generated from the engine 4.5 x 107 Watts
Hydraulic diameter 0.05

020079-3
The inlet mass flow rate for Model II is half of the total flow rate passing through the two inlets of the canopy
model. The system is modeled using the K-Ɛ turbulence model and the energy equation. In cell conditions, the
engine has a source term which is the heat generation. The heat generated by all components of the genset is
accounted for from the engine in this analysis.

Optimization Set-Up

In Design Xplorer module of ANSYS, the first tab to be run is the Design of Experiments (DOE). Here the user
is supposed to provide the input for the type of DOE generation method. Central Composite Design (default
method) is used as the scheme to automatically use the available space efficiently. The DOEs are set for two
ranges of angles i.e. 300- 450 and 450- 600. The generated DOEs for the two sets of ranges are shown below in
FIGURE 5 & FIGURE 6.

FIGURE 5: Design Points for 300 – 450 angle range. FIGURE 6: Design Points for 450 – 600 angle range.

Once the DOE set up is updated, the design points in the parameter table of the parameters tab are updated with
the values generated by the Central Composite Scheme in DOE tab. Once the values for the pressure drop and the
temperature rise for the 16 design points are generated. The Response surface optimization tab has to be updated
after the user inputs the objective functions i.e. the temperature rise and pressure drop in this case. Further, the
objective function could be constrained, but in this optimization, no constraints have been applied. The method
used for Response surface optimization, in this case, is the 2nd Order Polynomial method.
After the value for the response, the surface is obtained, the surrogate model is optimized using a simple screening
method by generating 50 sample design points to minimize pressure drop and maximize heat transfer.

RESULTS AND ANALYSIS

The result from the path line analysis of Model I in ANSYS Fluent is as depicted in FIGURE 7 & FIGURE 8.
The pressure and temperature near the louvered side of the canopy are measured as an average of the surface
pressure and temperature respectively. This is measured by creating a plane at either end of the inlet and outlet
sides of the canopy. The pressure drop across the canopy is observed to be 3 inches of H 2O and the temperature
rise is noted to be 640C.

FIGURE 7: Static Pressure Pathline for Model I.

020079-4
 FIGURE 8: Total Temperature Pathline for Model I.

Model II

The result from the contour analysis of Model II in ANSYS Fluent is as depicted in FIGURE 9 &
FIGURE 10. The pressure and temperature near the louvered side of the canopy are measured as an average of
the surface pressure and temperature respectively. This is measured by creating a plane at either end of the two
inlets and two outlet sides of the canopy. The pressure drop across the canopy is observed to be 2.16 inch of H2O
and the temperature rise is noted to be 470C.

FIGURE 9: Static Pressure Contour for Model II. FIGURE 10: Total Temperature Contour for Model II.

The response surface that is obtained from the design points for the range of 300 – 450 for Model I is shown in
FIGURE 11 which is observed to be at inlet and outlet angles of 37.50. The temperature rise and pressure drop at
the response surface point are observed to be 61.1 K and 1720 Pa. The local sensitivity of each parameter i.e. the
inlet and outlet louver angles, on the temperature rise and the pressure drop, is depicted below in FIGURE .11

FIGURE 11: Response Point and Local Sensitivity of Output Parameters for 300 - 450 angle range.

020079-5
The response surface that is obtained from the design points for the range of 45 0 – 600 for Model I is shown in
12which is observed to be at an inlet angle of 54.750 and outlet angles of 52.50 and 53.40. The temperature rise
and pressure drop at the response surface point are observed to be 73.23 K and 1014 Pa. The local sensitivity of
each parameter i.e. the inlet and outlet louver angles, on the temperature rise and the pressure drop, is depicted
below in Figure .12

FIGURE 12: Response Point and Local Sensitivity of Output Parameters for 450 – 600 angle range.

The goodness of fit from the observed values of the design points and the predicted values of the response surface
in 300- 450 range of angles is plotted in FIGURE .13

FIGURE 13: Goodness of Fit for 300- 450 range. FIGURE 14: Goodness of Fit for 450- 600 range.

The goodness of fit from the observed values of the design points and the predicted values of the response surface
in 450- 600 range of angles is plotted in FIGURE 14. The optimization of the objective functions
that are minimizing the pressure drop and maximizing the heat transfer across the canopy is carried out using
sample screening where the solution converged after 50 iterations for both ranges of angles for canopy Model I.
Three candidate points were selected which provided the best solution that satisfied the objective functions for
the range of 300- 450. The best solution was the three louvers at 45 0 angles satisfying the objective function by
giving a pressure drop of 1305 Pa and temperature rise of 65.43 0C followed by the second and third candidate
points at equity. The second candidate point showcased a pressure drop value of about 1391 Pa and temperature
rise of 65.620C while the third candidate point showcased a temperature rise of about 64.74 0C and pressure drop
of about 1396 Pa.

020079-6
 FIGURE 15: Candidate Points for 300- 450 range. FIGURE 16: Candidate Points for 450- 600 range.

Three candidate points were selected which provided the best solution that satisfied the objective functions for
the range of 450- 600 as shown in FIGURE 15 and FIGURE 16 respectively. The best solution was the three
louvers at 600 angles satisfying the objective function by giving a pressure drop of 868.8 Pa and temperature rise
of 77.750C followed by the second and third candidate points at equity. The second candidate point showcased a
pressure drop value of about 927 Pa and a temperature rise of 76.970C while the third candidate point showcased
a temperature rise of about 74.870C and pressure drop of about 919 Pa.

DISCUSSION OF RESULTS

From the results of Fluent analysis for Model I, the pressure drop across the canopy is observed to be 3 inch of
H2O which is more than the experimental value of 1inch of H 2O. This is because of the simplified model which
has less complicated structures and the inlet to the canopy has a single window shaped louver which increases the
pressure build-up at that location. Hence producing a greater pressure drop than the experimental value. Similarly,
the temperature rise across the canopy is more than the values obtained through the experiment because of the
limited opening area in Model I compared to the canopy used in the experimental test. Further, the heat radiated
from the surface of the components, conduction heat transfer, convection at the corners of the genset parts and
radiation from the surrounding was not taken into consideration resulting in a discrepancy of the obtained values.
For Model II, there was an increased number of louvers with a larger opening area which was set at 45 0 angle,
which was same as that of Model I. The increased opening area reduced the pressure drop to 2.16 inch of H2O
and 470C of temperature rise which was close to the experimental data in comparison to Model I.

For the multi-objective optimization using RSM (Response Surface Method), it was observed that with increasing
angles the restriction was reduced and the temperature rise was increased. The optimized value for pressure drop
and temperature rise in the 300 – 450 and 450 – 600 angle ranges was 5.24 inch of H2O, 65.430C and 3.48 inch of
H2O, 77.770C respectively. It can also be concluded that the observed values of the design points and predicted
values of the response surface were close enough which gave a good fitness curve; hence proving the robust nature
of the optimization tool.

CONCLUSION

In the current scenario with new emission laws i.e. CPCB IV soon to be introduced into the markets, it is necessary
that current design needs to meet the rules and regulations passed by the regulatory councils. ANSYS Fluent
Design Xplorer is an excellent tool that would help in optimizing the design of the present components, assemblies
or sub-assemblies to meet the requirements of the changing constraints and factors. Here, in this study, the design
of the canopy was optimized by varying the angles of the louvered canopy as well as the opening area. It was
observed that with an increase in the opening area, the pressure drop and temperature was optimized and met the
experimental data i.e. the pressure drop was noted to be 2.16 inch of H 2O and 470C at 450 angle opening of the
louver. The greater the angle of opening for the louver, the lesser was the restriction and the greater was the rate
of heat transfer. But taking the environmental conditions such as rain and dust into consideration, it is best to keep
the angle between 300 – 450. Design Xplorer is a robust tool which uses different schemes for generating design
points which are further used in building a meta-model which forms a base for connecting the actual model for
the optimization process.

020079-7
 REFERENCES

[1] P. P. Pandav, S. Barhatte and N. Gokhale, “Thermal optimization of Genset Canopy using CFD,” International
Journal of Mechanical and Production, Vol. 5, Issue 3, Jun 2015, pp.19-26.

[2] G. K. Babar, “CFD Driven Compact and Cost-effective Design of Canopy,” SAE Technical Paper 2017-26-
0254, Oct. 15, 2017.

[3] A. Gholapa and J. Khan, “Design and multi-objective optimization of heat exchangers for refrigerators,”
Applied Energy, no. 84, p. 1226–1239, 2007.

[4] L. Shabliy and A. Cherniaev, “Optimization of Gas Turbine Compressor Blade Parameters for Gas-dynamic
Efficiency under Strength Constraints,” in 4th International Conference on Simulation and Modeling
Methodologies, Technologies and Applications (SIMULTECH-2014), 2014.

[5] D.S. Mole et al. “Design Optimization of Acoustic Enclosure for Noise Reduction of Diesel Generator Set,”
SAE International Paper 2013-26-0254, Jan 2013.

[6] R. Waheed, et al., “An integrated experimental and computational approach to material selection for
soundproof thermally insulated enclosure of a power generation system,” in IOP Conference Series: 14th
International Symposium on Advanced Materials, 2016.

[7] P. Yadav, H. Bankal and N. V. Karanth, “Acoustic Enclosure Optimization for a Higher Capacity Diesel
Generator Set Using Statistical Energy Analysis (SEA) based Approach,” SAE Technical Paper 2017-26-
0188, Oct. 1, 2017.

[8] X. Gong and Y. L. Y. Zhang, “Structural Statics Analysis and Optimization Design of Regulating Device for
Air Conveyer Outlet in Coal Mine,” in E3S Web of Conferences 38, ICEMEE 2018, 2018.

[9] S. J. Winne and M. Chandrasekaran, "Optimization of Nozzle: Convergence using Ansys with RSM,
MOGA," ARPN Journal of Engineering and Applied Sciences, vol. X, pp. 5486-5489, 2015.

020079-8
View publication stats