Downloaded from SAE International by Univ of California Berkeley, Saturday, July 28, 2018

Acoustic Enclosure Optimization for a Higher Capacity 2017-26-0188

Diesel Generator Set Using Statistical Energy Analysis Published 01/10/2017

(SEA) Based Approach

Prasad Yadav, Harshal Bankar, and Nagesh Voderahobli Karanth
Automotive Research Association of India

CITATION: Yadav, P., Bankar, H., and Karanth, N., "Acoustic Enclosure Optimization for a Higher Capacity Diesel Generator Set
Using Statistical Energy Analysis (SEA) Based Approach," SAE Technical Paper 2017-26-0188, 2017, doi:10.4271/2017-26-0188.

Copyright © 2017 SAE International and Copyright © 2017 SAEINDIA

Abstract The DG set have an engine working as a prime mover and alternator
assembly to convert energy from mechanical to electrical form.
Diesel powered electric generators are used in a variety of
Engine alternator assembly is wrapped in a partially closed enclosure
applications, such as emergency back-up power, temporary primary
with some openings for ventilation and accessibility requirements.
power at industrial facilities, etc. As regulatory and customer
Enclosure has to perform two functions first to shield the engine noise
requirements demand quieter designs, special attention is given to the
and second to maintain proper ventilation for smooth functioning.
design of acoustic enclosures to balance the need of noise control
Both these functional necessities are contradictory in nature.
with other performance criteria like ventilation and physical
Challenge in design is to tradeoff between both requirements without
protection. In the present work, Statistical Energy Analysis (SEA)
affecting the performance. Present study uses source- path-receiver
approach augmented by experimental inputs is used to carry out
model approach to control noise emission outside enclosure with
Vibro-acoustic analysis of an enclosure for higher capacity Diesel
focus on noise source isolation and modification of noise emitting
generator set. The exterior sound radiated from an enclosed generator
path using SEA [2] solution technique.
is predicted and further enclosure is optimized for an improved
sound-suppression. The airborne sources such as engine, alternator,
Most of available literature is focused on sealed enclosures. Jackson
radiator fan and exhaust are modelled explicitly using experimental
[3] modeled the source and enclosure as two infinite parallel flat
noise source characterization. Structure borne inputs are also captured
panels separated by some distance. One panel is defined as a constant
in the test for improving modelling accuracy. The vibroacoustic
volume velocity source and the other as the vibrating enclosure wall.
performance of the enclosure and sound absorbers are modelled
Junger [4] refined this by modeling the enclosure as a simply
through Biot parameters and experimentally validated with a good
supported finite rectangular panel. The source was modeled as a
agreement between test and simulated results within 2 dB for overall
piston generating a uniform pressure field with addition of absorbing
levels. Further, path contribution and sensitivity analysis has been
material in the air gap. Oldham and Hillarby [5,6] modeled source as
done on base enclosure to find out major contributors to exterior
vibrating flexible plate to give a more realistic values. Agahi et al. [7]
radiated noise. Optimization is done using parameters like: noise path
improved the model further by a using a coupled structural finite
modifications (ducts, partitions), leakage minimization and acoustic
element and acoustic boundary element model at low frequencies.
louver designing based on preceding analysis results. Optimized
Ver [8] gave a simple analytical model to predict partial or leaky
enclosure has resulted in 3 dB reduction in overall noise level
enclosure performance. Munjal [9] used one dimensional code based
on transmission loss theory to model enclosure along with thermal
Introduction simulation through CFD. In recent times with advancements in
computational resources virtual Product development tools for
Diesel Generator (DG) sets are commonly used to fulfill demand of
vibroacoustic modelling has become vital for getting quicker and
electric power for variety of applications in today’s modernization. In
reliable NVH solutions. Boundary Element Method (BEM) approach
India MOEF put strict regulations to limit noise pollution for DG sets
was used by Seybert [10] for capturing acoustic resonant behavior of
with imposed noise limit being 75 dB(A) at one meter for DG sets up
the enclosure at low frequencies. BEM suffers from limitation of
to 1000 kVA as per test procedure mentioned in ISO 8528-10 [1].
excessive solution time and non-applicability for high frequency
There is an ominous need to minimize noise emission by DG sets to
solution. Both these can be overcome by an energy based novel
fulfill customer expectations and legislation requirements.
approach named Statistical Energy Analysis (SEA). Zhen and Copley
[11] have recently used Hybrid SEA approach for a small capacity
trolley mounted enclosure. Present study has attempted use of SEA
 Downloaded from SAE International by Univ of California Berkeley, Saturday, July 28, 2018

technique augmented with experimental test to correctly characterize Present work makes use of critical path analysis through simulation
the Enclosure of higher capacity Genset and subsequent parametric model supported by test inputs for above thus reducing number of
studies for design optimization. experimental trails. Pure SEA approach is used for building enclosure
model capturing all noise sources, enclosure geometry details,
interior and exterior sound fields. Baseline DG set measurements are
Problem Definition done for getting near field source noise levels and radiated noise
In present study 280 kW diesel engine with 1500 rpm speed is a levels at 1 meter. Sound Absorbing materials are modelled using Biot
prime mover for a higher capacity 300 kVA DG set .It has a parameters .Simulation model of baseline enclosure is validated with
rectangular sheet metal enclosure volume of 18.7 m3 with Length test and used for further sensitivity and critical noise path analysis
being highest dimension followed by height and width. Enclosure has study. Modifications are suggested from the inputs of preceding
two doors on each side, one door have a glass window for visual analysis for achieving targeted noise reduction. Finalized
inspection of generator controls. (Refer figure 1.) Inside the enclosure modifications are fabricated and tested for noise reduction.
there are several acoustic cavities separated by semi partitions,
louvers and distinct interior geometry changes. Air and sound can
pass through each cavity. Cool air enters from front and side panel
Measurements
openings and hot air leaves from top opening. Interior Cavities are Baseline measurements on existing DG set were done for getting
shown in figure 2. noise source characteristics required for building SEA model source
data and obtaining exterior radiated noise values through evaluation
The major noise sources were alternator, engine, cooling fan and the of sound pressure and vibration parameters.
exhaust. Engine exhaust is routed through a muffler and out from
exhaust port; it is located above the enclosure top. Engine, alternator DG set have both air borne and structure borne sources of the two
and radiator fan assembly rest on a frame structure made of thick airborne are dominant. The major airborne noise sources are engine,
metal and it is isolated from the base frame by use of elastomeric alternator, and radiator fan and exhaust tail pipe .Sound powers of all
mounts. The muffler was only noise source located outside the these sources are evaluated through near field pressure measurements
enclosure. 100 mm thick Rockwool is placed on enclosure walls for while structure borne sources are characterized by vibration
noise insulation. Objective of the study was to design the quiet measurements. The exterior radiated noise emission levels at one meter
enclosure to reduce overall noise by 3 dB(A) without increasing the are evaluated at 75% rated load as per ISO 8528-10 requirements using
existing enclosure volume. standard microphones and Data Acquisition System .Enclosure
dimensions were such that measurement consisted of nineteen locations
as shown in figure 3. Sound Pressure Levels (SPL) measured at all
locations are averaged logarithmically to arrive at final SPL value. For
sound absorption materials Biot parameters like porosity, Flow
resistivity, tortuosity, characteristic lengths are evaluated using
impedance tube measurements and inverse characterization.

Figure 1. Schematic of DG set component arrangement layout with enclosure

size (5 m x 1.7 m x 2.2 m)

Figure 3. Schematic of SPL setup as per ISO 8528-10

SEA Simulation
SEA is energy based method which calculates the response of
coupled structures and acoustic subsystems at high frequencies,
Figure 2. Schematic of enclosure interior air cavities where a large number of uncorrelated modes are well-suited to use
statistical concepts and representation of the subsystems and
couplings. SEA solves for coupled structural and acoustical
Methodology
subsystems simultaneously. It generates a virtual Transfer Path
Critical phase in any noise reduction project is identification of noise Analysis (TPA) indicating the flow of vibration or noise energy
sources, noise propagation paths and their relative ranking for noise through the system. These energy flows are useful for understanding
attenuation. Traditionally this step is done using experimental tests. the main noise contributors at target locations and for performing
 Downloaded from SAE International by Univ of California Berkeley, Saturday, July 28, 2018

sensitivity studies to reduce noise to meet target levels. SEA uses Baseline Results and Analysis
analytical and simplified representation elements for modelling the
Figure 6 shows comparison between predicted and measured sound
transmission of complex multi-source and multi-path noise through
pressure levels. A very good correlation is seen above 160 Hz with
the cavities and structure of enclosure. A model consists of flexural
overall levels within 2 dB tolerance which proves validity of baseline
plates and cavities, connected by line, point and area junctions. Mass
enclosure SEA model.
law, leak modelling and user defined Coupling Loss Factors based on
test measurements are also employed. Enclosure Panel are modelled Further path contribution analysis is done using simulation model to
as flat plate elements and air cavities as acoustic air cavity element find relative contribution from major paths to exterior sound field by
with application of various sources as shown in figure 4 and figure 5. just allowing noise propagation through one path and blocking
The model comprised more than 75 subsystems and 200 connections, remaining paths through use of heavy material. As seen from figure 7
considerably fewer elements than an FEA model. major contributors are inlet, outlet openings and side panel leaks.

Commercially available code VA-one was used to build the model.

SEA calculates power flow between coupled structures and acoustic
subsystems. It also infers reverberant fields. For the case of DG set
enclosure the bounded interior acoustic cavities, structural panels and
their interactions were modelled with standard SEA acoustic cavity
and flexural flat plate elements. However sound radiation from
enclosure panels to exterior sound filed and radiation from openings
(louvers, holes) to exterior sound field does not fit within standard
SEA approach as field is unbounded showing no resonance .Hence
acoustic radiation efficiency approximations are used through use of
Semi-Infinite Fluid (SIF) element which represent energy sink
Figure 6. Comparison of test vs simulation average SPL for Baseline enclosure
behavior. SIF is connected to exterior cavities attached to panel
surfaces to represent free field radiation. For modelling openings
Further Energy flow analysis is done at critical frequencies as shown
through louvers and ducts transmission loss at area junctions was
in figure 8 to understand relative strength of sound energy throughout
used .Its calculation are based on theoretical and empirical values.
the enclosure and identify the path to be modified for noise reduction
Damping is an important parameter in a SEA model. As design and
assembly factors can influence damping, it is critical to obtain in situ
damping loss factors. The in situ damping loss factors of the plate
subsystems were obtained via experimental SEA. The damping of the
acoustic cavities was modeled by adding absorption materials to the
interior surfaces as per the baseline design. Absorbing material lining
is modelled as a noise control treatment applied to panel using Biot
parameters which are experimentally evaluated. Experimental SEA
was used to verify proper modal density for major subsystems.

Figure 7. Path contributions to exterior noise field

A typical flow is shown at 1000 Hz in figure 8 showing majority of

energy flow through top outlet opening, near internal partition and
side door opening

Figure 4. Existing Enclosure SEA structure model

Figure 5. Baseline Enclosure SEA interior cavity with sources Figure 8. Sound Energy flow at 1600 Hz inside baseline enclosure
 Downloaded from SAE International by Univ of California Berkeley, Saturday, July 28, 2018

Design Modifications References

As seen from path contributions and energy flow analysis major 1. ISO 8528:Part 10, ”, International Organization of
sound energy flows via top outlet opening hence a lined duct was Standardization, Geneva, 1993.
proposed as shown in figure 9 to attenuate this energy. Duct is lined 2. Lyon Richard H. and DeJong Richard G., Theory and
with sound absorbing material with thickness of 50 mm. Application of Statistical Energy Analysis, Butterworth-
Heinemann, (1995).
3. Jackson, R. S., “Some Aspects of the Performance of Acoustic
Hoods,” Journal of Sound and Vibration, Vol. 3, No. 1, pp. 82-
94 (1966).
4. Junger, M. C., “Sound Transmission through an Elastic
Enclosure Acoustically Coupled to a Noise Source,” ASME
Paper No. 70-WA/DE-12 (1970).
5. Oldham, D. J. and Hillarby, S. N., “The Acoustical Performance
of Small Close Fitting Enclosures. I. Theoretical Models,”
Journal of Sound and Vibration, Vol. 150, pp. 261-281 (1991).
6. Oldham, D. J. and Hillarby, S. N., “The Acoustical Performance
of Small Close Fitting Enclosures. II. Experimental
Investigation,” Journal of Sound and Vibration, Vol. 150, pp.
Figure 9. Schematic of a proposed lined duct on a radiator side opening 283-300 (1991).
7. Agahi, P., Singh, U. P., Hetherington, J. O., Numerical
Next dominant contribution paths were front, rear inlet openings and Prediction of the Insertion Loss for Small Rectangular
louver opening via inlet partition; to restrict the direct energy flow Enclosures”, Noise Control Engineering Journal, Vol. 47, No. 6,
inlet partition height was increased and louvers were lined with sound pp. 201-208 (1999).
absorption lining thickness such that not to restrict the air flow. Third
8. Ver, I. L., Enclosures, Cabins, and Wrappings, Noise
highest path was leaks through side doors. Proper sealing through bid
andVibration Control Engineering, Principles andApplications,
tapes was suggested to reduce these leakages. Modified canopy
Ver, I. L. and Beranek, L. L. (editors), Wiley, Hoboken, New
prototype was built as per above suggestions and tested for noise
Jersey, pp. 517-556 (2006)
levels. Overall noise reduction of 3 dB was achieved as shown in
figure 10. 9. Ravikanth G S G, Munjal M L,” Design and thermo-fluid dynamic
analysis of an acoustic enclosure for a diesel Genset,Journal of
Acoustic society of India, Vol-33,pp 458-463, 2005
10. Martinus, F., Quabili, Z., and Seybert, A., "Sound Through
Partial Enclosures with Louvers," SAE Technical Paper 2001-
01-1525, 2001, doi:10.4271/2001-01-1525.
11. Zhen, J., Copley, D., Londhe, N., and Fredrickson, S.,
"Obtaining Structure-borne Input for Hybrid FEA/SEA
Engine Enclosure Models through a Simplified Transfer Path
Analysis," SAE Int. J. Commer. Veh. 8(2):253-257, 2015,
doi:10.4271/2015-01-2349.

Contact Information
Figure 10. Comparison of modified enclosure noise levels with baseline Mr. Prasad S Yadav
enclosure Manager, NVH Lab
Automotive Research association of India, Pune, India
[email protected]
Conclusions
Contact No- +912030231490
A SEA based vibro-acoustic model was used to predict noise
performance of an enclosed DG set. The SEA simulation
methodology proved adequate and showed excellent correlation with Acknowledgments
experimental results (within 2 dB) hence it is used further for design Authors would like to acknowledge The Director, Automotive
optimization. Some deviations may be attributed to inherent SEA Research Association of India for her support and granting permission
assumptions and model simplifications. Design modification to publish this paper. The Authors would like to thank all their
suggested through this approach yielded considerable noise reduction colleagues at NVH lab for their esteem co-operation during this work.
of 3 dB on a proto type level proving efficacy of the novel approach.
 Downloaded from SAE International by Univ of California Berkeley, Saturday, July 28, 2018

Definitions/Abbreviations
CFD - Computational Fluid dynamics
dB - decibels
FEA - Finite Element Analysis
NVH - Noise Vibration and Harshness

The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session organizer. The process
requires a minimum of three (3) reviews by industry experts.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of SAE International.

Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper.

ISSN 0148-7191

http://papers.sae.org/2017-26-0188