:$%&2

DYNAMIC SIMULATION OF A PISTON FOR COMPRESSOR APPLICATION

AND CORRELATION WITH BENCH TESTS
ALFREDO R. CABRERA
ELIAS B. RAMOS
Pistones Moresa S. A. de C. V.
Celaya, Gto, México

FOLKHARD HOELZEL
Wabco Auto
Hannover, Germany

YNGVE TORNQVIST
Wabco Compressor
Charleston, SC, USA

Abstract
This paper presents an analysis through dynamic simulation of a piston for a compressor
application used in the brake system of Commercial Vehicles Class 5 through Class 8. Several
modifications of the piston profile and ovality were done to optimize the results in product
validation and improve product performance for compressor applications, the result was an good
correlation between the dynamic simulation outputs with respect to the product validation tests.

Introduction
The application of dynamic simulation on pistons for compressor applications has not been
utilized because the operating conditions did not require a specialized design with respect to
temperature, speed and mechanical loads, however, in some applications, the operating conditions
have increased leading to severe demands during validation.

The present work was developed for single dual cylinder compressors for WABCO, used in the
brake system of Commercial Vehicles Class 5 through Class 8. Initially, the piston for this
application was designed considering conservative operating conditions that showed good results
during initial testing, according to additional testing requirements, the research of new piston skirt
shapes was developed in terms of profile and ovality, the results from the dynamic analysis as
well as the correlation with compressor bench tests are described in this paper.

Initial Design
Initially, the operating conditions considered were relatively conservative design criteria with
pressures and operating temperatures, speeds varying from 1000 to 2500 rpm and assembly
clearances, resulting in an initial profile and ovality skirt geometry based on the piston dynamic
analysis using the code Pisdyn developed by Ricardo Software [1], the model used predicts the
skirt to cylinder hydrodynamic behavior, as well as the distribution and magnitude of the asperity
contact pressures in selected operating conditions according to the test requirements. The
resulting initial proposal showed acceptable wear results between the piston and cylinder sliding
surfaces in compressor bench tests according to WABCO test criteria.

Up to this point, there was no indication of failure using the initial test conditions, in order to
complete the validation program, an additional test was selected demanding higher operating
conditions for the piston.

Test Conditions
According to WABCO’s complete test validation schedule, a test with higher speeds, thermal and
mechanical loads should also be accomplished, these extreme operating conditions resulted in a
compressor failure due to a piston scuffing. Once the test’s operating conditions were obtained,
showing cylinder pressures lower compared to a combustion engine [2], a new model was
generated in the piston dynamic analysis in accordance to the test data as shown in Table 1.

TEST CONDITIONS
Condition 1 Condition 2 Condition 3 Condition 4
Speed [RPM] 3000 3000 3600 4500
Press Press 01 Press 03 Press 02 Press 02

Table 1 Extreme compressor operating conditions.

The testing cycle does not correspond to the order presented in Table 1, it is a combination of
each resulting in the complete test program, the endurance of each condition varies as well
between each other, and are established by the customer. The pressure curves corresponding to
the test condition are represented in Figure 1, each of these conditions were simulated, however, a
failure analysis done to the scuffed pistons reveled that the operating condition with Press 3 @
3000 rpm (Condition 2) was the failure condition in the piston.

COMPRESSOR CYLINDER PRESSURE

20

18
100 -

PRESS 01 16
PRESS 02
83 -
PRESS 03 14
PRESSURE [%]

71-
12

10

0
-200 -100 0 100 200 300 400 500 600

CRANCK ANGLE [deg]

Figure 1 Cylinder pressure curves during extreme test cycle.

Once evaluating Condition 2 in the dynamic analysis to the skirt initial design, the hydrodynamic
behavior of the piston showed the asperity contact location under the test loads. A parametric
study with different skirt profile and cam shapes was done to develop a proposal that meets the
test requirements. The new proposal resulted in increasing the skirt top clearance as well as
increasing the piston ovality to avoid the asperity contact that lead to the piston failure. The
results from the numerical simulation, according to the lubrication model used in Pisdyn [3-4]
that were used to select the modified skirt shape were through the skirt extent of lubrication,
location and magnitude of the asperity contact, frictional power loss and wear load.

Figure 2 shows a comparison between the original skirt profile shape and the proposal resulting
from the parametric analysis on skirt shapes for the speed, load and test endurance of test
Condition 2.

PROPOSAL SKIRT SHAPES

40

35

PROPOSAL 3
30 OVALITY 3
SKIRT HEIGHT [mm]

INITIAL
25 OVALITY 1

20

15

10

0
0.020 0.000 -0.020 -0.040 -0.060 -0.080 -0.100 -0.120 -0.140

DIAMETRAL REDUCTION [mm]

Figure 2 Profile shapes analyzed for test Condition 2.

Results
Results with the initial profile for the over speed and load tests were unacceptable, in
consequence of the different piston failures, considering the different scuffed pistons, a
reconstruction of the failure sequence was established. Appendix A shows progressively the
failure sequence presented during test Condition 2. It is evident how a contact on the top of the
piston skirt in the thrust plane is generated, this results in scuffing of the same skirt area, as the
failure progresses, the following skirt areas to scuff are on the bottom of the skirt near the piston
side panels, further on, both top and lateral damaged areas are joined together to finally damage
the complete area of the skirt.

These results were predicted in the dynamic analysis using the operating failure conditions,
Appendix B shows the hydrodynamic pressure distributions indicating high oil film pressures
near the piston side panel, as well as asperity contact right were the piston starts to scuff. The
highest asperity contact pressure is presented in the top of the skirt on the thrust plane, although
no high oil film pressure is presented in this region, a lack of lubrication is assumed in this area,
on the other hand, the asperity contact shown on the side panel is presented as an effect of the oil
film breakage in consequence of a combination between the hydrodynamic pressures and the high
stiffness level in this area on both side panels. The wear load plots show a clear definition of the
skirt supporting areas in the complete cycle. This analysis confirms that the piston failure begins
on the skirt areas described above, it is not expected that the code reflects the failure sequence as
shown in Appendix A, owing to an effect of the failure progression with respect to the time the
areas in contact are exposed.

From the previous analysis, profile Proposal 3 and Ovality 3 were selected from various skirt
designs, which were evaluated considering mainly the four critical over speed and load test
conditions, even Condition 2 presented acceptable Moresa design criteria in the dynamic analysis.
Appendix C presents the skirt to cylinder hydrodynamic oil film distribution with reduced
pressures avoiding localized lubricated areas during the complete compressor cycle. The
validation of this skirt design was obtained once the testing program was completed to the release
of the piston. The results from the optimized piston design after extreme operating condition test
is presented in Appendix D.

Conclusions
The inclusion of complete and accurate test data in the numerical simulation resulted in a design
proposal that meets the requirements of a complete validation program.

The prediction of the asperity contact location in the dynamic analysis that led to the piston
failure in the initial proposal was detected once the test data for the failure condition was
included, resulting in a good correlation between the analysis and the test results.

The failure sequence is not followed in the numerical results, yet, it is important to detect the
failure initiation that results in a piston proposal that meets design goals.

This study showed that the piston dynamic simulation may not only be used in combustion engine
applications with severe operating conditions, but also to applications like compressors that
although have different level of thermal and mechanical loads, a good correlation is obtained.

Acknowledgements
The authors would like to thank WABCO and Pistones Moresa management for their support and
permission to publish, as well as the development team including Alberto Soria, Fernando Nito
and Samuel Regalado.

References
1. Ricardo Software, Pisdyn Manual, Documentation / User’s Manual, Version 2.7, USA, 2000.

2. Elias Ramos et al, “From Concept to Practice”, Fourth Ricardo Software International User
Conference, Detroit, USA, 1999.

3. Rifat Keribar and Zafer Dursunkaya, “A Comprehensive Model of Piston Skirt Lubrication”,
SP- 919, Engine Tribology, SAE Paper 920483.

4. Zafer Dursunkaya and Rifat Keribar, “Simulation of Secondary Dynamics of Articulated and
Conventional Piston Assemblies”, SP- 919, Engine Tribology, SAE Paper 920484.
Appendix A
Piston skirt failure sequence under over speed and load test during Condition 2.

Contact initiation Scuffing progressing from the skirt to the land area

Beginning of skirt contact in piston side panels Failure progression joining initial contact areas

Scuffing is extended to the center skirt area Complete skirt failure

Appendix B
Dynamic analysis results for the initial proposal in the over speed and load failure test condition
(Condition 2).

Hydrodynamic pressure distribution, localized near the side panels; low hydrodynamic oil pressure and
distribution is presented at the top of the piston skirt in the thrust plane.

Asperity contact magnitude and distribution at different crank angle degrees demonstrating the areas of
failure initiation.

Skirt and cylinder wear load in the complete Condition 2 cycle

Appendix C
Simulation results of the new skirt design including over speed and load test requirements
(Condition 2).

Hydrodynamic pressure distribution showing no localized oil film pressures and an improvement on the
skirt extent of lubrication.

Asperity contact pressure not shown in the complete Condition 2 Cycle.

Resulting skirt and cylinder wear load.

Appendix D
Results after compressor bench test under the over speed and load operating conditions
(Condition 1 to 4) with the optimized skirt profile and ovality in agreement with the numerical
simulation results.

Results of central wear patch for 2 compressors after over speed and load test (Condition 1 to 4)