Abstract
Automotive turbochargers are known to operate into the self-excited unstable region. In the past these instabilities have been accepted as unavoidable, but recent developments in analysis and instrumentation may make it possible to reduce or eliminate them. A test stand has been developed at Virginia Tech to measure the vibrations of a 3.9 liter diesel engine stock turbocharger with both stock floating ring journal bearings and also custom design fixed geometry bearings. Vibration spectrum content clearly identifies the shaft instabilities and provides the basis for additional evaluation of future improved bearing design modifications. The current results, for custom fixed journal bearings, have clearly revealed a distinct jump with associated shift in the spectrum frequency content. This paper will document the recent tests of custom design fluid film bearings that have experienced this nonlinear jump condition.
1. Introduction
High speed turbochargers are intended to increase the power of internal combustion engines. Turbocharging was not widely accepted in the early years, but recently turbocharging has become standard for most diesel engines and is used in many gasoline engines as well. Since the earliest turbocharger prototypes, researchers have attempted to improve turbocharger reliability and increase turbocharger life while keeping the cost per unit at a minimum (Born, 1987). Since vibration-induced stresses and bearing performance are major failure factors, rotordynamic analysis should have been an important part of the turbocharger design process. However, a thorough rotordynamic investigation was very difficult and only recently have researchers started to document their findings (Andres and Kerth, 2004; Holmes et al., 2004).
Advances in rotordynamic analysis computer programs have now made the analysis of a turbocharger rotor-bearing system a reality (Gunter and Chen, 2000). Manufacturers have begun using these tools to better understand the dynamics of high speed turbochargers. Design improvements, however, cannot depend on computational analysis alone and on-engine test data are still required for these still very difficult analytical predictions.
A previous investigation used a commercial finite element analysis (FEA) computer program to model the dynamics of the turbocharger (Alsaeed, 2005; Kirk et al., 2007). That investigation demonstrated how linear analysis can be beneficial for understanding the basic experimental dynamic performance of the turbocharger rotor bearing system; this current experimental research extends that work with on-engine testing that demonstrates the nonlinear jump for certain types of fixed geometry bearings now under test.
2. Description of the engine test stand
The test turbocharger was installed on a 3.9 liter 130 HP 4 cylinder diesel engine. Design and setup of the test stand has been performed by senior students of mechanical engineering as an undergraduate design project for the past five years (Kirk et al., 2008, 2010). The engine is installed, using its stock mounts, on a heavy cast-iron test base. It can be coupled to a chassis dynamometer with a flywheel adapter plate and a shaft floating between two universal joints. The current purpose of the dynamometer is to control turbocharger top speed by controlling engine speed and load, not to measure engine performance. Fuel, coolant, exhaust, and control connections are made as required. The initial testing of new custom bearings do not require the engine to be under full load, because the no load condition can drive the turbocharger past the onset speed of the two basic modes of instability of concern for the current design turbocharger.
The turbocharger has a special target nut on the compressor end of the rotor which permits non-contact measurement of the rotor vibration relative to the turbo stator casing. The turbochargerisbolted to the engine exhaust manifold where the engine exhaust is directed through the turbo and exits into an exhaust line that has suction to pull the gases from the basement to the discharge stack on the roof of the facility. The turbocharger has an open inlet which permits an optical speed pickup to detect the running speed and a second sensor detects the engine crank speed. The data is collected with the use of special signal conditioning instruments and the Bently Nevada, PC based, ADRE monitoring system.
3. Discussion
3.1. Initial test results with custom bearings at both bearing locations
The initial testing on the stock floating bush journal bearings have all revealed that two modes are easily excited. From very low speed the 1st mode with a conical, near-rigid body, mode shape is observed followed by the onset of a second cylindrical mode with slight bending, which is excited at turbocharger speeds just over 60,000 rpm. The first mode reduces in amplitude as the second mode is excited, and both are usually evident all the way to full speed, no-load conditions. When the six axial groove custom bearings were installed on both compressor and turbine bearing positions, without the free floating outer film, but only with a single oil film per bearing, the basic modes were similar but were found to occur at a reduced frequency. In addition, the amplitude jump was noted to occur just past the normal design speed of the engine. The first instance of the jump is documented in Figures 1 and 2, where the cascade plot and then the pre- and post-jump shaft target nut orbits are shown.
Cascade plot showing nonlinear 1st mode jump, with the 8.1 mil-pp at 10.5 kcpm changing to 13.1 mil-pp at 12.75 kcpm, at near top turbo speed for test run S1005 (six axial groove custom design at both compressor and turbine bearing locations). Shaft orbit at target nut (a) pre-jump and (b) post-jump operating condition for test run S1005.
This first occurrence was so unusual that caution was used near this speed until the turbocharger was disassembled and checked. No damage was evident to the bearings or the rotor shaft. This build was with six axial groove bearings with o-rings on the outer diameter and pressed into the bearing housing. Both bearings were of the same custom design. The unexpected jump occurred near the engine speed of 2500 cpm, which is the design speed of the diesel engine.
3.2. Modified test plan
As a result of the unexpected nonlinear jump the test plan was modified to apply the custom bearings to the compressor end of the turbocharger only. The compressor end has a gravity load of only 0.4448 N (0.1 lb) upward, while the turbine end bearing has a downward load of 4.448 N (1 lb). The decision was made to keep the stock floating ring design at the turbine end, since no overload condition had been observed with that stock design. The goal is now to gain confidence with the capability of the bearing design analysis tools and the accuracy of the prediction of the onset speed for instability. The actual magnitude of the bearing loading is unknown for the loaded condition and by having a marginal load capacity it may be possible to better estimate the magnitude of the aerodynamic loads (volute loads for the centrifugal compressor and the turbine stage). This is a topic of another ongoing investigation at the Virginia Tech Rotor Lab and beyond the scope of the current discussion.
3.3. Results for no-Load operation with axial groove and floating ring bearing combination
For the no load condition testing, both the six axial groove and eight axial groove bearing design proved to have acceptable but larger than desired vibration levels for the first mode of instability. Normally, the first mode decreases when the second mode becomes unstable, but for these runs the first mode remained strong and at slightly reduced frequencies, compared with the stock bearing results.
The 10 axial groove bearing was installed on the compressor end and run with caution. The expected jump did occur and was even more violent than the previous jump shown in Figure 1 for the six axial groove bearing design for both bearings. The frequency spectrum of Figure 3 clearly shows the jump in frequency and amplitude for this 10 axial groove bearing design. The overall radial amplitude was in excess of 30 mil pp at the target nut probe location.
Cascade plot showing nonlinear 1st mode jump, with the 6.6 mil-pp at 12 kcpm changing to 21 mil pp at 18 kcpm with the turbo operating at 84 kcpm shaft speed; test run S1013.
The shaft orbit just prior to the jump was approximately 14 mil pp as shown in Figure 4 and the amplitude just after the jump was in excess of the 28 mil pp shown in Figure 5. The timing marks indicated by the bold black dots on these figures are for every third cycle of shaft spin. The frequency noted as 27.4 kcpm in Figure 5 is actually one third the actual frequency of 82,200 cpm.
Orbit just prior to jump, 10 axial groove compressor end and stock floating bush on turbine end. Orbit just after jump, 10 axial groove on comp end plus stock floating bush on turbine end.
3.4. Results for full-Load operation with axial groove and floating ring bearing combination
The results of the no-load testing indicated that it may be possible for either the six or eight axial groove bearing design to provide an enhanced stability under added engine load. The previous loading procedure had used the top engine speed of 2800 rpm speed to have the dyno load to pull the engine down to 2500 rpm, the spec design speed. It was desirable to have more control over the load on the engine so a fixed speed loading sequence was tried. The engine load was increased in fixed increments to the maximum load of the engine, while the speed was held at 2000 rpm crank speed. Then the speed was increased slowly to 2500 rpm under the full-load condition. This resulted in a smoother increase of the turbocharger speed and less added stress on the system.
The results are given in Figure 6, for a full load, with six axial groove bearing design on the compressor end, which was without a violent jump; but as the load increments, a small change in frequency can be identified and the first mode initially reduces in amplitude, but then increases a modest amount. The second mode is observed to reduce in amplitude until the load is dropped and the deceleration begins. It is interesting that this 140,000 rpm speed is the highest turbocharger speed to date, and with a non-floating ring bearing on the compressor end of the rotor. The downside of this run was the larger than desired overall amplitudes of 20 mil pp at idle speed.
Six axial groove on compressor end and floating ring on turbine end; loaded at engine speed of 2000 rpm to a load of 379.6 Nm (280 ft-lb), then engine speed increased slowly to 2500 rpm.
The eight axial groove bearing was the next full load test bearing. The results of Figure 6 seemed encouraging, because the eight axial groove bearing had been slightly better than the six axial groove bearing for the no load condition. The hopes were high as the loading at 2000 rpm fixed speed produced smaller amplitudes up to the full load of 379.6 Nm (280 ft-lb), then as the engine speed was increased to nearly 2300 rpm, the nonlinear jump occurred, going to 20 mil pp and with the frequency changing from about 15 kcpm to 20 kcpm, as shown in Figure 7. It drops out immediately with a small drop in speed, and as speed increases, it jumps once again. The shaft orbit prior, near and in full jump can be observed in Figures 8–10. No damage was evident on inspection of the bearings.
Eight axial groove loaded to 379.6 Nm (280 ft-lb) at 2000 rpm, then accelerated to 2300 engine speed cpm when the jump in amplitude and frequency occurs for test run S1021. Orbit prior to jump, full load, eight axial groove on compressor end, stock floating ring at turbine end. Orbit at start of jump, full load, eight axial groove on compressor end, floating ring on turbine end. Orbit in full jump, full load, eight axial groove on compressor end, floating ring on turbine end.
4. Summary and conclusions
The current tests for the 3.9 liter 130 HP diesel engine with stock turbocharger operating at no load and full load have provided valuable information and insight into this particular turbocharger's dynamic performance for a series of custom, fixed bore, bearing designs. The conclusions from the current experimental research results are as follows:
The floating ring bearing has superior load capacity when compared to standard axial groove bearing designs. The six isolated oil-in holes in the stock floating ring bearing are not to be approximated as a six axial groove bearing. The gravity load is not the only load to be considered when designing bearings for turbochargers. A nonlinear jump can occur in a low load capacity bearing under no-load or loaded conditions. The fixed six axial groove bearing did not seem to give added drag on the rotor system, but this must be confirmed by the same loading sequence for the double floating ring bearing build. That is the test currently in progress.
Additional testing is currently underway to do the same loading sequence on the stock floating ring bearings. The close evaluation of all of the loading results will hopefully provide additional clues as to the volute loading force levels in this classic design turbocharger.
Footnotes
Acknowledgments
The authors wish to thank Dr. John Nicholas of Lufkin-RMT for donating the custom bearings to conduct these tests and also the 2010 ME 4016 Turbo Project Team for their assistance in operating the test stand.
Funding
This work was sponsored by the Virginia Tech Rotordynamics Industry Affiliates Group.