eJournals International Colloquium Tribology 23/1

International Colloquium Tribology
ict
expert verlag Tübingen
125
2022
231

Quantifying Wet Brake Chatter Using an Accelerometer

125
2022
Michael Botkin
Caroline Mueller
ict2310407
23rd International Colloquium Tribology - January 2022 407 Quantifying Wet Brake Chatter Using an Accelerometer Michael Botkin Southwest Research Institute, San Antonio, Texas, USA Corresponding author: michael.botkin@swri.org Caroline Mueller Southwest Research Institute, San Antonio, Texas, USA 1. Introduction Wet brakes are common in off-road equipment and other high-torque applications where high levels of heat must be dispersed. Under the right conditions, wet brakes are susceptible to brake chatter caused by a “stick-slip” action which induces noise and vibration in the machine. For many years, this action was thought to be a result of the ratio of static to dynamic coefficient of friction [1]; however, research has shown that this action may also be triggered by a sufficiently negative friction velocity curve and/ or machine design [2]. Brake chatter, defined as an audible noise induced under braking, can be quantified via sound pressure level, variations in braking torque, or vibration measured using an accelerometer [1]. To measure sound, a technician listens for brake chatter in the audible range, or a microphone is used. This provides a good indication of what a driver would experience, but is sensitive to differences in observer and environment, resulting in poor repeatability. Strain gauges on the shaft driving the brake can be installed to measure braking torque. When chatter occurs, it is seen in the measurement as a variation in the measured torque [3,4]. Larger variations in torque produce more severe chatter, and there is a threshold torque variation value below which no audible chatter occurs. For a more accurate measurement, the strain gauges should be placed as close to the brake as possible, but this requires extensive modifications and maintenance since the brake is an internal, rotating component. Using an accelerometer to measure brake chatter allows for an easy test setup, as it can be mounted on the external surface of the machine. Furthermore, the accelerometer can pick up vibrations that occur even when little or no noise is produced—this provides a potential method to differentiate between fluids based on non-audible vibration. 2. Test Rig and Method To develop an accelerometer-based measurement of brake chatter, an off-road axle test installed at Southwest Research Institute was selected. The test rig was already fitted with internally mounted strain gauges to measure torque variation at the brake. An accelerometer was mounted on the external casing of the axle, directly above the brake disk, measuring in a direction tangent to brake disk rotation at that point. The accelerometer output was recorded at a DAQ rate of 2,000 Hz. A Fourier Transform was conducted on the accelerometer data during brake engagement to convert it into the frequency domain; for engagements that produced audible brake chatter, a resonant frequency of 340 Hz to 420 Hz was identified for this system. To quantify the severity of the chatter event for a single brake engagement, the integral of the frequency domain data from 340 Hz to 420 Hz was approximated using a midpoint Riemann sum. These values were then plotted against the magnitude of torque variation during the same engagement to establish a correlation between the two measurements. For an individual system, threshold values may be established through repeated testing to indicate when brake chatter occurs. 3. Data and Analysis 1,200 brake engagements were performed on various test fluids at different axle speeds, brake application pressures, fluid temperatures, and states of disk wear. The resulting dataset, shown in Figure 1, was generated by concurrently recording data through the strain gauges and the accelerometer. The magnitude of torque variation measured by the strain gauges is plotted against the magnitude of the integral of the resonant frequency range measured by the accelerometer. From historical testing, it is known that audible chatter on this test rig occurs if the magnitude of torque variation during a brake engagement exceeds 875 Nm. This is shown as a horizontal dashed line in Figure 1 and is known as the threshold value. Values above this horizontal line are understood to produce audible chatter, while values below this line do not. The audibility threshold value for the accelerometer is represented by the vertical dashed line on the plot. Its location was determined by minimizing the number of points that fell within regions of disagreement between the strain gauges and the accelerometer. The areas marked Type 1 and Type 2 in Figure 1 indicate a disagreement between the strain gauges and the accelerometer and differentiate between types of disagreement—a false positive chatter identification by the accelerometer or a false negative. For this test rig, the audibility threshold value for the accelerometer was determined to be 2.10. Using this threshold, the strain gauges and accelerometer were in agreement for 97.8% of brake applications. 408 23rd International Colloquium Tribology - January 2022 Quantifying Wet Brake Chatter Using an Accelerometer Figure 1: The magnitude of torque variation during a brake engagement on a text axle was plotted against the magnitude of the integral of the resonant frequency range measured by the accelerometer. Each marker color represents a test run. The dashed lines represent the threshold values above which audible chatter occurs, determined by repeated experiment. 4. Conclusions The method described above can reliably differentiate audible brake chatter from quiet brake engagements in a wet brake off-highway axle stand at different wheel hub speeds, brake engagement pressures, fluid sump temperatures, and states of disk wear. The integral area under the frequency-domain curve in the resonant frequency range is characteristic of brake chatter intensity and independent of fluid or operational conditions tested. A threshold value can be determined from experimentation that distinguishes audible brake chatter from quiet brake engagements. Finally, the method described herein is a suitable replacement for the strain gauge method used previously on the described axle. References [1] Anleitner, M.A., “Vibration and Noise in Oil- Immersed Friction Couples-A Basic Discussion,” SAE Technical Paper 861202, 1986, https: / / doi.org/ 10.4271/ 861202. [2] Friesen, T.V., “Chatter in Wet Brakes,” SAE Technical Paper 831318, 1983, https: / / doi.org/ 10.4271/ 831318. [3] Cave, W. and Lochte, M., “Development of an Updated Brake Chatter Test for Anti-Brake Chatter Transmission/ Hydraulic Fluids,” SAE Technical Paper 961817, 1996, https: / / doi.org/ 10.4271/ 961817. [4] Michael, R.A., “Key Elements of Wet Brake and Clutch Design,” SAE Technical Paper 921660, 1992, https: / / doi.org/ 10.4271/ 921660.