EXPERIMENTAL INVESTIGATION OF HOT JUDDER CHARACTERISTICS IN PASSENGER CARS

Transcription

1 EB2015-VBN-006 EXPERIMENTAL INVESTIGATION OF HOT JUDDER CHARACTERISTICS IN PASSENGER CARS Xu, Xinfu * ; Winner, Hermann Technische Universität Darmstadt, Germany KEYWORDS hot judder, driving test, order analysis, transfer behavior, resonance frequency ABSTRACT Hot judder is a forced, brake-induced, wheel rotational speed-dependent vibration, typically occurs during light to moderate brake applications at high speeds (1). It is mainly a driving comfort influencing factor caused by brake system, which is usually perceived by the driver as minor to severe vibrations transferred through the chassis during braking. Currently, most of the hot judder researches are executed on test bench, the discrepancies and correlations between bench test and vehicle test as well as the characteristics and transfer behaviors of hot judder in passenger cars are still not well known. In order to explore these issues, preliminary driving tests are carried out, and hot judder is mainly identified and investigated with accelerometers. This test method is firstly validated by coherence analysis among acceleration at brake caliper carrier bracket, acceleration at brake caliper, brake torque variation (BTV), brake pressure variation (BPV), disk thickness variation (DTV) and disk waviness acquired with dynamometer tests. The validation test results from dynamometer are also applied as comparison analysis with driving test. It is found by driving tests that hot judder is strongly amplified by some resonance frequencies in the car. By further analysis of dynamometer test results, amplification effect of resonance frequency on hot judder is also detected. Current test results show that the higher dominant order found in drag braking is mainly caused by the resonance frequency and hot judder increase rates for those orders that pass through the resonance frequency are greatly raised in stopping brake applications. However, this conclusion needs to be further examined with specific tests. Besides, reproducibility of hot judder in driving tests and influences of braking deceleration, initial braking temperature and speed on the occurrence and intensity of hot judder are roughly studies. Transfer behaviors of hot judder from wheel brake to car body and the influences of higher hot judder orders on driving comfort are preliminarily discussed. INTRODUCTION Brake induced vibration and noises Figure 1 shows a classification of vibration and noises induced by braking system according to their occurring frequencies (2, 3). Unlike other brake noises, brake judder is mainly forced, low frequency vibration, which occurs during light to moderate ( g) brake from higher speeds (mostly>120 km/h) to lower speeds (around 80 km/h) (2). The vibration frequency of brake judder is directly proportional to the wheel speed. Order analysis is commonly used in

2 the data processing of brake judder problems. The order of brake judder is calculated with equation (1). Judder order Frequency Hz (1) Wheel rotational speed RPS Brake judder can be classified into two distinct subgroups in terms of different causes: cold judder and hot judder. Cold judder is caused by static Lateral Run-Out (LRO) of brake disk which is either from manufacture or assembly, and wear induced irreversible DTV which exists and happens under off-brake status. This static LRO and DTV are normally dominant with the first and second order. As a result, cold judder occurs at a lower frequency range from 10 Hz to 100 Hz (4). While hot judder is caused by thermally induced reversible dynamic LRO and DTV, it is combined with uneven temperat Regalen ure distribution on the friction surfaces of brake disk, which is known as hot spots. This dynamic LRO and DTV are typically dominant with higher orders. Therefore, hot judder occurs with a larger frequency range of 10 Hz to 600 Hz, since the thermal increase of cold judder is also thought to be caused by hot judder. Furthermore, due to the non-uniform thermomechanical stress, a variation of friction coefficient (µ-variation) on the friction ring of brake disk can emerge as a result of the microstructural transformation, which is mainly the formation of martensite. This is another cause for hot judder (1). Figure 1: Classification of brake induced vibration and noises (2, 3) Effect chain and transfer paths of brake judder In a disk brake, brake torquet is calculated by: Where, T F r 2 F r 2 p A r (2) Tan eff Cl eff B Pis eff F Tan N Tangential force at disk/pad interface r eff m Effective friction radius p B bar Brake pressure

3 F Cl N Clamping force - Coefficient of friction A Pis m 2 Piston area Because the piston surface area A Pis is constant, it can be seen from equation (2) that brake torque T can be influenced by variation of friction coefficient, BPV p B as well as variation of effective brake radius reff. Because BTV is normally calculated per revolution, assuming that the effective friction radius keeps constant within one revolution, BTV is mainly determined by BPV and µ-variation. Nevertheless, BPV is mainly caused by LRO and DTV of brake disk. Processes in the orange frame in Figure 2 present the effect chain of brake judder. Processes in the green frame in Figure 2 are the transfer paths of brake judder. Through braking system, BPV is finally transferred to brake pedal and causes pedal pulsation, which might be perceived via the foot by driver. For the transmission of BTV, on the one hand, it is transferred to the tire/road interface through the torque balance between brake disk and tire, where causes brake force variation and then the vibration of wheel carrier in the longitudinal direction of vehicle; on the other hand, it is directly transferred to wheel carrier through brake caliper for fixed caliper or brake caliper carrier bracket for fist caliper. Afterwards, vibrations of wheel carrier are transmitted to steering wheel through steering system. The rotating oscillations of steering wheel might be perceived by the hand of driver. And through spring and damper as well as different arms that are connected with car body, vibrations of wheel carrier are transmitted to car body. Vibrations in car body, especially at dashboard, are sometimes accompanied with a low frequency noise which is called hum (5). They might be haptically or acoustically noticed by the driver and passengers. Concrete transfer paths of brake judder differ in different cars depending on the types of their suspensions and steering systems. Examples of brake judder transfer path investigation can refer to (2) and (6). Figure 2: Effect chain and transfer paths of brake judder Motivation Nowadays, driving comfort and driving safety are gaining more and more attention in the development of automobile. Hot judder as one source of vibration generated by coupling among individual components in wheel brake primarily affects driving comfort but could, when confronting an inexperienced driver for the first time, lead to faulty reactions and reduced driving safety. In Addition, in some critical conditions, hot judder can also lead to disk cracking due to high temperature gradients.

4 Because of road excitation influence, less controllability of testing conditions and complexity of applying some sensors in driving test, currently most investigations on hot judder are based on dynamometer test. Correlations or differences between bench test and vehicle test as well as hot judder behaviors in real driving conditions are still not well known. In addition, a great deal of hot judder tests have been carried out on the dynamometer at Institute of Automotive Engineering (FZD) of Technische Universität Darmstadt (1, 3). High dominant orders of hot judder were detected in most of the bench tests. According to the transmissibility of the above mentioned transfer paths found by cold judder researches (2, 6, and 7) and the sensitivity of human to different vibration frequencies, it should be difficult for the high-frequency hot judder to be perceived by driver. Transfer behaviors of high-frequency hot judder as well as its influences on driving comfort are still not clarified. As initial part of the whole research project, objectives of this work are to identify hot judder behaviors in passenger cars and to preliminarily investigate the transfer behaviors of hot judder from wheel brake to car body. METHODOLOGY Validation of identification method based on coherence analysis Because it is not yet well known which production cars have significant hot judder problem. In order to investigate hot judder behaviors in vehicles, target car should be firstly selected by massive tests with passenger cars. At FZD-Dynamometer, five characteristic parameters, as listed in Table 1, are used to measure and evaluate the intensity of hot judder as defined by Sardá (3) and Fischer (1). Since strain gauge, pressure transducer, capacitive sensor as well as infrared camera are difficult to be amounted in the car without constructive changes, accelerometer was employed to identify the situation of hot judder in passenger cars. Because it can be easily stuck on the surface of brake components with glue. Table 1: Characteristic parameters of hot judder at FZD-dynamometer Characteristic parameter Measured quantity Test equipment Hot spots Disk temperature distribution/ Thermal image Infrared camera BTV Brake torque Strain gauge BPV Brake pressure Pressure transducer DTV Disk waviness Disk surface geometry Capacitive distance sensor In order to determine the suitability of using accelerometer to test hot judder, validation tests with a reference wheel braking system that has already been testified with evident hot judder were carried out at dynamometer before driving tests. Test setup and sensor distribution are shown in Figure 3. Two accelerometers were adopted in the tests separately: one at brake caliper carrier bracket and the other at brake caliper, considering that some cars with large disks or other types of brake caliper do not always have enough space to stick accelerometer at the similar position.

5 Figure 3: Test setup and sensor distribution at dynamometer Figure 4 gives the order analysis of the hot judder characteristic parameters from a drag braking test. Vibration accelerations at caliper carrier bracket and caliper have the same 11 th dominant order as BTV, DTV, and Waviness and their development tendencies in time domain are very similar. However, amplitudes of lower orders of the accelerations are relative small compared with those of BTV, DTV and Waviness, on account of the dependence of acceleration on vibration frequency. Figure 4: Order analysis of characteristic parameters of hot judder, drag braking with 1870 RPM (equivalent velocity of 225km/h) and constant brake pressure of 25bar It can be seen from further analysis with Magnitude Squared Coherence (MSC) Estimate among these characteristic parameters indicated in Figure 5 that accelerations, BTV, DTV and Waviness correlate with each other very well at most of the orders, especially under 15 th order. The MSC Estimate is a function of frequency with values between 0 and 1 that indicates how well two signals correspond with each other at each frequency. It is calculated with: C xy 2 Pxy( f) ( f) (4) P ( f ) P ( f ) xy yy

6 Where x and y are two signals with the same length; f is frequency; Pxx( f ) and Pyy( f ) are the power spectral densities of x and y; Pxy( f) is the cross power spectral density of x and y. It might because that this dynamometer system is controlled with a brake pressure throttle valve in the brake line. BPV measured in the test does not correlate well with the other signals. It is not appropriate to demonstrate the characteristics of hot judder and is not further analyzed in the following. Figure 5: Magnitude Squared Coherence estimate among characteristic parameters of hot judder, drag braking with1870 RPM (equivalent velocity of 225km/h) and constant brake pressure of 25bar As conclusion, it is manifested from the dynamometer test that it is feasible to identify and evaluate hot judder with accelerometer, especially for the higher orders. In addition, accelerations at caliper and caliper carrier bracket can both reflect hot judder characteristics. So for the driving tests followed, only accelerometers were used to identify and investigate the occurrence of hot judder and its behaviors in passenger cars. Design and setup of driving test Since initial braking speed and temperature as well as braking deceleration are essential to the generation of hot judder, they were varied during the driving tests, as displayed in Table 2. Each test condition was repeated for two times to check the reproducibility. For each test, it was always speeded up with full acceleration to the initial braking speed and then directly slowed down with the aimed deceleration. The initial braking speed was limited to 180km/h by the length of the test track. Variable Table 2: Test design of driving tests Initial speed Initial temperature [km/h] [ C] Deceleration [g] Nr , , 0.3, , , 0.3, , , 0.3, 0.4

7 Figure 6 shows the sensors and their installation positions that were used in the tests. Temperature of brake disks was mainly monitored by thermocouple. Wheel rotational speed that is needed for order analysis was acquired by hall sensor which gives one impulse per revolution. Vibrations at wheel brake and car body were measured with accelerometers with different ranges. Number 3 and 4 are the main mounting positions of accelerometers on wheel brakes based on the caliper type and available mounting space. To measure vibrations of car body, an accelerometer was screwed to the seat track. Main specifications of these sensors are found in Table 3. Figure 6: Sensor setup for driving tests Deceleration was controlled by experienced driver by means of ADMA (Automotive Dynamic Motion Analyzer) or Smartphone with an App that is developed by FZD. Table 3: Specifications of applied sensors Measured quantity Sensor (Producer, Type) Effective range Disk surface temperature Thermocouple, type K 0 to +900 C (Continuous) Rotational speed Hall sensor, Hamlin impulse per revolution Acceleration car body Kistler, 8612B5 ± 5 g Acceleration wheel brake Kistler, 8610B50 ± 50 g RESULTS Hot judder identification Altogether, all disk brakes of four passenger cars (vehicle A, B, C and D) were tested. As examples, spectral analyses of the test results of front left wheel brake from vehicle A and D are illustrated in Figure 7 and Figure 8. Positions of accelerometers are found in Number 3 and 4 of Figure 6 separately. In Figure 8, 2 th to 12 th and even higher orders of vibration can be easily found, by which occurrence of hot judder is testified. However, only 2 th, 3 th and 4 th orders can be dug out from figure 7. No noticeable rotational speed dependent vibrations are found from all the other

8 Acceleration [m/s 2 ] Acceleration [m/s 2 ] Acceleration [m/s 2 ] wheel brakes. It is determined that only front brakes of vehicle D were manifested with obvious hot judder, and only this car was used for further hot judder investigation. Figure 7: Spectral analysis of vehicle A, Figure 8: Spectral analysis of vehicle D, with v 0 =180 km/h, a B =3.1 m/s 2, T 0 =185 C with v 0 =180 km/h, a B =3.3 m/s 2, T 0 =205 C Reproducibility and influence of initial conditions Since the braking conditions as well as the test track in driving tests are difficult to be precisely controlled, reproducibility is critical for this experimental research of hot judder. However, three consecutive brakes under similar braking conditions illustrated in Figure 9 show that hot judder orders and their amplitudes are reproducible with acceptable tolerance in the tests. Figure 9: Comparison of three tests under similar testing conditions By further driving tests and dynamometer stopping brake tests, influences of initial braking conditions on hot judder intensity are found as follows: Initial braking speed is decisive for the occurrence of hot judder. Hot judder can only be detected from initial speed of 150 km/h, and it grows with rising initial speed. Initial braking temperature has only slight influence, with 200 o C stronger than 100 o C. Dynamometer tests coincide with driving tests. Influence of braking deceleration is evident. In driving tests, it intensifies with incising deceleration, from 0.2, 0.3 to 0.4 g. However, it peaks at 2.5 g in dynamometer tests.

9 Characteristics of hot judder From Figure 8 it can be seen that there are four resonance frequencies within 600 Hz, by which (especially the first three) vibrations are strongly amplified. Some resonance frequency bands can also be detected in Figure 7. In order to inspect the hot judder behaviors during stopping brake application at test bench, whether they are also impacted by resonance frequencies, stopping brake tests with constant brake pressures of 5 bar, 10 bar, 15 bar, 20 bar, 30 bar (equivalent decelerations about 0.10, 0.20, 0.25, 0.35 and 0.45 g) were carried out at dynamometer with the same test setup as shown in Figure 3. An example of test results is demonstrated in Figure 10. Obvious amplification bands at round 350 Hz of vibration acceleration at brake caliper carrier bracket, BTV, Waviness and DTV are found. Besides, combing the driving test in Figure 8 and dynamometer stopping brake test in Figure 10, it is found that there is no single high dominant order for hot judder in stopping brake application. All orders are amplified when passing through the resonance frequencies. Figure 10: Stopping brake application at dynamometer with v= km/h, a B =2.5 m/s 2 Moreover, it can be seen from Figure 4 that the 11 th dominant order happens also at round 350 Hz. Besides, by checking other drag braking tests with different equivalent velocities of 175 km/h and 200 km/h, it is found that the dominant orders increase with dropping velocities, but all dominant orders occur at around this resonance frequency of 350 Hz. It is deduced that the higher dominant order in drag braking at dynamometer is basically produced by the resonance frequency in wheel brake. Amplification effect of resonance frequencies on hot judder By comparing dynamometer test results of drag braking application from Figure 4 with stopping brake application from Figure 10, it is found that amplitudes of higher dominant orders of hot judder characteristic parameters in drag braking are much bigger than those in stopping brake applications, whereas lower orders do not show much difference. That is to say, amplification extent of hot judder orders in drag braking applications by resonance frequency bands is much higher than that in stopping brake applications. Because rotational speed is constant in drag braking, resonance caused by the orders around resonance frequency band will keep increase until new energy balance is achieved. While in stopping brake

10 application, the time for each order passing through the resonance frequency bands is very short due to high speeds and hence the amplification effect is not so strong. Consequently, hot judder characteristics in drag braking and stopping brake applications are intrinsically different. It is necessary to combine drag braking and stopping brake application for the experimental investigation of hot judder. For further research of the dependence of hot judder increase rate on resonance frequency, amplitude enveloping lines of hot judder characteristic parameters for some lower orders and higher orders from above mentioned drag braking and stopping brake application at dynamometer are respectively shown in Figure 11 to Figure 14. For drag braking in Figure 11 and Figure 12, the increase rate of the dominant 11 th order is much greater than the other orders. For stopping brake application in Figure 13 and Figure 14, there is no single higher dominant order and all the orders that across the resonance frequency are greatly strengthened. It is deduced that vibrations caused by hot judder would be much smaller if there was no resonance frequency in hot judder frequency range. Resonance in brake components is essential to the development of higher orders of hot judder. Figure 11: Amplitude of lower orders, drag braking with v= 225km/h, p B = 25bar Figure 12: Amplitude of higher orders, drag braking with v= 225km/h, p B = 25bar Figure 13: Amplitude of lower orders, stopping brake Figure 14: Amplitude of higher orders, stopping brake application with v= km/h, a B =2.5 m/s 2 application with v= km/h, a B =2.5 m/s 2 Transfer behaviors of hot judder to car body Through the transmission of suspension system, vibration is weakened in car body confirmed by the vibration acceleration at seat track in longitudinal direction in Figure 15 (Sensor position see Number 5 in Figure 6). However, 8 th to 12 th orders are still noticeable. Crests of these orders at seat track in the neighborhood of the same resonance frequencies of about 80 Hz and 240 Hz as at caliper are easily detected in Figure 16. It is manifested that higher orders of hot judder as well as the amplification effect of resonance frequency can be transferred to car body, which reveals the potential influence of higher hot judder orders on driving comfort.

11 Acceleration [m/s 2 ] Figure 15: Vibration at seat track, Figure 16: Transfer behaviors of hot judder, with v 0 =180 km/h, a B =3.3 m/s 2, T 0 =205 C with v 0 =180 km/h, a B =3.3 m/s 2, T 0 =205 C CONCLUSION AND OUTLOOK By the preliminarily driving tests and comparison analysis of dynamometer tests, certain conclusions about hot judder characteristics in passenger cars could be drawn and further corresponding research subjects are raised. Hot judder is basically reproducible in driving tests and initial braking speed and initial disk surface temperature as well as braking deceleration are influencing factors for the occurrence and development of hot judder. But their influences in different wheel brake systems or between driving tests and dynamometer tests are different. o Mechanisms of the influencing factors should be researched in depth, which would be much helpful to the investigation of occurrence and development mechanism of hot judder. There are some resonance frequencies within the range of hot judder, by which hot judder is extraordinarily amplified both in drag braking and stopping brake applications. There is no single high dominant order in stopping brake application, and high dominant order in drag braking at dynamometer is caused by resonance in the wheel brake system. o Examinations of the transferability of these conclusions by further more specific driving tests with other different passenger cars and dynamometer tests with other different wheel brake systems are to be executed. o Sources of the resonance frequencies are to be found out. Higher orders of hot judder and the same resonance bands were found in car body, which demonstrate the potential influences of hot judder on driving comfort. o Transfer behaviors of hot judder from wheel brakes to the driver through suspension system, steering system and braking system as well as the influences on driving comfort are to be systematically investigated. REFERENCES (1) Fischer, S., Sardá, A., Winner, H.: Effects of different friction materials on hot judder an experimental investigation. EuroBrake 2013 EB2013-TE-005, 2013

12 (2) Bittener, C.: Reduzierung des Bremsrubbelns bei Kraftfahrzeugen durch Optimierung der Fahrwerkslagerung, Dissertation TU München, 2006 (3) Sardá, A.: Wirkungskette der Entstehung von Hotspots und Heißrubbeln in Pkw- Scheibenbremsen. Fortschritt-Berichte VDI Reihe 12 Nr. 704, ISBN , Düsseldorf, 2009 (4) Meyer, R.: Brake Judder - Analysis of the Excitation and Transmission Mechanism within the Coupled System Brake, Chassis and Steering System. In 23 rd Annual Brake Colloquium & Exhibition, Orlando, Florida, October, 2005 (5) Helena, J.: Brake judder. Dissertation, Chalmers University of Technology, 2001 (6) Engel, H, G.: Systemansatz zur Untersuchung von Wahrnehmungen, Übertragung und Anregung bremserregter Lenkunruhe in PKW. Fortschritt-Berichte VDI Reihe 12, Nr Düsseldorf, 1998 (7) Schlecht, A.: Minimierung der Schwingungsempfindlichkeit von Kraftfahrzeugvorderachsen. Dissertation, Technische Universität München, 2012.