Sign up for our weekly email to stay on top of the latest news and insights!
Source: Applus IDIADA
This TBR Technical Corner is the second in a series of three articles on the impact that different friction materials have on corrosion-induced brake judder —both at system and vehicle levels— during the removal of rust by Narcis Molina, Project Manager, Braking Systems in Applus IDIADA.
The first part of this series can be viewed by clicking HERE.
As explained in the first part of this TBR Technical Corner, this article presents a methodology that incorporates both dynamometer and vehicle tests into the investigation of the impact that different friction materials have on corrosion-induced brake judder —both at system and vehicle levels— during the removal of rust. In that first part, the problem has been presented and the methodology, introduced.
Results
Figures 3 and 4 show the evolution of the initially-corroded disc friction ring surface for both sets 2 and 4 —NAO material, respectively tested in the dynamometer and the vehicle. Note that, after the first snub, the oxide layer is significantly removed —indeed, no visual differences are observed from the 5th / 10th stop onwards.
Figure 5 shows braking torque and hydraulic pressure, in the time domain, for the first and last (30th) snubs of the dynamometer and vehicle tests. Note the oscillations that both signals present over the steady-state section of the braking applications; at the end of the corrosion-cleaning test, their amplitudes are significantly reduced.
It is precisely the evolution of the observed fluctuations that needs to be studied in detail. The analysis of the main signals —basically, hydraulic pressure and braking torque— can be broken down into three main steps:
- Pre-processing of the signal (Figure 6):
- rpm low-pass filtering.
- Signal offset correction.
- Signal high-pass filtering.
It allows the steady-state oscillation to be isolated from the braking build-up section.
2. Order analysis (Figure 7), where the frequency associated with the rotational speed of the corrosion-induced vibration —forced by the irregular shape of the rotor— is studied.
It allows the contribution of each order to the global value of the response to be determined.
3. Frequency spectrum analysis (Figure 8), which is based on the Fast Fourier Transform (FFT) of the given signal. Results are plotted as colourmaps, a three-dimensional representation in which the X-axis is the signal frequency (Hz), the Y-axis shows the brake rotor rotational speed (rpm) and the Z-axis indicates the vibration amplitude (e.g. for braking torque, Nm).
The frequency spectrum analysis allows any type of vibration to be investigated —not only accelerations, but also the fluctuations already observed in pressure or torque, commonly known as Brake Pressure Variation (BPV) and Brake Torque Variation (BTV).
Table 10 details the main settings for the FFT. Resolution and overlap percentage have been chosen to guarantee an appropriate number of FFTs (about 25) throughout each brake snub, while keeping a good resolution of the colourmap. Given that all tests have been performed from 60 km/h to 0 km/h, the scale frequency has been set up to 50 Hz.
The FFT uses the Root Mean Square (RMS) amplitude format. Furthermore, for each window (i.e. increment of rotational speed), the RMS value of the corresponding spectrum —called the overall level— is calculated to obtain the overall level of energy across the frequency range of interest. This overall level, tracked versus speed, elucidate how the amount of energy in the signal changes during the braking maneuver.
Figure 9 shows the results for the BTV over the 3rd snub of the 4th set (NAO; vehicle testing). Given that wheel torque transducers are mounted on both sides of the front axle, the analysis is presented in parallel for the FL and FR corners. Figures 9a and 9b represent the colourmap, whilst 9c and 9d show the distribution of orders. Note, in magenta, the mentioned overall order —i.e. the RMS value for each window.
Accordingly, for every single snub, the presented vibration amplitude corresponds to the maximum value of the RMS overall level —regardless of the rotational speed. For instance, in the example shown in Figure 9, the maximum values are 31.3 and 16.5 Nm (respectively, FL and FR), which average 23.9 Nm for the front axle. This is precisely the BTV magnitude reported in Figure 10d (set 4; 3rd snub).
Figure 10 shows the BTV and BPV evolution throughout the corrosion-cleaning test for all 4 sets. In terms of dynamometer testing, a sharp decay is observed in BPV and BTV during the first snubs, while all vehicle-based results tend to diminish more smoothly, approximately, till the 10th stop. A similar tendency is observed for the longitudinal acceleration (Figure 10e) of the front calipers —accelerometers are only instrumented in the vehicle.
An expert driver evaluates the main rust-induced judder vibrations sensed across the vehicle. Figure 11 presents the subjective evolution for each set—where the rates given to the steering wheel, the brake pedal and the floor pan have been averaged. A minimum acceptability threshold is stablished at 7.0. Set 3 (LS) is consistently perceived as performing better, with an overall grade of 8.8 (whereas set 4, NAO, averages 7.6).
As seen in Figures 10b, 10d and 10e, set 4 (NAO) repeatedly suffers from larger pressure, torque and acceleration fluctuations throughout the test, a trend that agrees with the corrosion-induced judder subjective evaluation presented in Figure 11. This relationship is confirmed in Figure 12, where both BPV and BTV are correlated against the averaged vibration ratings.
About Applus IDIADA
With more than 25 years’ experience and 2,450 engineers specializing in vehicle development, Applus IDIADA is a leading engineering company providing design, testing, engineering, and homologation services to the automotive industry worldwide.
Applus IDIADA is located in California and Michigan, with further presence in 25 other countries, mainly in Europe and Asia.
