Eventos Anais de eventos
DINAME 2023
XIX International Symposium on Dynamic Problems of Mechanics
Experimental Estimation of Brake Pad Wear for Passenger Cars Based on Longitudinal Dynamics
Submission Author:
Kenneth Jensen , Hovedstaden , Denmark
Co-Authors:
Kenneth Jensen, Ilmar Santos
Presenter: Kenneth Jensen
doi://10.26678/ABCM.DINAME2023.DIN2023-0103
Abstract
This paper considers the problem of estimating brake pad wear of a passenger car, based on usage data and vehicle longitudinal dynamics. To broaden the applicability of the proposed methods, only basic usage data is considered, using inertial measurement unit (IMU), global navigation satellite system (GNSS), and standard vehicle on-board diagnostics (OBD) data. This excludes the use of manufacturer specific vehicle data. The paper considers the following problems: (I) estimation of vehicle drag, (II) estimation of engine braking, (III) prediction of vehicle braking state, and (IV) estimation of brake pad wear. Results from (I) and (II) are used for (III), predicting the braking state by comparing measured vehicle acceleration with accelerations estimated from longitudinal dynamics. The brake pad wear is estimated using linear regression with a single feature: the product of distance travelled and estimated longitudinal braking force during braking. Data was sampled from a modern passenger car, for fitting and validation of models. For the wear estimation (IV), the thickness of a brake pad of the car was measured. A total of 18 measurements over the course of approximately 6,000 km were carried out. The drag and engine braking estimates from (I) and (II) show good agreements with measurements. The predicted braking states were also validated experimentally, using object tracking of the brake pedal. Finally, the linear regression model appears to indicate an almost linear relation between the used feature and the measured wear with a R2=0.97.
Keywords
Wear prediction, Longitudinal dynamics, parameter estimation, maintenance, experimental validation
