The design of a model-based fault detection and diagnosis system is discussed. It was illustrated how the functional integration of analytical process knowledge can expand the functionality of a mechatronic system by providing the driver with information about the state of the braking system and by warning him of arising errors, which might eventually lead to accidents. First, a model of the hydraulic braking system was derived. This development was based on the analogy between hydraulic and electric circuits and resulted in an eighth order non-linear system. All parameters have been identified from measurements at a braking system test-bed. Comparisons of simulations and experimental data have proven the high fidelity of the model. Based on this model, a fault detection and diagnosis system has been implemented, which allows early detection, diagnosis and localization of small, drifting faults such as leakages or air entrapments in the hydraulic system. This fault detection and diagnosis procedure supervises the amount of brake fluid consumed by the brake system. The reference value is generated by utilizing the aforementioned brake system model, which is fed by pressure sensors mounted at the master brake cylinder chambers. The effective amount of fluid consumed is calculated from the displacement of the vacuum brake booster diaphragm. Two common types of faults are classified, air inclusions and leakages. Classification is accomplished by employing a correlation analysis scheme. The main advantage of this method is the small computational expense resulting in real-time capability of the algorithm. The supervision system only requires data of sensors which are already available in modern passenger cars. Due to the likeness of hydraulic and electro-hydraulic braking systems, it is easily possible to use the presented methods to cover electro-hydraulic braking systems as well. Here, one can use more affluent sensor signals, since all four wheel brake pressures and the pressure and volume stored in the accumulator are sensed. These additional information allow wheel-individual fault detection and diagnosis. Upon detection, diagnosis and localization of faults, the system can initiate countermeasures such as for example inform the driver about necessary maintenance or repair work.
A model based supervision system for the hydraulics of passenger car braking system
Modellbasierte Prozessüberwachung der Bremshydraulik in einem Personenkraftwagen
2002
6 Seiten, 7 Bilder, 9 Quellen
Conference paper
Storage medium
English