Options
1990
Conference Paper
Titel
Requirements for new diagnostic methods for engines in automotive industry
Abstract
The methods for analysis of angular acceleration of pressure and vibration signals presented in this paper have the advantage of non-contacting signal acquisition and of simple methods for signal processing and feature extraction. The analysis of angular acceleration provides a high sensitivity for the detection of a number of mechanical faults and of faults in the combustion process. It is well suited for the detection of misfiring. In this paper only global features defined over whole working cycles were used. By extracting features from sections of the working cycle a cylinder-specific recognition of misfunctions is feasible. Again statistical methods are well suited. The acceleration signal is a good starting point for the detection of malfunctions, but for the differentiation between several possible causes additional faultspecific methods have to be applied. A wide range of diagnostic methods is known and applied in automotive industry. Generally the exploitation of pressure and vibration signals for diagnostic purposes can be ameliorated. These signals are helpful e.g. in the recognition of faulty valve clearance, of valve leakage and of mechanical faults at injection nozzles. Valve leakage and faults at injection nozzles are factors which also influence the angular acceleration. Therefore methods for the recognition of these faults only were decribed. They have in common a processing strategy based on synchronization to the crankshaft's angular position. Synchronous averaging of signals results in noise reduction for the recognition of fine details. The replacement of comblicated static leakage tests by features derived from pressure sensors brings the advantage of simplified tests with acquisition of statistical variances.