Engine Fault Detection Using Wavelet Packet Transforms 2013-01-1705
The engine fault detection problem is of special interest in the automotive industry due to its direct implementation in the diagnostics of typical engine problems as well as engine preventive and predictive maintenance measures. The procedure of identifying the type and location of an engine fault can be very costly and time consuming. This paper is dedicated to the identification of more than 26 different engine faults using multi-level signal processing technique and an exhaustive search in a predefined database. This database can be easily modified and/or updated to include new types of faults; it also gives the flexibility of providing warning signals of unidentified problems whenever it detects one.