Characterization of Detonation Phenomenon Signal Using Adaptive Filtering and Power Estimation

The automotive area has improved its technology in order to make combustion engines achieve higher rates of delivered power, fuel economy and lower pollution emissions. Such factors are directly influenced by the detonation phenomenon or knock, that happens under high temperature and torque requests and it may be detected by a knock sensor and proper signal processing techniques. The accurate identification of detonation, such as its intensity and time length, allows a better identification of threshold conditions that cause this phenomenon. Moreover, a special case is the flex fuel engines (work with ethanol and/or gasoline), in which the detonation identification may show which kind of fuel is being used, allowing an optimized engine management. This work focus on the identification and characterization of the knock phenomenon signal utilizing power estimators and adaptive filters. First, a simple signal model is considered. In the sequel, it is show that adaptive techniques lead to efficient detectors.