Misfire Fault Diagnosis Of Engine Based On EEMD

In recent years,the automobile ownership of our country presents sustained and rapid growth situation.The popularity of automobiles brings convenient to people,at the same time,also brings problems,such as traffic safety,energy consumption and environmental pollution.The engine is the core part of the automobile.The engine misfire fault causes the insufficient of power and directly affects the driving safety.The oil and gas mixture without sufficient combustion will cause fuel waste and excess emissions when misfire happens.Therefore,the engine misfire fault diagnosis has important practical significance.The crankshaft speed signal is relatively easy to get,and it contains misfire fault information.The misfire fault diagnosis can be implemented by analyzing of the crankshaft speed signal.However,the crankshaft speed signal is non-stationary,and its statistical properties change over time.Moreover,the crankshaft speed signal is a result of several physical processes,and it also contains other interference.Therefore,it is difficult to extract the fault information from it.Aiming at the above situation,firstly it is proposed in this paper that the empirical mode decomposition(EEMD)is used to analysis the crankshaft speed signal and extract the fault information.The EEMD method is suited to process the non-stationary signal,and it effectively resolved the multi-mode problem so that it has strong anti-interference ability.Secondly,the mode mixing problem existing in EEMD easily causes the loss of fault information.Therefore,the method for multi-mode recognition and mode reconstruction is proposed in this paper,which can resolve the multi-mode problem so as to extract the fault information more completely.Finally,since the fault characters change with the engine speed in nonlinear form and the mapping relation from fault information to fault types is also nonlinear,a nonlinear estimation model should beestablished in order to realize the misfire fault diagnosis under any speed condition.Therefore,a nonlinear fault classifier is built by the least squares support vector machine(LSSVM).The simulation experiments prove that EEMD is suitable for processing the non-stationary crankshaft speed signal,and it can be used as an off-line misfire fault detection method;At the same time,they also verify that the misfire fault diagnosis method proposed in this paper has better accuracy,anti-interference performance and generalization ability,which can be used as an on-line misfire fault diagnosis method under any speed condition.