Model Based Approach for Fault Detection and Isolation in a Four Stroke Engine Using an Acoustic Signal
Pages : 1342-1347
The paper deals with the problem of fault detection in an automobile engine employing an artificial neural network (ANN). The Fault Detection and Isolation (FDI) is not an easy task for an inexperienced mechanic or driver because it needs a lot of knowledge and experience. Many times, the trial and error approach is seen to be applied to detect the fault and because of that, the engine may get more damaged instead of getting repaired. To overcome such type of problem, the new approach has been suggested to diagnose the fault correctly without opening the engine. Therefore, this paper presents the model based technique to detect the Air Filter Fault (FF) and Spark Plug Fault(SP) in a Four Strokes Engine using a single sensor. The Hero Honda Passion Four Stroke (HHPFS) engine is used for experimentation. The Artificial Neural Networks have been employed to classify the faults correctly. Performances of Multilayer Perception Neural Network (MLP NN) and Support Vector Machine (SVM) have been compared on the basis of Average Classification Accuracy (ACA) and finally, the optimal Neural Network has been designed for the best performance.
Keywords: Automobile Engine, Artificial Neural Network, Classification Accuracy, MLP & SVM.
Article published in International Journal of Current Engineering and Technology, Vol.3,No.4(Oct- 2013)