FTA-SVM-based fault recognition for vehicle engine

Engine testing technology has made great development and gathering the engine data of failure becomes more and more easily. The engine fault recognition method based on data driving has made rapid development. The support vector machine (SVM) is currently a well-known machine learning technique. It has been applied to deal with range of fault recognition problems due to its unique advantages. There must to have two elements before use SVM: the one is a definite fault pattern, the other one is the mapping model of feature data and fault pattern for training SVM model. To tackle this problem, this paper raised a new method of engine fault recognition which combines the support vector machine (SVM) with fault tree analysis (FTA). Experimental results show that the method of FTA-SVM-based fault recognition is more effective and reasonable.