Research On Aircraft Engine Fault Diagnosis Based On Support Vector Machine

Aircraft engine is a complex mechanism which works in high rotate speed, high load andhigh temperature circumstance. The possibility of engine faults increases with theaccumulation of time. Diagnosing the faults of the engine timely and efficiently could notonly reduce the maintenance cost and ensure engine working stable and reliable, but alsoreduce accidents and avoid larger personnel and property losses. It has become one of the hotspot of current research to diagnose the faults with the machine learning method which basedon artificial intelligence to construct classification learning models.At present, the neural network is one of the artificial intelligence application in aircraftengine fault diagnosis, but neural network training needs a mass of fault samples. And neuralnetwork has high dimension and local minima problems which influence aircraft engine faultdiagnosis ability. An aircraft engine fault diagnosis method based on support vector machines(SVM) is presented. After extracting optimized parameters of kernel principal componentanalysis (KPCA) via particle swarm optimization algorithm from a certain type data ofturbofan engine sensors, the kernel principal component analysis (KPCA) method is adoptedto reduce its data dimension and obtains the fault features. The support vector machine (SVM)fault diagnosis model is established by using the fault features. The diagnosis effect could betested through the test sample. The experimental results show that this support vector machine(SVM) method has high diagnostic accuracy and a certain anti-interference ability. Therefore,it can be applied to the actual aircraft engine fault diagnosis research.The system of Aircraft engine the fault diagnosis is designed based on Microsoft VisualC++6.0and SQL server2005database in this paper. Three different modules are included inthis system, including data management module, Diagnosis algorithm training module andfault diagnosis module. The diagnosis algorithm is designed by the C++and correspondingfault data are saved in SQL server database. The interface of the system is designed by VisualC++6.0.