A fault prognosis scheme based on SVM for the flight control system and PHM system

Flight control system is a kind of complicated and various control system, in which some faults cannot be easy found. Moreover the fault trend is likewise not predictable. Using Support Vector Machines (SVMs), the faults not only are diagnosed but also are prognosed correctly. SVMs have originally been used for classification purposes but their principles can be extended easily to the task of regression and time series prediction. In this paper, the fault prognosis scheme of the flight control system was established through Support Vector Regression (SVR) in the solution follows from the fault detection scheme of the modern fighter plane. This scheme which consisted of the modules of the actuator, the steering face and fault diagnosis is discussed to overcome the problem of fault trend. It shows how to obtain weighted estimates for regression by applying SVM/SVR. The paper also discusses a framework combined SVM with HMM for the Prognostic and Health Management (PHM) of the flight control system. This method of this paper is illustrated for non-stability fault signals and demonstrates how to obtain predicted estimates with selection of an appropriate HMM units, in the case of future PHM design.