Artificial Intelligent Fault Detection of A Turboshaft Engine for Smart UAV Using SIMULINK Performance Model

摘要

It is not easy that all engine faulted conditions can be monitored only by the conventional model based fault detection approach such as the GPA (Gas Path Analysis) method because type and severity of the fault are various and complex. Therefore, in this study, a model-based diagnostic method using the Neural Network algorithms was proposed for the fault detection of the PW206C turbo shaft engine, which was selected as a power plant for the tilt rotor type Smart UAV (Unmanned Aerial Vehicle). A precise performance model is needed to perform the model-based diagnostics. In order to obtain the component maps for the performance model, a new component map generation method which may identify component characteristics conversely from the limited performance deck data provided by the engine manufacturer using the system identification method and the genetic algorithms was used. The PW206C turbo shaft engine performance model with the generated component maps using SIMULINK carried out the performance analysis at various operating conditions such as flight altitude, flight Mach number and gas generator rotational speed changes. After obtaining the performance deterioration learning data of compressor, compressor turbine and power turbine based on performance data at various operation conditions, it was trained by the BPNN (Back Propagation Neural Network) method. According to the fault detection analysis results, it was confirmed that the proposed fault detection method could find well the fault of compressor, compressor turbine and power turbine at on-design point as well as off-design point conditions.