A Study on Practical Condition Monitoring System for a 2-Spool Turbofan Engine Using Artificial Intelligent Algorithms

摘要

The engine performance and health monitoring systems using the artificial intelligent methods such as Neural Networks, Fuzzy Logic and Genetic Algorithms have been developed to monitor the engine condition. Among them the Neural Networks is mostly used to engine fault diagnostic system due to its good learning performance, but it has a drawback due to low accuracy and long learning time to build learning data base, and the Fuzzy logic has good characteristics to identify the fault patterns but a difficulty to quantify the degradation. In practical use of the health monitoring systems, the sensor noise and biases reduce greatly the diagnostic system's accuracy. However the intelligent condition monitoring method using Genetic Algorithms (GA) has high accuracy to diagnose the engine faults even though the measurement signals containing sensor noise and biases. Therefore this work applies the GA to diagnose 2-spool turbofan engine AE3007H for a High Altitude Long Endurance (HALE) UAV. Through comparative studies between the intelligent diagnostic method using GA and the model based diagnostic methods using linear and non-linear Gas Path Analysis (GPA) methods, it is found that the proposed GA diagnostic method has much higher accuracy on all types of fault cases not only single and multiple fault cases but also consideration of sensor noise and biases.