【作者】 宋述芳； 何入洋；

【Author】 Shufang Song;Ruyang He;School of Aeronautics,Northwestern Polytechnical University;

【机构】 西北工业大学航空学院；

【摘要】 航空发动机的正常运行是飞行安全的重要保障。然而航空发动机结构复杂,其失效模式多样,存在竞争关系及相关性,这使得航空发动机性能退化程度的预测困难重重。本文利用某航空发动机的性能监测数据,考虑了涡轮后燃气温度、燃油消耗量偏差、高压转子转速、滑油压力、高\低压转子振动值、在翼时间等因素的影响,采用BP神经网络算法对航空发动机性能退化程度进行分析和预测。依据各因素的重要性分析结果,仅用重要因素作为输入变量,可得到与工程实际一致的发动机性能退化结果,准确、客观地反应了航空发动机性能变化。更多还原

【Abstract】 Aero-engine is the most important subsystem of the airplane. The normal operation of aero-engine can ensure the safety of the airplane. There are many kinds of failure mode with correlation for the aero-engine. Thus, it is difficult to predict the performance degradation of aero-engine. In the paper, Back Propagation(BP) neural networks are employed to predict the performance degradation of aero-engine on the basic of the detection data with seven influence factors, such as exhaust gas temperature, the deviation of fuel consumption and so on.Importance measure analysis is presented to reduce the dimension of influence factors. The results demonstrate that BP neural networks can give accurate prediction values for the performance degradation of aero-engine and describe the change of the aero-engine’s performance parameter accurately and objectively.更多还原