【作者】 施浩；

【导师】 刘成良； 黄亦翔； 林志刚；

【作者基本信息】 上海交通大学 ， 机械工程（专业学位）， 2017， 硕士

【摘要】 随着我国民航机队规模不断的扩大,民用航空的安全问题变得越发重要。飞机发动机的健康性能评估对飞行安全至关重要。但飞机发动机工况复杂,往往难以对其进行准确的性能评估。本文提出了一种基于多工况聚类的飞机发动机健康状态评估方法。首先对工况进行自动划分,再从大量传感器数据中选取与发动机性能衰退趋势密切相关的传感器变量进行分析并提取特征,之后定义Logistic模型,最后给出对飞机发动机的健康状态评价。该方法在NASA提供的飞机发动机传感器数据上进行验证,不仅能够对其健康性能状态进行评估,而且可以对发动机故障进行预测和故障排除。结果表明该方法不仅结果可靠,而且具有一定实用性,可为飞机发动机的剩余使用寿命预测提供一种可行且有效的理论方法和手段,对提高飞机发动机的安全性和可靠性具有重要价值。更多还原

【Abstract】 With the fast and steady expansion of China’s civil aviation fleet,Civil aviation security issues become increasingly important.The performance of the health assessment of aircraft engine is of vital importance to the safety of flights.But the aircraft engine condition is complex,and often too difficult to be evaluated accurately.This paper presents a health evaluation method based on automatic clustering of of the aircraft engine conditions.Firstly,from a large number of sensor data,the sensor variables closely related to the engine performance deterioration trend are selected for analysis and feature extract io.Secondly,the Logistic model is defined to evaluate the aircraft engine health condition.Finally,the method for evaluating the performance of the engine health status is validated on the aircraft engine sensor data through a NASA verification procedure.It can not only evaluate the the performance of the health status,but also on the engine fault prediction and fault elimination.The accurate prediction of the remaining useful life of the engine has important value to improve the safety and reliability of aircraft engine.更多还原