论文部分内容阅读
引言本文故障诊断的目的是在故障参数 X(性能参数)和发动机测量参数或测量参数的某种形式 Y 之间建立联系。X 和 Y 之间是复杂的非线性关系。神经网络有描述复杂非线性关系的能力,能够在 X 和 Y 之间建立直接的网络映射,这样,不需要依靠数学仿真模型就可直接进行诊断求得 X。目前主要是利用 BP 神经网络对发动机的稳态测量参数进行定性分析,隔离故障到单元体。例如美国 F-16飞机上已采用 BP 神经网络进行发动机故障的定性诊断。但在实际问题中,还要知道发动机故障的严重程度,才能更好地进行视情维
Introduction The purpose of this troubleshooting is to establish a link between the failure parameter X (performance parameter) and the engine measurement parameter or some form of measured parameter Y. Between X and Y is a complex non-linear relationship. Neural networks have the ability to describe complex nonlinear relationships and establish direct network mappings between X and Y so that X can be directly diagnosed without relying on a mathematical simulation model. At present, the main use of BP neural network qualitative analysis of the steady-state measurement parameters of the engine, fault isolation to the unit body. For example, the United States F-16 aircraft has been using BP neural network for qualitative diagnosis of engine failure. However, in practical problems, it is also necessary to know the severity of the engine failure in order to be better visual emotion