Abstract:

The health assessment method of oil system based on SOM neural network is researched. SOM neural network is used to define the health parameters under different system conditions comparing with the health parameters under fault conditions. Different health levels are divided to provide a basis for health judgement of oil system in a timely manner. The normal samples and fault samples are used to verify effectiveness of SOM neural network.SOM neural network shows lots of advantages when it is applied to health assessments such as good generalization ability,fast learning rate,intuitive result of output layer and so on,providing new methods and ideas for health assessment of oil system.

Key words: aero-engine oil system, SOMneural network, health assessment, QAR