航空发动机气路传感网络的故障节点挖掘仿真

Aeroengine Gas Path Sensor Network Fault Node Mining Simulation

航空发动机的气路传感网络故障检测,是通过与发动机气路传感网络产生的基准线进行比较,产生的偏差作为阀值,作为检测故障的基础。由于航空发动机气路传感网络的高度非线性、噪声、测量不足等因素的影响,检测的基准线也呈现波动状态,传统的故障挖掘方法,需要通过做一些必要的假设,设置各种可能故障和性能恶化的范围,规定故障挖掘空间以降低波动干扰,但是上述假设降低了检测的精度。提出一种航空发动机气路传感网络的故障节点挖掘方法。计算故障传感节点的特征,为故障传感节点的挖掘提供数据基础。利用神经网络建立故障传感节点挖掘模型,完成故障传感节点的挖掘。实验结果表明,利用改进算法进行航空发动机气路传感网络的故障节点挖掘,能够有效提高检测的准确性和挖掘效率。

The paper proposed a failure sensor node mining method for aero-engine pneumatic sensor network. The characteristics of the failure sensor nodes were calculated to provide data basis for failure sensor node mining. Neural network was used to establish failure sensor node mining model, to complete the failure sensor node mining. Experimental results show that using the improved algorithm for failure sensor node mining of the aero-engine pneu- matic sensor network, can effectively improve the detection accuracy and the mining efficiency.