压气机失速检测的神经网络反面选择模型  

A Neural Network Model of Compressor Stall Detection Based on Negative Selection Principle

作　　者：侯胜利[1] 王威[1] 乔丽[1] 史霄霈[1] 毕志蓉[1] 

机构地区：[1]徐州空军学院,江苏徐州221000

出　　处：《电光与控制》2010年第5期38-41,共4页Electronics Optics & Control

基　　金：国家自然科学基金资助项目(60672179)

摘　　要：提出了一种航空发动机压气机失速故障检测的神经网络反面选择模型。该模型利用人工免疫系统的反面选择原理来构建神经网络检测器,通过训练将压气机的异常模式信息存储在分布的检测器中,根据检测器的激活来发现故障。通过混沌时间序列的异常检测仿真实验,研究了模型参数对故障检测性能的影响。某型涡喷发动机失速检测实验表明,该方法对压气机失速信号的模式特征具有较强的分辨能力,同时证实神经网络检测器比常规的二进制编码检测器具有更好的故障识别能力。A neural network model of aeroengine compressor stall detection based on negative selection principle was proposed.The principle and structure of the model were presented,and its training algorithm was derived.Through neural network training,the information of abnormal patterns was stored in the distributed neural network-based detectors.This model had distinguished capability of adaptation,which was well suited for dealing with practical problems under time-varying circumstances.A fault can be found out through the relevant activated detectors.Simulations of anomaly detection in chaotic time series were carried out to investigate the effect of model parameters on the capability of fault detection.Experiments of compressor stall detection in certain type of turbo engine demonstrated that the proposed method can achieve precise discrimination to pattern features of stall signals,which also testify that neural network-based detectors have better recognition ability than binary encoding detectors.

关 键 词：航空发动机 压气机 故障检测 反面选择原理 人工神经网络 

分 类 号：V235.113[航空宇航科学与技术—航空宇航推进理论与工程]