摘要
通过优化故障检测的方法,可以提高大推力氢氧补燃发动机故障检测的精确性能;BP神经网络是基于梯度的方法确定权值与阈值,而梯度下降法本身就很容易受到局部极小点的影响;遗传算法是一种多点搜索的优化方式,具有良好的全局寻优能力,可以优化BP网络的不足;基于发动机试车过程中测得到的流量、温度、压力等信号,应用GA-BP神经络构建发动机工作过程的非线性辨识模型,对大推力氢氧补燃发动机可能的运行故障进行检测,由试车数据的仿真结果可知,该算法达到了较好的故障检测效果。
In order to improve the performance of fault detection of high-thrust LOX/LH2 staged combustion cycle engines,a prediction method is proposed.BP neural network is based on the gradient method to determine the weight and threshold,and the gradient descent method itself is easy to be affected by local minimum.Genetic algorithm(GA)is a multi-point search optimization method,which has good global optimization ability and can optimize BP neural network.Based on the flow,temperature,pressure and other signals measured during the engine test run,the GA-BP neural network is used to construct a nonlinear identification model of the engine working process,to detect the possible operating faults of the high-thrust hydrogen-oxygen supplementary combustion engine.The simulation results show that the algorithm achieves a better fault detection effect.
作者
李宁宁
武小平
薛薇
胡慧
徐志强
LI Ningning;WU Xiaoping;XUE Wei;HU Hui;XU Zhiqiang(Beijing Aerospace Propulsion Institute,Beijing 100076,China)
出处
《计算机测量与控制》
2022年第8期14-18,43,共6页
Computer Measurement &Control
关键词
大推力氢氧补燃发动机
故障检测
神经网络
遗传算法
high-thrust LOX/LH2 staged combustion cycle engines
fault detection
neural network
genetic algorithm
作者简介
李宁宁(1993),男,河南商丘人,硕士研究生,工程师,主要从事航天发动机控制及健康管理系统设计方向的研究。
