CPSO优化PNN的陀螺故障诊断方法

A gyroscope fault diagnosis method based on PNN optimized by CPSO

针对惯性导航系统中的陀螺仪输出信号非线性、故障特征不明显的问题,为提高惯导系统中惯性器件的故障诊断正确率,提出一种基于改进粒子群算法(PSO)优化概率神经网络(PNN)的陀螺信号故障诊断方法。首先,针对光纤陀螺运行过程中常见的四种故障信号,建立数学模型并进行小波变换提取其故障特征系数;其次,使用Cubic混沌映射以及非线性递减的惯性权重系数对粒子群进行粒子更新,并用于概率神经网络的最优平滑因子选择;最后,训练概率神经网络对陀螺仪故障信号进行分类和诊断。离线测试结果表明,CPSO算法优化的PNN网络针对四种故障分类的平均正确率达到95.8%。

To address the problem that the output signal of the gyroscope in an inertial navigation system is nonlinear and that the fault characteristics of this signal are not obvious and improve the correct rate of fault diagnosis accuracy of inertial devices in inertial navigation systems,a gyroscope fault diagnosis method based on improved particle swarm optimization algorithm(PSO)and probabilistic neural network(PNN)is proposed.Firstly,for the four common fault signals during the operation of gyroscope,mathematical models are established and fault feature coefficients are extracted by wavelet transform.Secondly,the particle update of particle swarm uses Cubic chaos mapping and nonlinear decreasing inertia weight coefficients in a way,and the improved particle swarm optimization algorithm is used for optimal smoothing factor selection of probabilistic neural networks.Finally,the probabilistic neural network is trained to classify and diagnose the gyroscope fault signals.The offline test results show that the CPSO-PNN network achieves an average correct rate of 95.8%for the four fault classifications.