基于GA-BP神经网络的轨道结构病害诊断方法研究

Research on track structure disease diagnosis method based on GA-BP neural network

针对BP神经网络在运算过程中极易陷入局部极小值,且对轨道结构病害识别准确率不稳定的问题,提出一种基于遗传算法优化BP神经网络的轨道结构病害诊断方法。通过建立车轨耦合动力学模型,仿真得到正常、轨枕空吊、道床松散和道床板结4种服役状态信号,利用变分模态分解和多尺度排列熵方法对振动信号进行特征提取,并组建高维特征向量,作为BP神经网络模型的输入。通过遗传算法优化BP神经网络模型,对比优化前后的识别准确率,充分证明了基于遗传算法优化BP神经网络的方法,在轨道结构病害识别及诊断上的有效性。

Aiming at the problem that BP neural network is easy to fall into local minimum in the operation process and the accuracy of track structure disease identification is unstable, a track structure disease diagnosis method based on Genetic Algorithm Optimized BP neural network is proposed. By establishing the vehicle rail coupling dynamic model, the four service state signals are obtained by simulation. The characteristics of vibration signals is extracted by using variational modal decomposition and multi-scale arrangement entropy method, and a high-dimensional feature vector is established as the input of BP neural network model. The BP neural network model is optimized by genetic algorithm, and the recognition accuracy before and after optimization is analyzed, which fully proves the effectiveness of the method of Optimizing BP neural network based on Genetic Algorithm in track structure disease recognition and diagnosis.