论文部分内容阅读
针对时差定位法受不同模式波速度差异及波形传播畸变等因素影响的问题,将神经网络技术应用到声发射源定位中。在通常的BP小波神经网络中,BP算法实质上是一种基于梯度下降法的局部搜索算法,易使网络陷入局部最小值而使得搜索成功概率较低。作为改进,利用粒子群算法对小波神经网络中的参数进行优化,然后再利用基于粒子群优化的小波神经网络进行声发射源定位。仿真实验结果表明,选择合适的网络结构和输入参数,粒子群优化算法可以准确定位碰摩位置,且计算更加简单有效,具有良好的应用前景和进一步研究的价值。
Aiming at the problem that the time-difference positioning method is affected by the difference of wave velocity and wave propagation distortion, the neural network technology is applied to the location of acoustic emission sources. In the usual BP neural network, the BP algorithm is essentially a local search algorithm based on the gradient descent method, which easily causes the network to fall into a local minimum and makes the probability of successful search lower. As an improvement, the particle swarm optimization is used to optimize the parameters of the wavelet neural network, and then the wavelet neural network based on PSO is used to locate the AE sources. The simulation results show that the particle swarm optimization algorithm can accurately locate the rubbing position and select the appropriate network structure and input parameters, and the calculation is more simple and effective, and has a good application prospect and further research value.