带收缩因子的PSO算法在齿轮箱故障诊断中的应用

应用带收缩因子的粒子群优化算法训练神经网络的权值矩阵，使神经网络的收敛速度大大提高，避免了其陷入局部最优解的缺陷；根据振动实验室齿轮箱实验数据，分析研究故障信号的特点，提取相应的特征参数，应用训练后的神经网络诊断齿轮箱的故障，实验表明故障诊断率较高。

Application of PSO Algorithm with Compression Factor in Gear-box Faults Diagnosis

The paper has trained the weight matrix of neural network by PSO algorithm with compression factor. The method makes the convergence rate of the neural network greatly enhance and avoids involving local extremum. According to empirical data of gear-box in the vibration lab,the fauh signal characteristic is researched and the characteristic parameter is abstracted. Applying trained neural network to diagnose gear-box faults gets good effects.