中文摘要：

在简要介绍了Volterra级数基本理论的同时，应用遗传算法对Volterra级数的核进行了辨识，并取得了良好的辨识效果。与传统的基于信号处理的诊断方法不同，该方法将基于系统模型的Volterra级数非线性故障诊断方法应用于转子轴承系统的故障诊断，研究了转子在正常状态和碰摩状态下的起车过程中Volterra级数核的变化。实验结果验证了该方法的可行性与有效性。

英文摘要：

This paper briefly introduced the basic theory of Volterra series. Then, it identified the Volterra kernels by using genetic algorithm （GA） and achieved a satisfactory identification result. Being different from the traditional diagnosis schemes based on signal processing approach, the Volterra series method based on system modeling was employed in the area of fault diagnosis for rotor --bearing system. We investigated the changes of Volterra kernels under normal state and rubbing state during run--up stages. Finally,experimental results certify its feasibility and effectiveness.