摘要
本文运用时间序列方法研究工作状态下回转式压缩机的振动故障。对压缩机的振动信号进行建模，并提出用特征值归一化残差平方和（ＮＲＳＳ）作为识别故障状态的指标。实验结果证明：ＮＲＳＳ指标对回转式压缩机故障诊断是十分有效的。
This paper used time series method to analyze the vibration faults of working compressor, and the time series model of vibration signal was setup in the paper. The characteristic normal residue square sum (NRSS) is used to identify the fault state. The result of experiment shows that NRSS is effective for compressor fault diagnasis.
