中文摘要：

齿轮传动作为机械传动主要形式之一,应用极为广泛,设备在高速运转下,一旦发生齿轮断齿故障将会带来巨大的经济损失以及人员伤害,为让损失降到最低,需要做到故障强度早知道,因此设备故障强度预测显得尤为重要。单通道预测方法由于获取振动信息不完善,导致预测结果一致性差,从而不能很好地实现故障强度的预测。通过全矢谱获得的频谱结构具有唯一性的特点,能够很好地弥补单通道的不足,在此基础上,将时序预测方法 ARMA模型与全矢谱技术相结合,提出了全矢-ARMA模型预测方法,并把该方法应用到齿轮断齿故障强度预测研究中。实验表明,该方法预测齿轮断齿故障强度结果与实际较吻合。

英文摘要：

Gear transmission, one of main mechanical transmission, is extensively used in mechanical industry. In the event of gear broken fanlt under high-speed operation, which will bring huge economic losses and personal injuries. In order to reduce losses to a minimum and know fault early, equipment fault strength prediction is particularly important. Considering incomplete vibration information that leading to poor consistency of predictive results, single-channel prediction method cannot realize accurate prediction of machine fault strength. While by obtaining spectral structure with unique characteristics, full vector spectrum （FVS） can well make up for the deficiency of single-channeL On this basis, the prediction method of FVS-ARMA model was proposed, which combined ARMA model with full vector spectrum technology. It was applied to predict the gear broken fault strength. Experiments show that prediction results of this method were identical with the practical effects.