工业机器人传动轴振动信号多源融合GRP-SVM诊断

Multi-Source Fusion GRP-SVM Diagnosis of Industrial Robot Transmission Shaft Vibration Signals

工业机器人传动系统的运行稳定性直接影响到使用效率,进而节约工业成本。利用高斯随机映射-支持向量机(GRP)来实现故障,达到对多源信息降维效果。综合考虑按照支持向量机(SVM)方法实施分类,设计了相应的实验方案,经验证可知上述方法满足实际可行性。研究结果表明:当样本数量达到800~2100时,获得理想的故障诊断结果。振动信号融合可以获得比Z方向更高诊断精度,获得更准确预测结果,精度达到了99.89%。该研究对提高工业机器人传动系统运行稳定性具有很好的理论支撑,易于推广开来。

The operating stability of the drive system of industrial robots directly affects the efficiency of use,and then saves industrial costs.The random mapping algorithm is used to realize the fault and achieve the dimensionality reduction of multi-source information.After considering the classification according to SVM method,the corresponding experimental scheme is designed.It is proved that the above method meets the practical feasibility.The results show that when the number of samples reaches 800~2100,the ideal fault diagnosis results can be obtained.The vibration signal fusion can obtain higher diagnostic accuracy than the Z direction,and obtain more accurate prediction results,with an accuracy of 99.89%.This research has a good theoretical support for improving the operational stability of industrial robot drive system and is easy to be popularized.