城市轨道交通道岔转辙机动作功率参考曲线推荐方法研究

Research on Action Power Reference Curve Recommendation Method of Urban Rail Transit Switch Machine

介绍了一种根据道岔转辙机历史动作功率曲线,通过无监督机器学习自动生成转辙机动作功率参考曲线的方法。该方法首先通过提取历史曲线特征,然后形成自适应分类成簇,最后以簇为单位生成推荐的参考曲线。测试验证表明,该方法可以降低设备维护人员工作量,提高参考曲线设置的准确度,提升转辙机设备的智能化维护。

A method to automatically generate the reference curve of switch machine action power by unsupervised machine learning according to the historical action power curve of switch machine is introduced.Firstly,the feature of history curve is extracted,and then clustering is formed by adaptive classification.Finally,the recommended reference curve is generated by clustering.Test results show that this method can reduce the workload of equipment maintenance personnel,improving the accuracy of reference curve setting and the intelligent maintenance of switch machine equipment.