融合粗糙集与灰色模型的道岔故障预测

Turnout fault prediction based on Rough Set and Grey Model

以高铁常用S700K型转辙机为例,通过预测电流特征实现对未发生故障的预先检修。结合粗糙集与灰色理论提出一种新方法用于道岔故障预测,该方法通过粗糙集的知识获取与规则约简,获得最小诊断规则;通过离散灰色预测模型的建模方法,实时建立不同故障类型的预测模型。随机抽取30组故障进行诊断,其中96.67%与实际情况相符,可满足诊断准确率的要求;随机抽取1组预测情况,预测值与实际值之间的残差较小,可满足预测准确率的要求。所提出方法能够有效减少由故障带来的安全问题,可操作性高,更具实用性。

Taking S700K switch machine which is often used in high speed railway as an example,by predicting current characteristics,the fault of turnout which has not happened can be examined and repaired in advance.A new method for fault prediction of turnout based on Rough Set and Grey Theory was presented.And the minimum diagnosis rules were obtained by knowledge acquisition and rule reduction methods which are based on Rough Set.Then through modeling method,discrete grey prediction models of different fault types were established in real time.Thirty groups of faults were randomly selected for diagnosis,and 96.67%of them were consistent with the actual situation,which could meet the requirements of diagnostic accuracy.A group of cases were randomly selected for prediction,and it could be known that the residual between the predictive value and the actual value is small,which could meet the requirement of predictive accuracy.The proposed method can effectively reduce accident rate caused by faults,and also have high operability and practicability.