基于数据挖掘技术的CT设备故障预警研究

Research on fault early warning of CT equipment based on data mining technology

目的:基于数据挖掘关联规则算法构建CT设备故障特征信息模型,为CT设备运行和维护中的故障预警提供决策支撑。方法:根据CT设备实时监控数据信息和故障数据特点,分析研究Apriori算法中的缺陷,对CT设备故障特征信息进行特征构建,进而提出改进Apriori算法,以0~1矩阵为故障数据映射方案,按照矩阵设置规则进行数据计算处理,分析故障数据交际的频繁关联选项,研究CT设备多次故障与运行参数设置的关系,构建CT设备故障特征信息模型并进行验证,对比采用线性回归模型、物联网检测模型和故障特征信息模型的故障检测率和精确率以及故障特征信息模型应用前后的CT设备开机率。结果:CT设备故障特征信息模型的CT设备故障检测率和精确率分别为82.7%和94.2%,均高于线性回归模型和物联网检测模型;根据模型预测结果将CT设备预防性维护保养调整至故障高峰发生前1个月,模型应用后试点设备开机率比模型应用前提高5.7%。结论:CT设备故障特征信息模型能够降低传统算法的工作复杂性,提高CT设备故障检测准确性,为CT设备故障预警和预测提供良好的决策支持。

Objective:To build a CT equipment fault characteristic information model based on the data mining association rule algorithm to solve the problem of fault early warning in the operation and maintenance of CT equipment.Methods:According to the real-time monitoring data information and fault data characteristics of CT equipment,the defects in the Apriori algorithm were analyzed and studied,and the characteristics of CT equipment fault characteristic information were constructed,and then an improved Apriori algorithm was proposed.Taking the 0-1 matrix as the fault data mapping scheme,the data were calculated and processed according to matrix setting rules,of the frequent association options for fault data communication were analyzed,the relationship between multiple faults of CT equipment and operating parameter settings were researched,and the CT equipment fault characteristic information model was constructed and validated.The fault detection rate and precision rate of linear regression model,Internet of Things(IoT)detection model and fault characteristic information model were compared,and the startup rate of CT equipment before and after the application of fault feature information model were compared.Results:The fault detection rate and precision of CT equipment based on fault characteristic information model were 82.7%and 94.2%,respectively,which were higher than the linear regression model and the IoT detection model;according to the model prediction results,the CT equipment preventive maintenance can be adjusted to one month before the failure peak occurs,the equipment startup rate increased by 5.7%compared with hat before the application of the model.Conclusion:The fault characteristic information model of CT equipment can reduce the work complexity of traditional algorithms,improve the accuracy of CT equipment fault detection,and provide good decision support for CT equipment fault early warning and prediction.