首页> 外国专利> Processing method of predictive models for mechanical faults

Processing method of predictive models for mechanical faults

机译:机械故障预测模型的处理方法

摘要

The present invention relates to a technology for predicting a mechanical defect occurring in a facility, and a method of processing a predictive model for a failure prediction of a target facility by a failure prediction system is a time series representing the state of the facility from the start of the facility operation to the occurrence of the failure. Clustering or classifying the entire section as a sub-degradation stage based on the change pattern of the indicator data, extracting the prediction factor, and generating a plurality of prediction models that predict the time of failure, and the entire section The probability of the new time series index data to be predicted in the segmented sub-sections is estimated using the classifier generated when classifying for, and the identifier of the section with the highest estimated probability among the sub-sections for one observation event. Is recorded, the recorded identifier is stored as a sequence according to the time series of the indicator data, and the monotony of the sequence is evaluated for each prediction model to derive a prediction model that is most similar to the new time series indicator data.
机译:本发明涉及一种用于预测设施中发生的机械缺陷的技术,以及通过故障预测系统处理用于目标设施的目标设施的故障预测的预测模型的方法是表示设施状态的时间序列开始设施运作到失败的发生。基于指示符数据的更改模式,提取预测因子,并生成预测失败时间的多个预测模型,以及整个部分的概率,将整个部分分类为子降级阶段。使用在分类时生成的分类器和具有一个观察事件的子部分之间具有最高估计概率的分类器来估计要预测的分段子部分中的时间序列数据。记录,记录的标识符作为指示符数据的时间序列存储为序列,并且对每个预测模型评估序列的单调,以导出与新时序列指示符数据最相似的预测模型。

著录项

  • 公开/公告号KR102249849B1

    专利类型

  • 公开/公告日2021-05-10

    原文格式PDF

  • 申请/专利权人

    申请/专利号KR1020200188899

  • 发明设计人 최진혁;조원정;이진우;최소담;

    申请日2020-12-31

  • 分类号G05B23/02;

  • 国家 KR

  • 入库时间 2022-08-24 18:46:55

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号