发明名称 Automated root cause analysis
摘要 A method for performing computerized automated root cause analysis for a complex product, such as a turbine, and computer-readable medium encoded with programming instructions for performing such a root cause analysis are provided. The method may include: monitoring the product; detecting a product misbehavior; forming a misbehavior pattern with timeslots of normal operation and timeslots of misbehavior operation; comparing timeslots of the misbehavior pattern to corresponding timeslots of channels of information in at least one database relating to the product; measuring how close the misbehavior pattern numerically fits to the information; and automatically identifying a best numerical fit.
申请公布号 US9606533(B2) 申请公布日期 2017.03.28
申请号 US201114233744 申请日期 2011.12.09
申请人 SIEMENS AKTIENGESELLSCHAFT 发明人 Pedersen Henrik
分类号 G05B23/02;F01D17/02;F01D21/00;F02C9/00;G06Q10/06;G07C3/00 主分类号 G05B23/02
代理机构 Beusse Wolter Sanks & Maire 代理人 Beusse Wolter Sanks & Maire
主权项 1. A method comprising: monitoring, by a turbine controller, technical parameters of a turbine; detecting a turbine misbehaviour including abnormal turbine behaviour below a fixed alarm limit based on the monitoring of the turbine; performing computerized root cause analysis based on the detected turbine misbehaviour to determine a root cause of the turbine misbehaviour, the root cause analysis comprising: forming a misbehaviour pattern predicting failure of the turbine with timeslots of normal operation of the turbine and timeslots of the detected turbine misbehaviour of the turbine wherein the misbehaviour pattern is based on a cluster of misbehaviour;comparing a timeslot of the misbehaviour pattern to corresponding timeslots for each information associated with a type of the turbine misbehaviour in at least one database relating to the turbine;calculating an indicator, for said each information, indicative of a fit of the misbehaviour pattern;ordering into a list the calculated indicator for said each information to prioritize a representation of a root cause of the turbine misbehaviour; andautomatically identifying a best fit from the list; and taking preventive actions to prevent the failure associated with the root cause associated with the best fit.
地址 Munich DE