水电机组故障诊断专家系统运行状态智能预测及故障识别  

Title Intelligent Prediction of Operating State and Fault Identification of Hydropower Unit Fault Diagnosis Expert System

作　　者：张明儒 孙永鑫[2] 赵日升 刘忠仁 于爽 韩毅[3] 梁彬[3] ZHANG Ming-ru;SUN Yong-xin;ZHAO Ri-sheng;LIU Zhong-ren;YU Shuang;HAN Yi;LIANG Bin(HebeiFengning Pumped Storage Limited Company,Chengde 068300,China;Harbin Electric Machinery Company Limited,Harbin 150040,China;Harbin Electric of Large Electric Machinery,Harbin 150040,China)

机构地区：[1]河北丰宁抽水蓄能有限公司,河北承德068300 [2]哈尔滨电机厂有限责任公司,黑龙江哈尔滨150040 [3]哈尔滨大电机研究所有限公司,黑龙江哈尔滨150040

出　　处：《节能技术》2024年第5期463-469,共7页Energy Conservation Technology

基　　金：国网新源集团有限公司科技项目资助:河北丰宁抽水蓄能电站机组智能运检故障诊断系统研究与应用技术开发服务。

摘　　要：对水电设备进行智能预测可以将故障诊断提前至故障发生前,由此可预防不稳定运行对机组设备的损害,延长机组寿命。本文基于ARMA模型、匹配矩阵模型(MM)和Hu不变矩原理,针对水电机组需要重点监控的三类数据:自相关性较强参数、自相关性弱参数、轴心轨迹参数,开发了一套水电机组智能预测和故障识别系统。本系统自2019年实际装备以来预测结果和实际数据拟合良好,将各类潜在故障数据输入系统,系统均可预测并准确判断故障类型。综上,本系统可实现对故障的提前预测和诊断,有力保障水电机组的正常稳定发电。Intelligent prediction of hydropower equipment can advance fault diagnosis to before the occurrence of faults,thus preventing the damage of unstable operation to the unit equipment and prolonging the life of the unit.In this study,based on ARMA model,matching matrix model(MM)and Hu invariant moment principle,a set of intelligent prediction and fault identification system for hydropower units is developed for the three types of data that need to be focused on the monitoring of the hydropower units:strong autocorrelation parameter,weak autocorrelation parameter,and axial trajectory parameter.This system has been actually equipped since 2019 the prediction results and the actual data fit well.Inputting the various types of potential fault data into the system,the system can predict and accurately determine the type of fault.In summary,this system can realize the advance prediction and diagnosis of faults,which can strongly guarantee the normal and stable power generation of hydropower units.

关 键 词：ARMA模型 故障预测 轴心识别 HU不变矩 水电机组 

分 类 号：TK730.8[交通运输工程—轮机工程]