基于时序贴近度与改进SVM的水机轴心轨迹诊断

发布时间：2018-06-17 11:53

本文选题：水电机组 + 轴心轨迹　；参考：《排灌机械工程学报》2017年12期

【摘要】：为了提高水轮机组诊断的精确性,提出应用时间序列模糊贴近度特征提取轴心轨迹特征参数,通过改进SVM模型并引入故障分类准确性判定因子对参数化的水电机组轴心轨迹开展了智能诊断.应用改进SVM对时间序列特征引入正确率、错误分类率计算方法,从而对诊断后轴心轨迹分类准确性进行判定,由此促进运行状态设备智能诊断,提高故障诊断系统的自动诊断水平及准确率;引入多类分类支持向量机算法、分类准确度判断解决异常状态下机组轴心轨迹特征参数无法识别、识别率低的问题.通过对改进扩展时序距离时间序列贴近度度量算法的应用解决了水电机组实时轴心轨迹特征参数准确性差和实时性差的问题.该方法提高了检测精度,同时增强了人机交互性,具有重要的理论意义和实用价值.
[Abstract]:In order to improve the accuracy of hydraulic turbine diagnosis, the feature of fuzzy closeness degree of time series is applied to extract the characteristic parameters of axis locus. By improving the SVM model and introducing the accuracy factor of fault classification, the intelligent diagnosis of the axis locus of the parameterized hydropower unit is carried out. The improved SVM is used to calculate the correct rate and error classification rate of the time series features, so as to judge the accuracy of the axial trajectory classification after diagnosis, thus promoting the intelligent diagnosis of the running state equipment. To improve the automatic diagnosis level and accuracy of fault diagnosis system, the multi-class classification support vector machine algorithm is introduced to determine the classification accuracy to solve the problem that the characteristic parameters of the axis track of the unit can not be identified and the recognition rate is low under abnormal condition. Through the application of the improved time series closeness measurement algorithm of extended time series, the problems of poor accuracy and real-time performance of the characteristic parameters of the real time axis trajectory of hydropower units are solved. This method improves the accuracy of detection and enhances the human-computer interaction, which has important theoretical significance and practical value.
【作者单位】：兰州工业学院电气工程学院;国网青海省电力公司电力科学研究院;青海师范大学;
【基金】：国家自然科学基金资助项目(51769012) 甘肃省科技计划资助项目(1506RJZA059)
【分类号】：TV738

【相似文献】