基于大数据的变压器油中溶解气体关键状态量动态预警研究

Research on dynamic early warning of key state quantities of dissolved gas in transformer oil based on big data

针对传统方法难以解决变压器故障诊断中精确度不高、无法实现故障预警的问题,本文利用大数据分析方法,提出一种变压器油中溶解气体关键状态量动态预警方法。该方法采用了高斯混合聚类模型对设备的正常、亚健康和异常状态进行评价,并利用了隐马尔科夫转移矩阵提取色谱演化过程的动态特征参量,实现了亚健康状态下变压器设备状态的短期预测,实现了变压器亚健康状态的动态预警,突破了个性化运行环境下设备亚健康状态的实时诊断及剩余寿命预测等技术瓶颈。经过对案例库中的变压器进行实证分析,本文提出的方法能够反映出气体增长速率与变压器亚健康程度之间的关系,能够实现过热缺陷设备提前100天左右的短期动态故障预警,在故障动态预警方面具有实用价值。

Aiming at the problem that the accuracy of transformer fault diagnosis is not high due to the lack of data,in order to realize the dynamic early warning of transformer fault,this paper proposes a dynamic early warning method of key state quantity of dissolved gas in transformer oil by using big data analysis method.In this method,Gaussian mixture clustering model is used to evaluate the normal,sub-health and abnormal state of the equipment,and hidden Markov transfer matrix is used to extract the dynamic characteristic parameters of chromatographic evolution process.The short-term prediction of transformer equipment status under sub-health status is realized,and the dynamic early warning of transformer sub-health status is realized,which breaks through the sub-health equipment subhealth status under personalized operation environment Real time diagnosis of health status and prediction of residual life are the technical bottlenecks.The results of empirical analysis show that the method proposed in this paper can reflect the relationship between gas growth rate and sub-health degree of transformer,and realize short-term dynamic fault warning of overheated defect equipment 100 days in advance,which has practical value in fault dynamic early warning.