智能变电站机电设备巡视数据缺失恢复建模研究

Research on Modeling of Missing Inspection Data Recovery of Electromechanical Equipment in Smart Substation

现有的数据恢复方法存在相对误差较大、建模时间较长的问题。为此,针对智能变电站机电设备,提出了一种新的巡视数据缺失恢复建模方法。通过矩阵填充的方式采集机电设备巡视数据,并对巡视数据实施预处理,通过一阶偏导实施对张量的PARAFAC(PARAllel FACtors,平行因子)分解,通过梯度形式得到缺失数据的二阶、三阶张量集合的一阶偏导推导结果,再将结果推广至N阶张量,从而构建智能变电站机电设备巡视数据缺失恢复模型。测试结果表明:所提方法的巡视数据缺失恢复相对误差较低、运行较快。

The existing data recovery methods have the problems of large relative error and long modeling time.Therefore,a new modeling method of patrol data loss recovery is proposed for electromechanical equipment in smart substation.The inspection data of mechanical and electrical equipment are collected by means of matrix filling,and the inspection data are preprocessed.The parafac decomposition of tensors is implemented through first-order partial derivative,and the first-order partial derivative derivation results of second-order and third-order tensors of missing data are obtained through gradient form,and then the results are extended to the n-order tensors.Thus,a data loss recovery model for patrol data of mechanical and electrical equipment in smart substation is constructed.The test results show that the method in this paper has low relative error and fast operation.