基于改进数据驱动模型的低压台区拓扑关系自动识别技术研究

针对传统技术在识别低压台区拓扑关系时存在用户相位识别精度和接入表箱识别精度低的问题，研究了基于改进数据驱动模型的低压台区拓扑关系自动识别技术。从历史用电数据中采集低压台区用户的电压数据，在标准化处理后得到电压矩阵方程式；在初始化纠错参数的基础上，通过求解用户关系方程得到低压台区配电器与用户间的关系，通过关系校验组建低压台区拓扑关系的识别方程式，结合差值向量的取值范围完成纠错处理；利用改进数据驱动模型求得二次规划模型的矩阵，并构建低压台区拓扑关系识别模型。实验结果表明，该技术可以提高用户相位识别精度和接入表箱识别精度，证明其具有更好的识别性能。

Research on automatic identification of low-voltage substation topology based on improved data-driven model

In order to solve the problem of low user phase recognition accuracy and low access box recognition accuracy in the recognition of low-voltage substation area topological relation，an automatic recognition technology of low-voltage substation area topological relation based on improved data-driven model was studied.The voltage data of low-voltage substation area users are collected from the historical power consumption data, and the voltage matrix equation is obtained after standardized processing.On the basis of initializing error correction parameters, the relationship between low-voltage substation area distributor and users is obtained by solving the user relationship equation, and the identification equation of low-voltage substation area topological relationship is established by relationship verification,and error correction is completed by combining the value range of difference vector.The matrix of the quadratic programming model is obtained by using the improved data-driven model,and the identification model of data-driven substation topological relationship is constructed.Experimental results showed that this technology can improve the user phase recognition accuracy and access table box recognition accuracy，which proves that it has better recognition performance.