Distribution Network Topology Analysis and Parameter Calculation Method Based on Pseudo Measurement Data

Parameter estimation and topology identification of distribution networks are the foundation of distribution network planning, operation analysis, and safety control. Traditional linear regression methods have high requirements for measurement data errors or noise data. Only in the absence of noise can the estimation be accurate. However, both the actual input measurement values (such as voltage amplitude and phase angle) and the output measurement values (such as active and reactive power) contain noise data. For topology estimation, even if the measurement error is small, regression methods cannot obtain accurate topology. In response to the above issues, the impact of measurement errors on line parameter estimation and topology recognition was quantitatively analyzed, and the pseudo measurement data was cleaned and preprocessed; On this basis, a basic model for estimating distribution network parameters was constructed, and a line parameter estimation model considering bilateral measurement errors was established. In response to the difficulty in solving the problem caused by its non-convexity, an improved immune clone cuckoo algorithm is used to solve the equivalent transformation and obtain a multi parameter mathematical model that is easy to solve. Finally, based on the IEEE 33 node system, simulation analysis was conducted and compared with traditional linear regression and least squares methods to demonstrate that the proposed method still has good estimation accuracy even when the measurement error reaches 10%.