Data-driven partition control strategy for harmonic distortion and voltage unbalance

Local controls for harmonic distortion and voltage unbalance have become inapplicable due to the increasing penetration and decentralisation of disturbance sources. The complication and expansion of distribution networks cause difficulties in measuring certain system parameters, which reduce the practicability of global optimisation controls. Accordingly, this study introduces a partition control strategy driven by the disturbance data of nodes, which is used to allocate active power filters (APFs) to individually or simultaneously suppress harmonic distortion and voltage unbalance in distribution networks. A multidimensional piecewise linear representation based on local extreme points is proposed to extract the trend features of univariate disturbance time series (UDTS) and multivariate DTS (MDTS). Trend features, disturbance severities, and subsection lengths are used to establish a parameter matrix as the pattern representation of DTS. A feature distance (FD) method is developed to implement pattern matching on UDTS or MDTS to measure the coupling degrees of single or integrated disturbance among nodes. Area division and APF allocation are performed based on the coupling degrees. Case studies on a 14-bus distribution system are conducted to verify the rationality of the partition control strategy and the accuracy of the FD method.