Studies On Technologies Of Reconfiguration And Coordinated Reactive Power Optimization In Active Distribution Grid

Distribution generators (DG) has been continuously integrated into distribution grid, which makes the traditional problems, such as topology optimization and reactive power control, face new challenges. This paper focused on reconfiguration and coordinated reactive power control. In the studies, diverse optimization methods has been combined and improved for efficiency enhancement.Firstly, a new intelligent random algorithm, harmony search algorithm (HSA), was investigated. The method was illustrated based on the prototype of HSA system, namely the natural harmony improvising process. Through derivation, the convergence ability of HSA has been quantified. Comparisons were made to prove HSA’s solution performance.Secondly, this paper investigated optimal grid reconfiguration problem considering data uncertainties, with the integration of DG in smart distribution grid. By considering the objectives of minimal line loss cost, minimum Expected Energy Not Supplied (EENS) and minimum switch operation cost, the proposed grid reconfiguration formulation was formed as a multi-objective optimization problem. Through normalization of each objective function, the multi-objective optimization was transformed to single-objective optimization. The optimization problem was further solved by combining Binary Particle Swarm Optimization (BPSO) and HSA considering DG integration. Interval analysis was applied to deal with parameter uncertainties. Test cases showed the applicability of the proposed method, and comparisons with different methods and scenarios were provided.Finally, this paper focused on coordinated reactive power control considering renewable DG. The problem was solved by dynamic programming (DP) and HSA, respectively. In process of DP method, a target function considering minimum power loss and minimum operations for traditional reactive power devices was generated. Constraints including power flow equations and bus voltage limits were involved. For wind turbines and photovoltaic devices, the renewable DG probability model was built as dispatchable and non-dispatchable type respectively, and reasonably transformed to equivalent pattern. Test case on 10kV standard system showed the method’s ability to solve the optimization problem. In the process of HSA, the target function was revised by attaching complicated constraints. The DG modelling was further improved by piecewise, which kept it consistent with probability model. Besides, the classical HSA was improved to solve mixed-variable problem and adjust its iteration step size automatically. Numerical tests provided convincing results for applicability and efficiency of the proposed method. Comparisons and performance analysis with other algorithms including DP method have also been illustrated.