Stochastic Volt-Var optimization function for planning of MV distribution networks

Volt/Var Optimization (VVO) function is a key element in operation of distribution networks and major part of advanced Distribution Management Systems (DMS). From planning prospective, VVO function can be used to optimize reactive power flow in distribution network to recommend the best operating condition for the control equipment in a predefined period of time in future (i.e. 24 hour). In fact VVO minimizes the total system loss for a forecasted set of load and computes the optimized setting for transformer on-load tap changers (OLTC), Voltage Regulators (VR), and Capacitor Banks (CB), while system voltage profile is maintained within its limits. In this paper we use a full mixed integer linear programming (MILP) model for solving VVO problem for a planning application. The objective of this paper is to develop a planning VVO engine which can calculate the most probable expected loss of the network for the next 24 hours, and can recommend the best expected operating condition for the control equipment. To model the uncertainty of load, an ARMA model is applied to create several forecasted load scenarios to feed them into the VVO engine (which is implemented in a commercial solver GAMS (General Algebraic Modeling System). The implemented models have been tested on a real distribution network in southern Sweden and results are presented.