Optimal location and dimensioning of capacitors in microgrids using a multicriteria decision algorithm

Abstract

This work presents a methodology for optimal compensation of reactive power in Electric Microgrids using a multicriteria decision algorithm based on heuristic methods. It was verified the optimal location and dimensioning of fixed capacitor banks in a microgrid with 14 buses with objective functions based on cost, efficiency and quality criteria in a maximum demand operating condition and also considering the constraints of objective variables in minimum demand scenarios. The proposed capacitance to be installed was considered as a discrete variable. The location of discretized reactive capacitances was simulated at candidate nodes analyzing the power flow in the case study, resulting in a wide solution space that was tackled by means of multicriteria optimization, using dominance elimination techniques and the weighted sum method for decision making. The variables analyzed were: cost, maximum and average deviations of the voltage profile, power factor, total losses in the lines of the system and THD. All these variables were also verified in minimum demand scenarios. The proposed solution provides significant improvements in the variables analyzed and verifies the optimal performance of the technique. The mathematical analysis demonstrates the need of addressing the reactive compensation problem by means of multicriteria decision and the proposal provides a very novel tool for calculating the location and dimensioning of reactive compensation devices in distribution systems and microgrids. The programming was done in the Matlab environment with simulations using Simulink. The case study analyzed is a very novel validated Microgrid system of which it is known the variables that take part of this analysis, as an approximation of the study of a very real Microgrid.

Abstract

Microgrids, Voltage profile, Line losses, Power factor, Capacitor, Reactive power compensation, Multicriteria decision, THD, Matlab-Simulink.