Multi-Objective Reactive Power Optimization Method for the Distribution Grid with Distributed Generator

To enhance the economic efficiency and operational security of distribution grids, this paper develops a reactive power optimization model that incorporates distributed power sources. The model aims to minimize the costs of reactive-load compensation equipment, reduce voltage deviations, and lower network losses while satisfying operational constraints. To overcome the common drawbacks of the standard genetic algorithm—such as limited optimization precision and a tendency to converge to local optima—four improvement strategies are introduced. These include an enhanced encoding scheme, an initial population generated via opposition-based learning, an elite retention strategy, and the adaptive adjustment of crossover and mutation rates. Together, these modifications strengthen the algorithm’s global search capability.

The proposed approach is validated using the IEEE30 node system. Compared with both the conventional genetic algorithm (GA) and an adaptive genetic algorithm, the improved method demonstrates faster convergence and a more robust ability to escape local optima. Simulation results indicate that the suggested algorithm effectively reduces voltage fluctuations and power losses in the network while improving the overall cost-efficiency of grid operation.