Optimal Power Flow Based on Genetic Algorithms and Clustering Techniques

Abstract

Optimal power flow problems have been studied extensively for the past decades. Two approaches for solving the problem have been distinguished: mathematical programming and evolutionary algorithms. The first is fast but is not converging to a global optimum for every case. The second ones are robust but time-consuming. This paper proposes a method that combines both approaches to eliminate their flaws and take advantage of their benefits. The method uses properties of genetic algorithms to group their chromosomes around optima in the search space. The centers of these groups are identified by clustering techniques and furthermore used as initial points for gradient based search methods. At the end, the proposed method finds global optimum and its closest local optima. Continuous Newton-Raphson method is used to overcome ill-conditioned points in search space when calculating power flows. The proposed method is compared against similar methods showing considerable improvement.

Fault-Current Injection Strategies of Inverter-Based Generation for Fast Voltage Recovery

“…Therefore, gradient-based optimization methods would be prone to failing in finding optimal values of u p . As an alternative, the optimization method based on Genetic Algorithms and clustering techniques described in [15] has been used in this study.…”

“…Desirable parameter space represents a convex polytope which wraps the cluster of points in the search space whose relative value of the objective function is not much worse than the one of the optimum point. Since for the optimization part of the analysis in this paper the genetic algorithm combined with clustering techniques [15] is used, the polytope wraps the cluster of chromosomes containing the globally best chromosome.…”

A Fast Decomposition Method to Solve a Security-Constrained Optimal Power Flow (SCOPF) Problem Through Constraint Handling

“…Both the OPF and SCOPF are non-convex problems. Therefore, it is not possible to ensure a global minimum through mathematical programming [26]. Different attempts have been made to reach solutions close to the global minimum through genetic (GA), metaheuristic (MA), and machine learning algorithms.…”

“…Several works have tried to solve non-convex problems through a hybrid optimization strategy. For example, the OPF problem in [26] was solved through GA to group the chromosomes in a search space close to the absolute minimum, and then a continuous Newton-Rhapson method was used to mathematically reach the global minimum; however, overload constraints were not considered. In [7], ALM and ADMM were used to solve the SCOPF in DC.…”