Pseudo-dynamic substation expansion planning using hybrid heuristic and genetic algorithm

The substation expansion planning (SEP) is a complicated problem with a wide variety of its constraints. This paper presents a novel and efficient method to solve optimal SEP problem to determine optimal number, location, capacity, timing, and associated service area of each substation under all technical constraints over the planning period at minimum costs. In this respect, the genetic algorithm (GA) based k-means algorithm is developed to overcome some disadvantage of k-means algorithm to find optimal locations of new substations without any predetermined candidate locations. In this case, the GA is adapted based on k-means to place the substations in the centroid of their loads while location of existing ones is fixed. Moreover, allowable loading of substation is considered by a determined penalty cost in object function of GA to prevent placing substations where they may be overloaded. Then, the greedy heuristic algorithm is used to solve dynamic programming for optimal load assignment based on priority of load connection. The technical constraints such as permissible voltage drop and maximum loading substations, thermal limit of MV feeder, radial network is considered at this stage. The proposed method is used to solve optimal pseudo-dynamic SEP problem for Tabriz, which is the capital city of East Azerbijan province of Iran to verify efficiency and capability.