Gravitational Search Algorithm And Its Application In Fault Diagnosis And Reconstruction Of Distribution Network

Artificial intelligence is a new branch of study, which is researched and developed for simulation, extension and expansion of human and natural mature and intelligent theories, methods, and applications of the system, one of the most important parts is the theory research of artificial intelligence algorithm and application. As a branch of artificial intelligence algorithm, gravitational search algorithm is a novel and unique algorithm, which has been concerned by scholars both at home and abroad because of its superior capability and convergence speed. So in this paper, the current research situation of the gravitational search algorithm, a more in-depth research and analysis, and put forward the corresponding improvement measures and its application in the power distribution network fault diagnosis, reconstruction.First of all, the basic theory of gravitational search algorithm, in this paper, based on extensive reading, learning the existing literature, in-depth study of the theory of gravitational search algorithm and related algorithms module, and the simulation results of standard test functions show that the feasibility of the algorithm and its superior performance on part of the function.Secondly, owing to poor performance in high dimensional multimodal functions, the improved gravitational search algorithm is proposed. The improved algorithm is proposed by reverse learning initialization, makes the algorithm initial solution with high probability into optimal solution region, for subsequent optimization lay the foundation; In the later stage of the algorithm, the conjugate gradient method is introduced to overcome the difficulty in the precise search of the particle resultant force due to the late Euclidean distance of algorithm; the running process of the algorithm in the synchronization of chaotic mapping, makes the iterative algorithm stagnation has certain probability randomly generated solutions and keep the diversity algorithm. Simulation results show that the improved algorithm is more applicable to the peak of multidimensional functions.Finally, the improved gravitational search algorithm is used to solve the single objective and multi-objective model. For distribution network fault diagnosis is to construct the single objective optimization model and apply the algorithm to carry out the simulation experiment. The results show that this algorithm is feasible for fault diagnosis; for the practical problem that is power supply need to be restored timely in the Non fault area after the fault diagnosis of distribution network, this paper constructs a multi-objective model considering the network loss and the switching action of the power distribution network reconfiguration, using algorithm to verify the feasibility.