Distribution network planning is a complex multi-object and multi-restrictionproblem. This paper mainly studies the model and algorithm of substation planning. Onthe basis of Particle Swarm Optimization algorithm, use Adaptive Random Weight toimprove the control capability of inertial weight, and also introduce immune theory torealize the dynamic regulation between particle diversity and group convergence. By thisway, the innovational algorithm successfully avoid the problem of prematureconvergence and particle assembling occurred in traditional one, so greatly improve theconvergence precision of it.In this paper, the differences of land price and geography factors are taken intoconsideration in the substation planning, so the economy and feasible of it can beguaranteed, What's more, the initial algorithm is used to improve the calculation velocityand efficiency, and the Negative Grads method is firstly proposed to solve the problemof unavailable result in the process of calculation. Finally, the algorithm is practiced inthe example of real distribution network.
|