Power system planning method based on machine learning

Abstract

The invention discloses a power system planning method based on machine learning. The method includes: determining the number of newly-built substations and the capacity of each substation by utilizing a greedy method; and on the basis, solving the position of the newly-built transformer substation and the power supply range of each transformer substation in each stage by utilizing a Hopfield neural network algorithm, then reducing the capacity of the transformer substation according to the actual power supply condition of each transformer substation, finally determining that the optimal solution is met, establishing a model and optimizing. According to the invention, the artificial neural network applied to power grid planning is mainly a Hopfield network; the artificial neural network has excellent characteristics of strong nonlinear mapping capability, large-scale synergistic effect, clustering effect, parallelism, fault tolerance, robustness, no need of data normalization processing and the like, is suitable for solving large-scale combination optimization problems such as similar power grid planning, and has relatively high execution efficiency.

Description

technical field [0001] The invention relates to the technical field of power systems, in particular to a machine learning-based power system planning method. Background technique [0002] Power planning is a multi-objective, multi-stage, nonlinear, constrained mixed integer programming problem. Mathematically speaking, this is an operations research problem. However, traditional operations research methods, such as linear programming, integer programming, and mixed integer programming, have strict mathematical principles, but they often have problems such as wrong search direction and iterative divergence when solving substation planning problems. When the number of variables increases, it often falls into a "combination explosion", and it is difficult for these methods to obtain an optimal solution for the problem. At present, various optimization algorithms based on artificial intelligence technology have gradually become the mainstream algorithm for solving the mathemat...

AI Technical Summary

Problems solved by technology

However, traditional operations research methods, such as linear programming, integer programming, and mixed integer programming, have strict mathematical principles, but they often have problems such as wrong search directions and iterative divergence when solving substation planning problems.

When the number of variables increases, it often falls into a "combination explosion", and it is difficult for these methods to obtain an optimal solution for the problem

Images