Power distribution network voltage regulation method based on deep reinforcement learning algorithm

Abstract

The invention discloses a power distribution network voltage regulation method based on a deep reinforcement learning algorithm, and the method comprises the steps of connecting an energy storage system for the auxiliary services, such as system voltage regulation, etc., to the tail end of a power distribution network through knowing the influence factors of the voltage operation level of the power distribution network; so that the system voltage operation level problems caused by high intermittency of distributed renewable energy sources and load demand fluctuation can be effectively solved.According to the method, the operation of battery energy storage is modeled as a Markov decision process, the subsequent regulation and control capability is considered, and the Q deep neural networkis embedded to approach the optimal action value, so that the problem of overlarge state space is solved. The energy storage state of charge, the renewable energy predicted output and the load level form a state feature vector to serve as the input of the Q network, the optimal discretization charging and discharging action for improving the voltage operation level is output, training is carried out through a playback strategy, and the energy storage control method tending to the optimal voltage regulation strategy is obtained.

Description

【Technical field】 [0001] The invention belongs to the technical field of power system automation, and relates to a distribution network voltage regulation method based on a deep reinforcement learning algorithm. 【Background technique】 [0002] With the continuous improvement of the penetration rate of clean energy in the distribution network, its strong volatility and high uncertainty have an increasing impact on the safe and economic operation of the distribution network. When a large amount of renewable distributed generation (Renewable Distributed Generation, RDG) is connected to the distribution network, the fluctuation of its output will also have a negative impact on the voltage operation level of the distribution network, and even cause the voltage to exceed the limit. According to the non-decoupling characteristics of active power and reactive power in the distribution network, it can be known that controlling the balance of active power in the distribution network c...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to solve the problems of complex uncertainty modeling and poor convergence in the prior art, and the difficulty in solving problems caused by too large state space, and to provide a distribution network voltage regulation method based on deep reinforcement learning algorithm

Images