Integrated energy online scheduling method based on mixed time scale DRL

Abstract

The invention relates to an integrated energy online scheduling method based on a mixed time scale DRL. The method comprises the following steps: constructing a combined heat and power generation unit model, an electric boiler model and a gas boiler model of an integrated energy system; synchronously coordinating multiple time scales into a mixed time scale, and establishing a mixed time scale IES environmental economic scheduling model; respectively defining a mixed time scale state space, a mixed time scale action space and an environmental economy normalization reward function; and adopting a near-end strategy optimization algorithm to realize online decision making based on mixed time scale real-time feedback. According to the invention, the DRL is applied to solve the IES environmental economic scheduling problem, and a feasible scheme can be provided for promoting the ''dual-carbon'' target; and multiple time scales are synchronously coordinated into a mixed time scale, and the vacancy of the application of the DRL method in multi-time scale optimization scheduling is filled.

Description

technical field [0001] The invention belongs to the technical field of power transmission and distribution, and relates to an integrated energy online scheduling method based on a mixed time scale DRL. Background technique [0002] The integrated energy system (IES) can effectively utilize multiple energy sources in stages and vigorously promote the consumption of new energy sources, which is an important direction to achieve the goal of "double carbon". IES optimal scheduling is the premise to realize the balance between supply and demand of multiple energy sources in the system. The dispatching process mainly faces the following difficulties: First, the source-load uncertainty caused by the intermittency and volatility of renewable energy and the error of load forecasting brings great challenges to IES optimal dispatching. Second, the transmission speed of gas and heat in IES is much slower than that of electric energy. When the power system reaches a steady state, gas an...

AI Technical Summary

Problems solved by technology

The following difficulties are mainly faced in the dispatching process: First, the source-load uncertainty caused by the intermittent and fluctuating nature of renewable energy and the error of load forecasting has brought great challenges to IES optimal dispatching

Second, the transmission speed of gas and heat in IES is much slower than that of electric energy. When the power system reaches a steady state, gas and heat are still in a dynamic process. The inconsistent scheduling time scales of different subsystems increase the difficulty of IES scheduling.

[0005] However, the existing literature on IES optimal scheduling is based on a single time scale, ignoring the differences in the scheduling time scales of different energy systems in the actual process

Due to this difference, the state variable dimensions of different energy systems at the same time are inconsistent, and IES cannot cope with the defects of multiple state variable dimensions, which makes it difficult for the DRL method to deal with multi-time scale optimization problems

Images