Scheduling method and device for energy storage system of wind power plant integrated with prediction and decision

Abstract

The invention provides a scheduling method and device for an energy storage system of a wind power plant integrated with prediction and decision. The method comprises the following steps of a sample accumulation step: inputting the state st of awind power plant to an evaluation network, outputting Q values of all actions in an action space A, and determining a dispatching instruction at of the energy storage system by an Epsilon-greedy strategy after the energy storage system executes the dispatch instruction, calculating the returned reward rt and observing the state of the wind power plant in the next period st + 1, storing (st, at, rt, st + 1) as a sample in a buffer, and repeating the above process until the number of samples in the buffer reaches the preset upper limit; a Q value iteration step; and a network training step; a learning finish juding step: if gains acquired by the wind power plant do not increase and fluctuate within a pre-set range, indicating the evaluation network converges with the current scheduling instruction being optimal; otherwise, repeating the above steps till the evaluation network converges and outputting the optimal scheduling instruction.

Description

technical field [0001] The disclosure belongs to the field of wind farm energy storage system optimization, and in particular relates to a wind farm energy storage system dispatching method and device integrating prediction and decision-making. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] The output of wind turbines and the market price of electricity are directly related to the income of wind farms. The output of wind turbines is uncertain and unschedulable, and these characteristics make the income of wind farms fluctuate and uncontrollable. In addition, the electricity price released by the electricity market includes information on the balance of supply and demand in the market, transmission congestion, fuel costs, etc. Adjusting the electricity price is a means for the market to guide and restrict power generators to provide high...

AI Technical Summary

Problems solved by technology

The inventor found that the separation of prediction and decision-making caused the process to lose many effective decision-making basis contained in the original data, and additionally introduced error interference caused by the prediction algorithm

Images