黑龙江科学2024,Vol.15Issue(8) :49-53,57.

基于深度强化学习的储能调度优化研究

Research on Energy Storage Scheduling Optimization Based on Deep Reinforcement Learning

张杰 方伟 李思莹 程先龙
黑龙江科学2024,Vol.15Issue(8) :49-53,57.

基于深度强化学习的储能调度优化研究

Research on Energy Storage Scheduling Optimization Based on Deep Reinforcement Learning

张杰 1方伟 1李思莹 1程先龙1
扫码查看

作者信息

  • 1. 云南电网有限责任公司红河供电局,云南蒙自 661100
  • 折叠

摘要

可再生能源增长将成为我国未来能源发展的首要任务.但在大力发展可再生能源的同时,弃风弃光的问题随之而来.为了能够更充分利用可再生能源,构建了基于深度强化学习的储能调度模型.由于实际复杂情景中大多数指标参数为连续变量,为了更好地模拟现实情况,提出了一种基于连续动作空间的受阻能源混合储能模型,在考虑风能的基础上进一步纳入太阳能.引入碱性水电解器,将电能转化为氢气进行储存,丰富了电能的去向和存储形式,使电场盈利方式更加多元化.该模型在计算收益和成本时将设备的维修费用和相关金融成本也考虑在内,使模型更加贴近真实生产生活.通过比较二者在训练过程及测试结果的异同,验证其在面对复杂情境下进行储能调度的可行性.

Abstract

Renewable energy growth will become the top priority of China's future energy development.However,the problem of abandoning wind and light occurs with vigorous development of renewable energy.Since most of the index parameters in the actual complex scenario are continuous variables,in order to better simulate the real situation,the study proposes the hybrid energy storage model of hindered energy based on continuous action space,and further includes solar energy on the basis of considering wind energy.The introduction of alkaline water electrolyser converts electric energy into hydrogen for storage,which enricfies the destination and storage form of electric energy,and makes the profitability of electric field more diversified.The model also takes the maintenance cost and related financial cost into account when calculating the income and cost.This makes the model more close to the real production and life.By comparing the similarities and differences in the training process and test results,the study verifies the feasibility of energy storage scheduling in complex situations.

关键词

混合储能模型/碱性水电解器/TD3算法/PPO算法

Key words

Hybrid energy storage model/Alkaline water electrolyzer/TD3 algorithm/PPO algorithm

引用本文复制引用

出版年

2024
黑龙江科学
黑龙江省科学院

黑龙江科学

影响因子:1.014
ISSN:1674-8646
参考文献量13
段落导航相关论文