| 引用本文: | 郭光华,亓新宏,王瑞琪,樊相臣,程浩原,艾芊,孙树敏,邢家维.基于联邦学习的综合能源系统集成需求响应机制[J].电力自动化设备,2023,43(8):33-39 |
| GUO Guanghua,QI Xinhong,WANG Ruiqi,FAN Xiangchen,CHENG Haoyuan,AI Qian,SUN Shumin,XING Jiawei.Integrated demand response mechanism for integrated energy system based on federated learning[J].Electric Power Automation Equipment,2023,43(8):33-39 |
|
| 摘要: |
| 随着能源集成技术的发展，需求响应已经逐渐进化到集成需求响应(IDR)，同时用户对于隐私保护的关注也日益增长。针对拥有冷热电设备和负荷的配电网侧综合能源系统，建立了多能源耦合交互模型，以反映不同能源消费行为之间的相互影响，并以运行成本最低为目标，以设备出力特性和多能源负荷特性为约束，设计了IDR优化模型。为了保护用户隐私，提出了联邦学习(FL)架构，重写IDR模型并将其置于该FL架构中进行迭代计算。仿真结果表明所提计算方法与不考虑耦合的传统需求响应方案相比，具有较好的成本优势；与其他分布式需求响应算法相比，计算效率也有所提升。 |
| 关键词: 综合能源系统 多能源耦合 集成需求响应 成本优化 联邦学习 |
| DOI：10.16081/j.epae.202301009 |
| 分类号:TM732 |
| 基金项目:山东省重大科技创新工程资助项目(2019JZZY010903) |
|
| Integrated demand response mechanism for integrated energy system based on federated learning |
|
GUO Guanghua1, QI Xinhong1, WANG Ruiqi1, FAN Xiangchen1, CHENG Haoyuan2, AI Qian2, SUN Shumin3, XING Jiawei3
|
|
1.State Grid Shandong Integrated Energy Services Co.,Ltd.,Jinan 250001, China;2.School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China;3.State Grid Shandong Electric Power Research Institute, Jinan 250002, China
|
| Abstract: |
| With the development of energy integration technology, demand response(DR) has gradually evolved into integrated demand response(IDR),and users are increasingly concerned about privacy protection. Aiming at the integrated energy system at the distribution network side with cooling, heating and power equipment and loads, the multi-energy coupling interaction model is established to reflect the interaction effects between different energy consumption behaviors. Taking the minimum operating cost as the objective and the equipment output characteristics and multi-energy load characteristics as constraints, the IDR optimization model is designed. To protect users’ privacy, the federated learning(FL) architecture is proposed, the IDR model is rewrote and it is put in this FL architecture for iterative computation. The simulative results show that the proposed calculation method has better cost advantage compared with the traditional DR scheme that does not consider coupling. Meanwhile, the computational efficiency is also improved compared with other distributed DR algorithms. |
| Key words: integrated energy system multi-energy coupling integrated demand response cost optimization federated learning |
