基于MDP-ADMM的数据中心储能系统优化运行方法

doi:10.19799/j.cnki.2095-4239.2023.0018

摘要/Abstract

摘要：

为了缓解数据中心综合运行成本高、电网侧负荷峰谷差大的问题，提出一种基于马尔可夫决策过程与交替方向乘子法结合的数据中心储能系统优化运行方法。首先，分析了数据中心的基本结构，并对数据中心主要设备的功耗特性以及数据中心负荷在时间维度上灵活转移的特性进行分析；其次，考虑数据中心供电可靠性、储能系统充放电功率限制、充放电损耗成本等因素，建立以综合运行成本最小和削峰填谷效果最佳为目标的储能系统优化运行模型；然后，针对模型具有多时段耦合的特点，采用马尔可夫决策过程对其解耦重构，利用交替方向乘子法对重构后的问题进行迭代求解；最后，在MATLAB仿真环境下，对某大型数据中心进行仿真分析，仿真结果验证了所提优化运行方法的有效性与合理性。

关键词: 数据中心, 储能系统, 负荷转移特性, 马尔可夫决策过程, 交替方向乘子法, 优化运行

Abstract:

This study proposed an optimal operation method for energy storage systems in the data center, based on the Markov decision process, and an alternating direction multiplier method to alleviate the problems of the high comprehensive operating cost of the data center and large load fluctuations on the grid side. First, the basic structure, the power consumption characteristics of the primary devices, and the load time-shifting characteristics of the data center are analyzed. Second, considering the power supply reliability of the data center, the charge and discharge power limit, and the loss cost of the energy storage system, an optimal operation model is established to minimize the peak-to-valley load difference and comprehensive operating cost. Then, given the multi-time coupling characteristic, the Markov decision process is used to reconstruct the model, and the alternating direction multiplier method is used to solve the reconstructed problem. Finally, the simulation analysis of a large-scale data center is conducted in MATLAB, and the obtained results verify the rationality and effectiveness of the proposed optimal operation method.

Key words: data center, energy storage system, load shifting characteristic, Markov decision process, alternating direction multiplier method, optimization operation

中图分类号:

TM 911

陈绪昌, 王育飞, 薛花. 基于MDP-ADMM的数据中心储能系统优化运行方法[J]. 储能科学与技术, 2023, 12(6): 1890-1900.

Xuchang CHEN, Yufei WANG, Hua XUE. Optimal operation method of energy storage system in a data center based on MDP-ADMM algorithm[J]. Energy Storage Science and Technology, 2023, 12(6): 1890-1900.

图/表 12

图1

图2

图3

图4

表1

图5

表2

图6

图7

图8

图9

表3

不同预测误差下的优化结果"

预测误差	求解算法	综合运行成本/元	负荷峰谷差/kW
$σ = 0.1$	PSO	30741	1037
	ADMM	28164	967
	MDP-ADMM	25374	768
$σ = 0.2$	PSO	31943	1186
	ADMM	29759	1028
	MDP-ADMM	26637	807

表3

参考文献 27

1	JI H R, CHEN S R, YU H, et al. Robust operation for minimizing power consumption of data centers with flexible substation integration[J]. Energy, 2022, 248: doi: 10.1016/j.energy.2022. 123599.
2	中国信息通信研究院. 数据中心白皮书(2022年)[EB/OL]. [2022-11-07]. http://www.caict.ac.cn/kxyj/qwfb/bps/202204/t20220422_400391.htm.
	China Academy of Information and Communications Technology. White paper of data centers (2022)[EB/OL]. [2022-11-07]. http://www.caict.ac.cn/kxyj/qwfb/bps/202204/t20220422_400391.htm.
3	CHEN M, GAO C W, SHAHIDEHPOUR M, et al. Internet data center load modeling for demand response considering the coupling of multiple regulation methods[J]. IEEE Transactions on Smart Grid, 2021, 12(3): 2060-2076.
4	JIN C Q, BAI X L, YANG C, et al. A review of power consumption models of servers in data centers[J]. Applied Energy, 2020, 265: doi: 10.1016/j.apenergy.2020.114806.
5	吕佳炜, 张沈习, 程浩忠, 等. 集成数据中心的综合能源系统能量流-数据流协同规划综述及展望[J]. 中国电机工程学报, 2021, 41(16): 5500-5520.
	LYU J W, ZHANG S X, CHENG H Z, et al. Review and prospect on coordinated planning of energy flow and workload flow in the integrated energy system containing data centers[J]. Proceedings of the CSEE, 2021, 41(16): 5500-5520.
6	郭斌, 邢洁, 姚飞, 等. 基于双层规划模型的用户侧混合储能优化配置[J]. 储能科学与技术, 2022, 11(2): 615-622.
	GUO B, XING J, YAO F, et al. Optimal configuration of user-side hybrid energy storage based on bi-level programming model[J]. Energy Storage Science and Technology, 2022, 11(2): 615-622.
7	ZHANG S, SHEN Z R, ZHANG G L. Cost minimization for geo-distributed data centers with renewable resources and energy storages[C]//2019 IEEE Global Communications Conference (GLOBECOM). Waikoloa, HI, USA. IEEE, 2020: 1-6.
8	DABBAGH M, HAMDAOUI B, RAYES A, et al. Shaving data center power demand peaks through energy storage and workload shifting control[J]. IEEE Transactions on Cloud Computing, 2019, 7(4): 1095-1108.
9	LI J, QI W B. Toward optimal operation of Internet data center microgrid[J]. IEEE Transactions on Smart Grid, 2018, 9(2): 971-979.
10	GUO C S, LUO F J, CAI Z X, et al. Energy management of Internet data centers in multiple local energy markets[J]. Electric Power Systems Research, 2022, 205: doi: 10.1016/j.epsr.2021. 107760.
11	陈敏, 高赐威, 陈宋宋, 等. 考虑数据中心用电负荷调节潜力的双层经济调度模型[J]. 中国电机工程学报, 2019, 39(5): 1301-1313.
	CHEN M, GAO C W, CHEN S S, et al. Bi-level economic dispatch modeling considering the load regulation potential of Internet data centers[J]. Proceedings of the CSEE, 2019, 39(5): 1301-1313.
12	高晋坤, 余娟, 刘珏麟, 等. 考虑多时段设备耦合的数据中心能效优化方法[J]. 电力系统自动化, 2022, 46(15): 153-161.
	GAO J K, YU J, LIU J L, et al. Optimization method for energy efficiency of data center considering multi-period equipment coupling[J]. Automation of Electric Power Systems, 2022, 46(15): 153-161.
13	COMPARE M, MARELLI P, BARALDI P, et al. A Markov decision process framework for optimal operation of monitored multi-state systems[J]. Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 2018, 232(6): 677-689.
14	LU X, YIN H, XIA S W, et al. A real-time alternating direction method of multipliers algorithm for nonconvex optimal power flow problem[J]. IEEE Transactions on Industry Applications, 2021, 57(1): 70-82.
15	LIN S, FAN R D, FENG D, et al. Condition-based maintenance for traction power supply equipment based on partially observable Markov decision process[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(1): 175-189.
16	黄帅博, 陈蓓, 高降宇. 基于马尔可夫决策过程的电动汽车充电站能量管理策略[J]. 电力自动化设备, 2022, 42(10): 92-99.
	HUANG S B, CHEN B, GAO J Y. Energy management strategy of electric vehicle charging station based on Markov decision process[J]. Electric Power Automation Equipment, 2022, 42(10): 92-99.
17	张世旭, 苗世洪, 杨炜晨, 等. 基于自适应步长ADMM的配电网分布式鲁棒优化调度策略[J]. 高电压技术, 2021, 47(1): 81-92.
	ZHANG S X, MIAO S H, YANG W C, et al. Distributed robust optimal dispatch for active distribution networks based on alternative direction method of multipliers with dynamic step size[J]. High Voltage Engineering, 2021, 47(1): 81-92.
18	ATTARHA A, SCOTT P, THIÉBAUX S. Affinely adjustable robust ADMM for residential DER coordination in distribution networks[J]. IEEE Transactions on Smart Grid, 2020, 11(2): 1620-1629.
19	吴云芸, 方家琨, 艾小猛, 等. 计及多种储能协调运行的数据中心实时能量管理[J]. 电力自动化设备, 2021, 41(10): 82-89.
	WU Y Y, FANG J K, AI X M, et al. Real-time energy management of data center considering coordinated operation of multiple types of energy storage[J]. Electric Power Automation Equipment, 2021, 41(10): 82-89.
20	丁肇豪, 曹雨洁, 张素芳, 等. 能源互联网背景下数据中心与电力系统协同优化(一): 数据中心能耗模型[J]. 中国电机工程学报, 2022, 42(9): 3161-3176.
	DING Z H, CAO Y J, ZHANG S F, et al. Coordinated operation for data center and power system in the context of energy Internet(‍Ⅰ): Energy demand management model of data center[J]. Proceedings of the CSEE, 2022, 42(9): 3161-3176.
21	GUO C S, LUO F J, CAI Z X, et al. Integrated planning of Internet data centers and battery energy storage systems in smart grids[J]. Applied Energy, 2021, 281: doi: 10.1016/j.apenergy.2020.116093.
22	范睿, 孙润稼, 刘玉田. 考虑空调负荷需求响应的负荷恢复量削减方法[J]. 电工技术学报, 2022, 37(11): 2869-2877.
	FAN R, SUN R J, LIU Y T. A load restoration amount reduction method considering demand response of air conditioning loads[J]. Transactions of China Electrotechnical Society, 2022, 37(11): 2869-2877.
23	AHMED K M U, BOLLEN M H J, ALVAREZ M. A review of data centers energy consumption and reliability modeling[J]. IEEE Access, 2021, 9: 152536-152563.
24	蒋向兵, 汤波, 余光正, 等. 面向新能源就地消纳的园区储能与电价协调优化方法[J]. 电力系统自动化, 2022, 46(5): 51-61.
	JIANG X B, TANG B, YU G Z, et al. Coordination and optimization method of park-level energy storage and electricity price for local accommodation of renewable energy[J]. Automation of Electric Power Systems, 2022, 46(5): 51-61.
25	HE H W, SHEN H. Minimizing the operation cost of distributed green data centers with energy storage under carbon capping[J]. Journal of Computer and System Sciences, 2021, 118: 28-52.
26	林楷东, 陈泽兴, 张勇军, 等. 含P2G的电-气互联网络风电消纳与逐次线性低碳经济调度[J]. 电力系统自动化, 2019, 43(21): 23-33.
	LIN K D, CHEN Z X, ZHANG Y J, et al. Wind power accommodation and successive linear low-carbon economic dispatch of integrated electricity-gas network with power to gas[J]. Automation of Electric Power Systems, 2019, 43(21): 23-33.
27	江苏省发改委. 省发改委关于进一步完善分时电价机制有关事项的通知[EB/OL]. [2022-10-25]. http://fzggw.jiangsu.gov.cn/art/2021/12/27/art_72382_10232339.html.
	Jiangsu Development and Reform Commission. Notice of the Jiangsu Development and Reform Commission on further improving the TOU price mechanism[EB/OL]. [2022-10-25]. http://fzggw.jiangsu.gov.cn/art/2021/12/27/art_72382_10232339.html.

求解算法	最优适应度值	平均适应度值	标准差	达到最优解次数占比/%
PSO	14.2	15.9	1.3	20
ADMM	12.6	14.1	1.1	35
MDP-ADMM	12.4	12.7	0.6	55

[1]	孙晓霞, 桂中华, 高梓玉, 周冰倩, 刘夏, 张新敬, 郭欢, 李文, 盛勇, 朱阳历, 周健, 徐玉杰. 压缩空气储能系统动态运行特性[J]. 储能科学与技术, 2023, 12(6): 1840-1853.
[2]	许永红, 吴玉庭, 张红光, 杨富斌, 王焱. 基于气动马达的微型压缩空气储能系统的试验研究[J]. 储能科学与技术, 2023, 12(6): 1854-1861.
[3]	管敏渊, 沈建良, 徐国华, 汤舜, 张炜鑫, 曹元成. 锂离子电池储能系统靶向消防装备设计与性能[J]. 储能科学与技术, 2023, 12(4): 1131-1138.
[4]	武鑫, 尚文举, 马志勇, 滕伟, 张爽, 罗海荣. 抽水蓄能-飞轮混合储能系统协调控制方法[J]. 储能科学与技术, 2023, 12(2): 468-476.
[5]	陈海东, 蒙飞, 王庆, 侯峰, 王亦, 张志华. 储能系统和新能源发电装机容量对电力系统性能的影响[J]. 储能科学与技术, 2023, 12(2): 477-485.
[6]	侯美倩, 牛启帆, 邢洁, 单英浩. 计及可靠性的含源配电网储能系统的优化配置[J]. 储能科学与技术, 2023, 12(2): 504-514.
[7]	张华民. 全钒液流电池的技术进展、不同储能时长系统的价格分析及展望[J]. 储能科学与技术, 2022, 11(9): 2772-2780.
[8]	张从佳, 施敏达, 徐晨, 黄震宇, 慈松. 基于动态可重构电池网络的储能系统本质安全机制及实例分析[J]. 储能科学与技术, 2022, 11(8): 2442-2451.
[9]	郭雨涵, 郁丹, 杨鹏, 王子绩, 王金涛. 基于贪婪算法的分布式储能系统容量优化配置方法[J]. 储能科学与技术, 2022, 11(7): 2295-2304.
[10]	吴田, 林闽城, 海浩, 孙海渔, 温兆银, 马福元. 面向一次调频的镍氢电池系统开发[J]. 储能科学与技术, 2022, 11(7): 2213-2221.
[11]	张平, 康利斌, 王明菊, 赵广, 罗振华, 唐堃, 陆雅翔, 胡勇胜. 钠离子电池储能技术及经济性分析[J]. 储能科学与技术, 2022, 11(6): 1892-1901.
[12]	李建林, 张则栋, 李雅欣, 周毅, 岳云力. 碳中和目标下移动式储能系统关键技术[J]. 储能科学与技术, 2022, 11(5): 1523-1536.
[13]	李昊, 刘畅, 苗博, 张静. 考虑冷热电互补及储能系统的多园区综合能源系统协调优化调度[J]. 储能科学与技术, 2022, 11(5): 1482-1491.
[14]	陶飞跃, 王焕然, 李瑞雄, 赵静, 葛刚强, 贺新, 陈昊. 利用环境再冷的二氧化碳储能热电联产系统及其热力学分析[J]. 储能科学与技术, 2022, 11(5): 1492-1501.
[15]	田峰, 程志江, 杨涵棣, 杨天翔. 模块化多电平复合变换器电池储能系统容错控制策略[J]. 储能科学与技术, 2022, 11(5): 1583-1591.