This study presents a model predictive control method based on mixed-integer linear programming, taking the chilled water storage cooling system of a data center in Guangzhou as the research object. The optimization objective of the method is to minimize the energy consumption of the cooling system. By modeling the cooling system and environmental conditions and considering energy costs and cooling system efficiency, the optimal operation strategy for chillers and the scheduling arrangement for the chilled water storage cooling system are determined. During the optimization process, this research takes into account the influence of the minimum continuous operation time of chillers on the energy consumption of the cooling system and determines the optimal value to improve stability and reduce energy waste caused by frequent chiller start-ups and shutdowns. Through an annual simulation, this method reduces the total energy consumption by 6.52% and the total operating cost by 6.93%, compared to the baseline strategy.
闫军威等[4]以冷水机组总能耗最小为目标,通过建立各台冷机运行能效模型,基于遗传算法提出了具有多台冷机的冷却系统运行优化策略.Abou-Ziyan等[5]分析了不同冷机开启数量对机组综合COP的影响,发现最佳负载率分配策略比传统规则控制(Rule Based Control, RBC)方法能效提高了22%~33%.
因此,本文在充分考虑数据中心冷机序列控制、部分负载运行效率和水蓄冷冷却系统运行策略复杂性的前提下,提出了一种新的MPC方法.该方法使用长短期记忆(Long Short Term Memory, LSTM)神经网络预测未来的环境参数,建立了冷却系统能耗模型,将各参数代入使数据中心运行能耗最低的滚动优化模型,然后使用混合整数线性规划(Mixed Integer Linear Programming,MILP)算法对模型进行求解.通过比较不同冷机最小连续运行时间取值,选择最优值.这种方法可以实现冷水机组运行与负荷需求的部分解耦,使冷水机组能够始终以最高效率运行,从而降低冷却系统的能耗.
本文使用基于高斯过程回归(Gaussian Process Regression,GPR)的非线性回归的方法来拟合冷水机组的性能模型.GPR模型可以有效地处理复杂的非线性关系并提供预测的不确定性估计.此外,GPR模型可以处理噪声数据,这对于实际应用非常重要.将模型输入变量定义为冷水机组的PLR和冷却水进水温度Tcwi,输出变量定义为冷水机组性能系数COP.GPR模型的数学形式如下:
China Electronic Information Industry Development Research Institute. Research report on the development of new data infrastructure[N]. China Information World, 2022-09-05(8).
[3]
ZHANGH N, SHAOS Q, XUH B,et al .Free cooling of data centers:a review[J]. Renewable and Sustainable Energy Reviews,2014, 35: 171-182.
[4]
SAIDUR R. Energy, economics and environmental analysis for chillers in office buildings[J]. Energy Education Science and Technology Part A Energy Science and Research, 2010, 25(1/2): 1-16.
YANJ W, CHENC, ZHOUX, et al. Simulation study on optimal load distribution strategy of multiple chillers[J]. Heating Ventilating & Air Conditioning, 2016, 46(4):98-104.(in Chinese)
[7]
ABOU-ZIYANH Z, ALAJMIA F. Effect of load-sharing operation strategy on the aggregate performance of existed multiple-chiller systems[J]. Applied Energy, 2014, 135:329-338.
[8]
LIAOJ J, XIEX, NEMERH,et al .A simplified methodology to optimize the cooling tower approach temperature control schedule in a cooling system[J]. Energy Conversion and Management, 2019, 199: 111950.
[9]
GUOY B, TANZ H, CHENH X,et al .Deep learning-based fault diagnosis of variable refrigerant flow air-conditioning system for building energy saving[J]. Applied Energy,2018,225:732-745.
[10]
KATHIRGAMANATHANA, DE ROSAM, MANGINAE,et al .Data-driven predictive control for unlocking building energy flexibility:a review[J]. Renewable and Sustainable Energy Reviews, 2021,135:110120.
[11]
ASTEN, MANFRENM, MARENZIG .Building Automation and Control Systems and performance optimization:a framework for analysis[J].Renewable and Sustainable Energy Reviews,2017,75: 313-330.
[12]
SOYGUDERS, ALLIH .Predicting of fan speed for energy saving in HVAC system based on adaptive network based fuzzy inference system[J].Expert Systems with Applications,2009,36(4):8631-8638.
[13]
QURAT-UL-AIN, IQBALS, KHANS A,et al .IoT operating system based fuzzy inference system for home energy management system in smart buildings[J].Sensors,2018,18(9):2802.
[14]
AFRAMA, JANABI-SHARIFIF, FUNGA S,et al .Artificial neural network (ANN) based model predictive control (MPC) and optimization of HVAC systems:a state of the art review and case study of a residential HVAC system[J].Energy & Buildings,2017,141:96-113.
[15]
MAY D, BORRELLIF, HENCEYB,et al .Model Predictive Control of thermal energy storage in building cooling systems[C]//Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference.Shanghai: IEEE,2009:392-397.
[16]
SHANK, FANC, WANGJ Y .Model predictive control for thermal energy storage assisted large central cooling systems[J].Energy,2019,179:916-927.
[17]
DENGK, SUNY, LIS S, et al. Model predictive control of central chiller plant with thermal energy storage via dynamic programming and mixed-integer linear programming[J]. IEEE Transactions on Automation Science and Engineering, 2015, 12(2): 565-579.
[18]
CHANGY C, LINF A, LINC H. Optimal chiller sequencing by branch and bound method for saving energy[J]. Energy Conversion and Management, 2005, 46(13/14): 2158-2172.
LIAOY D, HUANGG S .A hybrid predictive sequencing control for multi-chiller plant with considerations of indoor environment control,energy conservation and economical operation cost[J].Sustainable Cities and Society,2019,49:101616.
[21]
ZOUW K, SUNY J, GAOD C,et al .Robust enhancement of chiller sequencing control for tolerating sensor measurement uncertainties through controlling small-scale thermal energy storage[J].Energy,2023,280:128152.
[22]
LIAOY, HUANGG, DINGY F,et al .Robustness analysis and enhancement of chiller sequencing control under uncertainties[J].Procedia Engineering, 2017, 205: 1878-1885.
[23]
ALRAWASHDEHK, PURDYC. Toward an online anomaly intrusion detection system based on deep learning[C]//2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA).Anaheim,CA: IEEE, 2016: 195-200.
[24]
BERNIERM A, BOURRETB. Pumping energy and variable frequency drives[J]. ASHRAE Journal, 1999, 41(12): 37-40.
基金资助
国家自然科学基金资助项目(52178073)
National Natural ScienceFoundation of China(52178073)
郴州国家可持续发展议程创新示范区建设专项资助项目(2022SFQ28)
Chenzhou National Sustainable Development Agenda Innovation Demonstration Zone Construction Project(2022SFQ28)
江西省教育厅科学技术研究资助项目(GJJ2200683)
Science and Technology Research Project of Jiangxi Provincial Department of Education(GJJ2200683)