Optimizing Control of Total Heat Supply Based on Machine Learning

doi:10.16182/j.issn1004731x.joss.201803046

Abstract

Abstract: The central heating system has complex structure, along with the characteristics of hysteresis, strong coupling and nonlinear. Contraposing the problem that the process is difficult to be identified and controlled by the mechanism modeling, an optimal control method of heat source total heat production based on machine learning is proposed. The heat source model of central heating system is established by BP neural network and long short-term memory neural network. Under the premise of meeting the demand of heating quality, with the total energy consumption as the optimization objective, the optimal control sequence of water supply temperature and water flow at heat source is obtained by the action-dependent dual heuristic programming (ADDHP) algorithm. The simulation analysis shows that, the established heat source model can effectively identify the heat source production process, and the ADDHP control method can achieve the optimal control of total heat production of heat source.

Key words: machine learning, long short-term memory neural network, action-dependent dual heuristic programming, heat source, optimal control

CLC Number:

TP301

Li Qi, Hu Xingqi, Zhao Jianmin. Optimizing Control of Total Heat Supply Based on Machine Learning[J]. Journal of System Simulation, 2018, 30(3): 1134-1143.

References 19

[1]	贠海. 城市集中供热问题分析与探索[J]. 科技创新导报, 2013(16): 131. Yun Hai. The city central heating problem analysis and exploration of [J]. Technology Innovation Herald, 2013 (16): 131.
[2]	O'Dwyer E, Tommasi L D, Kouramas K, et al. Modelling and Disturbance Estimation for Model Predictive Control in Building Heating Systems[J]. Energy & Buildings(S0378-7788), 2016, 130: 532-545.
[3]	魏宇杰. 集中供热系统控制策略研究[D]. 山西: 太原理工大学, 2015. Wei Yujie. Study on control strategy of central heating system [D]. Shanxi: Taiyuan University of Technology, 2015.
[4]	王源. 城市集中供热系统建模及其控制方法研究[D]. 南京: 东南大学, 2013. Wang Yuan. Study on Modeling and control method of urban central heating system [D]. Nanjing: Southeast University, 2013.
[5]	楼海雷, 高前东. 质量并调运行方式的实现与节能效果分析[J]. 区域供热, 2012, 1(5): 63-66.Lou yuan, Gao Qian Dong. Realization of quality adjustment and operation mode and analysis of energy saving effect [J]. regional heating, 2012, 1(5): 63-66.
[6]	李丽. 集中供热系统的研究与优化[D]. 保定: 华北电力大学, 2009. Li Li. Research and optimization of central heating system [D]. Baoding: North China Electric Power University, 2009.
[7]	郑泽宇, 顾思宇. TensorFlow实战Google深度学习框架[M]. 1版. 北京: 电子工业出版社, 2017: 199-210. Zheng Zeyu, Gu Siyu. TensorFlow, practical Google, deep learning framework, [M]. first edition. Beijing: Publishing House of electronics industry, 2017: 199-210.
[8]	林晓峰, 宋绍剑, 宋春宁. 基于自适应动态规划的智能优化控制[M]. 北京: 科学出版社, 2013: 77-121. Lin Xiaofeng, Song Shaojian, Song Chunning. Intelligent optimization control based on adaptive dynamic programming[M]. Beijing: Science Press, 2013: 77-121.
[9]	赵冬斌, 邵坤, 朱圆恒, 等. 深度强化学习综述: 兼论计算机围棋的发展[J]. 控制理论与应用, 2016, 33(6): 701-717. Zhao Dongbin, Shao Kun, Zhu Yuanheng, et al. A survey of depth reinforcement learning: also on the development of computer weiqi [J]. control theory and application, 2016, 33(6): 701-717.
[10]	段艳杰, 吕宜生, 张杰, 等. 深度学习在控制领域的研究现状与展望[J]. 自动化学报, 2016, 42(5): 643-654. Duan Yanjie, Lü Yisheng, Zhang Jie, et al. Deep learning research status quo and Prospect [J]. Journal of automation in the field of control, 2016, 42(5): 643-654.
[11]	杨建坤. 北方小城镇供热模式分析与热网优化控制的研究[D]. 上海: 同济大学, 2007. Yang Jiankun. Analysis of heating mode and optimization control of heat supply network in small towns in North China [D]. Shanghai: Tongji University, 2007.
[12]	周守军. 基于管网动态模型的城市集中供热系统参数预测及运行优化研究[D]. 山东: 山东大学, 2012. Zhou Shoujun. Research on parameter prediction and operation optimization of urban central heating system based on pipe network dynamic model [D]. Shandong: Shandong University, 2012.
[13]	冯小峰, 刘明波. 基于执行依赖启发式动态规划的紧急电压控制[J]. 华南理工大学学报(自然科学版), 2014, 42(4): 19-25, 58. Feng Xiaofeng, Liu Mingbo. Emergency voltage control based on execution dependent heuristic dynamic programming [J]. Journal of South China University of Technology (natural science edition), 2014, 42(4): 19-25, 58.
[14]	Wang F Y, Zhang H, Liu D.Adaptive Dynamic Programming: An Introduction[J]. IEEE Computational Intelligence Magazine(S1556-603X), 2009, 4(2):39-47.
[15]	魏庆来. 基于近似动态规划的非线性系统最优控制研究[D]. 沈阳: 东北大学, 2008. Wei Qinglai. Optimal control of nonlinear systems based on approximate dynamic programming Study [D]. Shenyang: Northeastern University, 2008.
[16]	Pirouti M, Bagdanavicius A, Ekanayake J, et al.Energy consumption and economic analyses of a district heating network[J]. Energy(S0360-5442), 2013, 57(8):149-159.
[17]	Prokhorov D V, Wunsch D C.Adaptive critic designs[J]. IEEE Transactions on neural Network(S1045-9227), 1997, 8(5): 997-1007.
[18]	Prokhorov D V, Wunsch D C.Adaptive critic designs.[J]. IEEE Transactions on Neural Networks, 1997, 8(5): 997-1007.
[19]	唐巍. 基于质量并调的集中供热系统总供热量优化研究[D]. 包头: 内蒙古科技大学, 2016. Tang Wei. The total heating central heating system and the amount of quality optimization research based on [D]. Baotou: Inner Mongolia University of Science and Technology, 2016.