考虑需求响应的变风量空调系统强化学习静压控制  

Reinforcement Learning Control of Static Pressure in Variable Air Volume Air-conditioning System with Demand Response

作　　者：韦应安 孟庆龙 辛东岳 任奕欣 杨洋 王钰翔 李彦鹏 WEI Yingan;MENG Qinglong;XIN Dongyue;REN Yixin;YANG Yang;WANG Yuxiang;LI Yanpeng(School of Civil Engineering,Chang'an University,Xi'an 710061,China)

机构地区：[1]长安大学建筑工程学院,西安710061

出　　处：《建筑科学》2024年第8期208-219,共12页Building Science

基　　金：陕西省自然科学基础研究计划资助项目(2023-JC-YB-335)。

摘　　要：需求响应(DR)策略被提出用于解决电网中的电力不平衡等问题。暖通空调(HVAC)系统的节能降耗是建筑节能的重要组成部分,变风量(VAV)空调系统的运行是1个复杂的动态过程,而强化学习(RL)算法在动态控制方面具有巨大潜力。本研究的重点是基于RL,结合DR策略,优化VAV空调系统的静压控制。基于上述背景,提出了基于DR-RL的VAV空调系统静压优化控制策略。建立了TRNSYS+Python联合仿真平台进行RL模型训练,并在物理实验平台上测试了基于DR-RL的静压优化控制策略。与传统的静压控制策略相比,在保证基本热舒适的前提下,所提策略有效实现了DR期间风机负荷削减及全天运行费用节省。Demand response(DR)strategies have been proposed to solve problems such as power imbalance in the grid.Energy saving in heating,ventilation and air-conditioning(HVAC)systems is an important part of building energy efficiency.The operation of variable air volume(VAV)air-conditioning systems is a complex dynamic process,and reinforcement learning(RL)algorithms have great potential for dynamic control.This study focused on optimizing the static pressure control of VAV air-conditioning systems based on RL,combined with DR strategies.Based on the above background,a DR-RL-based static pressure optimization control strategy for VAV air-conditioning systems was proposed.A joint TRNSYS+Python simulation platform was established for RL model training,and the DR-RL-based static pressure optimization control strategy was tested on a physical experimental platform.Compared with the traditional static pressure control strategy,the proposed strategy effectively achieves fan load reduction during DR and all-day operation cost saving under the premise of ensuring basic thermal comfort.

关 键 词：变风量空调系统 静压控制 强化学习 需求响应 控制策略优化 

分 类 号：TU831.3[建筑科学—供热、供燃气、通风及空调工程]