多区域建筑灰箱建模及参数辨识方法  

Multi-zone Building Grey-box Modeling and Parameter Identification Method

作　　者：赵安军[1] 米璐 于军琪[1] 查波 魏渊 ZHAO Anjun;MI Lu;YU Junqi;ZHA Bo;WEI Yuan(School of Building Services Science and Engineering,Xi'an University of Architecture and Technology,Xi'an 710055,China;China Northwest Architecture Design and Research Institute Co.,Ltd,Xi'an 710018,China)

机构地区：[1]西安建筑科技大学建筑设备科学与工程学院,西安710055 [2]中国建筑西北设计研究院有限公司,西安710018

出　　处：《建筑科学》2023年第12期222-231,254,共11页Building Science

基　　金：国家自然科学基金重点项目“含氢多能源供需系统协同运行的基础理论与关键技术”(62192750)。

摘　　要：智慧供热的发展允许建筑各区域能够独立监测及控制室温,为实现室温精确预测以制定各区域温度动态调节的最优控制策略,建立1种简化的多区域建筑灰箱模型并提出参数辨识方法。首先考虑相邻区域间的热耦合,基于等效电路法建立多区域灰箱模型结构;其次利用粒子群优化(Particle Swarm Optimization,PSO)算法辨识模型参数,同时引入人工蜂群(Artificial Bee Colony,ABC)算法改善PSO易陷入局部最优的缺陷,以此提高辨识精度;最后以西安市某高校7#宿舍楼为研究对象,基于相关供暖数据进行模型验证。实验结果表明,辨识得到的室温预测模型在不同天气和相邻房间温度发生变化的情况下,预测室温均能稳定跟踪实际室温的变化趋势,误差保持在±0.5℃之内,具有良好的预测精度和鲁棒性,可以满足实际工程需求。The development of intelligent heating allows each zone of the building to monitor and control room temperature independently.To achieve an accurate prediction of room temperature to develop the optimal control strategy for the regional temperature dynamic adjustment,a simplified grey-box model and parameter identification method for multi-zone building were proposed.The thermal coupling between adjacent zones was considered,and a multi-zone building grey-box model structure was established based on an equivalent circuit method.Particle Swarm Optimization(PSO)algorithm was then used to identify the model parameters,and Artificial Bee Colony(ABC)algorithm was introduced to improve the defect of PSO prone to fall into local optimal.Taking the 7#dormitory building of a university in Xi'an as the research object,the model was verified based on the measured heating data.Experimental results show that the identified room temperature prediction model can stably track the changing trend of actual room temperature under different weather conditions and the dynamic change of temperature in adjacent rooms,with the error kept within±0.5℃,indicating good prediction accuracy and robustness,and ability to meet the needs of practical engineering.

关 键 词：多区域建筑 灰箱模型 参数辨识 室温预测 粒子群优化算法 人工蜂群算法 

分 类 号：TU111[建筑科学—建筑理论]