空调系统节能调控中的室温大滞后及智能预测控制方法

1. Target. The proposal will focus on the method to forecast and implement intelligent control for the large time delay performance of indoor temperature during control process of variable air volume air conditioning system (VAV system) by theoretical and experimental researches..2. Significance. The research result will be important theoretical significance for the solution of the large time delay nonlinear control problem of the building thermal environment system. Also, it will be important for the energy-saving and emission reduction and application prospect by increasing the control performance of automation control system for building thermal environment system, increasing the function of buildings, decreasing building energy consumption, promoting the development of intelligent building. .3. Contents. 1) Research on dynamic response characteristics of the large time delay of indoor temperature during control process of VAV system. 2) Research on the method of neural network multistep prediction for large time delay performance of indoor temperature. 3) Research on vector based Takagi-Sugeno self-adaptive neural network fuzzy control model for building thermal environment system and the automatic acquisition by clustering of the fuzzy control rules. 4) Experiment of the neural network multistep prediction control for large time delay of indoor temperature of VAV system in laboratory. .4. Key issues. 1) Dynamic response characteristics and its mathematical description of the large time delay of indoor temperature to its influence factors. 2) The method of neural network multistep prediction for large time delay performance of indoor temperature based on multi-factor thermal resistance network topological structure of building thermal environment system. 3) Buildup of vector based T-S self-adaptive neural network fuzzy control model of building thermal environment system. 4) Automatic acquisition by clustering of the fuzzy control rules of building thermal environment system based on annual typical working conditions of VAV system..5. Methodology. Method of combining with theoretical research, control simulation, technological development and experiment in laboratory.

1. 目的：解决变风量空调系统节能调控中室温大滞后特性的预测和智能控制问题。.2. 意义：对解决建筑热环境系统大滞后非线性控制难题具有重要的理论意义；对提高建筑热环境控制系统的性能、提升建筑功能、降低能耗、促进建筑智能化发展，具有重要的节能减排意义和应用前景。.3. 内容：1）空调系统节能调控中室温大滞后动态响应特性；2）室温大滞后神经网络多步预测方法；3）建筑热环境系统向量基T-S型自适应模糊神经网络控制模型及规则聚类自获取方法；4）空调系统室温大滞后多步预测模糊神经网络控制试验。.4. 关键：1）室温对各类影响因素的大滞后响应特性及数学描述；2）基于多因素热阻网络的室温大滞后特性神经网络多步预测方法；3）向量基T-S型自适应模糊神经网络控制模型的建立；4）以全年典型空调工况为聚类中心的建筑热环境系统T-S型模糊规则聚类获取方法。.5. 方法：理论研究、控制仿真、技术开发和试验相结合。

变风量空调系统室温大滞后特性的预测和智能控制是大型公共建筑空调系统节能控制中的关键问题，本项目针对该问题，通过理论和试验研究，开展了以下研究工作。.首先，建立了多区域双风机变风量空调房间室温变化动态响应数学模型，分析了引起室温滞后响应的影响因素；试验研究了在多区域双风机变风量空调系统动态调节过程中室温变化对各类调节量的动态响应特性，依据换气次数提出了室温变化惯性时间系数与空调房间换气次数之间的反比例关系式。.其次，利用Elman网络动态记录历史信息的能力，提出了Elman网络模型参数在线优化方法，建立了室温纯滞后特性Elman网络多步预测模型，提出了变风量末端多区域双风机变风量空调室温Elman网络多步预测控制方法。.第三，提出了基于末端预测阀位的变静压模糊控制方法，通过对比试验，验证了提出的模糊预测控制方法的有效性及节能效果。.第四，为满足工程应用需求，本项目提出了变风量空调控制系统分布式总线网络架构和变风量空调系统节能群控模糊控制设计方法，完成了变风量空调系统监测控制硬件及软件系统的开发，建立了相应的网络化控制试验平台，开发了空调系统智能管理软件，部分技术在实际空调系统中得到了应用。.本项目发表学术论文31篇，其中，SCI检索论文14篇、EI检索论文7篇。获实用新型专利5项，软件著作权1项。出站博士后2名，培养博士生2名，硕士生5名（其中1人获校优秀硕士学位论文奖）。连续3年与美国宾夕法尼亚州立大学、科罗拉多大学博德分校和日本东京大学等举办建筑智能化国际学术活动周，取得了较好的国际交流与合作效果。.本项目研究成果对解决建筑热环境系统大滞后非线性控制难题具有重要的理论意义；对提高建筑热环境控制系统的性能、提升建筑功能、降低能耗、促进建筑智能化发展，具有重要的节能减排意义和应用前景。