基于改进ARX模型的房间冷负荷预测  

Prediction of Room Cooling Load Based on Improved ARX Model

作　　者：邰敏 李占培[1] 刘廷章[1] 金碧瑶[1] 薛凡 TAI Min;LI Zhan-pei;LIU Ting-zhang;JIN Bi-yao;XUE Fan(School of Mechanical Engineering and Automation,Shanghai University,Shanghai 200072,China)

机构地区：[1]上海大学机电工程与自动化学院

出　　处：《控制工程》2019年第12期2258-2263,共6页Control Engineering of China

基　　金：国家自然基金项目(61273190);上海市自然科学基金(13ZR1417000)

摘　　要：实时冷负荷的精确预测是优化空调系统运行的关键。针对传统的基于室外气象参数和历史冷负荷的ARX模型低普适性的问题,从变量区间划分出发,提出了带温度索引的ARX模型和基于最小二乘支持向量机(LSSVM)的ARX模型。仿真实验结果表明,提出两个模型相比于传统ARX模型,精度均有大幅提升。基于LSSVM的ARX模型具有最高的预测精度和普适性。Accurate prediction of real-time cooling load is the fundamental work for optimizing the operation of air conditioning systems. Inspired by interval partitioning of variables, two improvements of ARX model are proposed, which are based on temperature index and least squares support vector machine(LSSVM), to solve the problem that traditional ARX model based on outdoor weather parameters and historical cooling load has low universality. Compared with the traditional ARX model, simulation results show that accuracies of the two proposed models are both greatly improved. The ARX model based on LSSVM has the highest prediction accuracy and universality.

关 键 词：房间冷负荷 ARX 温度索引 LSSVM 建筑节能 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]