基于梯度提升回归树负荷预测的水蓄冷空调系统运行优化  

Operation Optimization for Chilled Water Storage Air Conditioning System Based on Load Prediction Using Gradient Boosting Regression Tree

作　　者：李玲荣 贾志洋 薛琪 吕远 王江情 晋欣桥[1] LI Lingrong;JIA Zhiyang;XUE Qi;LYU Yuan;WANG Jiangqing;JIN Xinqiao(School of Mechanical and Power Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学机械与动力工程学院,上海200240

出　　处：《制冷技术》2024年第5期9-15,共7页Chinese Journal of Refrigeration Technology

基　　金：国家自然科学基金(No.51776118)。

摘　　要：本文提出一种基于超参数优化负荷预测的水蓄冷空调系统运行优化策略。采用梯度提升回归树算法建立负荷预测模型,并利用基于贝叶斯优化的树状帕曾密度估计算法进行超参数优化;基于负荷预测结果,对机组启停组合和蓄/释冷量进行寻优,从而优化水蓄冷空调系统的运行。以某工厂空调系统为研究对象,增设水蓄冷系统,并对提出的优化运行策略进行了验证。与原系统相比,改造并运行优化后的系统,高、中和低负荷三个典型日的日平均性能系数分别提高了4.84%、3.55%和2.67%,节能率分别为4.60%、3.14%和2.48%。In this paper,an operation optimization strategy for water storage air conditioning system based on hyperparameter optimization load prediction is proposed.The Gradient Boosting Regression Tree algorithm is used to establish the load prediction model,and the Tree-structured Parzen Estimator algorithm based on Bayesian optimization is used for hyperparameter optimization.Based on the load prediction results,the start-stop combination of the chillers and the charging/discharging capacity are optimized,to optimize the operation of the water storage air conditioning system.Taking the air conditioning system of a factory as the research object,a water storage system is added,and the proposed optimization operation strategy is verified.Compared with the original system,the average daily performance coefficients of the three typical days of high,medium and low load are increased by 4.84%,3.55%and 2.67%,respectively,and the energy saving rates are 4.60%,3.14%and 2.48%,respectively.

关 键 词：水蓄冷 负荷预测 超参数优化 运行优化 

分 类 号：TB611[一般工业技术—制冷工程] TK02[动力工程及工程热物理]