项目名称: 基于数据挖掘方法及不确定性分析对公共建筑冷水机组群控策略的识别、评估和优化方法研究
项目编号: No.51508394
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 建筑科学
项目作者: 李铮伟
作者单位: 同济大学
项目金额: 20万元
中文摘要: 冷水机组的群控策略由于涉及到建筑能耗、机组寿命、及室内热环境的稳定性,是冷水机组运行中一个关键内容。学术界的研究表明,通过对冷水机组的群控策略进行优化,可以节约10%以上的能源。然而,实际工程领域对于群控策略的优化应用较少,一个重要原因在于冷负荷测量存在不确定性、优化算法可靠性不高。此外,由于冷水机组运行管理的复杂性,节能并不是其运行的唯一目标。随着大型公共建筑的用电分项计量在很多地区得到推广,对冷水机组群控策略进行在线优化操作上已经可行。因此,当前急需一套能够对群控策略进行科学、全面的评价,并提高其综合运行目标的方法。另一方面,随机优化算法适合于参数不确定条件下的优化,而数据挖掘则适合于在运行阶段动态识别设备的能效模型及运行策略,因此,本课题提出结合数据挖掘方法和随机优化算法,对冷水机组群控策略的识别、评估和优化方法进行研究。该课题的实施,对于推进公共建筑节能的工作具有重大意义。
中文关键词: 建筑节能;冷水机组群控;数据挖掘;不确定性分析
英文摘要: Among various chiller operation strategies, sequencing control is a key component due to its impact on energy consumption, chiller's health, and indoor thermal environment. Research has shown that, optimization of chiller sequencing control strategy can save energy by more than 10%. However, application of optimal chiller sequencing control in practice is rarely seen, an important reason is the existence of uncertainty of cooling load measurement, and the low reliability of optimal algorithm. Furthermore, as mentioned above, energy conservation is not the only target of chiller sequencing control. Along with the popularity of electricity sub-metering system in large public buildings, online optimization of the chiller sequencing control logic is practically feasible. Thus, a systematic method to analyze, evaluate, and optimize the sequencing control strategy is urgently needed. On the other hand, stochastic optimization method has been proved to be a suitable approach given uncertain parameters, and data mining method has been shown as an effective method in identifying the dynamic energy consumption model and control strategy of chillers. For this reason, it is proposed to combine data mining method and stochastic optimization method, to identify, evaluate, and optimize chiller sequencing control strategy. The outcome of this project could contribute significantly to energy conservation in the public building sector.
英文关键词: building energy conservation;chiller sequencing control;data mining;uncertainty analysis