基于贝叶斯网络的冷冻水系统运行诊断  

Bayesian Network-based Operational Diagnosis of Chilled Water Systems

作　　者：刘雪峰[1] 徐瑾蔓 林奇峰 许穗民[3] LIU Xuefeng;XU Jinman;LIN Qifeng;XU Suimin(School of electric power,South China University of technology,Guangzhou,Guangdong 510640;Guangdong Architectural Design and Research Institure Group Co.,Ltd.,Guangzhou,Guangdong 510000;China Light Industry Guangzhou Engineering Co.,Ltd,Guangzhou,Guangdong 510000;Guangdong Architectural Design and Research Institute Group Co.,Ltd,Ltd,Guangzhou,Guangdong 510000)

机构地区：[1]华南理工大学电力学院,广东广州510640 [2]中国轻工业广州工程有限公司,广东广州510000 [3]广东省建筑设计研究院集团股份有限公司,广东广州510000

出　　处：《制冷》2025年第1期37-42,共6页Refrigeration

摘　　要：大型中央空调系统涉及的变量数目庞大,已然成为多变量相互耦合、非线性的复杂系统。倘若系统存在运行故障,在不对故障问题进行改进的情况下一味的优化系统能效是毫无意义的。若要根据实时的数据准确的判断管网系统运行状况,这要求庞大的管网系统中众多节点具有数据监测能力。目前,大型系统可获得的实时数据远不能支撑实时运行调节的判断,因此本文通过冷冻水管网系统物理模型大数据模拟,通过人为制造不同扰动构建空调系统运行数据库,弥补了实际空调系统稳定运行带来的数据稀疏问题,解决了传统依赖于大量多种工况数据库的诊断分析。然后,阐述如何利用贝叶斯概率建立空调系统的运行诊断模型。最后,我们将通过实例验证贝叶斯概率在空调系统运行诊断中的有效性。Large central air-conditioning systems involve a large number of variables and have become complex systems with multiple variables that are coupled and non-linear.If the system has operational faults,it is pointless to optimize the system's energy efficiency without improving the faults.To accurately judge the operating conditions of a pipeline system based on real-time data requires the data monitoring capability of many nodes in a large pipeline system.At present,the real-time data available for large-scale systems is far from supporting the judgement of real-time operation and regulation,therefore,this paper builds an air-conditioning system operation database by creating dfferent perturbations artificially through the big data simulation of the physical model of the chilled water pipeline network system to make up for the data sparsity problem brought about by the stable operation of the actual air-conditioning system,and solves the diagnosis and analysis traditionally relying on a large number of multiple working condition databases.Then,we will explain how to use Bayesian probability to establish the operation diagnostic model of air-conditioning system.Finally,we will verify the effectiveness of Bayesian probability in the operation diagnosis of air-conditioning systems through examples.

关 键 词：中央空调 冷冻水系统 贝叶斯诊断 

分 类 号：TK1[动力工程及工程热物理—热能工程] TU111[建筑科学—建筑理论]