Bayesian Networks for Whole Building Level Fault Diagnosis and Isolation
Bayesian Networks for Whole Building Level Fault Diagnosis and Isolation
复制标题
DOI：
--
复制
发表时间：
2018-07
期刊：
影响因子：
--
通讯作者：
Y. Chen;Jin Wen;T. Chen
中科院分区：
文献类型：
--
作者：
Y. Chen;Jin Wen;T. Chen
在美国，建筑物消耗了大约40%的一次能源，而商业建筑中51%的一次能源消耗用于供暖、通风和空调系统。故障的传感器、组件和控制系统以及退化的HVAC和照明组件是导致能源浪费和室内环境不满意的主要原因。在建筑物HVAC系统中，一个部件或设备发生故障可能导致其他封闭子系统的异常。因此，系统级故障诊断方法有助于定位此类故障的根本原因。贝叶斯网络（BN）是故障诊断中的一种常用工具，可以处理不确定性的概率推理。本文提出了一种由故障层和故障征兆层组成的两层贝叶斯网络，用于诊断建筑暖通空调系统在制冷运行模式下影响多个子系统的系统级故障。基于天气/进度信息的模式匹配（WPM）方法被开发，创建基线数据，并为开发的BN生成泄漏概率。在一个校园建筑的BAS数据收集在冷却季节，以评估所提出的方法的有效性。
Buildings consume about 40% of primary energy in the U.S., and 51% of the primary energy usage in commercial buildings are consumed by heating, ventilation and air conditioning (HVAC) system. Malfunctioning sensors, components, and control systems, as well as degrading HVAC and lighting components are main the reasons for energy waste and unsatisfactory indoor environment. In building HVAC systems, faults occurring in one component or equipment can cause abnormality in other closed subsystems. Therefore, a system level fault diagnosis method is helpful to locate root-cause for such faults. Bayesian network (BN) is a prevalent tool in fault diagnosis which can handle probabilistic reasoning of uncertainty. In this paper, a two-layer Bayesian network which consists of fault layer and fault symptom layer is developed to diagnose system level faults that have an impact on multiple subsystems for building HVAC system during a cooling operation mode. Weather/schedule information based Pattern Matching (WPM) method is developed create the baseline data and to generate LEAK probabilities for the developed BN. BAS data from a campus building during the cooling season are collected to evaluate the effectiveness of the proposed method.