Development of Fault Detection and Diagnosis Tool for Building Air-conditioning Systems Based on Knowledge Database

建筑空调系统的故障检测与诊断(FDD)是调试过程中必须验证的重要项目之一。然而，如果没有实际使用的能力，开发FDD工具是非常不够的。本研究的目的是开发一种简单易用的FDD工具，帮助在整个建筑空调系统中找到故障位置。本研究取得的成果如下：1.编制了某实验小型建筑空调系统的模拟程序，作为FDD工具的发射器，并利用实测数据对模拟参数进行了标定。该仿真程序可以较高精度地模拟室内空气温度、系统COP和系统日耗电量。开发了原型FDD工具，并将其引入实验楼的BEMS中，验证了该工具的实用性。在夏季实验中，多次重现了室内温度设定故障、送风风机温度控制故障和HPC产生的冷热水温度设定故障，并尝试用所研制的检测工具对其进行检测。在测试结果中，该工具能够检测到故障的概率约为90%。通过对某办公楼实际空调系统的模拟案例研究，对该工具的节能效果进行了定量检验。结果表明，该仪器能较高概率地检测出夏、春、冬三季产生的故障，与不使用该仪器的情况相比，每年可节省5-10%的能耗。

Fault detection and diagnosis (FDD) in building air conditioning systems is one of essential items which should be verified in commissioning process. Development of a tool to FDD is, however, extremely insufficient without ability to use practically. The purpose of this study is to develop an easy-to-use FDD tool that helps to find the faulty place through the whole building air conditioning system. The results obtained in this study are as follows;1. The simulation program of an air-conditioning system introduced in an experimental small building has been developed as a launcher of FDD tool, and the simulation parameters have been calibrated by the measured data. The indoor air temperature, system COP and system daily electric power Consumption could be simulated in high accuracy by the simulation program.2. The prototype FDD tool has been developed, and verified the tool's practicability by introducing in BEMS of the experimental building. In summer experiments, three kinds of fault have been reproduced several times; fault to set up temperature in rooms, fault to control temperature at the fan for the air supply and fault of cold/hot water set up temperature generated by the HPC, and tried to detect them by the developed tool. In the results, the probability that the tool could detect the faults was about 90%.3. The tool's energy saving effect has been examined quantitatively by simulation case studies targeted an actual air-conditioning system in an office building. In the results, it has been clarified that faults generated in summer, spring and winter seasons could be detected in high probability, and the tool could save energy consumption 5-10% yearly, compared to a case that the tool was not used.