A Fault Detection And Diagnosis Method Based On Kriging Model For Chillers

In daily life and working and learning processes, the role of heating, ventilation and air conditioning(HVAC) system cannot be replaced. However, more and more learners paid attention to it, as a major component in building, when considering energy saving at present. It is known from statistics that energy consumption of the system amount to more than 60 percent of total building energy consumption, where chillers account for 40 to 50 percent. Thus when fault occurs in chillers, it affects not only operation efficiency and safety, but also the thermal comfort of the indoor space,further increasing the energy consumption of the entire system. Therefore, from the facet of energy consumption, to ensure normal operation is really of great importance.What’s more, when faults occeured, timely and reliablly using fault detection and diagnosis(FDD) strategy can be practically meaningful in high-effiency chiller operation and energy saving area.After analysing a good number of relative researches that I have already read,this paper proposed Kriging model which is introduced to detect the faults in the chillers of building air-conditioning systems. This model is built and validated by using the normal data from ASHRAE RP-1043. The method of T-statistic and exponentially-weighted moving average(EWMA) are compared by the sensitivity of performance indexes. The results show that EWMA can achieve better performance sensitivity. Combined with EWMA and Kriging model and the rules of fault diagnosis,the chiller faults like condenser fouling, refrigerant overcharge, refrigerant leakage,non-condenser gas, reduced evaporator water flow rate and reduced condenser water flow rate are diagnosed using the measured data from ASHRAE RP-1043. The diagnosis results have shown that the chiller faults at different levels can be accurately and efficiently detected and diagnosed by using Kriging model. As it is known, correct rates and fault rates will be different with different diagnosis rules.Conventional two rules are compared in this article by applying Kriging model and EWMA method, and the result shows it is not obvious to tell that the rule based on sensitivity is much better than the rule based on changing direction of performance index.