データセンター用空調機を対象とする 凝縮器の診断

抄録

In this paper, diagnosis technique to estimate a clogging state of the condenser using the dynamic data of the air conditioner was proposed to optimize the cleaning time of the condenser in a data-center. Dynamic model of a condenser was determined based on the heat balance between the refrigerant - air - condenser in the condenser. Input data and output data measured under normal condition was used to determine model parameters. To estimate a clogging state of the condenser, target input data was inputted into the model and compared output data from the model with the measured output data. In the case that difference between model output and measured output was large, clogging occurred. In this paper, dynamic data of an air conditioner was gotten under the condition of outside temperature of 15 degrees Celsius and 35 degrees Celsius. Our proposed technique was checked using measured data and showed good performance regardless of outside temperature.