On the Accuracy of the State-Queuing Model for the Thermostatically Controlled Loads

The development of smart metering technologies makes it possible to apply the thermostatically controlled loads (TCLs) into the demand response (DR) programs such as direct load control (DLC). To characterize the dynamics of a large number of TCLs, the computationally efficient sate-queuing (SQ) model is designed in some recent research works. However, very few attentions have been given to its accuracy. This paper reviews the existing SQ modeling approach, and reveals its limitations in the accuracy by comprehensive comparison and analysis. The results show that the aggregate TCLs do not exactly have the Markov property, and the existing SQ modeling approach cannot accurately characterize the aggregate dynamics of the TCLs. The findings of this paper are very important for the management and control of large amount of TCLs.