 |
Center for Nonlinear Studies |
Recent works have shown that control strategies can be adopted to perform regulation and load-following in aggregate power demand of a population of thermostatically controlled loads (TCLs). Instead of traditionally used direct load control programs which act by directly interrupting power, the newly proposed hysteresisbased control works on manipulating the thermostat set points of the loads. While it’s difficult to keep track of the temperature and power demand of the individual loads in the population, the probability of the temperature of each load being in a given state (ON or OFF) can be estimated rather accurately. In order to control the aggregate power demand of such a population, it is necessary to analytically quantify the response of the population to set point disturbance. In our works we have established an analytical relation between the change in the aggregate power demand of the loads and a change in the set point. With the help of our model of the system suitable controllers can be designed to make the aggregate power track a desired power waveform, like that coming from a renewable energy source. We have designed a proportional-integral controller which made the aggregate power track reference signals in the form of step, ramp and sinusoid. While our analytical model assumes a low-diffusion, homogeneous system of loads, numerical studies have been done to assess the situation when there is higher noise and heterogeneity.
In collaboration with Dr. Nikolai Sinitsyn, Dr. Scott Backhaus & Dr. Misha Chertkov, LANL and Prof. Ian Hiskens, EECS, U of Michigan