Towards load control of populations of air conditioners with guaranteed comfort margins

The increasing penetration of renewable sources in electricity grids has motivated interest in controlling loads to compensate for variability of these generation sources. Air conditioners (ACs) are one type of loads that can be effectively controlled by broadcasting temperature setpoint offsets and using measurements of aggregate power demand for feedback, as shown in recent works. While such control approach can arbitrarily shape demand, it does so at the expense of end-use comfort, which may be a deterrent to participation in such programmes. This paper explores the use of an alternative to setpoint offsets to shape aggregate demand of ACs without violating end-use comfort levels. The proposed control consists in manipulating the width of the ACs' temperature regulation bands by raising the low temperature limit (or lowering the high temperature limit) to change aggregate demand while keeping AC temperatures within the specified ranges. We develop a deterministic mathematical model for the AC aggregate demand response to changes in the new proposed control input, and numerically validate it against a simulated population of ACs. The resulting model is amenable to systematic analysis and design for direct load control with guaranteed comfort margins. We illustrate this point by deriving a theoretical upper bound on the maximum energy that may be released from a given population of ACs controlled using the proposed control input.