Flexibility capacity of thermostatically controlled loads in a distribution network

Thermostatically Controlled Loads (TCLs) such as air conditioners and water heaters are a source of flexible power demand for the power grid: many different power consumption trajectories exist that can maintain consumers’ quality of service (QoS). An important problem is to quantify the flexibility capacity of a collection of TCLs. Prior work has either ignored the effect of the distribution network (such as voltage limits) in quantifying flexibility capacity of collections of TCLs, or considered only temperature constraints of the TCLs, ignoring the cycling constraints. This paper provides a computationally tractable characterization of flexibility capacity of TCLs spread over an unbalanced distribution network. This characterization can be used to schedule an optimal substation level demand deviation trajectory for the TCL collection in the distribution network: closest to what the grid operator needs but feasible for the collection, meaning no TCL needs to violate its local quality of service constraints in tracking that reference. Numerical verification of the method is provided in simulation with the IEEE 13 bus system.