Optimal unit commitment accounting for robust affine reserve policies

We describe a new approach to robust unit commitment for an electricity network, which couples the switching decisions to a set of time-coupled redispatch rules in order to minimize the expected cost of operation over a planning horizon. We assume bounded uncertainties arising from imperfect predictions of loads and intermittent renewable infeeds. We refer to the time-coupled redispatch rules as affine reserve policies, and they apply not only to generators but to other continuously-controllable devices such as energy storage units or demand response installations, which are modelled as generic mixed logical-dynamical systems. We use a lumped-parameter example to demonstrate that unit commitment decisions coupled with affine reserve policies can reduce the number of time periods in which expensive peaking plants need to be employed, and that the effect is present for a wide range of parameter values. An important benefit is that the approach allows the uncertainty of future energy storage levels to be managed more tightly than existing formulations would allow.