This report discusses the testing of the stochastic control algorithm S-TEO (Stochastic Technical Economic Optimization) for the calculation of the optimal production profiles of microgrid devices, in order to minimize next-day operating costs (typically for energy and gas). RSE’s Distributed Energy Test Facility (DER-TF) has been used for the testing.
Operating deterministically, such a type of scheduling assumes having reliable forecasts of exogenous variables such as PV production; however, even having reliable forecasts available, the unpredictability of weather phenomena is such that corrective actions often have to be taken the next day, thus leading to additional costs.
As opposed to deterministic management, stochastic management is based on the idea of explicitly considering the randomness of certain quantities in the optimization issue, such as local consumption and uncontrollable local production (e.g., from PV panels). In this way, the optimization issue may also include the computation of disturbance compensation strategies that will be followed on the next day, depending on the specific random quantities that will occur.
S-TEO was first tried in a simulated context, then implemented in the control system of the DER-TF, and finally tested in the field in different usage scenarios. The tests proved how effective the adopted control strategy were in terms of minimizing cost and error in power exchange compared to a deterministic strategy.
This report also discusses a control strategy for aggregates of prosumer systems equipped with generation equipment and storage systems to feed their loads but also available for participation in the dispatch service market (DSM).To this end, a system was implemented for optimal, real-time dispatch of aggregator setpoints to aggregate storage systems. A series of simulated tests demonstrates the effectiveness of the designed control strategy.