Data-driven Nonlinear Prediction Model for Price Signals in Demand Response Programs (Lærkevej)

In dynamic prices-based demand response (DR) schemes, electricity consumers receive a time-varying price signal by their home energy management systems (i.e., local controllers) to schedule their individual generation and consumption, aiming to minimize the overall cost while providing services to the grid. Although dynamic price-based DR programs neither affect consumers’ autonomy nor privacy, they expect DR operators to formulate electricity price signals. By influencing consumers’ aggregate price response, dynamic prices might also affect power system’ security. Therefore, a robust price generating model considering heterogeneous consumers’ dynamics and different DR capabilities needs to be developed. In this paper, we propose a nonlinear auto-regressive with exogenous inputs (NARX) model to predict dynamic electricity prices. The main objective of this study is to learn consumers’ flexibility behavior in relation to different factors. Such an understanding can be exploited by a DR operator to formulate adequate electricity prices that can achieve a certain change in load under dynamic prices-based schemes.