Price-directed demand in smart grids operating within deregulated electricity markets calls for real-time forecasting of the price of electricity for the purpose of scheduling demand at the nodal level (e.g., appliances, machines, and devices) in a way that minimizes energy cost to the consumer. In this paper, a novel hybrid methodology for electricity price forecasting is introduced and applied on a set of real-world historical data taken from the New England area. The proposed approach is implemented in two steps. In the first step, a set of relevance vector machines (RVMs) is adopted, where each RVM is used for individual ahead-of-time price prediction. In the second step, individual predictions are aggregated to formulate a linear regression ensemble, whose coefficients are obtained as the solution of a single objective optimization problem. Thus, an optimal solution to the problem is found by employing the micro-genetic algorithm and the optimized ensemble is employed for computing the final price forecast. The performance of the proposed methodology is compared with performance of autoregressive-moving-average and naïve forecasting methods, as well as to that taken from each individual RVM. Results clearly demonstrate the superiority of the hybrid methodology over the other tested methods with regard to mean absolute error for electricity signal pricing forecasting.
Integration of energy systems with information technologies has facilitated the realization of smart energy systems that utilize information to optimize system operation. To that end, crucial in optimizing energy system operation is the accurate, ahead-of-time forecasting of load demand. In particular, load forecasting allows planning of system expansion, and decision making for enhancing system safety and reliability. In this paper, the application of two types of kernel machines for medium term load forecasting (MTLF) is presented and their performance is recorded based on a set of historical electricity load demand data. The two kernel machine models and more specifically Gaussian process regression (GPR) and relevance vector regression (RVR) are utilized for making predictions over future load demand. Both models, i.e., GPR and RVR, are equipped with a Gaussian kernel and are tested on daily predictions for a 30-day-ahead horizon taken from the New England Area. Furthermore, their performance is compared to the ARMA(2,2) model with respect to mean average percentage error and squared correlation coefficient. Results demonstrate the superiority of RVR over the other forecasting models in performing MTLF.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.