Short-term price forecasting considering distributed generation in the price-sensitive environment of smart grids

In smart grids environment, all participants, including different load types, are able to utilize the network. Advances in measurement tools in these networks make it easy to move towards dynamic pricing and non-fixed electricity tariffs. In this environment, the forecasted electricity price is declared to wholesale and retail consumers via Advanced Measurement Instruments (AMI). Therefore, motivated by different factors such as optimizing the economic/environmental issues or increasing the reliability, the customers are able to react to prices and manage their consumption. This pattern of reaction brings about extensive changes in the load and price curves of the network. On the other hand, environmental concerns resulting from the use of fossil fuels have increased the exploitation of renewable energies. But, as the penetration of renewable energy sources increases, serious improvements and modifications for the existing electric grid are needed to accommodate and integrate these intermittent sources. In this paper, a hybrid model is presented for simultaneous short term forecasting of electricity prices considering Distributed Generation (DG) in the price-sensitive environment of smart grids. The proposed model combines the Support Vector Regression (SVR) network with an Adaptive Neuro Fuzzy Inference System (ANFIS) network, and it is capable of tracking customers' reaction to declared prices. This model is applied on the data of power markets of Nordpool regrion, Denmark, where the smart grids are very active. The results of short term price forecasting for a target day (1/1/2016) shows the accuracy of the model.