Day-Ahead Residential Electricity Demand Response Model Based on Deep Neural Networks for Peak Demand Reduction in the Jordanian Power Sector

Abstract

In this paper, a comprehensive demand response model for the residential sector in the Jordanian electricity market is introduced, considering the interaction between the power generators (PGs), grid operators (GOs), and service providers (SPs). An accurate day-ahead hourly short-term load forecasting is conducted, using deep neural networks (DNNs) trained on four-year data collected from the National Electric Power Company (NEPCO) in Jordan. The customer behavior is modeled by developing a precise price elasticity matrix of demand (PEMD) based on recent research on the short-term price elasticity of Jordan’s residential and the analysis of the different types of electrical appliances and their daily operational hours according to the latest surveys. First, the DNNs are fine-tuned with a detailed feature analysis to predict the day-ahead hourly electrical demand and achieved a mean absolute percentage error (MAPE) of 1.365% and 1.411% on the validation and test datasets receptively. Then the predictions are used as input to a detailed model of the Jordanian power grid market, where a day-ahead peak-time demand response policy for the residential sector is applied to the three distribution power companies in Jordan. Based on different PEMD analyses for the Jordanian residential sector, the results suggest a reduction potential of 5.4% in peak demand accompanied by a cost reduction of USD 154,505 per day for the Jordanian power sector.

Analyzing the Impacts of a Deep-Learning Based Day-Ahead Residential Demand Response Model on The Jordanian Power Sector in Winter Season

“…Once the optimal prices are calculated, they are announced to the Service providers at the start of the next day and the expected impact of the DR on the day-ahead demand is used to further optimize the day-ahead unit commitment scheduling. The profit optimization model used by the GO, considering its energy purchased from the PSs and energy sold to the SPs, is depicted as follows [7]:…”

“…The last part, which is not considered in the proposed model, depicts the monthly peak cost-income 𝑃𝐶𝐼 𝐶 from each bulk consumer, which is a function of the monthly demand peak 𝑑𝑝 𝑐 at specified peak times and the demand peak pricing 𝑑𝑝𝑝 𝑐 . The GO through DR can impact the day-ahead hourly demand per bulk consumer 𝒅𝒔 𝒄,𝒉 by changing the hourly selling price 𝒔𝒑 𝒄,𝒉 in a way that reduces the peak demand as follows [7]:…”

“…To run the day-ahead DR model, the day-ahead hourly demands should be obtained. In this work and relating to our previous work [7], a deep neural network (DNN) was trained on four years of the hourly electrical demand of Jordan's power sector was used. Fig.…”

“…This is done by using very costly and fast ramping energy generators, where a reduction in the peak demand can highly reduce the economic burden on the GO in the peak time. To achieve a reduction in peak demand and influence consumers to shift their demand outside of the peak period, a demand response (DR) system was proposed [7]. DR systems allow the participation of different types of electrical demand end consumers to contribute to the power market and balancing the electrical demandsupply.…”

“…In our previous work, a detailed day-ahead DR model was developed for the residential sector of Jordan's power sector, where a precise day-ahead demand prediction using a deep neural network was introduced, and a detailed day-ahead demand response model was discussed [7]. In this work, we present a detailed analysis of the impact of the proposed model on the winter month of December 2019 as follows:…”