Deep Learning-based Power Load Shedding Approach for Gaza's Electricity Grid

Abstract

Many developing countries suffer from a chronic electricity deficit. As a particular case, this shortage in electricity supplies in Gaza Strip undermines the already fragile conditions, and is considered as an obstacle to any economic development. In order to distribute the available amount of electricity evenly, the local utility executes plentiful on/off power load shedding daily. However, further complications are experienced because of variations in demands and sources daily. Here, a high-precise prediction model for load demand is proposed. This is to help staff in making quick and accurate decisions. The model is built using an artificial neural network (ANN), and trained based on supervisory control and data acquisition (SCADA)'s data including generation, temperature, humidity, and wind speed. The model can predict the energy consumption for given parameters, and hence identifies the power load shedding schedule needed for the given day. The proposed model was evaluated in terms of accuracy and has proven to result in a good estimation.