Smart grid energy management system for Industrial applications

Abstract

Over the years, energy needs have become complicated due to rapid industrialization, population growth and enforcement of stringent measures to reduce the carbon footprint globally. Industries being amongst the largest consumers of electricity generated worldwide, can realize huge savings in energy cost by implementing energy management programs. This research incorporates the aspects of a smart grid in designing an energy management system (EMS) where demand side management (DSM) is utilized to enable industrial users minimize their energy costs. A forecasting model for electricity prices and demand is developed using Long Short Term Memory (LSTM) - Recurrent Neural Network (RNN) technique. The predicted prices are used in load scheduling to realize potential energy cost savings. The non-priority loads are scheduled to leverage on low electricity prices during off peak times. The effectiveness of the designed energy management strategy is tested using an IEEE 30 bus system. A suitable operation schedule with committed units for each hour is given for one sample day. Using the test system with 20 loads yielded an annual energy cost saving of $2,961,169.20, a payback period (PBP) of 4 years 5 months, and a return on investment (ROI) of 22.78%. The ROI can be improved by considering savings from future years of having the EMS in place. Overall, industries can maintain a competitive edge by investing in an effective EMS that will enable them minimize their electrical energy costs which forms one of the top operating expenses.

Keywords: Smart Grid, Energy Management System, Long Short Term Memory, Recurrent Neural Network, Demand Side Management, Demand Response, Time of Use