SMART GRID ENERGY MANAGEMENT BASED ON IOT AND COMPUTATIONAL INTELLIGENCE

Abstract

Booming demand, depleting natural resources, deregulation, generation capacities, and grids are some of the factors that are a concern for energy efficiency. In Smart grid era, Demand Side Management (DSM) plays an indispensable role in development of sustainable cities and societies. Based on the current practice of utility system, the load shape objectives can be char acterized into six categories: peak clipping, valley filling, load shifting, strategic conservation, strategic load growth, and flexible load shape. Out of these six classical approaches: “peak clipping” and “load shifting” are most widely applicable and relevant for energy efficiency us ing DSM [121]. Home Energy Management System (HEMS) lies under the umbrella of DSM, it allows residential consumers to supervise and manage the power usage of their appliances to reduce their electricity bills. Since, residential consumers are more concerned about their energy bills as well as comfort, energy optimization with multiple objectives grips valuable and resourceful usage of electricity. Therefore, the usage awareness and scheduling optimization alone have the potential to reduce consumption by 15% in private households. The stochastic problem of scheduling optimization in HEMS involves arbitrary dynamics of renewable energy, consumer demand, consumer behavior, and electricity price. For con sumers, energy storage and load scheduling can provide effective means in reducing their elec tricity costs. Further, the limited battery capacity, the finite optimization time period and in teraction with energy storage devices complicates the energy scheduling and control decisions. DSM in HEMS using load shifting technique is a challenging optimization research problem, where the main aim is to have optimal utilization of available energy resources while reducing the electricity bills. Reducing the complexity of scheduling optimization helps in the judicial use of power consumption by the effective control of smart home appliances. The research work focuses around the development of scheduling algorithm for the con sumer which aims to minimize the electricity price without compromising the comfort time period of using various home appliances. First, a comprehensive survey regarding the major factors affecting the optimized management solution and consequent decision making in HEMS i has been performed. Second, a robust deep learning algorithm for solar irradiance forecasting has been developed. It can forecast GHI value using various weather parameters in different seasons as well as at different steps (1-step, 2-step, 12 steps ahead). Third, deep learning en semble model has been utilized to forecast appliances’ power utilization. Using dynamic Item set counting (DIC) algorithm, association between multiple appliances has been determined. Fourth, a real time scheduling algorithm has been proposed which takes into account the fore casted PV power, appliance power, battery constraints, various other constraints and forecast the 24- hour schedule of the appliances which minimizes the electricity price without compromis ing the comfort. Fifth, smart grid reliability problem has been assessed by modelling it using graph computational model and assessing the reliability through various indices .