AI Powered IoT based Home Energy Management System towards DG Integration

Abstract

This paper attempts to present an Internet of Things (IoT) based Home Energy Management System (HEMS) to accommodate Artificial Intelligence (AI) based Distributed Generation (DG) Integration in Smart Micro Grid environment. This work presents the feasibility of load and weather Big Data acquisition, online load forecasting /prediction, online load priority assignment, online peak clipping estimation on the cloud (Thingspeak). Load priority and peak clipping Demand Side Management (DSM) Techniques are used to run Load Management Algorithm (LMA) on the cloud which ultimately executes AI-based DG Integration at the load center. Based on the real time load and weather Big Data Load Priority Table (LPT) is generated on the cloud in order to assign the priority to the loads of the load center and online forecasting is also carried out on the cloud to estimate Total Power PT using Artificial Neural Network (ANN) and five different ML algorithms which are later utilized to estimate Threshold /Peak Clipping Power PTh. Using these LPT and PTh values LMA executes AI-based DG Integration. To that end, a novel IoT-based HEMS is designed and fabricated to run cloud-based LMA in order to accommodate proposed DG Integration. Amongst all the prediction algorithms trained on the cloud to estimate PTh, Linear Regression (LR) ML algorithm is witnessed to have lowest RMSE of 4.44e−13, MSE of 1.97e−25 and MAE of 2.11e−13. MATLAB 2018b licensed version is used to run ANN and ML algorithms based forecasting/prediction on the cloud.