Electric Load Forecasting for Internet of Things Smart Home Using Hybrid PCA and ARIMA Algorithm

Abstract

Many types of research have been conducted for the development of Internet of Things (IoT) devices and energy consumption forecasting. In this research, the electric load forecasting is designed with the development of microcontrollers, sensors, and actuators, added with cameras, Liquid Crystal Display (LCD) touch screen, and minicomputers, to improve the IoT smart home system. Using the Python program, Principal Component Analysis (PCA) and Autoregressive Integrated Moving Average (ARIMA) algorithms are integrated into the website interface for electric load forecasting. As provisions for forecasting, a monthly dataset is needed which consists of electric current variables, number of individuals living in the house, room light intensity, weather conditions in terms of temperature, humidity, and wind speed. The main hardware parts are ESP32, ACS712, electromechanical relay, Raspberry Pi, RPi Camera, infrared Light Emitting Diode (LED), Light Dependent Resistor (LDR) sensor, and LCD touch screen. While the main software applications are Arduino Interactive Development Environment (IDE), Visual Studio Code, and Raspberry Pi OS, added with many libraries for Python 3 IDE. The experimental results provided the fact that PCA and ARIMA can predict short-term household electric load accurately. Furthermore, by using Amazon Web Services (AWS) cloud computing server, the IoT smart home system has excellent data package performances.