Design and Implementation of a Factory Pollutant Monitoring System using Plasma by Self-Encoding Neural Network

Document Type : Original Article

Authors

Department of Software Engineering, Khorramabad branch, Islamic Azad University, Khorramabad, Iran

Abstract

A factory pollutant monitoring system using the Internet of Things (IoT), machine learning, and plasma technology has been designed and implemented. The goal is to effectively reduce pollution and absorb hazardous particles. So, an ESP32 electronic board has been used along with temperature and humidity (DHT22), carbon dioxide (MG812-CO2), and sulfur dioxide (MQ136-SO2) sensors to collect data and transmit it to the server via the IoT. The data is stored in Google Sheets and prepared for modeling after preprocessing. Then, an Autoencoder network has been used, due to its high ability to reduce dimensions and eliminate noise, reconstructs the data with minimal error. The Autoencoder extracts key features of the data to optimize the plasma level and pollution reduction. The results show that this system performs better in reducing pollution and energy consumption than traditional methods such as electro-filters. Also, high efficiency and cost reduction are advantages of this approach.

Keywords


  • Received Date: 10 September 2024
  • Received Date: 03 December 2024
  • Accepted Date: 28 December 2024
  • Published Date: 18 March 2025