Scholarlink

<p>Indoor Air Quality (IAQ) plays a crucial role in human health, with people spending over 90% of their time indoors. This study presents a compact, smart air purifier that integrates advanced filtration, real-time monitoring, and machine learning (ML)-based predictive control. The purifier uses a dual-stage filter system—an Intense Field Dielectric (IFD) filter for fine particulate matter (PM) removal and a zeolite-based filter for volatile organic compound (VOC) adsorption—optimized through Computational Fluid Dynamics (CFD) simulations. An Arduino-based controller, paired with high-accuracy AQI sensors, enables automatic operation based on pollution levels, achieving up to 30% energy savings. Experimental results demonstrated a 76% reduction in PM 2.5 and 65% in VOCs within 30&#xa0;min. Eleven ML regression models were evaluated for AQI prediction; Lasso Regression showed the best performance with an R2 of 0.9179, RMSE of 5.14, and MAE of 4.13. Feature importance analysis identified PM2.5, PM10, noise, and humidity as key predictors. The novelty of this work lies in the integration of IFD–zeolite filtration with ML-based AQI prediction, enabling anticipatory and energy-efficient air purification. This low-cost, IoT-compatible solution is scalable for smart homes, offices, and sustainable building environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-powered AQI forecasting in an intelligent air purifier featuring intense field dielectric and Zeolite filters

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月10日

查看全文
  • Pankaj Beldar,
  • Prashant Kushare,
  • Atulkumar Patil,
  • Gulshan Kumar

摘要

Indoor Air Quality (IAQ) plays a crucial role in human health, with people spending over 90% of their time indoors. This study presents a compact, smart air purifier that integrates advanced filtration, real-time monitoring, and machine learning (ML)-based predictive control. The purifier uses a dual-stage filter system—an Intense Field Dielectric (IFD) filter for fine particulate matter (PM) removal and a zeolite-based filter for volatile organic compound (VOC) adsorption—optimized through Computational Fluid Dynamics (CFD) simulations. An Arduino-based controller, paired with high-accuracy AQI sensors, enables automatic operation based on pollution levels, achieving up to 30% energy savings. Experimental results demonstrated a 76% reduction in PM 2.5 and 65% in VOCs within 30 min. Eleven ML regression models were evaluated for AQI prediction; Lasso Regression showed the best performance with an R2 of 0.9179, RMSE of 5.14, and MAE of 4.13. Feature importance analysis identified PM2.5, PM10, noise, and humidity as key predictors. The novelty of this work lies in the integration of IFD–zeolite filtration with ML-based AQI prediction, enabling anticipatory and energy-efficient air purification. This low-cost, IoT-compatible solution is scalable for smart homes, offices, and sustainable building environments.