Scholarlink

<p>Modern power systems are transitioning toward smart grids (SGs) to overcome the limitations of conventional grids, including inefficient energy distribution, high operational costs, and minimal consumer involvement. Effective demand-side management (DSM) is essential to optimize energy usage, reduce peak loads, and minimize emissions, particularly in IoT-enabled SG environments. This study proposes a hybrid artificial intelligence-based technique combining Binary Waterwheel Plant Optimization Algorithm (BWPOA) with Temporal Inductive Path Neural Network (TIPNN), referred to as BWPOA-TIPNN. The BWPOA optimizes the operational scheduling of renewable energy sources, while TIPNN predicts load demand with high temporal accuracy, enabling efficient energy forecasting and proactive DSM. The proposed technique was put into practice on the MATLAB platform and evaluated under dynamic pricing schemes such as Real-Time Pricing (RTP) and Critical Peak Pricing (CPP). Results show that BWPOA-TIPNN outperforms existing approaches (HOGKAN, GWO, AHHO, and PSO) by achieving the lowest Peak-to-Average Ratio (PAR) of 43.50%, energy cost of $2.4, carbon emissions of 0.241&#xa0;kg CO₂/kWh, and computational energy consumption of only 0.01&#xa0;J. Under RTP and CPP, it delivered the lowest operational costs of $1.876 and $2.545, respectively. The model also demonstrated superior predictive accuracy with RMSE of 0.504 and MAE of 0.421. The BWPOA-TIPNN model significantly enhances DSM performance in smart grids, balancing cost, emissions, and load variability. The findings support its deployment for sustainable, intelligent energy management and inform policy directions for AI-driven grid modernization.</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

Efficient demand side management in smart grids using internet of things -enabled energy management systems

  • 期刊论文

  • 发表日期：

    2025年9月5日

查看全文
  • S. Bhala Priya,
  • S. Jaganathan,
  • Sivaprakash P.

摘要

Modern power systems are transitioning toward smart grids (SGs) to overcome the limitations of conventional grids, including inefficient energy distribution, high operational costs, and minimal consumer involvement. Effective demand-side management (DSM) is essential to optimize energy usage, reduce peak loads, and minimize emissions, particularly in IoT-enabled SG environments. This study proposes a hybrid artificial intelligence-based technique combining Binary Waterwheel Plant Optimization Algorithm (BWPOA) with Temporal Inductive Path Neural Network (TIPNN), referred to as BWPOA-TIPNN. The BWPOA optimizes the operational scheduling of renewable energy sources, while TIPNN predicts load demand with high temporal accuracy, enabling efficient energy forecasting and proactive DSM. The proposed technique was put into practice on the MATLAB platform and evaluated under dynamic pricing schemes such as Real-Time Pricing (RTP) and Critical Peak Pricing (CPP). Results show that BWPOA-TIPNN outperforms existing approaches (HOGKAN, GWO, AHHO, and PSO) by achieving the lowest Peak-to-Average Ratio (PAR) of 43.50%, energy cost of $2.4, carbon emissions of 0.241 kg CO₂/kWh, and computational energy consumption of only 0.01 J. Under RTP and CPP, it delivered the lowest operational costs of $1.876 and $2.545, respectively. The model also demonstrated superior predictive accuracy with RMSE of 0.504 and MAE of 0.421. The BWPOA-TIPNN model significantly enhances DSM performance in smart grids, balancing cost, emissions, and load variability. The findings support its deployment for sustainable, intelligent energy management and inform policy directions for AI-driven grid modernization.