Scholarlink

Precise prediction of residential power consumption, and effective management of load are important tasks in smart grid. The current research proposes a novel hybrid model of ANN with NSGA-II to solve the multi-objective optimization problems for smart grid operations. The model incorporates four important inputs to simultaneously predict forecast demand and load management reliability: time-of-day, temperature, consumer type, and historical load. The ANN model optimized by NSGA-II offers improved forecasting, resulting in the best fitness value of 855.176 kWh, and the resulting high correlation coefficient R = 0.97432 for the load forecasting. Meanwhile, the model also maintained a high level of load management reliability as present an best Fitness 86.7012% and a correlation R = 0.93381. Pareto front analysis demonstrated a trade-off solution between forecast accuracy (855.928–855.934 kWh) and reliability (84.043% to 84.086%) and therefore it is flexible in advising grid operator. This NSGA-II-ANN hybrid approach has wide range of applications for real-time load prediction, and better resource allocation and control for increasing smart grid stability in dynamic operation condition.

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

Multi-Objective Forecasting and Load Management in Smart Grids Using NSGA-II Optimized ANN Model

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Nilesh Pandurang Dabe,
  • Gosavi Kirti Raghuvir,
  • Sunil S. Kadlag,
  • Ashish Dandotia,
  • Mukesh Kumar Gupta,
  • Amit Tiwari

摘要

Precise prediction of residential power consumption, and effective management of load are important tasks in smart grid. The current research proposes a novel hybrid model of ANN with NSGA-II to solve the multi-objective optimization problems for smart grid operations. The model incorporates four important inputs to simultaneously predict forecast demand and load management reliability: time-of-day, temperature, consumer type, and historical load. The ANN model optimized by NSGA-II offers improved forecasting, resulting in the best fitness value of 855.176 kWh, and the resulting high correlation coefficient R = 0.97432 for the load forecasting. Meanwhile, the model also maintained a high level of load management reliability as present an best Fitness 86.7012% and a correlation R = 0.93381. Pareto front analysis demonstrated a trade-off solution between forecast accuracy (855.928–855.934 kWh) and reliability (84.043% to 84.086%) and therefore it is flexible in advising grid operator. This NSGA-II-ANN hybrid approach has wide range of applications for real-time load prediction, and better resource allocation and control for increasing smart grid stability in dynamic operation condition.