Scholarlink

It is crucial to strive towards adopting more eco-friendly methods in response to technological progress. Our primary objective is to acquire further knowledge regarding the process by which solar panels produce electricity and to motivate future generations to utilize non-renewable resources with utmost frugality. Predicting power generation for solar PV panels is challenging due to the influence of various changing elements such as fog, mist, pollution, sudden cloud cover, and the presence of flies. These factors, in addition to irradiance, contribute to the complexity of the prediction process. Therefore, performing an analytical calculation of irradiance for a specific location (geographical position) and day number does not yield the appropriate power outputs from solar panels. Therefore, machine learning techniques are employed to address this nonlinear parametric problem, while considering different limitations, such as varying test-train ratios and data quantities. In this work, we have utilized an analysis that relies on the outcomes of five machine learning estimators, namely random forest, decision tree, support vector machine, naive Bayes, and logistic regression.

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

A Real-Time Machine Learning Based Statistical Approach for Power Generation Prediction for Solar PV Plant

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sovan Bhattacharya,
  • Dola Sinha,
  • Ayusman Maji,
  • Rajdeep Saha,
  • Pronoy Das Choudhury,
  • Saibal Majumder,
  • Chandan Bandyopadhyay

摘要

It is crucial to strive towards adopting more eco-friendly methods in response to technological progress. Our primary objective is to acquire further knowledge regarding the process by which solar panels produce electricity and to motivate future generations to utilize non-renewable resources with utmost frugality. Predicting power generation for solar PV panels is challenging due to the influence of various changing elements such as fog, mist, pollution, sudden cloud cover, and the presence of flies. These factors, in addition to irradiance, contribute to the complexity of the prediction process. Therefore, performing an analytical calculation of irradiance for a specific location (geographical position) and day number does not yield the appropriate power outputs from solar panels. Therefore, machine learning techniques are employed to address this nonlinear parametric problem, while considering different limitations, such as varying test-train ratios and data quantities. In this work, we have utilized an analysis that relies on the outcomes of five machine learning estimators, namely random forest, decision tree, support vector machine, naive Bayes, and logistic regression.