Scholarlink

The aim the current study is to estimate the power generation (PG) of off-grid solar system utilizing Autoregressive Integrated Moving Average (ARIMA), k-nearest neighbor algorithm (kNN), Elman neural network (ENN), and Extreme Learning Machine (ELM). To achieve this aim, the hourly weather data were considered as input variables for the developed models. According to the findings, the most reliable model for predicting electricity generation was ARIMA R-squared = 0.994, RMSE = 100.331 W and MAE = 64.533 W). The study highlights the significance of selecting the appropriate predictive model to improve the reliability and efficiency of solar power generation, which will help provide more sustainable and reasonably priced energy sources.

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

Machine Learning-Based Prediction of Hourly Solar PV Power Generation: A Case Study in Al Mahmra, Lebanon

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Youssef Kassem,
  • Hüseyin Gökçekuş,
  • Abdirahman Mohamed Ali

摘要

The aim the current study is to estimate the power generation (PG) of off-grid solar system utilizing Autoregressive Integrated Moving Average (ARIMA), k-nearest neighbor algorithm (kNN), Elman neural network (ENN), and Extreme Learning Machine (ELM). To achieve this aim, the hourly weather data were considered as input variables for the developed models. According to the findings, the most reliable model for predicting electricity generation was ARIMA R-squared = 0.994, RMSE = 100.331 W and MAE = 64.533 W). The study highlights the significance of selecting the appropriate predictive model to improve the reliability and efficiency of solar power generation, which will help provide more sustainable and reasonably priced energy sources.