Scholarlink

<p>The application of distributed photovoltaics (PV) systems is experiencing significant growth in response to the current energy crisis. However, dust accumulation on PV panels presents a common challenge that adversely impacts PV energy conversion efficiency. To precisely predict the degree of PV dust accumulation, a predictive model based on PV power loss is proposed, with the impact of natural rainfall considered. The novelty of the proposed model lies in combining direct and indirect PV power prediction for accurate assessments of power loss and dust accumulation levels. Furthermore, a combined prediction method is presented to enhance the precision of both direct and indirect PV output power predictions through self-adaptive variable weights. The superiority of the selected algorithm combination is validated through ablation studies and generalization experiments. Furthermore, an empirical study on soiling effects was conducted at a rooftop PV power station in Hebei Province. The influence mechanisms of dust concentration gradients on the power loss rate of the PV system were analyzed, and the correlation between the dynamic accumulation characteristics of surface contaminants on PV modules and their optimal cleaning cycles was investigated, providing theoretical support for formulating scientific cleaning schedules.</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

Journal of Electrical Engineering & Technology Research on Pv Dust Accumulation Prediction Based on Variable Weight Combination Algorithm

  • 期刊论文

  • 发表日期：

    2025年8月2日

查看全文
  • Xiaoguang Zhu,
  • Zhengtian Wang,
  • Qilin Zhang,
  • Yuling He,
  • Xuewei Wu,
  • Kai Sun,
  • Liqin Song,
  • Xiaodong Du

摘要

The application of distributed photovoltaics (PV) systems is experiencing significant growth in response to the current energy crisis. However, dust accumulation on PV panels presents a common challenge that adversely impacts PV energy conversion efficiency. To precisely predict the degree of PV dust accumulation, a predictive model based on PV power loss is proposed, with the impact of natural rainfall considered. The novelty of the proposed model lies in combining direct and indirect PV power prediction for accurate assessments of power loss and dust accumulation levels. Furthermore, a combined prediction method is presented to enhance the precision of both direct and indirect PV output power predictions through self-adaptive variable weights. The superiority of the selected algorithm combination is validated through ablation studies and generalization experiments. Furthermore, an empirical study on soiling effects was conducted at a rooftop PV power station in Hebei Province. The influence mechanisms of dust concentration gradients on the power loss rate of the PV system were analyzed, and the correlation between the dynamic accumulation characteristics of surface contaminants on PV modules and their optimal cleaning cycles was investigated, providing theoretical support for formulating scientific cleaning schedules.