Scholarlink

<p>This study evaluates the performance of the measure-correlate-predict (MCP) method in relation to measurement periods and wind data recovery rates using different machine learning algorithms and multiple scenarios. Two distinct sampling methods were used for the study: cumulative monthly data sampling from one to twelve months and monthly percentage data sampling between 10 % and 100 %. The met mast wind data over a full year and 16 year-long-term reanalysis data were collected and concurrent data points from the two datasets were sampled using the two methods. Then, four machine learning models were trained and their accuracies evaluated by test data. Finally, the performance of these models was evaluated for determining annual energy production (AEP). Results indicated that the random forest (RF) model had superior performance in predicting wind speeds, which was used for calculating AEP. The AEP error ranges for the two methods were revealed in this paper.</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

Effect of wind measurement period and data recovery rate on performance of measure-correlate-predict method with machine learning

  • 期刊论文

  • 发表日期：

    2025年8月4日

查看全文
  • Buyankhishig Amarzaya,
  • Kyungnam Ko

摘要

This study evaluates the performance of the measure-correlate-predict (MCP) method in relation to measurement periods and wind data recovery rates using different machine learning algorithms and multiple scenarios. Two distinct sampling methods were used for the study: cumulative monthly data sampling from one to twelve months and monthly percentage data sampling between 10 % and 100 %. The met mast wind data over a full year and 16 year-long-term reanalysis data were collected and concurrent data points from the two datasets were sampled using the two methods. Then, four machine learning models were trained and their accuracies evaluated by test data. Finally, the performance of these models was evaluated for determining annual energy production (AEP). Results indicated that the random forest (RF) model had superior performance in predicting wind speeds, which was used for calculating AEP. The AEP error ranges for the two methods were revealed in this paper.