Scholarlink

With the continuous development of nuclear technology, prediction and assessment of radionuclide diffusion in the ocean is becoming an important issue, especially after the Fukushima nuclear accident in Japan. Whether the radionuclide transport model is credible or not is the premise for the application of the nuclide transport model. Thus, it is of great significance to establish a liquid diffusion model for nuclear leakage accidents and build a model verification domain to verify its credibility. Based on some mature examples, such as the two-dimensional numerical simulation of the MIKE21 model, this paper calculates its correlation coefficient, Taylor inequality coefficient, angular cosine coefficient and Euclidean distance. Meanwhile, this paper explains the mathematical indicators for evaluating the credibility of machine learning methods. Current research can serve as an investigation on the usage of artificial intelligence methods in marine radionuclides transport models.

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

Verifying the Accuracy of Two-Dimensional Ocean Nuclide Transport Model Using Statistical Parameters and Machine Learning Methods

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Chenrui Fan,
  • Zhengzhe Qu,
  • Yu Wang,
  • Feng Xie,
  • Yongye Liu,
  • Ping Wang,
  • Xinhua Liu

摘要

With the continuous development of nuclear technology, prediction and assessment of radionuclide diffusion in the ocean is becoming an important issue, especially after the Fukushima nuclear accident in Japan. Whether the radionuclide transport model is credible or not is the premise for the application of the nuclide transport model. Thus, it is of great significance to establish a liquid diffusion model for nuclear leakage accidents and build a model verification domain to verify its credibility. Based on some mature examples, such as the two-dimensional numerical simulation of the MIKE21 model, this paper calculates its correlation coefficient, Taylor inequality coefficient, angular cosine coefficient and Euclidean distance. Meanwhile, this paper explains the mathematical indicators for evaluating the credibility of machine learning methods. Current research can serve as an investigation on the usage of artificial intelligence methods in marine radionuclides transport models.