Scholarlink

<p>This study is the first to develop and benchmark two distinct physics-informed neural networks (PINNs) targeting Nusselt number (Nu) and performance evaluation criterion (PEC) to evaluate twelve real-world shell&#xa0;and&#xa0;helical coil heat exchanger configurations and assess a variety of predictive models. These models include linear regressors (Ridge, Lasso), ensemble learners (AdaBoost), support vector regression, and deep learning models (LSTM, Bi-LSTM, and GRU). As the main novelty, two dedicated PINNs are developed: one for predicting Nu and another for PEC. Both models leverage physical domain knowledge by embedding analytical heat transfer constraints, such as empirical Nu correlations and laminar friction factor formulations directly into the loss function, guiding learning beyond conventional data fitting. Input parameters such as Reynolds number, Dean number, and curvature ratio serve as predictors for both target variables. To improve model reliability, noise-robust data augmentation and grid-based hyperparameter tuning are applied. The results show the proposed PINN delivers unbiased Nu predictions across the full range with markedly lower error than Dittus–Boelter and Gnielinski (RMSE 1.12 vs 4.75 and 3.20; <i>R</i><sup>2</sup> 0.998 vs 0.935–0.962). It also achieves the best PEC accuracy (RMSE 0.044; <i>R</i><sup>2</sup> 0.997) over raw and calibrated empirical correlations (RMSE 1.32 and 0.39). Taylor diagrams show the highest correlation and near-unity normalized variance (<i>r</i> ≈ 0.998–0.999; <i>σ</i>/<i>σ</i><sub>obs</sub>≈1.02–1.03), and MixUp tests confirm robustness (ΔRMSE &lt; 2%) exceeding all other tuned ML/RNN models. These features give the model a reliable capability to predict Nu and PEC with high accuracy and strong generalization for heat exchanger analysis and thermal system design.</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

Physics-Guided Machine Learning and Deep Neural Networks for Accurate and Robust Heat Transfer Predictions in Shell and Helical Coil Heat Exchangers

  • 期刊论文

  • 发表日期：

    2025年9月21日

查看全文
  • Ali Hussein Abdulkarim,
  • Orhan Nooruldeen,
  • Ahmed Ghareeb,
  • Doğan Akgül,
  • Hatice Mercan,
  • Ahmet Selim Dalkilic

摘要

This study is the first to develop and benchmark two distinct physics-informed neural networks (PINNs) targeting Nusselt number (Nu) and performance evaluation criterion (PEC) to evaluate twelve real-world shell and helical coil heat exchanger configurations and assess a variety of predictive models. These models include linear regressors (Ridge, Lasso), ensemble learners (AdaBoost), support vector regression, and deep learning models (LSTM, Bi-LSTM, and GRU). As the main novelty, two dedicated PINNs are developed: one for predicting Nu and another for PEC. Both models leverage physical domain knowledge by embedding analytical heat transfer constraints, such as empirical Nu correlations and laminar friction factor formulations directly into the loss function, guiding learning beyond conventional data fitting. Input parameters such as Reynolds number, Dean number, and curvature ratio serve as predictors for both target variables. To improve model reliability, noise-robust data augmentation and grid-based hyperparameter tuning are applied. The results show the proposed PINN delivers unbiased Nu predictions across the full range with markedly lower error than Dittus–Boelter and Gnielinski (RMSE 1.12 vs 4.75 and 3.20; R2 0.998 vs 0.935–0.962). It also achieves the best PEC accuracy (RMSE 0.044; R2 0.997) over raw and calibrated empirical correlations (RMSE 1.32 and 0.39). Taylor diagrams show the highest correlation and near-unity normalized variance (r ≈ 0.998–0.999; σ/σobs≈1.02–1.03), and MixUp tests confirm robustness (ΔRMSE < 2%) exceeding all other tuned ML/RNN models. These features give the model a reliable capability to predict Nu and PEC with high accuracy and strong generalization for heat exchanger analysis and thermal system design.