Scholarlink

<p>There are two unique dual perspectives of the study. One of them is to investigate active–passive control mechanisms with waste discharge concentration, and other is to employ physics-informed neural networks (PINNs) to obtain solutions. The key purpose of this attempt is to explore thermal and mass transport of ternary nanofluid with active–passive control mechanisms and external pollutant discharge concentration. This study involves a cone–disk combination as a geometry. It includes a detailed comparison of active–passive control characterized by constant surface concentration and Brownian/thermophoresis effects, respectively. The flow domain is modeled over a rotating cone–disk combination, a geometry relevant in engineering systems like rotating machinery, impinging jet, and bioreactors. A novel PINN framework is employed to solve the ordinary differential equations. To train PINN, the Limited-memory Broyden–Fletcher–Goldfarb–Shanno optimizer is implemented due to its superior convergence speed and reduced memory consumption. The external source pollutant parameter enhances source strength and leads to greater pollutant injections into system that raises concentration profile. The temperature increases, and active control mechanisms yield higher temperature profiles compared to passive control.</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-informed neural network framework using L-BFGS optimizer for active–passive thermal control pollutant dynamics in ternary nanofluid over cone–plate rheometer configuration

  • 期刊论文

  • 发表日期：

    2026年1月14日

查看全文
  • Saleh Chebaane,
  • Assad Ayub,
  • Syed Zahir Hussain Shah,
  • Alaa Dafhalla,
  • Leila Manai,
  • Hira Affan

摘要

There are two unique dual perspectives of the study. One of them is to investigate active–passive control mechanisms with waste discharge concentration, and other is to employ physics-informed neural networks (PINNs) to obtain solutions. The key purpose of this attempt is to explore thermal and mass transport of ternary nanofluid with active–passive control mechanisms and external pollutant discharge concentration. This study involves a cone–disk combination as a geometry. It includes a detailed comparison of active–passive control characterized by constant surface concentration and Brownian/thermophoresis effects, respectively. The flow domain is modeled over a rotating cone–disk combination, a geometry relevant in engineering systems like rotating machinery, impinging jet, and bioreactors. A novel PINN framework is employed to solve the ordinary differential equations. To train PINN, the Limited-memory Broyden–Fletcher–Goldfarb–Shanno optimizer is implemented due to its superior convergence speed and reduced memory consumption. The external source pollutant parameter enhances source strength and leads to greater pollutant injections into system that raises concentration profile. The temperature increases, and active control mechanisms yield higher temperature profiles compared to passive control.