Scholarlink

<p>Current study presents a novel application of an intelligent network for a second-order nonlinear ordinary differential equation that portrays the electrohydrodynamic (EHD) fluidic flow model in a cylindrical conduit with an ion drag configuration. An intelligent network-based numerical solver is designed via a neural network optimized with Levenberg-Marquardt back propagation (NNLMBP). The NNLMBP network is applied to achieve a precise solution to the EHD fluid flow-based boundary value problem. A dataset of the EHD model is formulated by exploiting the strength of the Runge Kutta method for implementing NNLMBP with evaluation of different values for nonlinearity constant and electric Hartmann’s number to analyze the radial flow velocity numerically. The approximate solutions of the NNLMBP networks-based solver for various scenarios and cases of the EHD ion drag flow model are through training, testing, and validation procedures and compared with reference solutions for the validation and correctness of the NNLMBP approach. The worth and value of the designed NNLMBP are recognized by regression analysis with <i>R</i> ~ 1, error-histogram plots with zero line error bin have central value around E-06 to E-07, and convergence curves having negligible mean square error in the range E-12 to E-10, for the exhaustive numerical experimentations.</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

Design of intelligent Levenberg-Marquardt back-propagated networks for the analysis of radial velocity of electrohydrodynamic ion drag flow model

  • 期刊论文

  • 发表日期：

    2025年10月15日

查看全文
  • Saeed Ehsan Awan,
  • Muhammad Kashif,
  • Muhammad Asif Zahoor Raja,
  • Ihtesham Jadoon,
  • Shujaat Ali Khan Tanoli,
  • Assad Hafiz

摘要

Current study presents a novel application of an intelligent network for a second-order nonlinear ordinary differential equation that portrays the electrohydrodynamic (EHD) fluidic flow model in a cylindrical conduit with an ion drag configuration. An intelligent network-based numerical solver is designed via a neural network optimized with Levenberg-Marquardt back propagation (NNLMBP). The NNLMBP network is applied to achieve a precise solution to the EHD fluid flow-based boundary value problem. A dataset of the EHD model is formulated by exploiting the strength of the Runge Kutta method for implementing NNLMBP with evaluation of different values for nonlinearity constant and electric Hartmann’s number to analyze the radial flow velocity numerically. The approximate solutions of the NNLMBP networks-based solver for various scenarios and cases of the EHD ion drag flow model are through training, testing, and validation procedures and compared with reference solutions for the validation and correctness of the NNLMBP approach. The worth and value of the designed NNLMBP are recognized by regression analysis with R ~ 1, error-histogram plots with zero line error bin have central value around E-06 to E-07, and convergence curves having negligible mean square error in the range E-12 to E-10, for the exhaustive numerical experimentations.