Marquardt Algorithm Research Articles

Comparative fitting methodology of cure kinetics models based on differential scanning calorimetry

The optimization of composite cure cycles is critical in materials engineering, requiring a deep understanding of the material’s cure kinetics. This study employs Differential Scanning Calorimetry (DSC) coupled with advanced fitting algorithms to determine the most accurate cure kinetics models for optimizing time-temperature profiles. Given the wide variability in model selection and fitting approaches in existing literature, our research systematically compares several cure kinetics models and their fitting methodologies to identify the most effective strategies. We highlight the advantages and limitations of each approach, with a particular focus on the robustness of the Levenberg–Marquardt (LM) algorithm compared to the Trust-Region-Reflective (TRR) algorithm. Given the existence of multiple local minima in the objective function, a global approach (genetic algorithm) is employed. Our findings reveal that, although the TRR and LM algorithms are faster, they are less accurate than the genetic algorithm. Moreover, the Kamal-Sourour model shows the highest prediction accuracy among the tested models. This comparative analysis provides a systematic approach and crucial insights into selecting the optimal cure kinetic model and corresponding parameters, based on experimental data, allowing for enhancing the efficiency and accuracy of composite curing processes.

Real-Time Nonlinear Image Reconstruction in Electrical Capacitance Tomography Using the Generative Adversarial Network

This study investigated the potential of the generative adversarial neural network (cGAN) image reconstruction in industrial electrical capacitance tomography. The image reconstruction quality was examined using image patterns typical for a two-phase flow. The training dataset was prepared by generating images of random test objects and simulating the corresponding capacitance measurements. Numerical simulations were performed using the ECTsim toolkit for MATLAB. A cylindrical sixteen-electrode ECT sensor was used in the experiments. Real measurements were obtained using the EVT4 data acquisition system. The reconstructed images were evaluated using selected image quality metrics. The results obtained using cGAN are better than those obtained using the Landweber iteration and simplified Levenberg–Marquardt algorithm. The suggested method offers a promising solution for a fast reconstruction algorithm suitable for real-time monitoring and the control of a two-phase flow using ECT.

An artificial neural network approach to characterizing the behavior of bioconvective nanofluid model using backpropagation of Levenberg–Marquardt algorithm

The study of the behavior of bioconvective nanofluid model (BC-NFM) was explored using bioconvection properties in the computational analysis of magnetized nanofluid flow that was convectively heated, which is critical for applications in energy systems and biomedical devices. We implemented a backpropagated Levenberg–Marquardt neural network approach (BLMNNA) to enhance the analysis and prediction accuracy of such fluid flows. Using the Adams numerical technique, we generated a comprehensive dataset for eight distinct scenarios by variation of thermophoresis parameter, Rayleigh number, Deborah parameter, Hartman number, Prandtl number, bioconvection Lewis number, and Schmidt number to train and test our model. The results demonstrate significant improvements in prediction accuracy and computational efficiency compared to traditional numerical solvers. The steps of training, testing, and validating the developed BLMNNA are used to get the desired numerical solutions of BC-NFM for various instances. The worth and accuracy of the stochastic numerical solutions of BC-NFM are established and authenticated by the outputs of the designed methodology BLMNNA through Adaptation graphs of Mean Square Error, regression studies, plots of error histogram and index of state transition. Excellent measurements of performance in terms of MEAN SQUARE ERROR are achieved at level 9.76E[Formula: see text], 1.41E[Formula: see text], 1.79E[Formula: see text], 1.22E[Formula: see text], 8.49E[Formula: see text], 7.19E[Formula: see text], 9.72E[Formula: see text], and 9.35E[Formula: see text] against 69, 73, 96, 76, 237, 86, 120, and 33 epochs. The significant connection between the proposed and reference findings demonstrates the validity of the BLMNNA based on error analysis, which ranges from E[Formula: see text] to E[Formula: see text] for all situations. The results show that the proposed neural network approach achieves a high level of accuracy, closely matching the outcomes obtained from the numerical solver. The method provides an efficient and precise solution for the analysis of complex fluid flow problems, representing a significant advancement over traditional numerical techniques.

Fast Two-dimensional Positioning Method of Crab Pulsar Based on Multiple Optimization Algorithms

In the two-dimensional positioning method of pulsars, the grid method is used to provide non-sensitive direction and positional estimates. However, the grid method has a high computational load and low accuracy due to the interval of the grid. To improve estimation accuracy and reduce the computational load, we propose a fast two-dimensional positioning method for the crab pulsar based on multiple optimization algorithms (FTPCO). The FTPCO uses the Levenberg–Marquardt (LM) algorithm, three-point orientation (TPO) method, particle swarm optimization (PSO) and Newton–Raphson-based optimizer (NRBO) to substitute the grid method. First, to avoid the influence of the non-sensitive direction on positioning, we take an orbital error and the distortion of the pulsar profile as optimization objectives and combine the grid method with the LM algorithm or PSO to search for the non-sensitive direction. Then, on the sensitive plane perpendicular to the non-sensitive direction, the TPO method is proposed to fast search the sensitive direction and sub-sensitive direction. Finally, the NRBO is employed on the sensitive and sub-sensitive directions to achieve two-dimensional positioning of the Crab pulsar. The simulation results show that the computational load of the FTPCO is reduced by 89.4% and the positioning accuracy of the FTPCO is improved by approximately 38% compared with the grid method. The FTPCO has the advantage of high real-time accuracy and does not fall into the local optimum.

Numerical Simulation and Neural Network Model for Hydromagnetic Nanofluid Convection in a Porous Wavy Channel with Thermal Non-Equilibrium Model

Abstract The present study aims to analyze the nonfluid MHD convective heat transfer in a porous wavy channel with a local thermal non-equilibrium (LTNE) model. Such a model finds applications related to enhancement in thermal performance, increasing the heat transfer coefficient in the compact design of heat exchangers for the aerospace and automotive industries and elevation in the efficiency of the solar collector. A sinusoidal porous wavy LTNE channel containing nanofluid and subjected to the induced and applied magnetic fields is considered. A uniform magnetic field is applied orthogonal to the channel and the induced magnetic field effects are considered due to the large magnetic Reynolds number. The momentum, continuity, energy, and nanoparticle volume fraction equations constitute the coupled nonlinear system of differential equations and are solved using the Galerkin finite element method. The reliability of the technique is assessed by comparing the proposed procedure with the results from earlier sources. A detailed analysis is presented to determine the effects of different physical parameters arising in the system on temperature, nanoparticle concentration, and velocity profiles. As an illustration, the findings exhibit that increasing the modified diffusivity ratio increases the values of the nanoparticle volume fraction whereas, reducing the modified diffusivity ratio enhances the temperature distribution. A higher value of thermal Rayleigh number presents a significant involvement of buoyancy forces, potentially resulting in the development of convective currents. A higher Nield number indicates more effective heat transport from the solid surface to the nanofluid. Consequently, there is a minimal thermal difference between the solid surface and the bulk nanofluid. Effective heat transmission enhances the nanofluid ability to absorb heat and generates a more consistent dispersion of temperature inside the fluid. The performance of the designed algorithms of the artificial neural network, namely, the Levenberg – Marquardt algorithm, in the problem under consideration is evaluated and the methodology is found reasonably precise with the matching of order around 6 to 7 decimal places of accuracy.

Machine learning approach for predicting the machining characteristics of additive manufactured ABS material

An Additive manufacturing process offers advantage for the production of intricate designs with increased efficiency and reduced material waste. In this research, an experimental and analysis work has been conducted to identify the significant 3-D printed input parameters effecting the output response characteristics as kerf width. The correlation between experimental and predicted values has been established through the application of Artificial Neural Network (ANN), Support Vector Machine Regression (SVMR), and Response Surface Methodology (RSM) techniques. The validation of predicted models conducted with the help of mean prediction error, maximum error and minimum error. The performance of ANN model was analyzed by feedforward-propagation method using Levenberg- Marquardt algorithm. The maximum prediction error investigated by ANN, SVM, and RSM were noted as 4.28, 5.18, and 5.84 respectively. The optical micrographs have been investigated to understand the hole and crater formation phenomenon.

Soft sensor modeling based on self-organizing fuzzy neural network with clustering, merging, and splitting scheme

Abstract As a pivotal role in the control, optimization, and monitoring of contemporary industrial processes, soft sensors are frequently employed in the prediction of key quality variables. To achieve accurate prediction of key quality variables in industrial processes, a soft sensor modeling method based on the self-organizing fuzzy neural network with the clustering, merging, and splitting scheme (SOFNN-CMS) is proposed. First, the supervised fuzzy C-means clustering algorithm is proposed to identify the appropriate initial center and width of the fuzzy neural network, obtaining appropriate initial fuzzy rules. Then, a neuron merging and splitting strategy is designed to adjust the structure of the fuzzy neural network, by merging and splitting the hidden neurons according to the distance of clusters, increasing the adaptability of the fuzzy neural network. Besides, to accelerate the convergence of estimation errors, an improved Levenberg Marquardt algorithm is utilized to update neural network parameters in the training phase, realizing the soft sensor modeling of key quality variables. The effectiveness of the proposed SOFNN-CMS neural network is demonstrated on two benchmark problems and an industrial debutanizer column. Finally, the experiments showcase that the proposed SOFNN-CMS neural network can obtain better soft sensor modeling performance with a compact structure.

Predicting heat transfer Performance in transient flow of CNT nanomaterials with thermal radiation past a heated spinning sphere using an artificial neural network: A machine learning approach

An efficient heat transfer phenomenon using nanofluid have greater challenges in various industries, engineering application the recent trend. Keeping this in present scenario, this study aims to optimize the heat transmission rate in the magnetized flow of nanomaterials through a rotating, spinning sphere. The heat transfer phenomena in the time-dependent fluid are enhanced by the incorporation of nonlinear radiation and a variable heat source. Additionally, the free convective flow is influenced by the effects of thermal buoyancy and a transverse magnetic field. The proposed model along with several factors is standardized through adequate transformation rules. Further, shooting-based Runge-Kutta technique is adopted with the help of built-in MATLAB function bvp4c for the solution of the transformed system. The prime focus of the proposed work is the optimizing heat transfer rate combined with regression analysis using artificial neural network and then it uses Levenberg Marquardt algorithm with well-posed training, testing, and validation data. The error analysis also presented briefly and the variation of characterizing parameters is depicted via graphs. Further, the important outcomes are; the particle concentration of carbon nanotubes contributes to decelerating the velocity profiles, leading to an increase in boundary layer thickness. In contrast, increasing magnetization has the opposite effect. Both nonlinear radiative heat and an additional heat source enhance the heat transfer phenomenon.

Globalization of convergence of the constrained piecewise Levenberg–Marquardt method

We develop linesearch algorithms intended for globalization of convergence of the piecewise Levenberg–Marquardt method for constrained piecewise smooth equation. Conditions ensuring global convergence properties and asymptotic superlinear convergence rate are proposed. The peculiarities of the global convergence results in the piecewise smooth case are discussed and illustrated by examples. We also provide numerical results for unconstrained and constrained reformulations of nonlinear complementarity problems, comparing the performance of the globalized piecewise Levenberg–Marquardt algorithm with some relevant alternatives.

Integrated neural network based simulation of thermo solutal radiative double-diffusive convection of ternary hybrid nanofluid flow in an inclined annulus with thermophoretic particle deposition

Numerical simulation of magnetohydrodynamic radiative double-diffusive convective heat and mass transfer of a ternary hybrid nanofluid with thermophoretic particle deposition in an inclined annulus is studied. Both sides of cylinders are preserved at uniform temperatures, while the remaining sides are thermally insulated. The coupled nonlinear partial differential equations are solved using a finite difference approach. The detailed computational results of heat and mass transfer rate, temperature, concentration and flow fields are presented and discussed for a distinct range of significant physical parameters. The results demonstrated that increasing the angle of the inclination factor increases fluid flow. In light of buoyancy, when the annular cavity is set downward, the enhanced shear velocity is transported upwards, where it reaches its optimal temperature. The higher temperature difference leads to augmented dynamic fluid motion, particularly in the radial direction, as the system seeks to balance the thermal energy through convective transfer. The Brownian motion parameter has abrupt mobility of nanoparticles in liquid, which promotes particle collisions with liquid molecules, yielding kinetic energy. Increased values of thermophoretic parameters augment the thermophoresis force. Thermophoretic particle deposition increases the concentration of nanoparticles in an annulus due to the migration of particles from hot to cold regions under the influence of temperature gradient. The heat and mass transfer characteristics of ternary hybrid nanofluid are forecasted through an artificial neural network-based Levenberg–Marquardt backpropagated algorithm. The built model shows the heat and mass transfer rate root mean squared values through the Levenberg–Marquardt algorithm as one and mean squared error values as 1e-07 and 1e-08 respectively.

A qualitative analysis of the artificial neural network model and numerical solution for the nanofluid flow through an exponentially stretched surface

This article aims to analyze the two-dimensional (2D) nanofluid (Ag/C2H6O2) flow past an exponentially stretched sheet. The magnetic field impact, heat source/sink, and convection in the thermal profile are taken into account. The complexity of the problem is reduced by introducing a dimensionless group of functions. The reduced model is transformed into a system of first-order ordinary differential equations (ODEs). This system is further analyzed with the artificial neural network (ANN), which is trained using the Levenberg–Marquardt algorithm. The whole dataset is sub divided into three parts: training (70%), validation (15%), and testing (15%). The impact of nonlinear heat source/sink parameter, magnetic parameter, volume fraction of nanoparticles, and Prandtl number is displayed through graphs. The heat source, volume fraction, and the Prandtl number cause an increase in the thermal profile with its larger values. The magnetic parameter causes a decline in both the thermal and momentum boundary layers with its higher values. The analysis shows that the thermal energy profile is enhanced with the larger values of the volume fraction of silver nanoparticles and heat source. For each case study, the residual error (RE), regression line, and validation of the results are presented. The performance of the proposed methodology is numerically tabulated for the nanoparticle volume fraction shown in Table 3, where the minimum absolute error (AE) is 5.3373e−11 at ϕ=0.05. Based on this, we recommend ϕ=0.05 for better performance. The AEs for the ANN and bvp4c are computed for the state variables in Tables for the magnetic parameter M=5,10, and 15. These tables show the overall performance of the ANN and further validate the present study. We have also validated the results of the ANN through the mean squared error graphically, where the accuracy of the proposed methodology is proven.

Machine learning model to predict the cooling load of mosque buildings during the design stage

The building sector accounts for almost 40% of the total global energy consumption. Mosques are among these buildings which found to consume huge amount of energy. There is an increased demand to establish new sustainable mosques. Although the design stage proves the superiority to reduce energy consumption up to 70%, majority of previous studies were found at operation and maintenance stages. This is due to the complexity of this stage, lack of information and support tools to assist designers. The current design tools and prediction models to estimate cooling load are complex, time-consuming and fail to meet the requirement of designers especially at the design stage. This paper aims to develop a machine learning prediction model to assist mosque designers in providing a range of lowest cooling load design alternatives. In developing the model, artificial neural network was applied including back-propagation strategy and Levenberg–Marquardt algorithm. The least mean square error was at 6.27 × 10 − 9 and the accuracy of the model reached at 99.88%. Further, nine experts were participated to estimate the effectiveness and ease of use of the prediction model. The model proved to be rapid, accurate and can be used by designers with no many simulation scenarios or complex calculations.

Design of two-loop FOPID-FOPI controller for inverted cart-pendulum system

The inverted cart-pendulum system (ICPS) consists in having a pendulum mounted on a sliding cart, with the pivot point fixed. This real time experiment indeed looks like a rocket and its functionality is akin to a rocket. These are the launchers and the missile guidance and control as well as construction anti-seismic measures also. The control aim in these systems is to maintain the inverted pendulum vertically stable. The system is causal but unstable and, therefore, has no minimum phase. Therefore, the right half plane pole and zero are close to each other. Therefore, the stability of the system can be considered as problematic at some points. Unfortunately, linear time- invariant (LTI) classical controllers are incapable of offering suffient loop robustness for such systems. This paper aims to project a two-loop fractional order controller (2-LFOC) design to stabilize a higher-order nonlinear inverted cart-pendulum system (ICPS). The modeling, linearization, and control of ICPS are demonstrated in this work. The control target is adjusted so that the inverted pendulum stabilizes in its upright state when the cart reaches the required point. To fulfill the control objective, two-loop FOPID-FOPI controllers are proposed, and the Levenberg Marquardt algorithm (LMA) is utilized to tune the controllers. A novel nonlinear integral of time-associated absolute-error (ITAE) based fitness formula considering the settling time and rise time is used to fit the controller parameters for 2-LFOC. A performance comparison with the PID controller in terms of different time domain parameters such as rise_time (T R ), peak_time (T P ), settling_time (T S ), maximum overshoot (OS M ), maximum undershoot (US M ) and steady-state error (E SS ) are investigated. Stability analysis using Riemann surface observation of the system compensated with the proposed controller is presented in this work. The robust behavior of the two-loop FOPID-FOPI controller is verified by the application of disturbances in the system and the Reimann surface observation.

Comprehensive analysis of normal shock wave propagation in turbulent non-ideal gas flows with analytical and neural network methods

Shock wave propagation in gases through turbulent flow has wide-reaching implications for both theoretical research and practical applications, including aerospace engineering, propulsion systems, and industrial gas processes. The study of normal shock propagation in turbulent flow over non-ideal gas investigates the changes in pressure, density, and flow velocity across the shock wave. The Mach number is derived for the system and explored across various gas molecule quantities and turbulence intensities. This study analytically investigated the normal shock wave propagation in turbulent flow of adiabatic gases with modified Rankine–Hugoniot conditions. Artificial neural network (ANN) techniques are used to estimate the solutions for shock strength and Mach number training validation phases of back-propagated neural networks with the Levenberg–Marquardt algorithm. The results reveal that pressure ratio with density ratio increase for higher values of increase in the turbulence level as well as intermolecular forces. A reverse trend is observed in velocity coefficient after shock in the presence of adiabatic gas. The regression coefficient values obtained using the network model ranged from 0.999 99 to 1, indicating an almost perfect correlation. These findings demonstrate that the ANN can predict the Mach number with high accuracy.

Entropy optimization in Casson tetra-hybrid nanofluid flow over a rotating disk with nonlinear thermal radiation: a Levenberg–Marquardt neural network approach

Abstract This research employs a neural network, specifically the Levenberg–Marquardt algorithm, to characterize the entropy optimization performance in the electro-magneto-hydrodynamic flow of a Casson tetra-hybrid nanofluid over a rotating disk. The problem was formulated mathematically using equations for momentum, continuity, and temperature. This study converts ordinary differential equations (ODEs) into partial differential equations (PDEs) by a self-similarity transformation. The equations are resolved via the fourth-order Runge-Kutta method in combination with a shooting technique for obtaining the required datasets. Using the Levenberg-Marquardt algorithm (LMA), these datasets are characterised as training, testing, and validation. The proposed outcomes are presented in multiple tables and graphs. This trained neural network is then utilized to predict the heat flow velocity and Nusselt number of the rotating disk. The developed model was evaluated using mean square error, error analysis, and regression analysis, thereby confirming the consistency, accuracy, and reliability of the designed technique. The best validation performance for skin friction and the Nusselt number for the Casson tetra-hybrid nanofluid flow across a rotating disk is 8752e-05 at epoch 95 and 0.00033239 at epoch 37. Training, validation, testing, and all performance metrics of the artificial neural network model are close to unity. As magnetic field strength increases, temperature profiles rise in di-hybrid, ternary-hybrid, and tetra-hybrid nanoparticle scenarios. Tetra-hybrid nanofluids are considered superior fluids when compared to di-hybrid, ternary-hybrid, and tetra-hybrid nanofluids. This optimization method holds promise for diverse applications in biotechnology, microbiology, and medicine, offering significant potential for various fields.

An experimental and theoretical study on the creep behavior of silt soil in the Yellow River flood area of Zhengzhou City

We took the silt soil in the Yellow River flood area of Zhengzhou City as the research object and carried out triaxial shear and triaxial creep tests on silt soil with different moisture contents (8%, 10%, 12%, 14%) to analyze the effect of moisture content on silt soil. In addition, the influence of moisture contents on soil creep characteristics and long-term strength was analyzed. Based on the fractional derivative theory, we established a fractional derivative model that can effectively describe the creep characteristics of silt soil in all stages, and used the Levenberg–Marquardt algorithm to inversely identify the relevant parameters of the fractional derivative creep model. The results show that the shear strengths of silt soil samples with moisture contents of 8%, 10%, 12% and 14% are 294 kPa, 236 kPa, 179 kPa and 161 kPa, respectively. The shear strength of silt soil decreases with increasing moisture content. When the moisture content increases, the cohesion of the silt soil decreases. Under the same deviatoric stress, the higher the moisture content of the silt soil, the greater the deformation will be. The long-term strength of silt soil decreases exponentially with the increase of moisture content. If the moisture content is 12%, the long-term strength loss rate of silt soil is the smallest, with a value of 32.96%. The calculated values of our creep model based on fractional derivatives have a high goodness of fit with the experimental results. This indicates that our model can better simulate the creep characteristics of silt soil. This study can provide a theoretical basis for engineering construction and geological disaster prevention in silt soil areas in the Yellow River flood area.

State of Charge Estimation in Batteries for Electric Vehicle Based on Levenberg–Marquardt Algorithm and Kalman Filter

A new optimization method for estimating the State of Charge (SOC) of battery charge state is proposed. This method incorporates the Levenberg–Marquardt Algorithm (LMA) for online parameter identification and the Extended Kalman Filter (EKF) for SOC. On the one hand, the LMA efficiently alleviates the ’Data saturation’ problem experienced by least squares methods by dynamically adjusting weights of data. On the other hand, the EKF improves the robustness and adaptability of SOC estimation. Simulation results under Hybrid Pulse Power Characteristic (HPPC) conditions demonstrate that this new approach offers superior performance in SOC estimation in batteries for electric vehicles compared to existing methods, with better tracking of the true SOC curve, reduced estimation error, and improved convergence.

Modeling the spatial distribution of soil physical properties in a semiarid tropical region

In semiarid regions, sustainably managing the limited available water resources is critical for food, water, and economic security. An important aspect of water management is monitoring, analyzing, and predicting soil moisture and hydrologic fluxes, especially under the threats of climate change. Unfortunately, since monitoring of soil moisture and water fluxes is often sparse in these regions, one needs to rely on hydrological modeling. The latter requires accurate knowledge of soil properties (e.g., texture, porosity, hydraulic conductivity), which in semiarid regions is also complicated by high level of spatial heterogeneity that is not captured by the low density of soil sampling efforts as well as by globally gridded soil databases. Using 300 monitoring sites across the Semiarid region of NE Brazil and a soil moisture model, this study aims to estimate soil properties at the sites through inverse-modeling and then spatially interpolation these properties via geostatistics. Inverse modeling with the Levenberg–Marquardt algorithm accurately captured soil properties across the sites, while clear spatial patterns in these properties allowed us to confidently interpolation them across the region. Comparing our results with a global soil dataset shows that the latter does not capture spatial variability in the region, underlining the need to leverage small scale information. The derived soil property datasets may then particularly be valuable for regional water and ecosystem management applications.

A Diffusion Model to Describe Water Absorption by Red Rice during Soaking: Variable Mass Diffusivity, Variable Volume, Use of Boundary-Fitted Coordinates

This article aims to carry out experiments on water absorption by husked red rice at constant temperatures of 28, 40, and 50 °C. The description of the absorption kinetics and the analyses of the water distribution and volumetric expansion of each grain, at a given instant, were carried out using a diffusion model. In order for the model to be as close as possible to the real physical situation, the mesh necessary to numerically solve the diffusion equation was generated from the photograph of a grain. Thus, the diffusion equation was written in Boundary-Fitted Coordinates (BFCs). The solution of the diffusion equation written in generalized coordinates was then discretized in space and time, using the Finite Volume method, with a fully implicit formulation, considering the variable volume and variable mass diffusivity as functions of the local moisture content. Optimizations based on the Levenberg–Marquardt algorithm make it possible to determine the parameters of an exponential function relating mass diffusivity and the local moisture content for each temperature. Statistical indicators (chi-square, determination coefficient, and Student’s t-test) allowed us to conclude that the proposed model was very satisfactory for all temperatures, making it possible to simulate the water absorption, water distribution, and volumetric expansion of the grain over time.

ANN‐driven insights into heat and mass transfer dynamics in chemical reactive fluids across variable‐thickness surfaces

AbstractThis study investigates the heat and mass transfer dynamics in exothermic, chemically reactive fluids over variable‐thickness surfaces using advanced numerical methods and artificial neural networks (ANN). The importance of understanding these processes lies in their significant industrial applications, such as in chemical reactors and heat exchangers. We transformed nonlinear partial differential equations into ordinary differential equations and used the bvp4c numerical method to generate a comprehensive data set. The ANN model, trained with the Levenberg–Marquardt algorithm, was evaluated for its accuracy in simulating complex fluid dynamics and thermosolutal transport phenomena. Our results revealed that increasing the second‐grade fluid parameter enhanced skin friction by 20.38%, heat transfer rate by 1.16%, and mass transfer rate by 4.06%. The ANN model demonstrated high predictive precision with a validation mean squared error of . These findings highlight the effectiveness of the ANN methodology in providing precise simulations of fluid dynamics, which is crucial for optimizing industrial processes.