Search Results (60)

Despite the fact that smoking is still a significant global public health concern, current mathematical models of smoking dynamics primarily depend on conventional numerical solvers. A particular five-compartment smoking dynamics model has not yet been solved using deep neural network (DNN) techniques. In order to fill this research gap, this work creates a unique DNN framework that can simulate nonlinear smoking dynamics in a computationally efficient manner. The Levenberg–Marquardt backpropagation technique is used to improve a dual-hidden-layer network consisting of 20 radial basis activation function (RBAF) neurons and 40 log-sigmoid activation function (LSAF) neurons. With a minimum mean squared error (MSE) of $1.865\times {10}^{-6}$ and a coefficient of determination $R^2$ equal to or near unity across all model variables, the trained DNN offers instantaneous predictions while maintaining superior accuracy, in contrast to traditional numerical methods that necessitate the explicit re-solving of differential equations for each parameter change. Crucially, our DNN-based framework is appropriate for automated public health decision-support systems since it functions independently and does not require human intervention during the prediction phase. Key smoking behaviors, such as initiation, quitting efforts, relapse dynamics, and long-term recovery patterns, are successfully replicated by the framework, while relapse dynamics are captured through the recovered-to-potential smoker pathway, consistent with the original model formulation. These findings show that the proposed DNN approach not only closes the methodological gap in the application of deep learning to smoking dynamics but also offers a dependable and computationally effective tool for quick evaluation of intervention scenarios, supporting evidence-based public health decision making without compromising accuracy. Full article

An artificial neural network (ANN) prediction model based on the Levenberg–Marquardt (LM) algorithm has been developed to predict the nonlinear heat and mass transfer characteristics of Cattaneo–Christov Carreau–Yasuda tri-hybrid nanofluid (CCHMF–THNF) flow over a porous stretching sheet. A mathematical model of the phenomenon was developed based on a number of elements, including the combined effect of magnetohydrodynamic forces, thermal and solutal relaxation and the influence of viscoelastic fluid behavior and is numerically analyzed utilizing MATLAB bvp4c software. A set of standard data was generated as a reference for developing the ANN-LM model with one hidden layer containing 10 neurons and log-sigmoid activation function, to achieve rapid predictions of velocity, temperature and concentration profiles from the identified data set. This study introduces a novel methodology to provide fast prediction capabilities for transport characteristics through integration of the ANN–LM model with the non-linear CCHMF-THNF model, producing computational savings by providing prediction accuracy of transport characteristics with MSE values on the order of $1.0\times {10}^{-10}$ using ANN–LM in place of repeated bvp4c solutions. Furthermore, the predictive capability of the developed ANN–LM framework may be beneficial in the areas of thermal management systems, polymer processing, energy transport applications, and magnetically controlled cooling technologies since they all share a need for fast access to transportation characteristic evaluation data. Full article

Predicting CO₂ absorption behavior in aqueous amine systems is a critical challenge for optimizing carbon capture technologies. This research develops a high-precision Artificial Neural Network (ANN) to simulate equilibrium data across various amine classes, including primary (MEA, DGA), secondary (DEA, DPA), and tertiary (MDEA) amines. The model architecture utilizes a Multi-Layer Perceptron (MLP) trained on a dataset split into 70% training, 15% validation, and 15% testing segments to prevent overfitting and ensure reliable generalization. By employing a Sigmoid activation function, the network achieved a coefficient of determination (R²) exceeding 0.98 and an absolute average relative deviation (AARD) below 5%. Furthermore, this study evaluates the efficacy of classical isotherms (Langmuir, Freundlich, and Temkin) strictly as empirical curve-fitting correlations for liquid-phase behavior. Results indicate that while these models are traditionally surface-adsorption based, the Langmuir form provides a mathematically robust fit for the tertiary amine MDEA (R² = 0.9673). Experimental observations indicate that Monoethanolamine (MEA) maintains the highest capacity for CO₂ uptake. Since the model relies on categorical descriptors for amine types, it offers a rapid and efficient framework for assessing specific solvents in post-combustion capture infrastructure. Full article

In the present work, a direct oil cooling strategy using a multi-nozzle configuration is proposed for the thermal management of high-power density electric machines. The stator and winding temperatures, heat transfer coefficient, injection pressure, and power consumption are investigated for different nozzle types, nozzle numbers, heights of nozzle combinations, and oil flow rates. In addition, an artificial neural network (ANN) model based on two algorithms is developed for predicting thermal performance under various operating conditions. The flat jet nozzle shows the lowest maximum winding temperature of 120.3 °C and a superior heat transfer coefficient of 3028.6 W/m²-K compared to both full cone nozzles. The power consumption for the flat jet nozzle is higher at 123.9 W compared to other nozzle types. The combination of four flat jet nozzles shows improved oil spray distribution and enhanced cooling compared to combinations of two and six flat jet nozzles. Further, the thermal performance of oil spray cooling with four flat jet nozzles improves when height and oil flow rate are increased. Oil spray cooling with the best configuration shows a winding temperature, heat transfer coefficient, and injection pressure of 98.9 °C, 3408.6 W/m²-K and 4.86 bar, respectively, at a flow rate of 20 LPM. The proposed neural network model with a Levenberg–Marquardt (LM) training variant and logarithmic–sigmoidal (Log) transfer function shows the lowest prediction error within ±2%. Full article

The effects of different activation functions (sigmoid and hyperbolic tangent), and node numbers in a hidden layer of artificial neural network (ANN) models on urban air quality forecasting were investigated in a coastal city of Korea. The ANN models of multilayer perceptron (MLP) with a back-propagation training algorithm for error calculation in cases of 13, 15, and 17 nodes in each hidden layer were performed using 15 input independent variables (PM, gas, and meteorological data of Gangneung city (Republic of Korea)), affected by PM and gas of an upwind Beijing city (China). Root mean square error (RMSE) and the coefficient of determination (R²; Pearson R) were evaluated to assess the two models’ forecasting abilities between the predicted and measured values. The values of R by ANN-sig (ANN-tanh) with 13, 15, and 17 hidden neuron numbers were 0.930 (0.950), 0.920 (0.947), 0.926(0.953) on PM₁₀, 0.953 (0.956), 0.927 (0.938), 0.949 (0.960) on PM_2.5, and 0.880 (0.959), 0.917 (0.886), and 0.882 (0.939) on NO₂. Regardless of node numbers and activation functions, the prediction abilities of the two models were excellent, showing the highest values of R in the ANN-tanh model with more neuron numbers in the hidden layer. Unlike previous studies’ insistence that smaller nodes (larger) in the hidden layer produce the overfit (underfit) result in the ANN model, the present study proves that more nodes in its hidden layer than the input layer can yield the best prediction than any other, as shown in their temporal distributions and scatter plots of predicted and measured data. Future-time air quality forecasting at Gangneung city can be calculated sequentially using its current time data and previous time data from Beijing city, using the suggested empirical formulas. Full article

Laser-induced breakdown spectroscopy (LIBS) has played a critical role in Mars exploration missions, substantially contributing to the geochemical analysis of Martian surface substances. However, the complex nonlinearity of LIBS processes can considerably limit the quantification accuracy of conventional LIBS chemometric methods. Hence chemometrics based on artificial neural network (ANN) algorithms have become increasingly popular in LIBS analysis due to their extraordinary ability in nonlinear feature modeling. The hidden layer activation functions are key to ANN model performance, yet common activation functions usually suffer from problems such as gradient vanishing (e.g., Sigmoid and Tanh) and dying neurons (e.g., ReLU). In this study, we propose a novel LIBS quantification method, named the Bayesian optimization-based tunable Softplus backpropagation neural network (BOTS-BPNN). Based on a dataset comprising 1800 LIBS spectra collected by a laboratory duplicate of the MarSCoDe instrument onboard the Zhurong Mars rover, we have revealed that a BPNN model adopting a tunable Softplus activation function can achieve higher prediction accuracy than BPNN models adopting other common activation functions if the tunable Softplus parameter β is properly selected. Moreover, the way to find the proper β value has also been investigated. We demonstrate that the Bayesian optimization method surpasses the traditional grid search method regarding both performance and efficiency. The BOTS-BPNN model also shows superior performance over other common machine learning models like random forest (RF). This work indicates the potential of BOTS-BPNN as an effective chemometric method for analyzing Mars in situ LIBS data and sheds light on the use of chemometrics for data analysis in future planetary explorations. Full article

The scrap-based electric arc furnace process is expected to capture a significant share of the steel market in the future due to its potential for reducing environmental impacts through steel recycling. However, managing impurities, particularly phosphorus, remains a challenge. This study aims to develop a machine learning model to estimate steel phosphorus content at the end of the process based on input parameters. Data were collected over one year from a steel plant, focusing on parameters such as the chemical composition and weight of the scrap, the volume of oxygen injected, injected lime, and process duration. After preprocessing the data, several machine learning models were evaluated, with the artificial neural network (ANN) emerging as the most effective. The Adam optimizer and non-linear sigmoid activation function were employed. The best ANN model included four hidden layers and 448 neurons. The model was trained for 500 epochs with a batch size of 50. The model achieves a mean square error (MSE) of 0.000016, a root mean square error (RMSE) of 0.0049998, a coefficient of determination (R²) of 99.96%, and a correlation coefficient (r) of 99.98%. Notably, the model was tested on over 200 unseen data points and achieved a 100% hit rate for predicting phosphorus content within ±0.001 wt% (±10 ppm). These results demonstrate that the optimized ANN model offers accurate predictions for the steel final phosphorus content. Full article

Recently there has been growing interest in dyeing biomaterials using natural sustainable plant extracts classified as eco-friendly. The microwave-assisted method provides fast heating and energy efficiency, more homogenous heat distribution in dyeing baths, less use of chemicals, and less heat loss, resulting in this method being greener—more sustainable and ecological. Artificial neural networks (ANNs) are used to predict the dyeing properties of fibers, which are often complex and dependent on multiple variables. This saves time and reduces costs compared to trial-and-error methods. This study presents the green dyeing of merino wool fiber with natural dye extracted from Aesculus hippocastanum (horse chestnut) shells using the microwave-assisted method. Before dyeing, the merino wool fiber underwent a pre-mordanted process with aluminum potassium sulfate with different concentrations using the microwave-assisted method. Spectrophotometric analysis of the light, washing, and rubbing fastness of the dyed merino wool fibers was performed. The color strength, light, washing, and rubbing fastness of the dyed merino wool fiber were developed using the pre-mordanting process. After the pre-mordanting process, the light fastness of the samples improved from 1–2 to 3, the color change increased from 2 to 3–4, and the rubbing fastness developed from 2–3 to 4 according to mordant concentration, mordanting time, and dyeing time quantities. The spectrophotometric analysis results indicate that color coordinates vary based on mordant concentration, mordanting, and dyeing duration. Furthermore, the results proved that microwave energy significantly shortened the mordanting and dyeing duration, resulting in an eco-friendly dyeing process. In this investigation, a feed-forward neural network (FFNN) model with sigmoid hidden neurons and a linear output neuron was used to predict the color strength dyeing property of merino wool fiber. Experimental results showed that the proposed model achieved a regression value of 0.9 for the color strength dyeing property. As demonstrated, the proposed FFNN model is effective and can be utilized to forecast the color strength dyeing properties of merino wool fiber. Full article

Conveyor belts are vital for material transportation in coal mines due to their cost-effectiveness and versatility. These belts endure significant wear from harsh operating conditions, risking substantial financial losses if they fail. This study develops five artificial neural network (ANN) models to predict conveyor belt damage using 11 parameters from the Belchatow brown coal mine in Poland. The models target five outputs: number of repairs and cable cuts, cumulative number of repairs and cable cuts, and their ages. Various optimizers (Adam, Nadam, RMSprop, Adamax, and stochastic gradient descent or SGD) and activation functions (ReLU, Swish, sigmoid, tanh, Leaky ReLU, and softmax) were tested to find the optimal configurations. The predictive performance was evaluated using three error indicators against actual mine data. Superior models can forecast belt behavior under specific conditions, aiding proactive maintenance. The study also advocates for the Diagbelt+ system over human inspections for failure detection. This modeling approach enhances proactive maintenance, preventing total system breakdowns due to belt wear. Full article

The coronavirus (COVID-19) pandemic continues to claim victims. According to the World Health Organization, in the 28 days leading up to 25 February 2024 alone, the number of deaths from COVID-19 was 7141. In this work, we aimed to model the waves of COVID-19 through artificial neural networks (ANNs) and the sigmoidal–Boltzmann model. The study variable was the global cumulative number of deaths according to days, based on the Peru dataset. Additionally, the variables were adapted to determine the correlation between social isolation measures and death rates, which constitutes a novel contribution. A quantitative methodology was used that implemented a non-experimental, longitudinal, and correlational design. The study was retrospective. The results show that the sigmoidal and ANN models were reasonably representative and could help to predict the spread of COVID-19 over the course of multiple waves. Furthermore, the results were precise, with a Pearson correlation coefficient greater than 0.999. The computational sigmoidal–Boltzmann model was also time-efficient. Moreover, the Spearman correlation between social isolation measures and death rates was 0.77, which is acceptable considering that the social isolation variable is qualitative. Finally, we concluded that social isolation measures had a significant effect on reducing deaths from COVID-19. Full article

The automobile industry relies primarily on spot welding operations, particularly resistance spot welding (RSW). The performance and durability of the resistance spot-welded joints are significantly impacted by the welding quality outputs, such as the shear force, nugget diameter, failure mode, and the hardness of the welded joints. In light of this, the present study sought to determine how the aforementioned welding quality outputs of 0.5 and 1 mm thick austenitic stainless steel AISI 304 were affected by RSW parameters, such as welding current, welding time, pressure, holding time, squeezing time, and pulse welding. In order to guarantee precise evaluation and experimental analysis, it is essential that they are supported by a numerical model using an intelligent model. The primary objective of this research is to develop and enhance an intelligent model employing artificial neural network (ANN) models. This model aims to provide deeper knowledge of how the RSW parameters affect the quality of optimum joint behavior. The proposed neural network (NN) models were executed using different ANN structures with various training and transfer functions based on the feedforward backpropagation approach to find the optimal model. The performance of the ANN models was evaluated in accordance with validation metrics, like the mean squared error (MSE) and correlation coefficient (R²). Assessing the experimental findings revealed the maximum shear force and nugget diameter emerged to be 8.6 kN and 5.4 mm for the case of 1–1 mm, 3.298 kN and 4.1 mm for the case of 0.5–0.5 mm, and 4.031 kN and 4.9 mm for the case of 0.5–1 mm. Based on the results of the Pareto charts generated by the Minitab program, the most important parameter for the 1–1 mm case was the welding current; for the 0.5–0.5 mm case, it was pulse welding; and for the 0.5–1 mm case, it was holding time. When looking at the hardness results, it is clear that the nugget zone is much higher than the heat-affected zone (HZ) and base metal (BM) in all three cases. The ANN models showed that the one-output shear force model gave the best prediction, relating to the highest R and the lowest MSE compared to the one-output nugget diameter model and two-output structure. However, the Levenberg–Marquardt backpropagation (Trainlm) training function with the log sigmoid transfer function recorded the best prediction results of both ANN structures. Full article

Nowadays, machine learning (ML) is a useful technology due to its high accuracy in constructing non-linear models and algorithms that can adapt to the complexity and diversity of data. Thus, the current work aimed to predict the soil quality index (SQI) from extensive soil data, achieving high accuracy with the artificial neural networks (ANN) model. However, the efficiency of ANN depends on the accuracy of the data that is prepared for training. For this purpose, MATLAB programming language was used to enable the calculation, classification, and compilation of the results into databases within a few minutes. The proposed MATLAB program was highly efficient, accurate, and quick in calculating soil big data for training the machine compared with traditional methods. The database contains 306 vector sets, 80% of them are used for training and the remaining 20% are reserved for testing. The optimal model obtained comprises one hidden layer with 250 neurons and one output layer with a sigmoid function. The ANN achieved a high coefficient of determination (R²) values for SQI estimation, with around 0.97 and 0.98 for training and testing, respectively. The results indicate that 36.93% of the total soil samples belonged to the very high quality class (C1). In contrast, the high quality (C2), moderate quality (C3), low quality (C4), and very low quality (C5) classes accounted for 10.46%, 31.37%, 20.92%, and 0.33% of the samples, respectively. The high contents of CaCO₃, pH, sodium saturation, salinity, and clay content were identified as limiting factors in certain areas. The results of this study indicated high accuracy of soil quality assessment using physical, chemical, and fertility soil features in regression analysis with ANN. This method, which is suitable for arid zones, enhances agricultural productivity and decision-making by identifying critical soil quality categories and constraints. Full article

The resistance spot welding (RSW) process is still widely used to weld panels and bodies, particularly in the automotive, railroad, and aerospace industries. The purpose of this research is to examine how RSW factors such as welding current, welding pressure, welding time, holding time, squeezing time, and pulse welding affect the shear force, micro-hardness, and failure mode of spot welded titanium sheets (grade 2). Resistance spot welded joints of titanium sheets with similar and dissimilar thicknesses of 1–1 mm, 0.5–0.5 mm, and 1–0.5 mm were evaluated. The experimental conditions were arranged using the design of experiments (DOE). Moreover, artificial neural network (ANN) models were used. Different training and transfer functions were tested using the feed-forward backpropagation approach to find the optimal ANN model. According to the experimental results, the maximum shear force was 5.106, 4.234, and 4.421 kN for the 1–1, 0.5–0.5, and 1–0.5 mm cases, respectively. The hardness measurements showed noticeable improvement for the welded joints compared to the base metal. The findings revealed that the 0.5–0.5 mm case gives the highest nugget and heat-affected zone (HAZ) hardness compared to other cases. Moreover, different failure modes like pull-out nugget, interfacial, and partial failure between the pull-out nugget and interfacial failure were noticed. The ANN outcomes based on the mean squared error (MSE) and coefficient of determination (R²) as validation metrics demonstrated that using the Levenberg–Marquardt (Trainlm) training function with the log sigmoid transfer function (Logsig) gives the best prediction, where R² and MSE values were 0.98433 and 0.01821, respectively. Full article

An optimisation model for small datasets was applied to thixocasted/compocasted composites and hybrid composites with hypereutectic Al-18Si base alloys. Composites were produced with the addition of Al₂O₃ (36 µm/25 nm) or SiC (40 µm) particles. Based on the design of experiment, tribological tests were performed on the tribometer with block-on-disc contact geometry for normal loads of 100 and 200 N, a sliding speed of 0.5 m/s, and a sliding distance of 1000 m. For the prediction of the tribological behaviour of composites, artificial neural networks (ANNs) were used. Three inputs were considered for ANN training: type of reinforcement (base alloy, Al₂O₃ and SiC), amount of Al₂O₃ nano-reinforcement (0 and 0.5 wt.%), and load (100 and 200 N). Various ANNs were applied, and the best ANN for wear rate (WR), with an overall regression coefficient of 0.99484, was a network with architecture 3-15-1 and a logsig (logarithmic sigmoid) transfer function. For coefficient of friction (CoF), the best ANN was the one with architecture 3-6-1 and a tansig (hyperbolic tangent sigmoid) transfer function and had an overall regression coefficient of 0.93096. To investigate the potential of ANN for the prediction of two outputs simultaneously, an ANN was trained, and the best results were from network 3-5-2 with a logsig transfer function and overall regression coefficient of 0.99776, but the predicted values for CoF in this case did not show good correlation with experimental results. After the selection of the best ANNs, the Taguchi grey multi-response optimisation of WR and CoF was performed for the same combination of factors as the ANNs. For optimal WR and CoF, the combination of factors was as follows: composite with 3 wt.% Al₂O₃ micro-reinforcement, 0.5 wt.% Al₂O₃ nano-reinforcement, and a load of 100 N. The results show that developed ANN, the Taguchi method, and the Taguchi grey method can, with high reliability, be used for the optimisation of wear rate and coefficient of friction of hypereutectic Al-Si composites. Microstructural investigations of worn surfaces were performed, and the wear mechanism for all tested materials was light abrasion and adhesion. The findings from this research can contribute to the future development of hypereutectic Al-Si composites. Full article

The limitations of existing commercial indirect liquid cooling have drawn attention to direct liquid cooling for battery thermal management in next-generation electric vehicles. To commercialize direct liquid cooling for battery thermal management, an extensive database reflecting performance and operating parameters needs to be established. The development of prediction models could generate this reference database to design an effective cooling system with the least experimental effort. In the present work, artificial neural network (ANN) modeling is demonstrated to predict the thermal and electrical performances of batteries with direct oil cooling based on various operating conditions. The experiments are conducted on an 18650 battery module with direct oil cooling to generate the learning data for the development of neural network models. The neural network models are developed considering oil temperature, oil flow rate, and discharge rate as the input operating conditions and maximum temperature, temperature difference, heat transfer coefficient, and voltage as the output thermal and electrical performances. The proposed neural network models comprise two algorithms, the Levenberg–Marquardt (LM) training variant with the Tangential-Sigmoidal (Tan-Sig) transfer function and that with the Logarithmic-Sigmoidal (Log-Sig) transfer function. The ANN_LM-Tan algorithm with a structure of 3-10-10-4 shows accurate prediction of thermal and electrical performances under all operating conditions compared to the ANN_LM-Log algorithm with the same structure. The maximum prediction errors for the ANN_LM-Tan and ANN_LM-Log algorithms are restricted within ±0.97% and ±4.81%, respectively, considering all input and output parameters. The ANN_LM-Tan algorithm is suggested to accurately predict the thermal and electrical performances of batteries with direct oil cooling based on a maximum determination coefficient (R²) and variance coefficient (COV) of 0.99 and 1.65, respectively. Full article