Biases Of Artificial Neural Network Research Articles

A CFD‐PBM‐ANN framework to simulate the liquid–liquid two‐phase flow in a pulsed column

AbstractCFD‐PBM numerical simulation is a powerful tool in the research of droplet swarm behavior. In this work, an artificial neural network (ANN) based droplet breakage frequency function is established based on the directly measured data from our previous studies. Then, the weights and biases of ANN are embedded into the CFD‐PBM code in the form of matrices and vectors. For the first time, a CFD‐PBM‐ANN simulation framework is established. Simulation results are in good agreement with the experimental data under different operation conditions. The cumulative droplet size distribution decreases with the increase of interfacial tension and pulse intensity. It is also found by the simulation that the droplet breakage frequency is relatively high at the edge of disc and doughnut plate, which is accordant with the distribution of turbulent energy dissipation and velocity gradient.

Predicting the thickness loss and recovery of acrylic carpet pile imposed to UV exposure using artificial intelligence

One of the metrics that establishes the carpet quality is its compress and decompress behaviors. In a research, addressed later in the paper, carpet samples with the acrylic pile were exposed to the ultraviolet (UV) exposure for 0, 8, 12, 16, and 20 h. The wear test was then carried out using drum revolutions of 0, 4000, 6000, and 8000. Following that, the samples were subjected to a static stress of 700 KPa for 2 h. After removing the load, the sample thickness was measured every 2 min. Finally, the thickness loss and recovery features were determined. Now in this study, the statistical analysis demonstrated the significance of recovery time, UV exposure, and drum revolutions on the features mentioned above, and in the next step, they were modeled using an artificial neural network (ANN) structure. The genetic algorithm, particle swarm optimization, and gray wolf optimization (GWO) were applied as the alternatives to the backpropagation (BP) algorithm for updating the weights and biases of ANN. In the end, it was found that the thickness loss and recovery could be predicted with 6.92 and 9.14% errors, using ANN-GWO and ANN-BP-GWO, respectively.

Development of ANN-based metaheuristic models for the study of the durability characteristics of high-volume fly ash self-compacting concrete with silica fume

The construction of durable and sustainable infrastructure requires the use of industrial byproducts such as fly ash (FA) and silica fume (SF) to enhance strength and durability. This study introduces novel machine learning models to forecast the results of rapid chloride penetration test (RCPT) for self-compacting concrete (SCC) containing high volumes of FA and SF. This study assessed the effects of supplementary cementitious materials (SCMs) and elevated temperature curing on the RCPT outcomes for SCC. Various metaheuristic algorithms including teaching–learning-based optimization (TLBO), ant colony optimization (ACO), the imperialist competitive algorithm (ICA), and shuffled complex evolution (SCE)—optimize the learning rate, weights, and biases of artificial neural network (ANN) models. A dataset of 360 experimental RCPT data points with seven input parameters was used to train and test the hybrid models. The accuracy of these models was assessed using eight performance indices, and the results were further analyzed through rank analysis, scatter plots, and error matrices. The training and testing sets for the AI models, specifically ANN-TLBO, exhibit a strong correlation between the experimental and predicted RCPT values, with R2 values of 0.9962 for training and 0.9676 for testing, compared to the R2 values of the other proposed models. Consequently, the results of this study suggest that the ANN-TLBO model is the optimal hybrid ANN model for predicting RCPT values in SCC. Verified experimental results and external validation indicate that the ANN-TLBO model is an effective alternative for accurately predicting real-time RCPT in SCC.

An artificial neural network model for determining stress concentration factors for fatigue design of tubular T-joint under compressive loads

PurposeThe stress concentration factor (SCF) is commonly utilized to assess the fatigue life of a tubular T-joint in offshore structures. Parametric equations derived from experimental testing and finite element analysis (FEA) are utilized to estimate the SCF efficiently. The mathematical equations provide the SCF at the crown and saddle of tubular T-joints for various load scenarios. Offshore structures are subjected to a wide range of stresses from all directions, and the hotspot stress might occur anywhere along the brace. It is critical to incorporate stress distribution since using the single-point SCF equation can lead to inaccurate hotspot stress and fatigue life estimates. As far as we know, there are no equations available to determine the SCF around the axis of the brace.Design/methodology/approachA mathematical model based on the training weights and biases of artificial neural networks (ANNs) is presented to predict SCF. 625 FEA simulations were conducted to obtain SCF data to train the ANN.FindingsUsing real data, this ANN was used to create mathematical formulas for determining the SCF. The equations can calculate the SCF with a percentage error of less than 6%.Practical implicationsEngineers in practice can use the equations to compute the hotspot stress precisely and rapidly, thereby minimizing risks linked to fatigue failure of offshore structures and assuring their longevity and reliability. Our research contributes to enhancing the safety and reliability of offshore structures by facilitating more precise assessments of stress distribution.Originality/valuePrecisely determining the SCF for the fatigue life of offshore structures reduces the potential hazards associated with fatigue failure, thereby guaranteeing their longevity and reliability. The present study offers a systematic approach for using FEA and ANN to calculate the stress distribution along the weld toe and the SCF in T-joints since ANNs are better at approximating complex phenomena than standard data fitting techniques. Once a database of parametric equations is available, it can be used to rapidly approximate the SCF, unlike experimentation, which is costly and FEA, which is time consuming.

A machine learning investigation of the ZnO–water nanofluid flow with magnetic field through convergent and divergent channels: a numerical study

This investigation focuses on how ZnO–H2O nanofluid flow affects the thermal processes in divergent/convergent channels under the effects of magnetohydrodynamics (MHD). Advanced machine learning techniques are utilised to comprehensively quantify the interactions among the implications of nanoparticles on thermal transmission and fluid actions. Our analysis encompasses the MHD and Joule emission effects, transforming the intricate system of nonlinear system of equations to ordinary differential equations (ODEs) following the strategy of non-dimensionalisation. Leveraging the integration of artificial neural networks (ANN) integrated a hybrid approach for numerical solutions. The optimum value of ANN-based fitness function is up to 5.53 × 10 − 5 . The optimal weights and biases of ANN through ABC–WCA are ranging from − 10 to 10 . These results are meticulously compared with shooting method and homotopy perturbation method (HPM) solutions. The absolute error between the ANN, shooting and HPM are equal up to four (4) decimal places. Further, the ANN-based fitness function optimised through ABC-WCA over 250 independent runs for the efficiency of algorithms. It is found that the current results are much more authentic and powerful than the existing less converging and slow analytic and numerical methods.

Machine learning intelligent based hydromagnetic thermal transport under Soret and Dufour effects in convergent/divergent channels: a hybrid evolutionary numerical algorithm

In this research, we analyze the complex dynamics of hydro-magnetic flow and heat transport under Sorent and Dofour effects within wedge-shaped converging and diverging channels emphasizing its critical role in conventional system design, high-performance thermal equipment. We utilized artificial neural networks (ANNs) to investigation the dynamics of the problem. Our study centers on unraveling the intricacies of energy transport and entropy production arising from the pressure-driven flow of a non-Newtonian fluid within both convergent and divergent channel. The weights of ANN based fitness function ranging from − 10 to 10. To optimize the weights and biases of artificial neural networks (ANNs), employ a hybridization of advanced evolutionary optimization algorithms, specifically the artificial bee colony (ABC) optimization integrated with neural network algorithms (NNA). This approach allows us to identify and fine-tune the optimal weights within the neural network, enabling accurate prediction. We compare our results against the established different analytical and numerical methods to assess the effectiveness of our approach. The methodology undergoes a rigorous evaluation, encompassing multiple independent runs to ensure the robustness and reliability of our findings. Additionally, we conduct a comprehensive analysis that includes metrics such as mean squared error, minimum values, maximum values, average values, and standard deviation over these multiple independent runs. The minimum fitness function value is 1.32 × 10−8 computed across these multiple runs. The absolute error, between the HAM and machine learning approach addressed ranging from 3.55 × 10−7 to 1.90 × 10−8. This multifaceted evaluation ensures a thorough understanding of the performance and variability of our proposed approach, ultimately contributing to our understanding of entropy management in non-uniform channel flows, with valuable implications for diverse engineering applications.

Prediction of Bearing Capacity of the Square Concrete-Filled Steel Tube Columns: An Application of Metaheuristic-Based Neural Network Models.

During design and construction of buildings, the employed materials can substantially impact the structures’ performance. In composite columns, the properties and performance of concrete and steel have a significant influence on the behavior of structure under various loading conditions. In this study, two metaheuristic algorithms, particle swarm optimization (PSO) and competitive imperialism algorithm (ICA), were combined with the artificial neural network (ANN) model to predict the bearing capacity of the square concrete-filled steel tube (SCFST) columns. To achieve this objective and investigate the performance of optimization algorithms on the ANN, one of the most extensive datasets of pure SCFST columns (with 149 data samples) was used in the modeling process. In-depth and detailed predictive modeling of metaheuristic-based models was conducted through several parametric investigations, and the optimum factors were designed. Furthermore, the capability of these hybrid models was assessed using robust statistical matrices. The results indicated that PSO is stronger than ICA in finding optimum weights and biases of ANN in predicting the bearing capacity of the SCFST columns. Therefore, each column and its bearing capacity can be well-predicted using the developed metaheuristic-based ANN model.

Initialization and Estimation of Weights and Bias using Bayesian Technique

Researches on artificial neural network models have shown that the method used to initialize and estimate weights and bias always determines the rate at which the network will converge and how efficient the model will perform. Although there are several methods that can be used to initialize and estimate network weights and bias, Bayesian approach is widely considered as a more efficient method for modelling artificial neural networks because it can easily compute the inverses of covariance matrices with high dimension that otherwise are computationally expensive. This study has developed a new filter, the First order Extended Ensemble Filter(FoEEF),that applies numerical solution in solving the inverses of high dimensional covariance matrices from I ^ to ′ s stochastic state-space dynamical models. The research applies the new FoEEF Filter to initializes and estimates the weights and bias of artificial neural network. Comparison on the performance of FoEEF filter in estimating weights and bias are done against the performance of EKF on function estimations by using two different functions, sin(x)+Q function and 3sin(x)3− 3sin(x)2 − sin(x) + 1 function, within 18 and 22 epochs. Emphasis of the performance is placed on the rate at which the models converge. This study gauges the performance by determining the minimum value of the mean square error produced from each epoch and average mean square error minimum value.The outcome from the study showed that FoEEF filter had the lowest value of mean square error and average mean square error which was achieved with the least number of epochs compared to EKF filter. This study concluded that the new FoEEF performed better than EKF and is a more suitable filter for that can be used for initialization and estimation of weights and bias in artificial neural network.

A Comparison Between Two Approaches to Optimize Weights of Connections in Artificial Neural Networks

Artificial neural networks (ANNs) have been used for estimation in numerous areas. Raising the accuracy of ANNs is always one of the important challenges, which is generally defined as a non-linear optimization problem. The aim of this optimization is to find better values for the weights of the connections and biases in ANN because they seriously affect the efficiency. This study uses two approaches to do such optimization in an ANN. For this aim, we create a feed-forward backpropagation ANN using the functions of MATLAB's deep learning toolbox. To improve its accuracy, in the first approach, we use the Levenberg - Marquardt algorithm (LMA) for training, which is available in MATLAB's deep learning toolbox. In the second approach, we optimize the values of weights and biases of ANN with Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), available in MATLAB's global optimization toolbox. Then, we assess the accuracy of estimation for the trained ANNs. In this way, for the first time in the literature, we compare these methods for the optimization of an ANN. The used data sets are also available in MATLAB. Based on the acquired results, in some data sets, training with LMA, and for some others training with PSO cause the best results, however, training with LMA is faster, significantly. Although the used approaches and the obtained conclusions are beneficial for researchers that work in this field, they have some limitations. For instance, since only the functions and data sets from MATLAB are used, it can only serve as an example for researchers.

Hybrid prediction strategy to predict agricultural information

The crop yield prediction (CYP) has a high significance in agriculture. Early crop yield predictions assist the farmers, decision-makers in making timely decisions during the actual growing season. In many developing countries such as India, the process of crop yield prediction is done manually, based on surveys and field visits which are time-consuming, expensive and prone to human error. To overcome these drawbacks, we propose a hybrid prediction strategy which can be applied to predict agricultural information such as crop yield and air temperature with a critical focus on crop yield prediction. The weighted principal component analysis (w-PCA) is used as the feature extraction strategy to extract the relevant features. A hybrid prediction strategy integrating artificial neural network (ANN) with modified-particle swarm optimization (m-PSO) is proposed. Initial parameters of ANN are selected using m-PSO with modified inertia weight and velocity update equations. This hybridized ANN then performs prediction on the selected features. This proposed prediction model which we call as hybrid-ANN (H-ANN) comprises of w-PCA as feature extractor, m-PSO for selecting initial weights and biases of ANN. Experiments were performed on eight real world and two benchmark agriculture data sets for crop yield and air temperature prediction. Results show that the proposed prediction model (H-ANN) performed with improvements in the range of 2 to 30%, 0.2 to 4% and 0.12 to 3% with respect to R-squared (R2), root mean square error (RMSE) and mean absolute error (MAE) respectively when compared to other prediction models such as ANN, ANN trained using GA (GA-ANN), ANN trained using standard PSO (SPSO-ANN), multiple linear regression (MLR), support vector regression (SVR) and ensemble of bagged regression trees (ET) on benchmark and real-world agricultural datasets.

Performance Evaluation of a GM-Type Double Inlet Pulse Tube Refrigerator Using Artificial Intelligence Approach with Experimental Validation

In the present article, a methodology is suggested to enhance the cooling performance of a double inlet pulse tube refrigerator (DIPTR), which is affected by multiple operating and geometrical parameters, using artificial intelligence methods. A procedure based on artificial intelligence method is adopted to generate the optimum range of inputs to achieve the maximum obtainable performance of the DIPTR. Artificial neural network (ANN) is developed using three different activation functions at the outer layer. It is observed that purelin and tansig activation functions predict the results (cooling power and percentage Carnot respectively) more accurately in accordance with the numerical results. In addition, it is observed that the particle swarm optimization (PSO) of the weights and bias of ANN is capable of representing the non-linear mathematical relationship among inputs and outputs more accurately in case of a DIPTR. It is further observed that the hybrid scheme of the artificial-neuro-fuzzy-inference system (ANFIS) can estimate both cooling capacity and percentage Carnot apparently more precise than the backpropagation algorithm. A numerical model is developed, based on the finite volume discretization of the governing equations with ideal gas assumption to generate the data matrix, which is essential to develop the ANN and ANFIS. Finally, an experimental analysis is conducted to validate the optimum range of inputs (low and high-pressures of compressor, waiting time of rotary valve and frequency) obtained from artificial intelligence models.