Artificial Neural Network Research Articles

Modelling and optimization for material removal rate while processing bio-compatible Ti-6Al-7Nb using wire electric discharge machining (WEDM)

Abstract This study utilizes machine learning methodologies, such as artificial neural networks (ANN) and TLBO (teaching Learning Based optimisation), to develop a model and Optimisation of the Material Removal Rate (MRR) in wire electrical discharge machining (WEDM) of Ti-6Al-7Nb. The material removal rate was determined by conducting WEDM experiments with different levels of control parameters, including spark on time, spark off time, peak current, servo voltage, and wire feed rate using a Full Factorial approach through 81 runs. The most effective architecture for the ANN model was 4–10–1, and the parameters were adjusted depending on R2. The Artificial Neural Network (ANN) predictions were compared to those produced by the Multiple Linear Regression (MLR) model. The performance of these models was assessed by calculating the correlation between the experimental values and predicted values by models (R2). MRR value is optimised using TLBO (Teaching Learning Based Optimisation), keeping the relation developed by MLR as the objective function and leading to an improved material removal rate. The proposed method ANN & TLBO would help accurately predict and optimise MRR while processing Ti-6Al-7Nb. These machine learning-based methods significantly enhance complex machining processes by providing predictive capabilities & optimizing parameters, hence playing a vital role in achieving higher efficiency, quality, and adaptability in manufacturing environments.

Acceleration of Modeling Capability for GDI Spray by Machine-Learning Algorithms

Cold start causes a high amount of unburned hydrocarbon and particulate matter emissions in gasoline direct injection (GDI) engines. Therefore, it is necessary to understand the dynamics of spray during a cold start and develop a predictive model to form a better air-fuel mixture in the combustion chamber. In this study, an Artificial Neural Network (ANN) was designed to predict quantitative 3D liquid volume fraction, liquid penetration, and liquid width under different operating conditions. The model was trained with data derived from high-speed and Schlieren imaging experiments with a gasoline surrogate fuel, conducted in a constant volume spray vessel. A coolant circulator was used to simulate the low-temperature conditions (−7 °C) typical of cold starts. The results showed good agreement between machine learning predictions and experimental data, with an overall accuracy R2 of 0.99 for predicting liquid penetration and liquid width. In addition, the developed ANN model was able to predict detailed dynamics of spray plumes. This confirms the robustness of the ANN in predicting spray characteristics and offers a promising tool to enhance GDI engine technologies.

Intercomparison of machine learning methods for statistical downscaling of daily temperature under CMIP6 scenarios: a case study from Iran

ABSTRACT General circulation models (GCM) represent a contemporary and advanced tool designed to simulate the response of climate system to alterations in greenhouse gas levels. Increasing spatial resolutions of the outputs of GCMs on a regional scale requires downscaling process. In this study, four machine learning (ML) models, namely, decision tree regression (DTR), support vector regression, artificial neural network, and K-nearest neighbors (KNN) were employed to downscale the outputs of four Coupled Model Intercomparison Project 6 climate models. The daily observations of temperature data at Nazmakan station in Kohgiluyeh and Boyer-Ahmad Province, Iran, were trained and tested for the periods 1995–2009 and 2009–2015, respectively. According to results of four metrics, DTR and KNN obtained the most valid evaluation in the training and test data, respectively. In addition, downscaling methods were used to predict future daily temperature for 2015–2045 under two Shared Socioeconomic Pathway (SSP2-4.5 and SSP5-8.5) scenarios. Mann–Kendall test results revealed an increasing trend of temperature in most of the coming years. The projected trend expects cooler summers and warmer winters in future. Furthermore, the predicted data under two climate scenarios were assessed in comparison with observed data for 2015–2022. Finally, this study demonstrates the strengths and weaknesses of nonlinear ML techniques in statistical downscaling.

Artificial Neural Networks: A New Frontier in Dental Tissue Regeneration

Artificial Neural Networks: A New Frontier in Dental Tissue Regeneration

Optimizing 3D Laser Foil Printing Parameters for AA 6061: Numerical and Experimental Analysis

Abstract This study utilizes a technology known as 3D laser foil printing (LFP) to create precise structures by layering metal foils using laser welding. Metal foils have the advantages of rapid cooling and efficient heat conduction, allowing for the formation of fine-grained structures. However, when dealing with materials like aluminum alloys in laser processes, defects can arise as a result of their high reflectivity. To address this challenge, laser circular oscillation welding (LCOW) is applied to the LFP process. LCOW's circular motions with higher scanning frequencies widen the keyholes and reduce some defects such as spattering, bubble formation, and microcracks. Simulation predictions with an error margin of approximately 10% in comparison to experimental results demonstrate the reliability of the model. Furthermore, the study integrates circular packing design with artificial neural networks to create comprehensive processing maps tailored to different criteria for extracting optimal welding parameters. As a result, for the optimized processing parameters screened using the above systematic process, no cracks were observed on the upper surface of the 3D LFP parts produced with a laser power of 800 W and a scanning speed of 550 mm/s, and only 0.12% porosity was observed from the cross section of the sample. Future research will focus on incorporating simulation results to model microstructures more precisely and continually refining LCOW parameters as new materials and technologies emerge, ensuring the ongoing enhancement of weld quality in the 3D LFP process.

Assessment of the Impact of Meteorological Variables on Lake Water Temperature Using the SHapley Additive exPlanations Method

The water temperature of lakes is one of their fundamental characteristics, upon which numerous processes in lake ecosystems depend. Therefore, it is crucial to have detailed knowledge about its changes and the factors driving those changes. In this article, a neural network model was developed to examine the impact of meteorological variables on lake water temperature by integrating daily meteorological data with data on interday variations. Neural networks were selected for their ability to model complex, non-linear relationships between variables, often found in environmental data. Among various architectures, the Artificial Neural Network (ANN) was chosen due to its superior performance, achieving an R2 of 0.999, MSE of 0.0352, and MAE of 0.1511 in validation tests. These results significantly outperformed other models such as Multi-Layer Perceptrons (MLPs), Recurrent Neural Networks (RNNs), and Long Short-Term Memory (LSTM). Two lakes (Lake Mikołajskie and Sławskie) differing in morphometric parameters and located in different physico-geographical regions of Poland were analyzed. Performance metrics for both lakes show that the model is capable of providing accurate water temperature forecasts, effectively capturing the primary patterns in the data, and generalizing well to new datasets. Key variables in both cases turned out to be air temperature, while the response to wind and cloud cover exhibited diverse characteristics, which is a result of the morphometric features and locations of the measurement sites.

High‐Gain Triple‐Band T‐Shaped Dielectric Resonator Based Hybrid Two‐Element MIMO Antenna for 5G New Radio, Wi‐Fi 6, V2X, and C‐Band Applications With a Machine Learning Approach

ABSTRACTIn this article, a rectangular dielectric resonator (DR) based multiple input multiple output (MIMO) hybrid antenna with a machine learning (ML) optimization approach for 5G New Radio (NR), WiMAX, WLAN, Wi‐Fi 6, Vehicle to Everything (V2X), and C‐band uplink applications are presented. The proposed antenna provides a triple band. It covers n1, n2, and n39 5G NR at a lower frequency (1.59–2.26 GHz), n48 5G NR at the middle frequency (3.1–3.87 GHz), and n46, n47, n96, WiMAX, WLAN, and C‐band uplink at higher frequency (5.25–6.91 GHz). It achieved more than 20% impedance bandwidth and 20 dB isolation in all frequency bands. The maximum gain of the proposed antenna is 15 dBi. The calculated MIMO parameters are within the accepted range. It is optimized through various ML algorithms, including random forest (RF), K‐nearest neighbor (KNN), artificial neural network (ANN), extreme gradient boosting (XGB) and decision tree (DT). The RF ML algorithm gives 98.09% accuracy compared to other ML algorithms for S parameter prediction.

Machine Learning Enabled Compact Frequency‐Tunable Triple‐Band Hexagonal‐Shaped Graphene Antenna for THz Communication

ABSTRACTIn this article, a compact triple‐band frequency‐tunable (FT) hexagonal‐shaped graphene antenna through a machine learning (ML) approach for terahertz (THz) application is presented. The proposed THz antenna is designed on a polyamide () substrate with a thickness of 10 μm, and graphene is used as an antenna radiator. The size of the substrate is 38 × 46 μm2. The FT is achieved by changing the chemical potential of graphene material. The performance of the proposed THz antenna has been investigated, and the impacts of several conducting materials like gold, aluminum, copper, and graphene and dielectric materials like Rogers RT/duroid 5880, polyamide, quartz, and SiO2 are explored. The proposed THz antenna provides three operating bands. The frequency of operation in Band‐1 is 2.51–5.05 THz, Band‐2 is 5.99–7.43 THz, and Band‐3 is 7.94–9.63 THz. The bandwidth in Band‐1, Band‐2, and Band‐3 are 2.54, 1.44, and 1.69 THz, respectively. The % of impedance bandwidth in Band‐1, Band‐2, and Band‐3 are 67.19%, 24.02%, and 21.28% respectively. The proposed antenna has a maximum peak gain of 5 dBi. The proposed antenna is optimized through various ML algorithms like random forest (RF), extreme gradient boosting (XGB), K‐nearest neighbor (KNN), decision tree (DT), and artificial neural network (ANN). The RF algorithm gives more than 99% accuracy compared to other ML algorithms and accurately predicts the S11 of the proposed antenna. The proposed THz antenna would be suitable for applications related to imaging, medical, sensing, and ultra‐speed short‐distance communication applications in the THz region.

Identification of optimal cutting parameters regarding lacquer wettability of CNC processed MDF panels by artificial neural network

Identification of optimal cutting parameters regarding lacquer wettability of CNC processed MDF panels by artificial neural network

Artificial Neural Network for Benchmarking the Dimensional Accuracy of the PLA Fused Flament Fabrication Process

Fused Deposition Modeling (FDM) is an additive manufacturing technique that uses a 3D printer to extrude molten filament through a nozzle, which moves along the X, Y, and Z axes to create parts with the desired geometry. FDM offers numerous advantages, especially for producing parts with complex shapes, due to its ability to enable rapid and cost-effective manufacturing compared to traditional methods. This study implemented an Artificial Neural Network (ANN) to optimize process parameters aimed at minimizing dimensional inaccuracies in the FDM process. Key parameters considered for optimization included the number of shells, infill percentage, and nozzle temperature. The ANN utilized three algorithms: Scaled Conjugate Gradient, Bayesian Regularization, and Levenberg-Marquardt. Model performance was evaluated based on dimensional deviations along the X and Y axes, with a hidden layer of 25 neurons. Among the algorithms, Scaled Conjugate Gradient provided the most accurate results in minimizing dimensional errors.

Predicting Chronic Liver Disease and Jaundice Detection an Integrated Approach Using Statistical Feature Extraction and Machine Learning Algorithms

An important health concern with chronic liver disease is early identification, which is essential for management and therapy. This work suggests an integrated method to identify tattoo-induced jaundice and forecast chronic liver disease by combining statistical feature extraction with machine learning techniques. First, we use statistical techniques based on projections to extract pertinent aspects from patient data, such as test results and clinical markers. We then use Artificial Neural Networks (ANN), KNN, and Naive Bayes to categorize individuals according to the probability that they have chronic liver illness and to pinpoint jaundice episodes associated with tattooing. The KNN technique offers interpretability, managing categorical data is made simple by Naive Bayes, and complicated patterns are captured by ANN using layers of neural networks. These models' performance is assessed using the following metrics: F1 score, recall, accuracy, and precision. Our research aims to improve prediction accuracy and offer practical advice for early diagnosis and individualized treatment plans. By utilizing advanced data analysis and machine learning instruments, the integrated strategy shows promise in enhancing healthcare outcomes.

Optimization of erosion performance of biomass and pet waste based composites using artificial neural network

Optimization of erosion performance of biomass and pet waste based composites using artificial neural network

Biofuel Production in Oleic Acid Hydrodeoxygenation Utilizing a Ni/Tire Rubber Carbon Catalyst and Predicting of n-Alkanes with Box–Behnken and Artificial Neural Networks

Oleic acid is a valuable molecule for biofuel production, as it is found in high proportions in vegetable oils. When used, oleic acid undergoes hydrodeoxygenation reactions and produces alkanes within the diesel range. These alkanes are free of oxygenated compounds and have molecular structures similar to petrodiesel. Our research introduces a novel approach incorporating oleic acid into the hydrodeoxygenation process of Ni/Tire Rubber Carbon (Ni/CTR) catalysts. These catalysts produced renewable biofuels with properties similar to diesel, particularly a high concentration of n-C17 alkanes. Moreover, our Ni/CTR catalyst produces n-C18 alkanes, but the generation of n-C18 alkanes typically requires more complex catalysts. Our procedure achieved 74.74% of n-C17 alkanes and 2.28% of n-C18 alkanes. We used Box–Behnken and artificial neural networks (ANNs) to find the optimal configuration based on the predicted data. We developed a dataset with pressure, temperature, metal content, reaction time, and catalyst composition variables as inputs. The output variables are the n-C17 and n-C18 alkanes obtained. ANN602020 was our best model for obtaining the peak response; it accurately forecasted the n-C17 and n-C18 generation with R2 scores of 0.9903 and 0.9525, respectively, resulting in an MSE of 0.0014, MAE of 0.02773, and MAPE of 2.03979%. The combined R2 score for both alkanes was 0.97139.

A hybrid algorithm for photovoltaic maximum power point tracking using Artificial Neural Network and Kinetic Gas Molecular Optimization

The quest for efficient photovoltaic (PV) energy conversion systems has led to the development of Maximum Power Point Tracking (MPPT) algorithms. In this paper, we propose a novel hybrid approach that combines the power of Artificial Neural Network (ANN) with the optimization capability of Kinetic Gas Molecular Optimization (KGMO) for MPPT. We present a high-level outline of the proposed algorithm along with example MATLAB code snippets for each step, highlighting its potential for improving PV system performance. This paper proposes a metaheuristic optimized multilayer feed‐forward artificial neural network (ANN) controller to extract the maximum power from available solar energy. Firstly, to improve the maximum power point (MPP) delivered by PV arrays and to overcome the drawbacks in the conventional MPPT method under irradiation variation, a hybrid MPPT controller is designed, in which the input parameters include the PV array voltage and current. The output parameter is the duty cycle of the DC/DC boost converter. The proposed approach abbreviated as ANN-KGMO MPPT controller is based on the Kinetic Gas Molecular Optimization (KGMO) Algorithm which is useful to train the developed ANN and to evolve the connection weights and biases to get the optimal values of duty cycle converter corresponding to the MPP of a PV array. Finally, the performance of the proposed control system is confirmed by simulation tests on a 2 kW PV system. In addition, the performance of an ANN-KGMO -based MPPT controller is also compared to the conventional perturb and observe (P&O) method. To analyze the results, simulations are performed by using MATLAB software.

Performing hardness classification using diffusive memristor based artificial neurons

Abstract Artificial neurons and synapses are the building blocks for constructing a neuromorphic system such as Spiking Neural Network (SNN) or Artificial Neural Network (ANN). Recently, there has been tremendous interest in using memristors to develop neuromorphic technologies that can be used in advanced SNNs and ANNs. Memristors, because of their simple device structure, easy and high-density fabrication, and integration with other semiconductor electronics are suitable candidates for the construction of neuromorphic concepts. However, not much has been discussed about using memristors for the development of sensors that can be utilized for object- classification especially their rigidity, shape and structure. In this article, we propose the application of memristors, specifically silver nanoparticle based diffusive memristor, in conjunction with a piezoelectric sensor within a robotics gripper, serving as one receptor (a tactile sensor) that triggers neuron circuitry with memristors to generate spikes. Furthermore, to perform hardness classification, we utilized various objects to collect data and generated multiple spikes corresponding to each object. This data was then utilized with a machine learning algorithm. The outcomes were compared with the accuracy of commercial FSR tactile sensors. Our approach demonstrated the capability of diffusive memristors in generating neuron spikes from tactile stimuli for hardness classification, achieving accuracy ranging from 82% to 100% during the validation of 20% test data across various algorithms, while the FSR sensors achieved an accuracy range of 95% to 98%.

Seasonal and annual trends in reference evapotranspiration and prediction using machine learning models across seven climatic zones of Bangladesh

ABSTRACT The primary aim of this investigation is to examine the historical (1989–2020) and future trends and magnitude of Reference Evapotranspiration (ET0) in terms of spatiotemporal measures, considering its significance as a hydro-meteorological parameter influenced by changing climate. The FAO-56 Penman-Monteith method is employed to analyze ET0, while the Modified Mann Kendall test is utilized to assess trends and Sen’s Slope Estimator is used for magnitude analysis. The future prediction is conducted using Support Vector Machine (SVM), Artificial Neural Network (ANN), and Random Forest (RF) models. Results show an increasing ET0 trend annually in the southeastern and northeastern zones, while decreased values are observed in other regions, particularly in the northwest (0.83 mm). Sen’s slope estimator reveals distinct fluctuations in ET0, with notable disparities in the Southeastern, Northeastern, Southwestern, and South-Central zones. Notably, Chattogram and Sitakunda experience an ET0 magnitude of 0.02 mm/Year, while other zones show a magnitude of 0.01 mm/Year. SVM outperformed other models, predicting rising ET0 in various seasons. These findings offer insights for optimizing irrigation systems and sustainable water management under changing climatic conditions.

Advanced Bearing-Fault Diagnosis and Classification Using Mel-Scalograms and FOX-Optimized ANN

Accurate and reliable bearing-fault diagnosis is important for ensuring the efficiency and safety of industrial machinery. This paper presents a novel method for bearing-fault diagnosis using Mel-transformed scalograms obtained from vibrational signals (VS). The signals are windowed and pass through a Mel filter bank, converting them into a Mel spectrum. These scalograms are subsequently fed into an autoencoder comprising convolutional and pooling layers to extract robust features. The classification is performed using an artificial neural network (ANN) optimized with the FOX optimizer, which replaces traditional backpropagation. The FOX optimizer enhances synaptic weight adjustments, leading to superior classification accuracy, minimal loss, improved generalization, and increased interpretability. The proposed model was validated on a laboratory dataset obtained from a bearing testbed with multiple fault conditions. Experimental results demonstrate that the model achieves perfect precision, recall, F1-scores, and an AUC of 1.00 across all fault categories, significantly outperforming comparison models. The t-SNE plots illustrate clear separability between different fault classes, confirming the model’s robustness and reliability. This approach offers an efficient and highly accurate solution for real-time predictive maintenance in industrial applications.

OLM-SLAM: Online lifelong memory system for simultaneous localization and mapping

Abstract Simultaneous Localization and Mapping is a fundamental task for robots in unknown environments. However, the poor generalization ability of learning-based algorithms in unknown environments hinders their widespread adoption. Additionally, artificial neural networks are subject to catastrophic forgetting. We propose a lifelong SLAM framework called OLM-SLAM that effectively solves the neural network catastrophic forgetting problem. To ensure the generalization of the neural network, this paper proposes a method for the sensitivity analysis of the network weight parameters. Meanwhile, inspired by human memory storage mechanisms, we design a dual memory storages mechanism that retains dynamic memory and static memory. A novel memory filtering mechanism is proposed to maximize image diversity within a fixed-size memory storage area addressing the problem of limited storage capacity of embedded devices in real-world situations. We have extensively evaluated the model on a variety of real-world datasets. Compared with CL-SLAM, the overall translation error of the test sequence is improved by 44.9%. The translation and rotation errors of Retention Ability were improved by 111.6% and 66.7%, respectively. The results demonstrate that OLM-SLAM can outperform previous methods of the same type, and OLM-SLAM has high retention ability when facing different sequences of the same type of environment.

Comparative Analysis of Machine Learning Models for Predicting Rice Yield: Insights from Agricultural Inputs and Practices in Rwanda

Food security is a global challenge, especially in developing countries like Rwanda. With a growing population and limited agricultural land, Rwanda struggles to meet increasing food demands. Rice is a staple food crop in Rwanda, playing a crucial role in the country’s food security. However, factors like climate variability, soil nutrient management, and limited access to high-quality inputs hinder rice yield optimization. This paper investigates the most effective machine learning model for predicting rice crop yield in Rwanda, using agricultural inputs and practices. The study used secondary datasets from the National Institute of Statistics of Rwanda (NISR) for rice yield prediction. Eight supervised machine learning algorithms were used, including Linear Regression, Random Forest, Gradient Boosting, Support Vector Machine, Artificial Neural Network, eXtreme Gradient Boosting Tree, and AdaBoost. The models were evaluated based on their accuracy in predicting rice yields, with RMSE, MAE, and Relative Error as primary metrics. The feature importance analysis was also conducted to identify significant factors influencing yield predictions. The study’s findings revealed that the Adaptive Boosting Tree model outperformed the other machine learning models in predicting rice yield. This model achieved RMSE, MAE and Relative Error of 0.69, 0.46 and 12.4%, respectively, indicating a high level of predictive accuracy. The feature importance analysis further highlighted the key factors that contributed to rice yield predictions, with the quantity of inorganic fertilizer, degree of erosion, season, and seed type emerging as the most influential variables. The study demonstrates the effectiveness of machine learning models, particularly the Adaptive Boosting Tree, in improving rice yield predictions and highlighting the crucial role of agricultural inputs like fertilizer and seed type, in influencing crop yields. The output from this study will help the farmers and stakeholders to make data-driven decisions about resource use and crop management.

Rainfall Projections for the Brazilian Legal Amazon: An Artificial Neural Networks First Approach

Rainfall in the Brazilian Legal Amazon (BLA) is vital for climate and water resource management. This research uses spatial downscaling and validated rainfall data from the National Water and Sanitation Agency (ANA) to ensure accurate rain projections with artificial intelligence. To make an initial approach, Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) were employed to forecast rainfall from 2012 to 2020. The RNN model showed strong alignment with the observed patterns, accurately predicting rainfall seasonality. However, median comparisons revealed fair approximations with discrepancies. The Root Mean Square Error (RMSE) ranged from 6.7 mm to 11.2 mm, and the coefficient of determination (R2) was low in some series. Extensive analyses showed a low Wilmott agreement and high Mean Absolute Percentage Error (MAPE), highlighting limitations in projecting anomalies and days without rain. Despite challenges, this study lays a foundation for future advancements in climate modeling and water resource management in the BLA.