Artificial neural networks application to predict the ultimate tensile strength of X70 pipeline steels

The paper presents some results of the research connected with the development of new approach based on the artificial neural network (ANN) of predicting the ultimate tensile strength of the API X70 steels after thermomechanical treatment. The independent variables in the model are chemical compositions (carbon equivalent), based upon the International Institute of Welding equation (CE_IIW), the carbon equivalent, based upon the chemical portion of the Ito-Bessyo carbon equivalent equation (CE_Pcm), the sum of the niobium, vanadium and titanium concentrations (VTiNb), the sum of the niobium and vanadium concentrations (NbV), the sum of the chromium, molybdenum, nickel and copper concentrations (CrMoNiCu), Charpy impact energy at ?10 °C (CVN) and yield strength at 0.005 offset (YS). For purpose of constructing these models, 104 different data were gathered from the experimental results. The data used in the ANN model is arranged in a format of seven input parameters that cover the chemical compositions, yield stress and Charpy impact energy, and output parameter which is ultimate tensile strength. In this model, the training, validation and testing results in the ANN have shown strong potential for prediction of relations between chemical compositions and mechanical properties of API X70 steels.     

Using Artificial Neural Networks to Investigate the Influence of Temperature on Hot Extrusion of AZ61 Magnesium Alloy

The hot extrusion process of magnesium alloy involves many processing parameters, billet temperature is one of the parameters that directly affect the tensile strength of finished product. Hot extrusion experiments of involving rectangular tubes are conducted at selected billet temperatures of 320, 350, 380 and 400 °C. Artificial neural networks (ANN) analysis then is performed at increments of 10 °C each time between the temperature 320 and 400 °C. Consequently, the magnesium alloy product can be obtained at the optimum tensile strength, as well as the most suitable temperature range for billet heating during hot extrusion process. This study mainly explores the relationship between the billet temperature and product tensile strength of the hot extrusion of magnesium alloy, and obtains the optimum temperature range through ANN analysis, and analyzes the relationship between the temperature and the tensile strength of a rectangular tube for various extrusion speeds and extrusion ratios. Subsequently, experiments are performed to confirm the accuracy of the results by using ANN analysis at different extrusion speeds and extrusion ratios. Finally, observing the microstructure enables researchers to acquire the relationship between the sizes of the crystalline grain of the magnesium alloy product at the different formation temperature.     

An expert system based on artificial neural network for predicting the tensile behavior of tailor welded blanks

The forming behavior of tailor welded blanks (TWB) is influenced by thickness ratio, strength ratio, and weld conditions in a synergistic fashion. In most of the cases, these parameters deteriorate the forming behavior of TWB. It is necessary to predict suitable TWB conditions for achieving better-stamped product made of welded blanks. This is quite difficult and resource intensive, requiring lot of simulations or experiments to be performed under varied base material and weld conditions. Automotive sheet part designers will be greatly benefited if an ‘expert system’ is available that can deliver forming behavior of TWB for varied weld and blank conditions. This work primarily aims at developing an artificial neural network (ANN) model to predict the tensile behavior of welded blanks made of steel grade and aluminium alloy base materials. The important tensile characteristics of TWB are predicted within chosen range of varied blank and weld condition. Through out the work, PAM STAMP 2G^® finite element (FE) code is used to simulate the tensile behavior and to generate output data required for training the ANN. Predicted results from ANN model are compared and validated with FE simulation for two different intermediate TWB conditions. It is observed that the results obtained from ANN are encouraging with acceptable prediction errors. An expert system framework is proposed using the trained ANN for designing TWB conditions that will deliver better formed TWB products.     

Predictions of toughness and hardness by using chemical composition and tensile properties in microalloyed line pipe steels

Artificial neural networks with multilayer feed forward topology and back propagation algorithm containing two hidden layers are implemented to predict the effect of chemical composition and tensile properties on the both impact toughness and hardness of microalloyed API X70 line pipe steels. The chemical compositions in the forms of “carbon equivalent based on the International Institute of Welding equation (CE_IIW)”, “carbon equivalent based on the Ito-Bessyo equation (CE_Pcm)”, “the sum of niobium, vanadium and titanium concentrations (VTiNb)”, “the sum of niobium and vanadium concentrations (NbV)” and “the sum of chromium, molybdenum, nickel and copper concentrations (CrMoNiCu)”, as well as, tensile properties of “yield strength (YS)”, “ultimate tensile strength (UTS)” and “elongation (El)” are considered together as input parameters of networks while Vickers microhardness with 10 kgf applied load (HV10) and Charpy impact energy at ?10 °C (CVN ?10 °C) are assumed as the outputs of constructed models. For the purpose of constructing the models, 104 different measurements are performed and gathered data from examinations are randomly divided into training, testing and validating sets. Scatter plots and statistical criteria of “absolute fraction of variance (R²)”, and “mean relative error (MRE)” are used to evaluate the prediction performance and universality of the developed models. Based on analyses, the proposed models can be further used in practical applications and thermo-mechanical manufacturing processes of microalloyed steels.     

Multi-objective optimisation of pulsed Nd:YAG laser cutting process using integrated ANN–NSGAII model

The paper presents an integrated model of artificial neural networks (ANNs) and non-dominated sorting genetic algorithm (NSGAII) for prediction and optimization of quality characteristics during pulsed Nd:YAG laser cutting of aluminium alloy. A full factorial experiment has been conducted where cutting speed, pulse energy and pulse width are considered as controllable input parameters with surface roughness and material removal rate as output to generate the dataset for the model. In ANN–NSGAII model, back propagation ANN trained with Bayesian regularization algorithm is used for prediction and computation of fitness value during NSGAII optimization. NSGAII generates complete set of optimal solution with pareto-optimal front for outputs. Prediction accuracy of ANN module is indicated by around 1.5 % low mean absolute % error. Experimental validation of optimized output results less than 1 % error only. Characterization of the process parameters in pareto-optimal region has been explained in detail. Significance of controllable parameters of laser on outputs is also discussed.     

Predicting the compressive strength of ground granulated blast furnace slag concrete using artificial neural network

In this study, an artificial neural networks study was carried out to predict the compressive strength of ground granulated blast furnace slag concrete. A data set of a laboratory work, in which a total of 45 concretes were produced, was utilized in the ANNs study. The concrete mixture parameters were three different water–cement ratios (0.3, 0.4, and 0.5), three different cement dosages (350, 400, and 450 kg/m³) and four partial slag replacement ratios (20%, 40%, 60%, and 80%). Compressive strengths of moist cured specimens (22 ± 2 °C) were measured at 3, 7, 28, 90, and 360 days. ANN model is constructed, trained and tested using these data. The data used in the ANN model are arranged in a format of six input parameters that cover the cement, ground granulated blast furnace slag, water, hyperplasticizer, aggregate and age of samples and, an output parameter which is compressive strength of concrete. The results showed that ANN can be an alternative approach for the predicting the compressive strength of ground granulated blast furnace slag concrete using concrete ingredients as input parameters.     

Neural network nonlinear modeling for hydrogen production using anaerobic fermentation

The potential of utilizing artificial neural network (ANN) model approach for simulate and predict the hydrogen yield in batch model using Clostridium saccharoperbutylacetonicum N1-4 (ATCC 13564) was investigated. A unique architecture has been introduced in this research to mimic the inter-relationship between three input parameters initial substrate, initial medium pH and reaction temperature (37 °C, 6.0 ± 0.2, 10), respectively, to predict hydrogen yield. Sixty data records from the experiment have been utilized to develop the ANN model. The results showed that the proposed ANN model provided significant level of accuracy for prediction with maximum error (10 %). Furthermore, a comparative analysis with a traditional approach Box–Wilson design (BWD) has proved that the ANN model output significantly outperformed the BWD. ANN model overcomes the limitation of the BWD approach with respect to the number of records, which is merely considering limited length of stochastic pattern for hydrogen yield (15 records).     

An artificial neural network model to characterize porosity defects during solidification of A356 aluminum alloy

In this study, a data-driven multilayer perceptron-based neural network model has been developed to predict the percentage of total porosity and mechanical properties, namely yield strength, ultimate tensile strength and percentage of elongation during the solidification of A356 aluminum alloy. Some of the important processing parameters such as cooling rate, solidus velocity, thermal gradient and initial hydrogen content have been considered as inputs to this model. The network training architecture has been optimized using the gradient-based Broyden–Fletcher–Goldfarb–Shanno training algorithm to minimize the network training error within few training cycles. Parametric sensitivity analysis is carried out to characterize the influence of processing parameters (inputs) on the behavior of porosity formation and simultaneously, the tensile properties of A356 alloy castings. It has been observed that initial hydrogen content in the melt and cooling rate has significant influence on the porosity formation and eventually on the tensile properties of the cast product. There has been an excellent agreement between artificial neural network predictions and the target (measured) values of porosity and the tensile properties as depicted by the regression fit between these values.     

ANN model to predict the effects of composition and heat treatment parameters on transformation start temperature of microalloyed steels

The paper presents some results of the research connected with the development of new approach based on the artificial neural network (ANN) of predicting the transformation start temperature of the phase constituents occurring in five steels after continuous cooling. The independent variables in the model are chemical compositions (C, Mn, Nb, Mo, Ti, N, Cu, P, S, Si, Al, V), austenitizing temperature, initial austenite grain size and cooling rate over the temperature range of the occurrence of phase transformations. For purpose of constructing these models, 138 different experimental data were gathered from the literature. The data used in the ANN model are arranged in a format of fourteen input parameters that cover the chemical compositions, initial austenite grain size and cooling rate, and output parameter which is transformation start temperature. In this model, the training and testing results in the ANN have shown strong potential for prediction of effects of chemical compositions and heat treatments on phase transformation of microalloyed steels.     

Artificial neural network to predict the effects of coating parameters on layer thickness of chromium carbonitride coating on pre-nitrided steels

Chromium carbonitride coatings were formed on plain carbon and alloy steels by pre-nitrocarburizing, followed by thermoreactive deposition and diffusion in a salt bath below 700 °C. In the present study, an artificial neural network-based model (ANNs) was developed to predict the layer thickness of pre-nitrided steels. Seventeen parameters affecting the layer thickness were considered as inputs, including the pre-nitriding time, salt bath compositions ratio, salt bath aging time, ferrochromium particle size, ferrochromium weight percent, salt bath temperature, coating time, and different chemical compositions of steels. The network was then trained to predict the layer thickness amounts as outputs. A 2-feed-forward back-propagation network was developed and trained using experimental data form literatures. Five steels were investigated. The effects of coating parameters on the layer thickness of steels were modeled by ANNs as well. The predicted values are in very good agreement with the measured ones indicating that the developed model is very accurate and has the great ability for predicting the layer thickness.     

Analytical modeling of tensile strength of functionally graded steels

In the present study, the tensile strength of ferritic and austenitic functionally graded steel produced by electroslag remelting has been modeled by artificial neural networks. Functionally graded steel containing graded layers of ferrite and austenite may be fabricated via diffusion of alloying elements during remelting stage. Vickers microhardness profile of the specimen has been obtained experimentally and modeled with artificial neural networks. To build the model for graded ferritic and austenitic steels, training, testing and validation using respectively 174 and 120 experimental data were conducted. According to the input parameters, in the neural networks model, the Vickers microhardness of each layer was predicted. A good-fit equation that correlates the Vickers microhardness of each layer to its corresponding chemical composition was achieved by the optimized network for both ferritic- and austenitic-graded steels. Afterward, the Vickers microhardness of each layer in functionally graded steels was related to the yield stress of the corresponding layer and by assuming Holloman relation for stress–strain curve of each layer, they were acquired. Finally, by applying the rule of mixtures, tensile strength of functionally graded steels configuration was found through a numerical method. The obtained results from the proposed model are in good agreement with those acquired from the experiments.     

Microstructure and mechanical properties of micro tensile specimens made of CuAl10Ni5Fe4 produced by micro casting

In this paper the results of micro tensile tests of specimens produced by vacuum pressure casting (VPC) and centrifugal casting (CC) with varied mould temperatures are described. The microstructure of VPC-micro specimens is clearly different from CC-micro specimens. Due to the large tolerance range of the chemical composition of the aluminum bronze CuAl10Ni5Fe4 and process instabilities the microstructure of batches of VPC-micro specimens can be very fine grained as well as very coarse grained even if the process parameters were nominally the same. As a consequence the mechanical properties of micro specimen show a wide range, e.g. the ultimate tensile strength (VPC, mould temperature T _m = 1,000°C) is between 670 and 1,000 MPa. Within one batch (ca. 15 specimens) the scatter of microstructure and mechanical properties is clearly smaller. If the chemical composition of the material is nearly constant for a larger number of batches, then the production process of vacuum pressure casting leads reproducibly to nearly identical microstructure and mechanical properties.     

Application of ANFIS for modeling of layer thickness of chromium carbonitride coating

In the present work, layer thickness of duplex coating made from thermo-reactive deposition and diffusion has been predicted by Adaptive network-based fuzzy inference systems (ANFIS). A duplex surface treatment on five steels has been developed involving nitrocarburizing and followed by chromium thermo-reactive deposition (TRD) techniques. The TRD process was performed in molten salt bath at 550, 625 and 700 °C for 1–30 h. The process formed a thickness up to 9.5 μm of chromium carbonitride coatings on a hardened diffusion zone. A model based on ANFIS for predicting the layer thickness of duplex coating of the specimens has been presented. To build the model, training and testing using experimental results from 84 specimens were conducted. The data used as inputs in ANFIS models are arranged in a format of twelve parameters that cover the chemical composition (C, Mn, Si, Cr, Mo, V, W), the pre-nitriding time, ferro-chromium particle size, ferro-chromium weight percent, salt bath temperature and coating time. According to these input parameters, in the Adaptive network-based fuzzy inference system models, the layer thickness of duplex coating of each specimen was predicted. The training and testing results in ANFIS models have shown a strong potential for predicting the layer thickness of duplex coating.     

Process optimization of SnCuNi soldering material using artificial parametric design

The European Union has implemented the directive restriction of hazardous substances (RoHS) prohibiting the uses of tin-lead solder. SAC305 (Sn96.5/Ag3.0/Cu0.5) has come into widespread use as a candidate soldering material in the electronics manufacturing industry. Nevertheless, the price of silver has increased dramatically in recent years. This study evaluates the feasibility of replacing the commonly used SAC305 with low cost SnCuNi (Sn99.25/Cu0.7/Ni0.05/Ge; SCN) solder alloy in wave soldering for high layer count printed circuit board. However, the melting temperature of SCN alloy is 227, 10  \(^{\circ }\text{ C}\) higher than SAC305. The objective of this research is to investigate manufacturing issues and propose an optimal process. Process parameters such as soldering temperature and dwell time are determined to achieve the desired quality levels. Multiple quality characteristics, namely assembly yield and solder joint pull strength, are considered. Thus, this study compares two approaches, integration of principal component analysis/grey relational analysis and artificial neural networks (ANN) combined with genetic algorithms (GA), to resolve the problems of multiple quality characteristics. The results of verification test shows that samples prepared with the process scenario suggested by the ANN combined with GA are superior. The process scenario with maximum desirability value is 268.64  \(^{\circ }\text{ C}\) soldering temperature and 7.42 s dwell time, indicating the recommended manufacturing process.     

Predicting roadheader performance by using artificial neural network

With growing use of roadheaders in the world and its significant role in the successful accomplishment of a tunneling project, it is a necessity to accurately predict performance of this machine in different ground conditions. On the other hand, the existence of some shortcomings in the prediction models has made it necessary to perform more research on the development of the new models. This paper makes an attempt to model the rate of roadheader performance based on the geotechnical and geological site conditions. For achieving the aim, an artificial neural network (ANN), a powerful tool for modeling and recognizing the sophisticated structures involved in data, is employed to model the relationship between the roadheader performance and the parameters influencing the tunneling operations with a high correlation. The database used in modeling is compiled from laboratory studies conducted at Azad University at Science and Research Branch, Tehran, Iran. A model with architecture 4-10-1 trained by back-propagation algorithm is found to be optimum. A multiple variable regression (MVR) analysis is also applied to compare performance of the neural network. The results demonstrate that predictive capability of the ANN model is better than that of the MVR model. It is concluded that roadheader performance could be accurately predicted as a function of unconfined compressive strength, Brazilian tensile strength, rock quality designation, and alpha angle R ² = 0.987. Sensitivity analysis reveals that the most effective parameter on roadheader performance is the unconfined compressive strength.     

Precipitate/austenite equilibrium in a titanium-niobium ultra low carbon steel

The variation in the equilibrium compositions of austenite and titanium-niobium-carbonitride precipitates in Fe-Ti-Nb-C-N-Mn-Si-Mo is considered over a range of temperatures between 1050 and 1500K and for an alloy composition which is typical of ultra-low carbon titanium-niobium microalloyed steels. The Kohler temperature dependent-subregular solution model is used to describe the austenite phase. The (Ti,Nb)(C,N)_x precipitate phase is described by the Hillert-Staffansson sublattice model for a four-component solution, with components mixed in pairs. The equilibrium between the two phases is explored while allowing for the nonstoichiometry of the precipitate phase. The results are shown to be in good agreement with experimental data on ultra-low carbon microalloyed steels.     

Estimation of FeB layer’s yield strength by comparison of finite element modeling with experimental data

In this study, low alloy steel substrates were borided by pack boriding process, for 2, 4 and 6 h at 900 °C. Microstructural observations were conducted by using SEM. The structural composition of layers consists of boron rich phase (FeB) and iron rich phase (Fe₂B). First, experimental indentation studies were carried out to determine the load–unload curves of FeB layers at different peak loads. Important parameters such as hardness and Young’s modulus of FeB layers, and contact area were obtained from experimental indentation test sample data. After the mechanical characterization of samples, finite element modeling was applied to simulate the mechanical response of FeB layer on low alloy steel substrate by using ABAQUS software package program. The unique contribution of this study different from previous methods is the estimation of the yield strength of FeB layer by combining the experimental indentation works and finite element modeling (FEM).     

Optimization of friction stir welding process using NSGA-II and DEMO

In welding processes, the selection of optimal process parameter settings is very important to achieve best weld qualities. In this work, neuro-multi-objective evolutionary algorithms (EAs) are proposed to optimize the process parameters in friction stir welding process. Artificial neural network (ANN) models are developed for the simulation of the correlation between process parameters and mechanical properties of the weld using back-propagation algorithm. The weld qualities of the weld joint, such as ultimate tensile strength, yield stress, elongation, bending angle and hardness of the nugget zone, are considered. In order to optimize those quality characteristics, two multi-objective EAs that are non-dominated sorting genetic algorithm II and differential evolution for multi-objective are coupled with the developed ANN models. In the end, multi-criteria decision-making method which is technique for order preference by similarity to the ideal solution is applied on the Pareto front to extract the best solutions. Comparisons are conducted between results obtained from the proposed techniques, and confirmation experiments are performed to verify the simulated results.