جستجوی مقالات مرتبط با کلیدواژه « شبکه عصبی مصنوعی » در نشریات گروه « فنی و مهندسی »

This research aims to simultaneously use geochemical modeling and geological parameters for gold grade estimation to identify promising zones of epithermal gold mineralization in the Zailik region, northwest of Iran. For this purpose, the employed geological evidence includes lithology and alterations like silicification, iron oxides, phyllic, and propylitic. For  geochemical modeling two methods were utulized: 1) artificial neural network (ANN),  2) integrating ANN with the Firefly algorithm. Geological evidence after quantification, along with the estimated amounts of gold in artificial intelligence methods, was entered into the hierarchical system in Expert Choice software for weighting. In this method, the weighting and determination of the degree of relative importance of geological parameters were attempted after consulting geological and exploration experts. Subsequently, artificial intelligence methods were also compared with each other using quantitative criteria such as the coefficient of determination and the root mean square error function. The results showed that the combined method of artificial neural networks with the Firefly algorithm provides better results due to the higher coefficient of determination (R2=0.643) and lower error function (RMSE=0.754). Therefore, it has a higher degree of importance to identify promising areas for mineralization. Finally, all the above parameters were combined with each other in the Arc GIS software using the fuzzy overlay method, and the optimal exploration targets were detected in the north and northeast of the region, enabling to continue the exploration targets along the root of gold mineralization in the neighboring areas according to the introduced model.

Increasing the depth of microchannels on metallic bipolar plates (BPPs) in PEM fuel cells leads to an increase in the efficiency. In this research, the stamping process has been applied for manufacturing of the BPPs made of commercially pure titanium with a direct parallel flow field. The effect of process parameters including die clearance, forming speed, and sheet/die friction coefficient on the filling rate and thinning of the BPPs was investigated. The required tests were designed via the response surface method (RSM), implemented by a validated finite elements (FE) model, and the desired outputs were extracted. Then, a quadratic equation was presented for predicting the filling rate based on the input parameters using the regression method. In the following, using the artificial bee colony algorithm, the coefficients of the mentioned equation were enhanced and its error was decreased almost by 53%. Finally, an artificial neural network (ANN) was used to predict the filling rate. The results demonstrated that the proposed ANN model is very effective and approximates the filling rate of the microchannel with high accuracy.

Short-term runoff forecasting is of particular importance due to its direct relationship with how managers interact with life risks caused by floods. In this research, by using artificial neural networks, simulation of rainfall-runoff process has been done on a daily basis in the Nasa Bam watershed. In order to predict the future process of using the water resources of the mentioned plain, different combinations of rainfall and temperature data and discharge and discharge difference of two consecutive days were used. The number of hidden layer neurons in the neural network varied between 2 and 10 neurons. The statistical criteria of root mean square error RMSE, mean absolute value of error MAE and correlation coefficient R were used to evaluate and compare the performance of neural networks in runoff forecasting. The results showed that by having 2 inputs and feedforward neural network or 1 input and newrbe network, the best performance was achieved and the rainfall-runoff process was predicted with higher accuracy.

Shear force is one of the most important influencing forces of structural beams and especially unreinforced deep concrete beams that cause sudden failure and sudden collapse of the structure and for this reason, it is of interest to designers and implementers of structures. The study of the conducted research shows that the proposed methods including the use of artificial neural network cannot predict Provide an accurate description of the behavior of deep reinforced concrete beams against shear force. This research, while specifying the cause of the error of the artificial neural network in estimating the shear strength of deep reinforced concrete beams, presents a solution to achieve accurate results by optimizing the artificial neural network by the meta-heuristic algorithm of particle swarm. For this purpose, 309 laboratory samples of deep concrete beams were collected from the research literature and the particle swarm algorithm was used to optimize the weights in the artificial neural network using MATLAB software. The comparison of the results of this method with the results of these letters and other existing methods that were examined in this research showed that the use of the combination of particle swarm algorithm and optimized artificial neural network to estimate the shear strength of unreinforced deep reinforced concrete beams provides more accurate answers.

The ground impedance measuring elements of distance relays operating on double-circuit transmission lines would reflect malfunctions when facing high-resistance as well as irrelevant-circuit SLG faults. In the first case, the element would undergo the under-reach (not detecting the faults), and in the second one undergo the over-reach (unplanned fault detection). With simultaneous online estimation of the fault distance and resistance through developing an adaptive σιγ-layer artificial neural network, this paper presents a ground element for the ground distance relaying of double-circuit transmission lines. The proposed element is fully immune against high-fault resistances and by providing a logical mechanism via estimated values, it would be also fully compensated for the over-reach arising from irrelevant-circuit faults. The implementation of the proposed approach is practically feasible due to its simple structure and operating principle. Further, it employs the measured data from one end of the line, hence operating reliably in the absence of a communication channel between the line ends. Performance of the proposed approach would be investigated via simulations in PSCAD/EMTDC and the results reveal that the SLG faults, regardless of the fault resistance values, are quickly detected and reliably tripped by the proposed element, without facing the over-reach in case of irrelevant-circuit faults.

In this paper, a new and efficient performance-based method is introduced for the optimal design of the tuned mass damper (TMD) to control the response of nonlinear structures. In this method, the FEMA-P58 probabilistic evaluation framework is used in the design and evaluation of TMD performance in order to reduce the repair cost and time caused by structural and non-structural damages. For this purpose, an innovative objective function has been defined based on structural responses resulting from different earthquakes, which is compatible with the probability of the repair cost and time exceeding a certain value in the FEMA-P58 method. Uncertainties due to earthquake records are directly considered in the introduced objective function. The genetic algorithm is used for the optimal design of TMD based on the proposed objective function. Also, due to the fact that considering the uncertainties in the input earthquake records leads to increasing the calculation time, therefore the artificial neural network technique is used as a fast estimator of the nonlinear dynamic response of the structure to reduce the calculation time. The probabilistic evaluation of the performance of the structure equipped with the proposed TMD shows the efficiency and effectiveness of the proposed design procedure in reducing the expected repair cost and time. So that the expected repair cost and time in the structure equipped with the proposed TMD under the design earthquake have been reduced by about 29% compared to the uncontrolled structure.

For the three-dimensional modeling of the S01 vein from the Zailik exploratory area, the sampled data of the trenches and boreholes of this vein were used, and the gold grade was estimated using ANN, ANN-PSO, and ANN-FFA methods. To check the accuracy of the modeling, it was compared with the estimate of grade using the ordinary geostatistical kriging method, as well as the geological evidence of the area, such as lithology and alteration.

Lozenge rug chaleshtor is in terms of artistic and aesthetic aspects, it is considered one of the consumer and capital goods that have international fame. From the past until now, this native art has been traditionally produced with the same design, pattern, and colour without paying attention to the tastes of its audience, which can be one of the reasons for the failure of this original art in the global rug markets shortly. Therefore, knowledge and awareness of consumers' taste, which is considered as one of the steps before production, and using modern science, can help to produce according to the taste of the audience and, accordingly, to be more successful in terms of sales and providing high export statistics. To achieve this, in this article, sub-branches of artificial intelligence were used to achieve the taste of Lozenge rug chaleshtor audiences. Three algorithms more related to the subject, 1. artificial neural networks, 2. decision tree learning and 3. support vector machine are compared and finally, the most suitable algorithm for this subject is the artificial neural networkss for receiving the taste of the audience of the Lozenge rug chaleshtor and providing suitable patterns in the field of design. , role, colour, dimensions, texture and price according to the audience's taste, and it was tried to answer the question of whether that artificial neural networkss can introduce the principles of the audience's taste of lozenge rug chaleshtor. in structural fields. For this purpose, primary data in the field of design and role, colour, raw materials, texture, dyeing, dimensions and price were collected through the contact questionnaire and then, using an artificial neural networkss, an algorithm was designed, the results showed that traditional lozenge designs with half motifs curved and quiet, with bright, limited, soft colours, natural dyeing, dimensions under 6 meters and preferably square with a price per meter of up to 10 million tomans is the final taste of the audience of this type of rugs.Keywords: Taste,Lozenge Rug, Chaleshtor, Artificial Neural Networks

As steel fibers are important reinforcement materials in concrete, in this study, the behavior of hook-shaped steel fibers from concrete is predicted through the use of artificial neural networks. In the absence of comprehensive laboratory data, data obtained from finite element analysis was used for modeling. The simulations are carried out using ABAQUS software's finite element method in 3D. Using the concept of the transition zone of the interface, whose parameters were obtained by inverse finite element analysis and experimental tests conducted on a sample of fibers, this model has been developed to simulate the interaction between fibers and concrete. On the basis of the results of the numerical model validated against the experimental results, the effective parameters of the fibers were extracted, and a neural network was then constructed based on the results. A multilayer forward perceptron artificial neural network and back-propagation training algorithm are used to predict pull-out force, with Marquardt-Lonberg optimization applied. The results demonstrate that the neural network model presented in this research is an effective method for predicting the pull-out force of fibers from concrete, in part because it allows the use of more variables in modeling, as well as delivering more accurate results.

Although rainfall-runoff modelling is not considered a big challenge recently, it is still one of the challenging issues for researchers in the basin or sub-basins without statistical data. One of the new methods in this field is the use of remote sensing techniques and the use of machine learning (artificial intelligence). In this research, to calculate the runoff in the basins without data, two basins were used including the basins for the Samian and the Amoghein hydrometric stations in Ardabil plain. The first station was chosen as the outlet of the Ardabil Plain basin for model training and calibration and the second station was used as the basin without data for verification and testing. Modelling was done using 9 input parameters including air pressure, vegetation cover index (low and high cover), soil temperature, ground surface temperature, soil water volume, runoff, evaporation potential and precipitation. Also, a parameter related to the observational data of the stations was used as an output. Modelling was done using four models of NARX, ANN-ACO, ANN-GA, ANN-PSO and the accuracy of the models were evaluated using MSE, R2, RMSE, NSE and MAE. The results showed that the NARX model is clearly superior to other models with an accuracy of 0.001, 0.86, 0.039, 0.855 and 0.015 respective to the above-mentioned measures. Remote sensing methods combined with artificial intelligence can respond to hydrologists’ challenges in data-scarce and ungauged basins around the world due to their ability to provide high-precision results.

Thermal conductivity of MWCNT(40%)-CuO(30%)-TiO2(30%)/Water nanofluid in different volume fractions and temperatures is modeled and analyzed by artificial neural network method. The type of artificial neural network is MLP. 48 experimental data series are used, 70%, 15%, and 15% are used for training, validation, and testing, respectively. The optimal neural structure has two hidden layers, in which there are 4 neurons in the first layer and 5 neurons in the second layer, with the transfer functions of logsig and tansig, respectively. Neural network training is done with Levenberg-Marquardt (ML) algorithm. The values of regression coefficient, R, and mean squared error, MSE, for the optimal structure are obtained as 0.9995753 and 2.8734E-06, respectively. The correlation equation is also presented to predict the thermal conductivity of nanofluid. The comparison between the correlation equation and the artificial neural network shows the superiority of the artificial neural network. MOD values for artificial neural network are also in the range of -3% to +7%.

This research highlights the potential of computer vision and machine learning algorithms to enhance weed control systems and decrease herbicide use in agriculture. The development of an efficient and cost-effective smart sprayer system has the capacity to not only benefit farmers financially, but also mitigate the environmental impact of herbicide application. Further investigation could explore the feasibility and practicality of implementing such systems on a larger scale in diverse crop fields. The recognition of weeds and crops based on their appearance characteristics is crucial for effective weed control systems, and the proposed BOVW algorithm demonstrated a high level of accuracy, reliability, and sensitivity in distinguishing between common weed species and sugar beet crops. The smart sprayer system, incorporating the BOVW algorithm, exhibited a high level of precision in identifying the product from weed species online. Notably, the provided sprayer system significantly reduced herbicide consumption by 78.93% and 69.38% in the best and worst mode of spraying, respectively. The findings suggest that the variable rate mode utilizing the BOVW detection algorithm represents the optimal mode of operation for the smart sprayer system.

Drilling rate of penetration is one of the most important parameters in optimization and cost reduction of drilling operation. In order to predict the rate of penetration with high precision, artificial intelligence methods are used in this paper. Drilling data of reservoir zone in one of the Maroun oilfield wells consisting of drilling operation parameters, drilling fluids properties, and rock mechanics properties, extracted from mud logging and petrophysics logs, is collected. 8 input parameters such as weight on bit, bit rpm, pump pressure, fluid specific gravity, fluid plastic viscosity, fluid yield point, shale volume, porosity, unconfined compressive strength, Young modulus and toughness are chosen as the most effective parameters on rate of penetration. Using these parameters, two artificial neural network models MLP-PSO and MLP-BP were developed to predict rate of penetration and their performances are evaluated. This study shows that artificial neural network trained by particle swarm optimization have better performance than other models in predicting rate of penetration. The performance of this model is compared with empirical models and the results show the superior performance of artificial neural network models over these models in predicting rate of penetration. Besides requiring more calculation time, these empirical models show lower precision in predicting drilling rate of penetration.

The role of cement fineness in the process of hydration and development of compressive strength in the early ages of cement-based materials is irrefutable and it requires that its effect be investigated by predicting models. Therefore, an extensive study including 640 cement composition (1920 cement mortar specimens) from a cement factory with different percentages of raw materials feeding to the cement kiln including SiO2, Al2O3, Fe2O3, CaO, MgO, SO3, K2O, and Na2O were used to predict the 7-day compressive strength of cement mortar by artificial neural network (ANN). To investigate the effect of cement fineness, two models have been developed in two states of with and without fineness effect. Results confirmed the significant role of cement fineness as an input parameter in the performance of predicting model. The findings of this research can be used in cement production facilities in order to reduce the laboratory costs.

In this research, the optimization of the artificial neural network (ANN) capability for predecting the tensile strength and elongation of friction stir welded Al-5083 (FS-welded Al-5083) was carried out. The effective parameters of ANN, such as the number of layers, number of neurons in hidden layers, transfer function between layers, the learning algorithm and etc. were investigated and the efficient neural network was determined to predict the tensile properties of FS-welded Al-5083. The investigations revealed that the perceptron neural network with two hidden layers and 17 neurons numbers, Lunberg-Marquardt training algorithm and Logsig transfer function for the intermediate layers and Tansig transformation function for the output layer is the most optimized neural network for the prediction. The optimized network has an optimal structure based on the minimum value of the mean square error of 0.05, the maximum total correlation coefficient of 0.93 and the line regression with an angle of 45 degrees between the actual and estimated values. Therefore, this network has a good performance for training, generalizing and estimating of tensile strength and elongation of FS-welded Al-5083.

Designing a model based on blockchain technology to strengthen cyber security in the banking industry is one of the new methods studied in the banking industry to strengthen cyber security. Accordingly, this study seeks to achieve the goal of evaluating a model based on blockchain technology to strengthen cyber security in the banking industry based on artificial neural networks. This model is based on a conceptual model used in an MLP neural network simulation that simulates a blockchain-like process. Also, the neural networks created in the block chain have a strong connection and the possibility of breaking them is low. The data became closer to normal distribution after learning, indicating that blockchain technology will be able to provide cyber security. The level of correlation and efficiency presented was also reported and the findings of the study showed that the efficiency related to blockchain technologies after learning reached the level of 770.57 units, which shows that using the MLP method to learn the process of blockchain technology can be Lead to greater efficiency for cyber security. Also, the value of variance is equal to 27.77 and the mean value of computational values is equal to 0.35 and the value of correlation is equal to 0.99.

Drought can mainly affect agriculture, especially rainfed agriculture due to its high dependence on precipitation, thereby compromising food security and social protection. In this study, the correlation between SPI and SPEI drought indices with rainfed wheat yield of 5 filds in Saqqez city during the period of 2001-2020 was investigated with neural network, random forest and support vector regression. TRMM precipitation and CRU evapotranspiration were used to calculate SPI and SPEI drought indices. The AquaCrop model was calibrated with observational data and then filds performance was simulated with the AquaCrop model for the period 2001-2020. The average yield of the filds was evaluated with the average yield of the entire Saqqez city, and the results showed that the data simulated with the model had a good correlation (R2=0.90) with the average rainfed wheat yield of Saqqez city. The results of evaluating the relationship between SPI and SPEI indices with rainfed wheat yield showed that the neural network and random forest method with a significant probability of 95% (P-value=0.0) and an explanatory coefficient of more than 0.70% and a high value of Nash Sutcliffe index And a small amount of underestimation has had a good estimate of the rainfed wheat yield and there is a significant relationship between SPI and SPEI drought indices with the rainfed wheat yield in the study area.