Feed Forward Back Propagation Algorithm Research Articles

Prediction of hopper discharge rate using combined discrete element method and artificial neural network

An Artificial Neural Network (ANN) was developed to predict the mass discharge rate from conical hoppers. By employing Discrete Element Method (DEM), numerically simulated flow rate data from different internal angles (20°–80°) hoppers were used to train the model. Multi-component particle systems (binary and ternary) were simulated and mass discharge rate was estimated by varying different parameters such as hopper internal angle, bulk density, mean diameter, coefficient of friction (particle-particle and particle-wall) and coefficient of restitution (particle-particle and particle-wall). The training of ANN was accomplished by feed forward back propagation algorithm. For validation of ANN model, the authors carried out 22 experimental tests on different mixtures (having different mean diameter) of spherical glass beads from different angle conical hoppers (60° and 80°). It was found that mass discharge rate predicted by the developed neural network model is in a good agreement with the experimental discharge rate. Percentage error predicted by ANN model was less than ±13%. Furthermore, the developed ANN model was also compared with existing correlations and showed a good agreement.

Growth profiling, kinetics and substrate utilization of low-cost dairy waste for production of β-cryptoxanthin by Kocuria marina DAGII.

The dairy industry produces enormous amount of cheese whey containing the major milk nutrients, but this remains unutilized all over the globe. The present study investigates the production of β-cryptoxanthin (β-CRX) by Kocuria marina DAGII using cheese whey as substrate. Response surface methodology (RSM) and an artificial neural network (ANN) approach were implemented to obtain the maximum β-CRX yield. Significant factors, i.e. yeast extract, peptone, cheese whey and initial pH, were the input variables in both the optimizing studies, and β-CRX yield and biomass were taken as output variables. The ANN topology of 4-9-2 was found to be optimum when trained with a feed-forward back-propagation algorithm. Experimental values of β-CRX yield (17.14 mg l−1) and biomass (5.35 g l−1) were compared and ANN predicted values (16.99 mg l−1 and 5.33 g l−1, respectively) were found to be more accurate compared with RSM predicted values (16.95 mg l−1 and 5.23 g l−1, respectively). Detailed kinetic analysis of cellular growth, substrate consumption and product formation revealed that growth inhibition took place at substrate concentrations higher than 12% (v/v) of cheese whey. The Han and Levenspiel model was the best fitted substrate inhibition model that described the cell growth in cheese whey with an R2 and MSE of 0.9982% and 0.00477%, respectively. The potential importance of this study lies in the development, optimization and modelling of a suitable cheese whey supplemented medium for increased β-CRX production.

ANN-based modelling and prediction of daily global solar irradiation using commonly measured meteorological parameters

The aim of this work is to develop an artificial neural network (ANN) based model for accurately predicting the daily global solar irradiation in the city of Fez. The potential of the developed model is verified and appraised through the local collected database for the period 2009-2015 from the radiometric station of the Faculty of Sciences and Technology of Fez. The obtained model is MLP with feed forward back-propagation algorithm containing three input parameters and a single hidden layer with nine neurons. Coefficient of determination R2, the mean absolute percentage error MAPE and the relative root mean square error RRMSE are respectively equal to 97.16%, 21.77% and 18.79%.

An artificial neural network approach to determine the rheological behavior of pickering-type diesel-in-water emulsion prepared with the use of β-cyclodextrin

With the use of β-cyclodextrin (β-CD), Pickering-type diesel-in-water emulsions were prepared based on the inclusion complex formed between diesel and β-CD which acted as an emulsifier. By using the artificial neural network (ANN), the rheological behavior of the emulsions was characterized using three input variables: diesel-to-water ratio, β-CD concentration, and shear rate and one-output variable as shear stress. Gradient descent (GD), conjugate gradient (CG), and quasi Newton (QN) were used as three different methods in the feed-forward back-propagation algorithm for network training. Hyperbolic tangent sigmoid and pure linear were the transfer functions used for transforming information between input and output through one hidden layer containing ten neurons. By dividing the experimental data into three sets of training, validation, and testing, the QN method in predicting shear stress was found to have performed better than the other two network learning techniques (R2=0.994 and MSE=0.006).

Prediction of hot deformation behavior of high phosphorus steel using artificial neural network

To predict the hot deformation behavior of high phosphorus steel, the hot compression experiments were performed with the help of thermo-mechanical simulator Gleeble® 3800 in the temperatures ranging from 750 °C to 1050 °C and strain rates of 0.001 s-1, 0.01 s-1, 0.1 s-1, 0.5 s-1, 1.0 s-1 and 10 s-1. The experimental stress-strain data are employed to develop artificial neural network (ANN) model and their predictability. Using different combination of temperature, strain and strain rate as a input parameter and obtained experimental stress as a target, a multi-layer ANN model based on feed-forward back-propagation algorithm is trained, to predict the flow stress for a given processing condition. The relative error between predicted and experimental stress are in the range of ±3.5%, whereas the correlation coefficient (R2) of training and testing data are 0.99986 and 0.99999 respectively. This shows that a well-trained ANN model has excellent capability to predict the hot deformation behavior of materials. Comparative study shows quite good agreement of predicted and experimental values.

Predicting residual compressive strength of self-compacted concreteunder various temperatures and relative humidity conditionsby artificial neural networks

Artificial neural network models can be successfully used to simulate the complex behavior of many problems in civil engineering. As compared to conventional computational methods, this popular modeling technique is powerful when the relationship between system parameters is intrinsically nonlinear, or cannot be explicitly identified, as in the case of concrete behavior. In this investigation, an artificial neural network model was developed to assess the residual compressive strength of self-compacted concrete at elevated temperatures (20-900oC) and various relative humidity conditions (28-99%). A total of 332 experimental datasets, collected from available literature, were used for model calibration and verification. Data used in model development incorporated concrete ingredients, filler and fiber types, and environmental conditions. Based on the feed-forward back propagation algorithm, systematic analyses were performed to improve the accuracy of prediction and determine the most appropriate network topology. Training, testing, and validation results indicated that residual compressive strength of selfcompacted concrete, exposed to high temperatures and relative humidity levels, could be estimated precisely with the suggested model. As illustrated by statistical indices, the reliability between experimental and predicted results was excellent. With new ingredients and different environmental conditions, the proposed model is an efficient approach to estimate the residual compressive strength of self-compacted concrete as a substitute for sophisticated laboratory procedures.

Predicting Force in Single Point Incremental Forming by Using Artificial Neural Network

In this study, an artificial neural network was used to predict the minimum force required to single point incremental forming (SPIF) of thin sheets of Aluminium AA3003-O and calamine brass Cu67Zn33 alloy. Accordingly, the parameters for processing, i.e., step depth, the feed rate of the tool, spindle speed, wall angle, thickness of metal sheets and type of material were selected as input and the minimum vertical force component was selected as the model output. To train the model, a Multilayer perceptron neural network structure and feed-forward backpropagation algorithm have been employed. After testing many different artificial neural network (ANN) architectures, an optimal structure of the model i.e. 6-14-1 was obtained. The results, with a correlation relation between experiments to predicted force,-0.215 mean absolute error, show a very good agreement.

Automated Classification Method for Early Diagnosis of Alopecia Using Machine Learning

Artificial neural networks are being used in many fields for different purposes. Medical diagnosis is one of the major purposes. In the field of medical, classification plays an important role as the main aim of the doctor is to classify whether a person is suffering from the disease or not. The objective of this paper is to evaluate neural network for the detection of alopecia in human beings and to find the accuracy. With the help of proposed model, the clinical experts will be able to get a second opinion that will help them to take proper decisions for diagnosing the presence of this disease in patients. This second opinion is crucial due to many factors while doing disease identification. These factors include increased population, environmental pollution, growing demands for proper medication, and less availability of medical experts to cope up with this increasing demand. Also, the dynamic nature of disease symptoms plays an important role in correct diagnosis of a certain disease. The proposed system uses a feedforward artificial neural network and backpropagation algorithm to classify patients with alopecia and without alopecia. The evaluation of the proposed system is done with the help of performance plot as well as regression plot. Experimental results show that the accuracy of proposed system is 91% which is reliable enough for a clinical expert to make his decisions.

Duration estimation model for bridge construction projects in Ghana

Purpose A major concern for construction professionals at the rural road agency in Ghana is the problem of fixing contract duration for bridge construction projects in rural areas. The purpose of the study was to develop a tool for construction professionals to forecast duration for bridge projects. Design/methodology/approach In all, 100 questionnaires were distributed to professionals at the Department of Feeder Roads to ascertain their views on the work items in a bill of quantities (BOQ) that impact significantly on the duration of bridge construction projects. Historical data for 30 completed bridge projects were also collected from the same Department. The data collected were executed work items in BOQ and actual durations used in completing the works. The qualitative data were analysed using the relative importance index and the quantitative data, processed and analysed using both the stepwise regression method and artificial neural network (ANN) technique. Findings The identified predictors, namely, in-situ concrete, weight of prefabricated steel components, gravel sub-base and haulage of aggregates, used as independent variables resulted in the development of a regression model with a mean absolute percentage error (MAPE) of 25 per cent and an ANN model with a feed forward back propagation algorithm with an MAPE of 26 per cent at the validation stage. The study has shown that both regression and ANN models are appropriate for predicting the duration of a new bridge construction project. Research limitations/implications The predictors used in the developed models are limited to work items in BOQs only of completed bridge construction projects as well as the small sample size. Practical implications The study has developed a working tool for practitioners at the agency to forecast contract duration for bridge projects prior to its commencement. Originality value The study has quantified the relationship between the work items in BOQs and the duration of bridge construction projects using the stepwise regression method and the ANN techniques.

Artificial neural network to predict the degraded mechanical properties of metallic materials due to the presence of hydrogen

Machine learning models were introduced to develop a relationship between the elemental composition and degraded mechanical properties in metallic materials due to the presence of hydrogen. Single layer and multilayer feed forward back propagation algorithm was developed as artificial neural network based machine learning models to predict the mechanical properties of hydrogen charged metallic materials. Multilayer feed forward back propagation model was used to predicts the tensile strength, had a network topology of 12-13-3-2. And the single layer feed forward back propagation model was employed to predict the percentage of elongation, has a network topology of 12-11-1. The developed models were validated and tested with unknown inputs and their capability was studied. The models were evaluated using Mean Absolute (MAE) value and represented the scatter diagram to demonstrate the efficiency of the models. The R-value for both the models seems to prove that the models are ready to be used in the practical applications.

Multi-response optimization and neural network modeling for parameter precision in heat reflux extraction of spice oleoresins from two pepper cultivars (Piper nigrum)

Black and white peppers are important oil bearing commodity crop in tropical areas. They are highly beneficial in food industries and herbal medicine, due to their amazing aroma and therapeutic activities. In this study, heat reflux technique was employed to extract medicinal oleoresin from the two peppercorns. For this purpose, various extraction parameters were considered viz: extraction time, particle size and feed-solvent ratio. These extraction parameters were employed to optimize the extraction yield and absorbed energy via Taguchi methodology. The established optimal condition values of the yield and absorbed energy from the parametric study were 13.22mg/g and 285.60J/min, respectively in black pepper heat refluxation. Moreover, in white pepper refluxation, the extraction yield and absorbed energy were 14.04mg/g and 264.82J/min, respectively. Artificial neural network (ANN) was used for prediction purposes. This was achieved by comparing two algorithms, transfer functions and neurons. A good training and better prediction of the experimental data were observed using the Levenberg Marquardt (LM) feed forward backpropagation algorithm with log sigmoid transfer function as hidden layer and 3-4-5-1 as model topology. Furthermore, a total of 19 and 25 bioactive compounds were identified in black and white pepper extracts, respectively. The results from the Scanning Electron Spectrometry (SEM) showed a remarkable morphological changes during the heat refluxation process.

Improvement of ε-polylysine production by marine bacterium Bacillus licheniformis using artificial neural network modeling and particle swarm optimization technique

ε-Polylysine is water soluble, biodegradable, edible and non-toxic homopolymer of l-lysine linked by the peptide bond between the carboxyl and ε-amino groups. ε-polylysine and its derivatives being used for past few decades for a broad range of industrial applications. However, the yields of ε-polylysine using wild type strains are comparatively low with respect to what desired for industrial production. Hence, in this study, an advanced modeling and optimization technique was applied to optimize medium parameters for enhanced ε-polylysine production by marine bacterium Bacillus licheniformis. The critical nutrients including glucose, yeast extract, magnesium sulphate and ferrous sulphate were incorporated in artificial neural networks (ANN) as input variables and ε-polylysine as the output variable. The ANN topology of 4-10-1 was found to be optimum upon training the model with feed-forward back propagation algorithm and on application of the developed model to particle swarm optimization (PSO) resulted in 3.56±0.16gL−1 of ε-polylysine under the following optimal conditions: glucose, 34gL−1; yeast extract, 2.3gL−1; magnesium sulphate, 0.44gL−1 and ferrous sulphate, 0.08gL−1. Thus, this optimization technique could significantly improve ε-polylysine by 196.7%, as compared to un-optimized medium. The potential significance of this study lies in the development of a suitable production medium for improved ε-polylysine production by an advanced optimization approach, ANN-PSO.

Noise Removal using Chebyshev Functional Link Artificial Neural Network with Backpropagation

In this proposed work, we have used an alternate ANN structure called Functional link ANN for image denoising. In contrast to other feed forward neural networks, the FLANN is a single layer structure, which never contain any hidden layer and non-linearity is introduced by enhancing the input pattern with a nonlinear function expansion called Chebyshev functional expansion. With the Chebyshev functional expansion, the network shows very good result in denoising the image corrupted by four different noise called Salt and Pepper noise, Gaussian noise, Speckle noise and Poisson noise. In this paper Gaussian noise is added to the speckle noise to give better result. In particular FLANN structure with Chebyshev functional expansion works best for Poisson noise suppression from an image. Here Back Propagation network is used to train the Chebyshev expanded image. BP network can be used to learn and store a great deal of mapping relations of input-output models, and no need to disclose in advance the mathematical equation that describes these mapping relations. Feed Forward Back Propagation (FFBP) algorithm performs quite well in the presence of different noise while preserving the image features satisfactorily

Prediction of Nephrolithiasis Based on Extracted Features of Ct-Scan Images using Artificial Neural Networks

Nephrolithiasis or Renal calculi or Kidney stones transpire in 1 in 10 populace at some time in their life. Renal calculi are a general cause of bloody urine and stern pain in the abdomen, flank, or groin. To locate the position, size and number of stones in the renal structure the patient is recommended to take computed tomography (CT-scan). As there is rapid raise in population the necessitate of more nephrologists is essential. So, the most important objective of this paper is to afford a supportive diagnosis system to the physician using neural networks in order to envisage nephrolithiasis based on extracted features of CT-scan. The proposed system incriminates with pre-processing, segmentation, feature extraction by applying neural network techniques and finally prediction of kidney stone is done. Artificial Neural networks are intermittently used as a dominant distinctive classifier for errands in medical diagnosis for premature detection of diseases. Here we introduced a Feed-Forward Backpropagation algorithm to lessen the diagnosis time and raise the accuracy of the system. The GLCM algorithm is introduced to extort the features which are used to train the network. The proposed system uses 22 input nodes, 10 hidden nodes and 1 output node. The network is trained until it targets its desired output. This trained network is used to cart out the classification automatically for a new model. The proposed system is tested with 50 real time samples, amid them 60% are used for training the network and 40% are used for testing the network. The intended system is implemented using MATLAB 8.5 software tool.

An improved large-signal model by artificial neural network for the MOSFETs operating in the breakdown region

An improved large-signal breakdown model establishment applying an artificial neural network (ANN) approach is presented for metal-oxide-semiconductor field-effect transistors (MOSFETs) operating in the breakdown regime for the first time. A neural network program with the feed-forward back propagation algorithm and Levenberg Marquardt optimization is utilized to obtain the MOSFET large-signal model for breakdown operation. When compared with the conventional model without the breakdown effects considered, more accurate large-signal characteristics in the breakdown regime can be obtained by incorporating the avalanche network using the ANN approach. The breakdown network is demonstrated to be significant for large-signal characterization at high bias according to the analysis of minimum acceptable error. Besides, the divergence problem due to the neglected avalanche network can be avoided during ANN training in the avalanche regime. The accuracy of the presented model can keep about 2% error. The presented ANN approach can be applied to device large-signal modeling in the impact ionization regime.

Development and Testing of an ANN Model for Estimation of Runoff from a Snow Covered Catchment

In this study, an attempt has been made to develop an ANN model to estimate runoff from a snow covered catchment of eastern Himalaya using feed-forward back-propagation algorithm with Levenberg–Marquardt optimization technique. The ANN model was programmed in C++ whereas a user-friendly GUI was developed in VB. The effects of past days rainfall and present day temperature data was observed on the performance of the selected ANN architecture in modelling snowmelt and monsoon season runoff. For this purpose, 8 years’ (2003–2010) daily data (rainfall, temperature, and discharge) were collected from CWC which were again divided into two parts (2003–2008 and 2009–2010) for training and testing of the ANN model, respectively. Initially it was found that the network can produce acceptable results with only rainfall data as input, but it needs at least past 3 days rainfall data to account for the antecedent moisture condition of the catchment. Networks 4-16-16-1 (with past 3 days rainfall) and 6-18-18-18-1 (with past 5 days rainfall) resulted modelling efficiency of 79.38 and 82.06% in training and 55.13 and 61.06% in validation, respectively. However, addition of present day temperature data as another input improved the performance in both training (ME 83.10 and 82.22%) and testing (ME 62.64 and 61.89%) marginally.

Modeling the wire-EDM process parameters for EN-8 carbon steel using artificial neural networks

Accompanying the evolution of mechanical industry, the need for alloy materials and ceramics as well, having high toughness, hardness & impact resistance are also increasing. Wire-EDM process is a quick fix to this box. As this WEDM process is proficient in machining these materials meticulously, regardless to their toughness & hardness. The present work discusses the experimental study on wire-cut machining of EN8 Carbon Steel. This paper presents an Artificial Neural Network approach in a multi objective optimization problem, for predicting the performance measure of a machining process. The input neurons were designed by Taguchi Orthogonal Array with 5 levels each with the help of software Minitab 17. In this research work, a comparative study for predicting multiple response parameters have been carried out using Feed forward and Cascade forward back propagation algorithms of ANN. This study was conducted by varying the network designing in each case. Two different back-propagation training algorithms were used in neural network designing such as Powell-Beale Conjugate Gradient (traincgb) and Levenberg-Marquardt (trainlm). Predictive accuracy of both the models was evaluated using Mean Square Error (MSE), Root Mean Square Error (RMSE) and coefficient determination (R2). The outcomes proved that both the algorithms were able to produce feasible models. Cascade model outperformed with all statistical tests between both.Keywords: Artificial Neural Network, Feed forward back propagation, Cascade forward back propagation

Predictive models for determination of compressive and split-tensile strengths of steel slag aggregate concrete

The current study adopts the results of completed laboratory experiments for modelling the strength properties of steel slag aggregate concrete using artificial neural network (ANN) technique. Note worthily, the considered experiments reported that strength properties increased with increasing curing age. The variation of factors for building the network includes: ground granulated blast furnace slag (GGBS) as partial replacement for granite at 20, 40, 60, 80 and 100%; water-cement ratio (w/c) at 0.5, 0.55 and 0.6; curing age at 7, 14 and 28 days. Other factors, such as cement and sand were kept constant. The input data were trained, learned and validated using the feed-forward back propagation algorithm. From the various trial and errors performed, the optimal ANN model which yielded the minimum mean square error and maximum absolute variance was 6-10-2. Therefore, based on the high confidence level of the model predictions, the models are recommended for predicting strength properties of steel slag aggregate concrete that falls within the limit of this study.

Artificial Neural Network Models for Predicting the Energy Consumption of The Process of Crystallization Syrup in Konya Sugar Factory

In this study, artificial neural network models have been developed from the sugar production process stages in Konya Sugar Factory using artificial neural networks to estimate the energy consumption of the process of crystallization syrup. Models developing specific enthalpy, mass and pressure as input layer parameters and consumption energy as output layer were used.124 different data are taken from Konya Sugar Factory during January 2016. Feedforward back propagation algorithm was used in the training phase of the network. Learning function LEARNGDM and the number of hidden layer kept constant as 2 and transfer functions are modified. In the developed 27 ANN model, 2-5-1 network architecture was determined as the best suitable network architecture and transfer function is determined logsig function as the optimal transfer function. Optimum results of the model taken in the coefficient of determination was found R = 0.98 neural network training, testing and validate was also found to be R = 0.98, the performance of the network for not shown data to network was found R=0,99.

Modeling of Rainfall-Runoff Correlations Using Artificial Neural Network-A Case Study of Dharoi Watershed of a Sabarmati River Basin, India

The use of an Artificial Neural Network (ANN) is becoming common due to its ability to analyse complex nonlinear events. An ANN has a flexible, convenient and easy mathematical structure to identify the nonlinear relationships between input and output data sets. This capability could efficiently be employed for the different hydrological models such as rainfall-runoff models, which are inherently nonlinear in nature. Artificial Neural Networks (ANN) can be used in cases where the available data is limited. The present work involves the development of an ANN model using Feed-Forward Back Propagation algorithm for establishing monthly and annual rainfall runoff correlations. The hydrologic variables used were monthly and annual rainfall and runoff for monthly and annual time period of monsoon season. The ANN model developed in this study is applied to Dharoi reservoir watersheds of Sabarmati river basin of India. The hydrologic data were available for twenty-nine years at Dharoi station at Dharoi dam project. The model results yielding into the least error is recommended for simulating the rainfall-runoff characteristics of the watersheds. The obtained results can help the water resource managers to operate the reservoir properly in the case of extreme events such as flooding and drought.