Cascade Feed-forward Neural Network Research Articles

Prediction of longitudinal and transverse profiles of pressure flushing cones using artificial intelligence and data pre-processing

AbstractOne of the most critical issues in dam reservoir management is the determination of sediment level after flushing operation. Artificial intelligence (AI) methods have recently been considered in this context. The present study adopts four AI approaches, including the Feed-Forward Neural Network (FFNN), Cascade Feed-Forward Neural Network (CFFNN), Gene Expression Programming (GEP), and Bayesian Networks (BNs). Experimental data were exploited to train and test the models. The results revealed that the models were able to estimate the post-flushing sediment level accurately. FFNN outperformed the other models. Furthermore, the importance of model inputs was determined using the τ-Kendall (τ–k), Random Forest (RF), and Shannon Entropy (SE) pre-processing methods. The initial level of sediment was found to be the most important input, while the orifice output flow rate was observed to have the lowest importance in modeling. Finally, inputs of higher weights were introduced to the FFNN model (as the best predictive model), and the analysis of the results indicated that the exclusion of less important input variables would have no significant impact on model performance.

Application of soft computing in water treatment plant and water distribution network

Analysis of traditional water distribution network (WDN) is more time-consuming and less effective to predict the problem related to water supply systems such as water quality, coagulant dose, and residual chlorine in developing countries. In the present paper water quality neural network, coagulation dose neural network, and residual neural network model were implemented. The performance of the Cascade Feed Forward Neural Network (CFFNN) and Feedforward neural network (FFNN) was excellent for the prediction of water quality parameters and residual chlorine respectively during the training and testing period. CFFNN water quality model (27-30-27) with R = 0.989 produced an excellent prediction of outlet water quality parameters. In coagulant dose modelling, CFFNN (2-40-1) yielded a good prediction with R = 0.947 for a broad range of turbidities as compared to other models. Similarly in residual chlorine modelling, FFNN (2-25-1) delivered the best prediction with R = 0.988 as compared to other models.

Exploring the relation between production factors, ore grades, and life of mine for forecasting mining capital cost through a novel cascade forward neural network-based salp swarm optimization model

This paper aims at exploring the relationship between production factors, ore grades, and life of mine for forecasting mining capital cost (MCC) for open pit mining projects. Accordingly, the relationship between annual mine and mill production (MineAP, MillAP), stripping ratio (SR), reserve mean grade (RMG), the life of mine (LOM), and MCC of 80 open pit mining projects were investigated and thoroughly evaluated. The dataset was then divided into two sections, with 56 observations used to develop the forecast models. The remaining 24 observations were used to test the accuracy of the developed models. Subsequently, the cascade feedforward neural network (CFNN) was developed to forecast MCC based on the influential parameters. In order to improve the accuracy of the CFNN model, the salp swarm optimization (SalpSO) algorithm was applied to train the CFNN model and optimize the weights of the model, called the SalpSO-CFNN model. The benchmark models which were developed in the previous studies, such as support vector machine (SVM), classification and regression tree (CART), and multiple layers perceptron (MLP) neural network, were also developed in this study to compare with the proposed SalpSO-CFNN model in terms of MCC forecast. The results revealed that production factors, ore grades, and LOM are closely related to MCC, and they are statistically significant. The forecast results also indicated that the proposed novel SalpSO-CFNN model provided a good accuracy with a mean absolute error (MAE) of 179.567, root-mean-squared error (RMSE) of 248.401, and determination coefficient (R2) of 0.980. This result is higher by 18% compared with the CART model and 2–6% compared with the remaining forecast models. A sensitivity analysis also indicated that MineAP, MillAP are the most influential parameters on the forecast of MCC, and they should be specially taken into account when forecasting MCC of open pit mining projects.

Monitoring the effect of surface functionalization on the CO2 capture by graphene oxide/methyl diethanolamine nanofluids

Different processes exist to capture carbon dioxide (CO2) and reduce its undesirable effects on the atmosphere. Stable suspensions of graphene oxide (GO) nanosheets in aqueous methyl diethanolamine (MDEA) solutions have recently attracted great attention as a potential CO2 absorption medium. Moreover, experimental analyses confirmed that GO surface functionalization positively affects the CO2 absorption by MDEA-based nanofluids. To the best of our knowledge, there are no mathematical models to investigate the effect of surface functionalization on the CO2 capture ability of GO-amine nanofluids. Artificial intelligence (AI) techniques are reliable methodologies for understanding behavior of even the most complex systems. Therefore, different AI models are designed to reveal the effect of GO surface functionalization on the CO2 capture ability of MDEA-based nanofluids. Our AI models use operating temperature, pressure, functionalized group, and GO dosage in the amine solutions to predict CO2 solubility in GO/MDEA nanofluids. The results confirm that the cascade feedforward (CFF) neural network is the most accurate AI paradigm for estimating CO2 solubility in aqueous GO/MDEA nanofluids in a wide range of operation conditions (i.e., AARD = 1.78%, MSE = 0.007, RMSE = 0.08, and R2 = 0.9906). Simulation results justified that the surface functionalization of the GO nanosheets by the NH2 group provides the most promising results for CO2 capture by the GO/MDEA nanosuspensions.

Performance Evaluation of Classifiers Created using Elman Back-Propagation and Cascade Feed-forward Neural Networks

Huge data being captured in our day-to-day activities are mostly imbalance. Such data therefore, calls for fast, accurate and robust techniques, through which they could be analyzed in order to fast-track early decision making. A Cascade Feed-forward Neural Networks and Elman Backpropagation are known techniques in neural network domain and their efficacies is therefore tested on separable data in this study. The objective of this study is to evaluate the performance of these techniques in solving a linear classification problem. The linear classification of data involves, splitting of separable data into two distinct clusters. In order to achieve the goal of this study, linear classifiers were created using the two aforementioned techniques. Both network structures were exposed to the same dataset and similar parameter configurations were set for each technique. The model that emanates through each technique was simulated using the set of untrained data. In order to determine the accuracy of each model created through each technique, their Mean Absolute Errors (MAE) was computed. The performance of each model was determined based on the value of MAE. The error computation for the simulated output reveals that, cascade neural network gives an error of 0.0928; while the model created using Elman Backpropagation network gives a relatively lower error of 0.0661. It can be inferred from this study, that, both techniques are capable of fitting accurate classifiers from dataset and specifically, both techniques are very suitable for binary classification of separable data.

Application of cascade feed forward neural network to predict coagulant dose

Inlet water quality fluctuations affect mainly coagulant dose, and outlet water quality of the water treatment plant (WTP). Many complex physical and chemical processes are involved in WTP and water distribution networks (WDN). These technologies show non-linear behavior, which is challenging to be described by linear mathematical models. Thus, there is a need to develop prediction models for coagulation dose. The present study involves the application of cascade feed-forward neural networks (CFFNN) to predict coagulant dose. CFFNN Model was developed by using the Levenberg-Marquardt Training Algorithm and Bayesian Regularization Training Algorithm to predict coagulant dose. During the development of these models, hidden nodes are varied from 15 to 60, and R is found between 0.914 and 0.947. The best results were obtained by the CFFNN model using the Bayesian Regularization Training Algorithm (CFNNCD2) with hidden node 40, where R = 0.945 for training and 0.947 for testing.

Modeling of CO2 capture ability of [Bmim][BF4] ionic liquid using connectionist smart paradigms

Abstract The burning of fossil fuels produces large amounts of exhaust gases containing carbon dioxide (CO2). The emission of CO2 into the atmosphere is widely known as the leading cause of global warming and climate change. The separation processes are responsible for capturing the CO2 to reduce its undesirable effects on the environment. Since the conventional processes have their drawbacks, it is crucial to find a more environment-friendly process for CO2 capture. Recently, ionic liquids (ILs) have become an interesting candidate for CO2 capture. In this study, the solubility of CO2 in the 1-n-butyl-3-methylimidazolium tetrafluoroborate ([Bmim][BF4]) is estimated using six different artificial intelligence (AI) techniques, including four artificial neural networks (ANN), support vector machines (LS-SVM), adaptive neuro-fuzzy interface system (ANFIS). The cascade feed-forward neural network has been found as the best model for the considered matter. This model predicts overall experimental datasets with excellent accuracy of AARD = 6.88%, MSE = 8 × 1 0 − 4 , and R 2 = 0 . 98808 . The maximum mole fraction of CO2 in the ionic liquid (i.e., 0.8) can be obtained at the highest pressure and the lowest temperature.

Reliability Prediction of CNC Machine Tool Spindle Based on Optimized Cascade Feedforward Neural Network

Aiming at the large error of traditional reliability prediction method, and the defects of BP neural network prediction method, a new method of optimized cascade feedforward neural network was proposed based on Adam algorithm to predict the reliability of CNC machine tool spindles. A three-layer optimized cascade feedforward neural network model for reliability prediction was established based on the first $\text{n}^{\mathrm {th}}$ reliability value and the mean time between failure corresponding to the ( $\text{n}+1$ ) $^{\mathrm {th}}$ reliability value as the input variables of the neural network. Besides, error comparison analysis was performed on the test set data using the existing relevant reliability data for simulation training. As per the research results, the absolute value of the maximum relative error of the reliability prediction value obtained by this method is 2.53% which is less than 3%, and the prediction method shows high accuracy. Compared with BP neural network, it has the advantages of faster learning speed and better nonlinear fitting ability. Therefore, this method is feasible to be utilized for the prediction of the reliability of CNC machine tool spindles, and provides a reference for improving the accuracy of CNC machine tool reliability prediction.

An efficient two-stage approach for structural damage detection using meta-heuristic algorithms and group method of data handling surrogate model

In this study, the performance of an efficient two-stage methodology which is applied in a damage detection system using a surrogate model of the structure has been investigated. In the first stage, in order to locate the damage accurately, the performance of the modal strain energy based index for using different numbers of natural mode shapes has been evaluated using the confusion matrix. In the second stage, to estimate the damage extent, the sensitivity of most used modal properties due to damage, such as natural frequency and flexibility matrix is compared with the mean normalized modal strain energy (MNMSE) of suspected damaged elements. Moreover, a modal property change vector is evaluated using the group method of data handling (GMDH) network as a surrogate model during damage extent estimation by optimization algorithm; in this part of methodology, the performance of the three popular optimization algorithms including particle swarm optimization (PSO), bat algorithm (BA), and colliding bodies optimization (CBO) is examined and in this regard, root mean square deviation (RMSD) based on the modal property change vector has been proposed as an objective function. Furthermore, the effect of noise in the measurement of structural responses by the sensors has also been studied. Finally, in order to achieve the most generalized neural network as a surrogate model, GMDH performance is compared with a properly trained cascade feed-forward neural network (CFNN) with log-sigmoid hidden layer transfer function. The results indicate that the accuracy of damage extent estimation is acceptable in the case of integration of PSO and MNMSE. Moreover, the GMDH model is also more efficient and mimics the behavior of the structure slightly better than CFNN model.

Prediction of viscosity of biodiesel blends using various artificial model and comparison with empirical correlations

From the perspective of renewability and environmental pollution, biodiesels are appropriate alternatives to conventional diesel fuels due to their proper combustion behavior and atomization characteristics, which can be influenced by the viscosity as an essential factor. Therefore, the viscosity prediction would be of importance for blend of biodiesel/diesel fuels. For biodiesel/diesel mixtures, the blended viscosity has been predicted with numerous empirical correlations available in the literature. In this work, the viscosity of fuel mixtures was evaluated through Generalized regression neural network (GRNN), Radial Basis Neural Networks (RBFNN), Multi-Layer Perceptron Neural Network (MLPNN), and Cascade Feed-forward Neural Network (CFNN), based on various experimental data. These developed models were then compared in terms of predictive accuracy and available empirical correlations to select the best model. Finally, the proposed model was compared with the most prominent biodiesel viscosity models confirming that the developed model has been superior in predicting the value of viscosity for biodiesel blends with reported values of 0.9997 and 0.87% for parameters of the coefficient of determination (R2) and absolute average relative deviation (AARD%), respectively.

Automated Recognition of Hand Grasps Using Electromyography Signal Based on LWT and DTCWT of Wavelet Energy

This paper presents a novel framework that automatically classifies hand grasps using Electromyogram (EMG) signals based on advanced Wavelet Transform (WT). This method is motivated by the observation that there lies a unique correlation between different samples of the signal at various frequency levels obtained by Discrete WT. In the proposed approach, EMG signals captured from the subjects are subjected to denoising using symlet wavelets, followed by Principal Component Analysis (PCA) for dimensionality reduction. Further, the important attributes of the signal are extracted using Lifting Wavelet Transform (LWT) and Dual Tree Complex WT (DTCWT). Multiple classifiers such as Feed Forward Neural Networks (FFNN), Cascaded Feed Forward Neural Networks (CFNN), Support Vector Machine (SVM) and Deep Learning Neural Network (DLNN) are used for classification. The simulation results are compared with various training algorithms and it is observed that DTCWT features combined with CFNN and trained with Gradient Descent with Adaptive Back Propagation (GDABP) algorithm achieved the best performance. The advantages of the proposed method were proved by comparing with the earlier conventional methods, in terms of recognition performance. These experimental results prove that the proposed method gives a potential performance in the recognition of hand grasps using EMG signals. In addition, the proposed method supports clinicians to improve the performance of myoelectric pattern recognition.

Application of artificial neural networks to predict flow velocity in a 180° sharp bend with and without a spur dike

This work has compared the performance of three well-known artificial neural network (ANN) approaches in turbulent flow pattern modeling based on geometric characteristics of the channel (angle of horizon, distance from outer bank of the bend, and distance from the bed) in a 180° sharp bend with and without a T-shaped spur dike. ANN methods discussed in this work are feed-forward neural network, cascade feed-forward neural network, and extreme learning machines. Acoustic Doppler velocimetry is used to measure the velocity components in x, y, and z directions. Conducting this research is innovative and significant from three aspects: First, the data of the flow pattern around T-shaped spur dikes in a 180° bend, given its high importance in rivers in nature, are very rarely available; second, application of neural network models for understanding the flow nature is highly efficient at other bend points where no experimental or field measurements have been conducted; and third, with these models, the cost and time required for conducting numerical and experimental modeling for prediction of the flow velocity significantly decrease. Outliers (abnormalities or anomalies) may be generated by various factors in the measured velocities. Before modeling by ANNs, the self-organizing map clustering method is used to detect outliers to obtain more accurate models. Performance of ANN models is evaluated using correlation coefficient (R) and root mean squared error criteria. In order to compare the performance of ANNs in flow velocity modeling at different locations of the bend, seven different data groups are considered using different combinations of samples by random sampling. Results of comparison of ANN models with experimental data indicate that ANNs provide reasonable results in most cases and may be employed successfully in estimating the velocity in sharp bends with and without the presence of a spur dike. Moreover, it may be concluded that ANN models without spur dikes are more accurate than those with spur dikes due to creating less turbulence flow.

Development of a modular system for estimating attenuation coefficient of gamma radiation: comparative study of different learning algorithms of cascade feed-forward neural network

In this paper, a modular system is developed for estimation of mass attenuation coefficient (MAC) of different materials/energies using artificial neural network (ANN). Cascade feed-forward neural network (CFFNN) as a type of ANN constructs mapping function between input patterns and the targets (i.e. MAC). Performance of different learning algorithms of CFFNN including gradient descent (GD), gradient descent with momentum (GDM), scaled conjugate gradient (SCG), Levenberg-Marquardt (LM), and Bayesian regularization (BR) are compared. For training, different categories of input patterns features are utilized to show the more appropriate one. Average mean relative error (AMRE) and cumulative distribution function (CDF) of the results indicate that BR learning algorithm accompany with the selected category of features (i.e. Z, E, and ρ) is more accurate for estimation of MAC (e.g. CDFAl (0.0069) = 0.99 and AMREAl = 0.0017). The advantages of the present method are: 1- Estimation of MAC is done fast (i.e. in comparison with Monte Carlo methods) and is done at a lower cost (i.e. without need to extra experiments) 2- Modular system reduces the risk of incorrect estimation 3- It is possible to extend the number of estimators for more materials/mixtures without unfavorably affecting the existing system.

Cascaded Feedforward Neural Networks for speaker identification using Perceptual Wavelet based Cepstral Coefficients

Cascaded Feedforward Neural Networks for speaker identification using Perceptual Wavelet based Cepstral Coefficients

Modeling of Chlorine and Coagulant Dose in a Water Treatment Plant by Artificial Neural Networks

Coagulation and chlorination are complex processes of a water treatment plant (WTP). Determination of coagulant and chlorine dose is time-consuming. Many times WTP operators in India determine the coagulant and chlorine dose approximately using their experience, which may lead to the use of excess or insufficient dose. Hence, there is a need to develop prediction models to determine optimum chlorine and coagulant doses. In this paper, artificial neural networks (ANN) are used for prediction due to their ability to learn and model non-linear and complex relationships. Separate ANN models for chlorine and coagulant doses are explored with radial basis neural network (RBFNN), feed-forward neural network (FFNN), cascade feed forward neural network (CFNN) and generalized regression neural network (GRNN). For modeling, daily water quality data of the last four years are collected from the plant laboratory of WTP in Maharashtra (India). In order to improve performance, these models are established by varying input variables, hidden nodes, training functions, spread factor, and epochs. The best models are selected based on the comparison of performance measures. It is observed that the best performing chlorine dose model using defined statistics is found to be RBFNN with R=0.999. Similarly, the CFNN coagulant dose model with Bayesian regularization (BR) training function provided excellent estimates with network architecture (2-40-1) and R=0.947. Based on the above models, two graphical user interfaces (GUIs) were developed for real-time prediction of chlorine and coagulant dose, which will be useful for plant operators and decision makers.

Comparative study of application of different supervised learning methods in forecasting future states of NPPs operating parameters

In this paper, some important operating parameters of nuclear power plants (NPPs) transients are forecasted using different supervised learning methods including feed-forward back propagation (FFBP) neural networks such as cascade feed-forward neural network (CFFNN), statistical methods such as support vector regression (SVR), and localized networks such as radial basis network (RBN). Different learning algorithms, including gradient descent (GD), gradient descent with momentum (GDM), scaled conjugate gradient (SCG), Levenberg-Marquardt (LM), and Bayesian regularization (BR) are used in CFFNN method. SVR method is used with different kernel functions including Gaussian, polynomial, and linear. RBN is used with radial activation function. Comparison of the results indicates that learning algorithms based on Gaussian distribution function (i.e. BR algorithm) and Gaussian kernel/activation functions are, in general, more precise for time series prediction. Moreover, learning methods based on Gaussian function lead to acceptable results in prediction of complicated time series, such as core inlet flowrate of large break loss of coolant accident (LBLOCA) which are changed irregularly and drastically. In other words, Gaussian learning algorithms/kernel functions/activation functions are appropriate choices for NPPs parameters forecasting.

Optimization-based method for structural damage detection with consideration of uncertainties- a comparative study

In this paper, for efficiently reducing the computational cost of the model updating during the optimization process of damage detection, the structural response is evaluated using properly trained surrogate model. Furthermore, in practice uncertainties in the FE model parameters and modelling errors are inevitable. Hence, an efficient approach based on Monte Carlo simulation is proposed to take into account the effect of uncertainties in developing a surrogate model. The probability of damage existence (PDE) is calculated based on the probability density function of the existence of undamaged and damaged states. The current work builds a framework for Probability Based Damage Detection (PBDD) of structures based on the best combination of metaheuristic optimization algorithm and surrogate models. To reach this goal, three popular metamodeling techniques including Cascade Feed Forward Neural Network (CFNN), Least Square Support Vector Machines (LS-SVMs) and Kriging are constructed, trained and tested in order to inspect features and faults of each algorithm. Furthermore, three well-known optimization algorithms including Ideal Gas Molecular Movement (IGMM), Particle Swarm Optimization (PSO) and Bat Algorithm (BA) are utilized and the comparative results are presented accordingly. Furthermore, efficient schemes are implemented on these algorithms to improve their performance in handling problems with a large number of variables. By considering various indices for measuring the accuracy and computational time of PBDD process, the results indicate that combination of LS-SVM surrogate model by IGMM optimization algorithm have better performance in predicting the of damage compared with other methods.

Prediction of coefficients of the Langmuir adsorption isotherm using various artificial intelligence (AI) techniques

Adsorption is a process that utilizes porous solid materials to separate some solutes from gas or liquid mixtures. The extent of this separation is often determined using the adsorption isotherms, i.e., semi-empirical correlation for relating the amount of adsorbed substances by the solid medium to its associated concentration in fluid phase at constant temperature. Prior to employing an adsorption isotherm, its coefficients should be adjusted using experimental data of a considered adsorption system. In this study, the coefficients of Langmuir model have been predicted using various types of artificial neural networks (ANNs), support vector machines, and adaptive neuro fuzzy interface systems, and coupled scheme of ANN-genetic algorithm. The employed ANN types are multi-layer perceptron neural network (MLPNN), radial basis function neural network, cascade feedforward neural network, and generalized neural network. The considered coefficients tried to be modeled as functions of temperature, pH, adsorbent density, and adsorbate molecular weight. Predictive accuracies of the AI techniques have been compared utilizing different statistical indices such as correlation coefficient (R2), mean square error, and absolute average relative deviation (AARD%). The results indicated that MLPNN was the most accurate model for predicting the coefficients of Langmuir isotherm, due to its AARDs of 24.64 and 22.40% for the first and second coefficients, respectively.

Empirical Mode Decomposition and Cascade Feed-Forward Artificial Neural Network Based Intelligent Fault Classifier

Recent years have witnessed a fast emergence of smart grids, since the availability of electric energy has become a crucial issue in power system engineering. In fact, smart grids will allow for good energy management while ensuring a better control and protection of electric power system components. In this respect, here we propose an intelligent fault classifier (IFC) based on the empirical mode decomposition (EMD) and cascade feedforward neural networks (CFFANN). The EMD is used to decompose sending-end voltages into intrinsic mode functions (IMFs). Then the Hilbert transform (HT) is utilized to extract features from these IMFs. After this step, we selected relevant attributes from the extracted features. CFFANN is used to make the learning process faster. This neural network is introduced to classify all the ten main faults that may occur in a transmission line. The IFC conception is validated using MATLAB/SIMULINK software. The obtained results show that the IFC is a reliable and robust system.

Artificial Neural Network Approach for Dynamic Modelling of Heat Exchanger for Data Prediction

Objective: Artificial Neural Networks became a powerful tool for dynamic modelling of non linear physical systems and for prediction of specific parameters of complex systems. This work aims to model a heat exchanger using different neural networks and to investigate their performance in predicting its outlet temperature. Methods: In this work four different neural networks namely Elman Recurrent Neural Networks (ERNN), Time Delay Neural Networks (TDNN), Cascade Feed Forward Neural Networks (CFFNN) and Feed Forward Neural Networks (FFNN) are modelled for the prediction of the outlet liquid temperature of a saturated steam heat exchanger from its liquid flow rate. A benchmark dataset consisting of 4000 tuples is used to train, validate and test the performance of each neural network model. Findings: All the four ANN models are trained, validated and tested which predicted the outlet temperature of the Heat exchanger with acceptable accuracy and the Elman Recurrent Neural Network is found to have the best accuracy by having the lowest Mean Square Error (MSE) and best Regression due to its feedback connections. Applications: The Artificial Neural Network Models simulated, especially the Elman Neural Networks are good at prediction of operational parameters of a physical system and hence can be used in prediction and forecasting of operational parameters of engineering, medical, financial and environmental systems.Keywords: Artificial Neural Networks, Elman Recurrent Neural Network, Heat Exchanger, Prediction