Multilayer Artificial Neural Network Research Articles

Identification of cavitation in centrifugal pump by artificial immune network

Reducing the cost of unscheduled shutdowns and enhancing the reliability of production systems is an important goal for various industries; this could be achieved by condition monitoring and artificial intelligence. Cavitation is a common undesired phenomenon in centrifugal pumps, which causes damage and its detection in the preliminary stage is very important. In this paper, cavitation is identified by use of vibration and current signal and artificial immune network that is modeled on the base of the human immune system. For this purpose, first data collection were done by a laboratory setup in health and five stages damage condition; then various features in time, frequency, and time–frequency were extracted from vibration and current signals in addition to pressure and flow rate; next feature selection and dimensions reduction were done by artificial immune method to use for classification; finally, they were used by artificial immune network and some other methods to identify the system condition and classification. The results of this study showed that this method is more accurate in the detection of cavitation in the initial stage compared to methods such as non-linear supportive vector machine, multi-layer artificial neural network, K-means and fuzzy C-means with the same data. Also, selected features with artificial immune system were better than principal component analysis results.

Real Time Control Application of the Robotic Arm Using Neural Network Based Inverse Kinematics Solution

Robotic arms are widely used in many industrial applications at present. The control of robotic arms involves position coordination Cartesian space by a forward/inverse kinematics solution method. The inverse kinematics is difficult for real-time control applications, computational requirements are intensive and the run-time is high. The traditional solution methods used geometric, algebraic, and numerical iterative techniques are inadequate and slow in the inverse kinematics solution. Recently, alternative solution methods based on artificial intelligence techniques have been developed to solve the inverse kinematics problem. In this study, a multi-layered feed-forward Artificial Neural Network model was developed to solve the inverse kinematics of the 5 degrees of freedom robotic arm. Using the Proportional-Integral control algorithm combined with this Artificial Neural Network model, the real-time position control of the robotic arm was accomplished. The obtained results were compared with the PI control supported by an analytical inverse kinematics solution in real-time. The results showed that the PI control combined with Artificial Neural Network has superior tracking ability, smaller control error, and better absolute fit to the reference.

Application of Artificial Neural Network for Prediction of Local Void Fraction in Vertical Subcooled Boiling Flow

This paper presents the feasibility study of potential application of multi-layer feed-forward Artificial Neural Networks (ANN) to predict local void fraction of subcooled boiling flows in vertical upward annular channel. A total of 638 experimental data points performed at KAERI and reported in literature was selected for training and testing ANN model. The seven basic parameters are chosen to be input variables and then the optimal structure of ANN which consist of two hidden layers with 131 neurons was determined based on traditional Trial-and-Error method after balancing the trade-off between the performance and training time. Results showed that the ANN model is capable to accurately predict the local void fraction with R2 value of 0.99931 for training data, R2 value of 0.99483 for testing data and R2 value of 0.99828 for all data. Also, it proved that the ANN training will be more effective with an extensive experimental database.

Adaptive signal control for bus service reliability with connected vehicle technology via reinforcement learning

This paper presents an adaptive signal controller for managing traffic delays and urban bus service reliability with fully adaptable acyclic timing plans. The signal controller is built upon a reinforcement learning framework that consists of a model-based and a data-driven component. The model-based component is represented by a hybrid kinematic wave traffic model that integrates macroscopic flow-based and microscopic vehicle-based state variables subject to stochastic demands and bus service status. To cope with the high dimensional solution space, the data-driven component is incorporated as a multi-layer artificial neural network and is used to approximate future traffic states and system performances with respect to prevailing control settings. Before the controller can be used, the neural network is to be trained through a series of realised dynamic state transitions via an on-policy temporal difference learning algorithm. The proposed control framework is tested over a real world corridor scenario in London, UK. The proposed controller is able to reduce both traffic delays and bus service variabilities subject to stochastic demands with acyclic timing plans that can be derived in short computational time. This study contributes to the design of adaptive network traffic control for multi-modal networks with connected vehicle technology and advanced learning-based optimisation techniques.

Classification of the Children with ADHD and Healthy Children Based on the Directed Phase Transfer Entropy of EEG Signals

Purpose: The present study was conducted to investigate and classify two groups of healthy children and children with Attention Deficit Hyperactivity Disorder (ADHD) by Effective Connectivity (EC) measure. Since early detection of ADHD can make the treatment process more effective, it is important to diagnose it using new methods. 
 Materials and Methods: For this purpose, Effective Connectivity Matrices (ECMs) were constructed based on Electroencephalography (EEG) signals of 61 children with ADHD and 60 healthy children of the same age. ECMs of each individual were obtained by the directed Phase Transfer Entropy (dPTE) between each pair of electrodes. ECMs were calculated in five frequency bands including, delta, theta, alpha, beta, and gamma. Based on ECM, an Effective Connectivity Vector (ECV) was constructed as a feature vector for the classification process. Furthermore, ECV of different frequency bands was pooled in one global ECV (gECV). Multilayer Artificial Neural Network (ANN) was used in the steps of classification and feature selection by the Genetic Algorithm (GA).
 Results: The highest classification accuracy with the selected features of ECV was related to theta frequency band with 89.7%. After that, the delta frequency band had the highest accuracy with 89.2%. The results of ANN classification and GA on the gECV reported 89.1% of accuracy.
 Conclusion: Our findings show that the dPTE measure, which determines effective connectivity between the brain regions, can be used to classify between ADHD and healthy groups. The results of the classification have improved compared to some studies that used the functional connectivity measures.

Prediction of PV Solar Panel Output Characteristics Using a Multilayer Artificial Neural Network (MLANN)

In this paper, PV solar collector was tested experimentally from 1st July to 31th August 2018 between 7:00 am and 6:00 pm under the weather conditions of Iraq. The PV output power was calculated by using the measured data of voltage and current obtained from experiments. To predict the PV solar output characteristics, four structures of a multilayer artificial neural network MLANN with Error Back-Propagation EBP were designed in MATLAB software. The MLANN structures have two inputs (temperature and irradiance) and three outputs (voltage, current and power). From experiment tests, a dataset of 434 hourly points was collected to investigate the structures of MLANN model. A 70% of the data used for training stage and 30% was distributed between the testing and validating stages. From test stage, the average of output value was taken for 14 numbers of data to compare with experimental values. The MLANN results show that the structures 2-4-4-1, 2-1-1-1 and 2-5-5-1 were the optimum testing model in the voltage, current and power output respectively with high accuracy and good agreement with experimental results.

Comparing the Accuracy of Artificial Neural Networks in Estimating the Weight of Cobb, Ross, and Arbor Acres Chicks using Video Image Processing Technology

This study aimed to compare the accuracy of artificial neural networks (ANNs) in estimating the weight of broilers using video image processing technology. A total number of 900 broiler chicks from three different strains (Ross 308, Cobb 500, and Arbor Acres) were fed on commercial diets and reared under standard situations for 42 days. Thirty male and female chicks from each strain were weighed randomly using digital scales every day while simultaneously filmed from top view using a Xenon camera (2MP 1080IP lens). In image processing, digital images initially were extracted from films and then each image was processed using GUI of MATLAB software. Sixteen morphological features extracted from images that significantly correlated with the chicks' weight, were used as inputs of the artificial neural network, and multilayer perceptron ANN was trained to predict the weight of chickens of each strain via an error propagation algorithm. The procedure was the same for all three strains. The accuracy of ANN models to predict the weight of chicks were 98.4% (with an average error of 7.9 g), 99.54% (with an average error of 0.37 g), and 99.67% (with an average error of 2 g) for Ross, Cobb, and Arbor Acres strains, respectively. In conclusion, a comprehensive intelligent model can be designed based on artificial neural networks and video image processing technology to estimate the weight of broiler chickens regardless of their strain type.

An improved aerodynamic performance optimization method of 3-D low Reynolds number rotor blade

Abstract In this paper, an improved aerodynamic performance optimization method for 3-D low Reynolds number (Re) rotor blade is proposed. A conventional optimization procedure of blade is usually divided into three parts, such as the parameterization method, the fitness value evaluation and the optimization algorithm. This work is mainly focused on the first two parts. The parametrization method, Camber-FFD, is presented based on the camber parametrization method and the free-form deformation algorithm (FFD). The shape of 3-D blade is parameterized by the incidence angles and the coordinates of the maximum camber points. The fitness value evaluation has been realized with the help of an adaptive topological back propagation multi-layer forward artificial neural network (BP-MLFANN). During the training of BP-MLFANN, the hybrid particle swarm optimization method combined with the modified very fast simulate annealing algorithm (HPSO-MVFSA) is adopted to determine the neural network topology adaptively. To verify the effectiveness of this aerodynamic optimization method, the aerodynamic performance of a 3-D low-Re blade, such as Blade D900, is optimized, and the results are compared and analyzed based on the experiments and simulations. It is proved that this aerodynamic optimization method is feasible.

An improved aerodynamic performance optimization method of 3-D low Reynolds number rotor blade

Abstract In this paper, an improved aerodynamic performance optimization method for 3-D low Reynolds number (Re) rotor blade is proposed. A conventional optimization procedure of blade is usually divided into three parts, such as the parameterization method, the fitness value evaluation and the optimization algorithm. This work is mainly focused on the first two parts. The parametrization method, Camber-FFD, is presented based on the camber parametrization method and the free-form deformation algorithm (FFD). The shape of 3-D blade is parameterized by the incidence angles and the coordinates of the maximum camber points. The fitness value evaluation has been realized with the help of an adaptive topological back propagation multi-layer forward artificial neural network (BP-MLFANN). During the training of BP-MLFANN, the hybrid particle swarm optimization method combined with the modified very fast simulate annealing algorithm (HPSO-MVFSA) is adopted to determine the neural network topology adaptively. To verify the effectiveness of this aerodynamic optimization method, the aerodynamic performance of a 3-D low-Re blade, such as Blade D900, is optimized, and the results are compared and analyzed based on the experiments and simulations. It is proved that this aerodynamic optimization method is feasible.

Modeling of Friction Phenomena of Ti-6Al-4V Sheets Based on Backward Elimination Regression and Multi-Layer Artificial Neural Networks.

This paper presents the application of multi-layer artificial neural networks (ANNs) and backward elimination regression for the prediction of values of the coefficient of friction (COF) of Ti-6Al-4V titanium alloy sheets. The results of the strip drawing test were used as data for the training networks. The strip drawing test was carried out under conditions of variable load and variable friction. Selected types of synthetic oils and environmentally friendly bio-degradable lubricants were used in the tests. ANN models were conducted for different network architectures and training methods: the quasi-Newton, Levenberg-Marquardt and back propagation. The values of root mean square (RMS) error and determination coefficient were adopted as evaluation criteria for ANNs. The minimum value of the RMS error for the training set (RMS = 0.0982) and the validation set (RMS = 0.1493) with the highest value of correlation coefficient (R2 = 0.91) was observed for a multi-layer network with eight neurons in the hidden layer trained using the quasi-Newton algorithm. As a result of the non-linear relationship between clamping and friction force, the value of the COF decreased with increasing load. The regression model F-value of 22.13 implies that the model with R2 = 0.6975 is significant. There is only a 0.01% chance that an F-value this large could occur due to noise.

A Flow Information-Based Prediction Model Applied to the Nonaxisymmetric Hub Optimization of a Centrifugal Impeller

Abstract A full-range prediction model for turbomachinery based on the flow-field information code is established in this article to solve the problems that traditional models do not have enough prediction accuracy and cannot reflect the complete performance characteristics of the impeller. The model, which can predict the complete performance curve of the impeller with higher accuracy, consists of two multilayer artificial neural network (ANN) submodels. Different from the traditional model, the ANN submodel uses the flow-field information code for pretraining layer by layer. The flow-field information code is the characteristic information extracted from the impeller flow field through the proper orthogonal decomposition (POD) method. By implicitly learning the flow-field information, the prediction error of the model is reduced by 29.7% compared with the single hidden layer ANN. Based on this model, the nonaxisymmetric, but periodic, hub optimization of a centrifugal impeller with 30 variables is carried out, with the goals of the higher efficiency and the wider flow range at the specified pressure ratio and the massflow rate at the design point. The result shows that, after the optimization, the isentropic efficiency at the design point increases by 1% and the flow range increases by 2% compared to the baseline.

Modeling of phenol removal from water by NiFe2O4 nanocomposite using response surface methodology and artificial neural network techniques

This study demonstrates the synthesis, characterization, and modeling of nickel ferrite nanocomposite, NiFe2O4 (NFC) as an adsorbent for the phenol contaminated aqueous environment. The characterization of the prepared NFC was performed with X-ray diffraction spectroscopy (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and vibrating sample magnetometer (VSM) techniques. The optimization and modeling of phenol removal using NFC was done through central composite design (CCD) and effective parameters of CCD were measured as input variables including the amount of NFC, pH, contact time and initial phenol concentration. The predicted results showed that the adsorption process using NFC as adsorbent had the maximum phenol removal (~99%) under predicted optimal conditions (pH = 7.67, NFC dosage = 0.15 g at room temperature), which also corresponded to the experimental values. In addition, a multilayer feed-forward artificial neural network (ANN) model was used to obtain a speculative phenol removal model. The network was trained for six replications after selection of the best neuron number for hidden layer. The value of MSE trained network was found to be 6.01718e-3 along with regression coefficient (R2 = 0.9934) that indicated satisfactory relationship. Isothermal modeling of phenol adsorption onto NFC was performed using well-known Temkin, Freundlich and Langmuir models and it was clear from the higher R2 value of 0.961 that the Langmuir model was significantly followed by experimental data. The maximum Langmuir adsorption capacity was found to be 274.72 mg/g at the optimal conditions. The obtained results prove that NFC could be an effective adsorbent for elimination of phenol contaminant from aqueous environment.

A Novel Genetic Algorithm Based Dynamic Economic Dispatch With Short-Term Load Forecasting

This article proposes an optimal energy scheduling method for power transmission networks using novel genetic algorithm (nGA) for solving the dynamic economic dispatch (DED) problem combined with machine learning based short-term load forecasting (STLF). The STLF is implemented based on a multilayer artificial neural network (MANN) to estimate the day-ahead variations in the demand. The efficacy of the proposed energy scheduling model together with the STLF is verified using a modified IEEE 30-bus system using real data of the power plants located in the Ereymentau region of Kazakhstan. The simulation results suggest that the proposed model offers a cost effective, reliable, and efficient dynamic energy scheduling in power transmission systems.

A Low-Cost System for Measuring Wind Speed and Direction Using Thermopile Array and Artificial Neural Network

Recent developments in wind speed sensors have mainly focused on reducing the size and moving parts to increase reliability and stability. In this study, the development of a low-cost wind speed and direction measurement system is presented. A heat sink mounted on a self-regulating heater is used as means to interact with the wind changes and a thermopile array mounted atop of the heat sink is used to collect temperature data. The temperature data collected from the thermopile array are used to estimate corresponding wind speed and direction data using an artificial neural network. The multilayer artificial neural network is trained using 96 h data and tested on 72 h data collected in an outdoor setting. The performance of the proposed model is compared with linear regression and support vector machine. The test results verify that the proposed system can estimate wind speed and direction measurements with a high accuracy at different sampling intervals, and the artificial neural network can provide significantly a higher coefficient of determination than two other methods.

Future ecosystem service value modeling with land cover dynamics by using machine learning based Artificial Neural Network model for Jashore city, Bangladesh

Land Use/Land Cover (LULC) provides provisional, supporting, cultural, and regulating ecosystem services that contribute to ecological environments, enhance human health and living, have economic advantages for sustaining living organisms. LULC transformation due to enormous urban expansion diminishing Ecosystem Services Values (ESVs) and discouraging sustainability. Though unplanned LULC transformation practice became more prevalent in developing countries, comprehensive assessment of LULC changes and their influences in ESVs are rarely attempted. This study aimed to illustrate and forecast the LULC changes and their influences on ESVs change in Jashore using remote sensing technologies. ESVs estimation and change analysis were conducted by utilizing -derived LULC data of the year 2000, 2010, and 2020 with the corresponding global value coefficients of each LULC type which are previously published. For simulating future LULC and ESVs, Land Change Modeler of TerrSet Geospatial Monitoring and Modeling Software was used in Multi-Layer Perceptron-Markov Chain and Artificial Neural Network method. The decline of agricultural land by 13.13% and waterbody by 5.79% has resulted in the reduction of total ESVs US$0.23 million (24.47%) during 2000–2020. The forecasted result shows that the built-up area will be dominant LULC in the future, and ESVs of provisioning and cultural services will be diminished by $0.107 million, $63400.3 by 2050 with the declination of agricultural, waterbody, vegetation, and vacant land covers. The study signifies the importance of a strategic rational land-use plan to strictly monitor and control the encroachment of built-up areas into vegetation, waterbodies, and agricultural land in addition to scientific mitigative policies for ensuring ecological sustainability.

Prediction of instantaneous real-world emissions from diesel light-duty vehicles based on an integrated artificial neural network and vehicle dynamics model

This paper presents a road vehicle emission model that integrates an artificial neural network (ANN) model with a vehicle dynamics model to predict the instantaneous carbon dioxide (CO2), nitrogen oxides (NOx) and total hydrocarbon (THC) emissions of diesel light-duty vehicles. Real-world measurement data were used to train a multi-layer feed-forward ANN model. The optimal combination of the various experimental variables was selected as the ANN input through a parametric study considering both practicality and accuracy. For CO2 prediction, two variables (engine speed and engine torque) are enough to develop an accurate ANN model. In order to achieve satisfactory accuracy for CO and NOx prediction, more variables were used for ANN training. The trained ANN model was used to predict road vehicle emissions by integrating the vehicle dynamics model, which was used as a supplementary tool to produce ANN input data. The integrated model is practical because it requires relatively simple data for input such as vehicle specifications, velocity, and road gradient. In the accuracy validation, the proposed model showed satisfactory prediction accuracy for road vehicle emissions.

Applying artificial neural networks to predict the enhanced thermal conductivity of a phase change material with dispersed oxide nanoparticles

The weak thermal conductance of a phase change material (PCM) can be intensified by dispersing nanostructured materials called nano-PCM. Accurate thermal conductivity (TC) prediction of nano-PCM is essential to evaluate heat transport during phase change processes, namely, melting and solidification. The present study develops an artificial neural network (ANN) to forecast the TC of n-octadecane as a PCM with dispersed oxide nanoparticles. A total of 122 experimental datasets from existing literature with a wide range of temperatures (5-60°C), nanoparticles (CuO, Al2O3, TiO2, and mesoporous SiO2), nanoparticle mass fractions (0.5-12 wt%) are compiled to train a multi-layered feed-forward ANN with Levenberg-Marquardt back-propagation algorithm. An optimal architecture of the neural network is acquired by varying the number of network hidden layers, the number of neurons in each layer, and the transfer function of layers. The minimum mean square error (MSE) of 1.3512 × 10−5 is obtained for the best developed ANN. Results show that average absolute deviation (AAD) of 0.002458, mean absolute percentage error (MAPE) of 0.8260%, and correlation coefficient (R) of 0.999964948 are achieved for training data. Moreover, MAPE, AAD, and R values are, respectively, 0.9478, 0.002167, and 0.9999715861 for testing data. The maximum percentage errors of ANN computed values are 2.31%, and 0.812% for liquid and solid phases, respectively. This indicates that the ANN model accurately predicts the enhanced TC of nano-PCM across various oxide nanoparticles, temperatures, and nanoparticle loadings.

Parametric analysis and minimization of entropy generation in bioinspired magnetized non-Newtonian nanofluid pumping using artificial neural networks and particle swarm optimization

Magnetohydrodynamic rheological bio-inspired pumping systems are finding new applications in modern energy systems. These systems combined the electrically conducting properties of flowing liquids with rheological behaviour, biological geometries and propulsion mechanisms. Further enhancements in transport characteristics can be achieved with the deployment of nanofluids. Second law thermodynamic analysis also provides a useful technique for optimizing thermal performance by minimizing entropy generation. In the present study, all these aspects are combined to analyze the heat transfer in magnetic viscoelastic nanofluid flow in a two-dimensional deformable channel containing a rigid porous matrix under peristaltic waves subject to a transverse magnetic field. The Williamson model is deployed for the nanofluid rheology and the Buongiorno model for nanoscale effects. Under lubrication approximations, the conservation equations for mass, momentum, energy and nanoparticle species are simplified. These partial differential equations are further non-dimensionalized using relevant transformation variables. The mathematical model is solved analytically by means of the Homotopy Analysis Method (HAM). Next, entropy generation is minimized by applying Particle Swarm Optimization (PSO) and Artificial Neural Networks (ANN). In the first phase, the equation for Entropy generation is derived as a function of temperature distribution, velocity profile utilizing geometrical and thermophysical parameters. The first step is to discover entropy generation to estimate some extraordinary influencing parameters. In the next step, some specific multi-layer perceptron ANNs are trained, which depend on the information from the first stage. In the last step, PSO in the considered peristaltic flow is used to minimize entropy generation. The optimized value (minimum) of entropy generation is 3.65 kJ/kg acquired at magnetic parameter (M) = 3, Brownian motion parameter (Nb) = 0.3, thermophoresis parameter (Nt)= 0.5 and Brinkman number (Br)=2 . Entropy generation is also very sensitive to both iteration number and magnetic field exhibiting a nonlinear topology.

Control of a hybrid upper-limb orthosis device based on a data-driven artificial neural network classifier of electromyography signals

This study presents the design of a functional electrical stimulation system (FES) for upper limb, including the construction and evaluation of an integral self-carrying rehabilitation device. The proposed device incorporated a hybrid system that performs the active mobilization control for the articulations or the assisted movement for the upper limb using electrical stimulation, as well as an on-line characterization of the electromyographic (EMG) signals captured in the trapezius and deltoid muscles. The orthosis was manufactured using a three-dimensional printer. The constructed device was electronically instrumented as a fully actuated robot, and it was controlled in a decentralized form by a set of state feedback (proportional-derivative) algorithms. This study proposed an interpolation method based on sigmoidal functions to solve the trajectory tracking for each actuated articulation. These algorithms used the estimated time-derivative of the tracking error provided by several explicit discretized super-twisting differentiators. The EMG signals were classified by a static multilayered artificial neural network trained with the Levenberg-Marquardt method, defining the movement intention triggered by the user. The device was tested in simulation software including the integration of several evaluation scenarios depending on the classified EMG signal and the tracking trajectory performance developed by the suggested state feedback controllers. The constructed orthosis was successfully evaluated with four volunteers showing the expected performance according to the proposed evaluation scenarios.

Leveraging artificial intelligence for more accurate and reliable predictions of anchors shear breakout capacity in thin concrete members

Understanding the shear behavior of post-installed anchors embedded in thin concrete members is considered complicated due to the many parameters involved including embedment depth, edge distance, concrete strength, anchor diameter, and member thickness. This study utilizes the Artificial Neural Network (ANN) concept to predict the capacity and behavior of adhesive and screw anchors with full-thickness embedment in thin concrete members and compare it with the adopted design model in ACI 318 standard. First, A multilayered feed-forward ANN model trained with Bayesian Regularization training algorithm is developed and compared against the proposed model based on the concrete capacity design method (CCD). Then, a parametric study is presented to evaluate the contribution of each variable to the shear capacity and ensure generalization. The study shows that ANN model provides a higher level of accuracy in predicting the shear capacity than CCD model by reducing the coefficient of variation from 18% to 11%. In addition, the parametric study shows that the CCD model does not capture the nonlinear effect of the variables on the capacity accurately and does not include the interaction effect between variables. For instance, CCD shows that the member thickness has a nearly linear effect on the capacity while ANN shows higher nonlinearity. In addition, the study presents a reliability analysis to assess the shift in safety level by adopting the ANN model. The ANN provided an average increase in reliability index by 32%. This study is considered the first study to utilize ANN to anchors embedded in thin concrete members with full embedment depth.