Feed Forward Back Propagation Research Articles

Feedforward neural network-based augmented salp swarm optimizer for accurate software development cost forecasting

This research proposes the use of feed-forward backpropagation neural networks (FFNN) to develop an accurate cost forecasting model in light of the challenges associated with forecasting software development costs (FSDC). The salp swarm algorithm (SSA) is first augmented and then employed to optimize the parameters of the developed FFNN predictor. A search enhancement mechanism and an elitism technique have been developed and incorporated into the SSA optimization process as two fresh and effective techniques for this goal. The search enhancement mechanism is employed to keep up a high rate of global exploration while also driving convergence towards the optimal area. Whereas elitism is used throughout the research phase to prevent stagnation in the local optima. Nineteen benchmark test functions and twelve benchmark software development cost forecasting data sets are utilized to assess the performance of the recommended enhancement techniques and developed algorithms. The results obtained from experiments show the superiority of the proposed techniques. In addition, the developed technique has been compared with many state-of-the-art methods, which demonstrates its durability. This is in addition to the statistical validation carried out on the results obtained, which also supports the robustness of the proposed technique.

Investigation of dual impact of nanoparticles-ethanol as additive to biodiesel-diesel fuel on an engine using artificial neural network prediction model

This paper investigates effects of the dual approach of adding nanoparticles and ethanol to palm biodiesel-diesel fuel at varied percentage as additive. Four samples namely: E10(100) B20, E15(100) B20, E10(100) B50, and E15(100) B50 were investigated on a single cylinder L70N Yanmar engine at constant speed of 2500 rpm under three load categories of 25 %, 50 % and 75 % for each sample. The experimental data was then feed-in as training data to develop an Artificial Neural Network (ANN) using the feed-forward back propagation algorithm (FFBP) and applying the Levenberg-Marquardt training pattern. Four (4) inputs were used viz: engine load, fuel, additive and cetane number. Correlation coefficient (R) and determination coefficient (R2) were used to validate performance of the network with two set of four hidden layers while targeting 5 outputs viz: BTE, HC, CO, NOx, and smoke. Experimental results indicate that at 25 % load the Brake Thermal Efficiency (BTE) values of samples E10(100) B20, E15(100) B20, E10(100) B50, and E15(100) B50 were 10.131 %, 14.972 %, 13.945 %, 14.091 % while at 50 % load it was 18.201 %, 25.101 %, 23.046 %, 23.779 % and at 75 % BTE values was 23.779 %, 32.011 % and 30.826 % which showed engine performance can be improved by magnetite nanoparticles addition with ethanol to biodiesel-diesel fuel up to 7.4 % in terms of BTE compared to only biodiesel-diesel blend fuel; emission was noticeably reduced for tested gases compared to previous literatures. The validated ANN results range 0.99350 to 0.99975 and indicated an acceptable level for the ANN prediction model for BTE. ANN exhaust emission prediction implicated smoke and NOx target performance was most accurate with R and R2 values at 0.9942, 0.9984, and 0.9980, 0.9960 respectively which indicated minimal error and hence the ANN model was satisfactory.

Boosting flood routing prediction performance through a hybrid approach using empirical mode decomposition and neural networks: a case study of the Mera River in Ankara

Abstract Flood routing is vital in helping to reduce the impact of floods on people and communities by allowing timely and appropriate responses. In this study, the empirical mode decomposition (EMD) signal decomposition technique is combined with cascade forward backpropagation neural network (CFBNN) and feed-forward backpropagation neural network (FFBNN) machine learning (ML) techniques to model 2014 floods in Ankara, Mera River. The data are split in order to avoid the underfitting and overfitting problems of the algorithm. While establishing the algorithm, 70% of the data were divided into training, 15% testing and 15% validation. Graphical indicators and statistical parameters were used for the analysis of model performance. As a result, the EMD signal decomposition technique has been found to improve the performance of ML models. In addition, the EMD-FFBNN hybrid model showed the most accurate estimation results in the flood routing calculation. The study's outputs can assist in designing flood control structures such as levees and dams to help reduce flood risk.

Appraisal of rock dynamic, physical, and mechanical properties and forecasting shear wave velocity using machine learning and statistical methods

Direct determination of shear wave velocity requires time, cost, and high accuracy due to the complexity of the rock texture. In current research statistical and intelligent approaches have been used to predict the shear wave velocity of rock samples. Also, a new correlation between dynamic and static rock properties was established and the shear wave velocity was estimated based on index tests using Gaussian process regression, multivariate linear regression, feedforward back-propagation artificial neural network, and K-nearest neighbor methods. In total, 120 data related to limestone and sandstone samples of the main projects were used for modeling. Water absorption, compressional wave velocity, and density were used as inputs. The outcomes revealed that the PW/SW ratio is equal to 1.69. Various statistics were used to check the method results. The statistical results showed that it is possible to forecast Ed, and SW with high accuracy. Also, the precision of the GPR was higher than the FBP-ANN, statistical analysis, and KNN. Estimation of SW by GPR showed R of 0.992, and RMSE of 0.06, respectively. These four methods were able to estimate the SW with a mean variation percentage of +0.19%. It's important to consider that these models are best suited for predicting SW when the predictor indicators fall within the same range as this study.

Machine learning-based modelling using ANN for performance prediction of a solar air heater design with jet impingement

The prediction of heat transfer performance for solar air heater (SAH) designs is important to assess its potential without relying on extensive experimental studies. This work aimed to predict the heat transfer characteristics of a SAH design with jet impingement using an extensive artificial neural network (ANN) study and covering a complete year-long ambient and various operating parameters (commensurate and non-commensurate). Total 84 ANN models were developed using different network types, transfer functions, and training functions. They were tested on 174 representative datasets collected throughout a year. The best ANN model was found to be the feed forward backpropagation model with tangent-sigmoid transfer function and LM training algorithm, which produced accurate predictions for thermal efficiency, exergy efficiency, and thermo-hydraulic efficiency, with sample errors within ±5 %. The numerical values of R2, MAE, RMSE, and COV from the ANN model results were respectively 0.9972, 0.9053, 1.1884, and 2.0618 for thermal efficiency, 0.9947, 0.2631, 0.3363, and 1.7157 for exergy efficiency, and 0.996, 0.0139, 0.0193, and 0.8076 for thermo-hydraulic efficiency. The study concluded that ANN models can accurately predict the heat transfer characteristics of an SAH design, requiring less time and less computational resources compared to conventional experimental or computational methods.

Jellyfish optimized recurrent neural network for state of health estimation of lithium-ion batteries

The state of health (SOH) of lithium-ion batteries is considered an important health indicator to evaluate different features such as efficiency, robustness, and accuracy. The SOH of the lithium-ion battery is investigated to determine the advent of failures to mitigate the battery risk, avoid uneven fault occurrences and ensure safe battery operation. Nonetheless, the accurate assessment of SOH is difficult due to the occurrence of capacity degradation and performance deviation with temperature and aging impacts. Therefore, this work presents an improved hybrid recurrent neural network (RNN) and jellyfish optimization (JFO) to estimate the SOH of lithium-ion batteries. A total of 124 lithium-ion battery cells classified in three batches, namely ‘2017-05-12’, ‘2017-06-30’, and ‘2018-04-12’ from the MIT Stanford- Toyota Research Institute battery database, are used for the proposed work. The systematic sampling method is utilized for developing a 31-dimensional data format. The 31-dimensional data format is developed considering various critical battery parameters such as capacity, voltage, current, and temperature. The JFO is applied to achieve the best suitable RNN model hyperparameters for accurate estimation outcomes. The validation of the hybrid RNN-JFO method is conducted with other JFO-optimized neural network (NN) models, such as feedforward neural network (FFNN) and backpropagation neural network (BPNN). Compared with FFNN-JFO and BPNN-JFO models, the experiment conducted with the RNN-JFO method delivered accurate outcomes. For the batches ‘2017-05-12’, ‘2017-06-30’, and ‘2018-04-12’, the average root mean square error (RMSE) error achieved with the RNN-JFO model was 0.0529, 0.2674, and 0.1638. The SOH estimation outcomes confirm the effectiveness of the RNN-JFO method for SOH estimation on different lithium-ion batteries.

A Meander Delay System Design Using Caps-Triple GAN Optimized with the Remora Optimization Algorithm for IOT Application

In this manuscript, a Meander Delay System design using Caps-Triple GAN optimized with the Remora Optimization Algorithm (ROA) is proposed for the Internet of Things (IoT) application. For designing the Meander Microstrip Delay Line (MMDL) Antenna, the Capsule Triple generative adversarial network (Caps-Triple GAN) is used. The Caps-Triple GAN error and delay parameters are optimized using the ROA. Hence, the designed MMDL Antenna is available in the IoT application. The proposed method is executed in the commercial Sonnet software package. The performance of the proposed method is examined under performance metrics, such as delay time with resistance characteristic, one-step delay time resistance characteristic, delay time, one-step delay time, error, and speed. The prediction results of the proposed MMDL-CAPS-Triple GAN-ROA method predict lower delay times of 14.77%, 15.89%, and 11.50%, higher speed of 32.57%, 16.25%, and 8.44% compared with the existing methods, such as Prediction of the meander delay system parameters for internet-of-things devices utilizing a Pareto-optimal artificial neural network and multiple linear regression (MDS-ANN-MLR), Predicting the frequency characteristics of hybrid meander systems utilizing the feed-forward back propagation network (MDS-FFBPN), Frequency Characteristic Analysis of Meander Structures with Different Connecting Electrodes (MDS-MoM), respectively.

Comparison of Swarm-based Metaheuristic and Gradient Descent-based Algorithms in Artificial Neural Network Training

This paper aims to compare the gradient descent-based algorithms under classical training model and swarm-based metaheuristic algorithms in feed forward backpropagation artificial neural network training. Batch weight and bias rule, Bayesian regularization, cyclical weight and bias rule and Levenberg-Marquardt algorithms are used as the classical gradient descent-based algorithms. In terms of the swarm-based metaheuristic algorithms, hunger games search, gray wolf optimizer, Archimedes optimization, and the Aquila optimizer are adopted. The Iris data set is used in this paper for the training. Mean square error, mean absolute error and determination coefficient are used as statistical measurement techniques to determine the effect of the network architecture and the adopted training algorithm. The metaheuristic algorithms are shown to have superior capability over the gradient descent-based algorithms in terms of artificial neural network training. In addition to their success in error rates, the classification capabilities of the metaheuristic algorithms are also observed to be in the range of 94%-97%. The hunger games search algorithm is also observed for its specific advantages amongst the metaheuristic algorithms as it maintains good performance in terms of classification ability and other statistical measurements.

A Discussion on Geodynamic Modeling Methodology: Inferences from Numerical Models in the Anatolian Plate

Numerical models have found widespread use in geosciences, mainly due to high-resolution datasets and the development of supercomputing facilities with powerful data processing and storage capabilities during the past two decades. Instantaneous and time-dependent geodynamic modeling studies were carried out in many regions of the Alpine-Himalayan orogenic belt, including the Anatolian Plate, to investigate mantle dynamics such as lower lithosphere deformation, upper mantle flow, and their surface implications. 
 This study focuses on the instantaneous numerical modeling technique by considering multidimensional thermomechanical models in the Central and East Anatolian plateaus. To this end, conventional geodynamic modeling processes are explained with a conceptual flow chart that shows a feed-forward backpropagation modeling architecture which is nonlinearly fed by a large parameter space. While addressing a complex natural phenomenon controlled by variables on a wide range of space-time scales, the limitations as well as advantages of numerical models are analyzed. 
 In addition to conventional techniques, systematic data improvement is discussed as a new strategy in data/parameter-dependent numerical model design through an iterative process based on the Grounded Theory method for the construction of an explanatory theory from the model. This involves not refinement but (re)construction of the data (i.e., measurement/analysis/scaling) as an effective way to reveal theory/information grounded in data. 
 It is speculated that this procedure, which includes problem-oriented data reconstruction accompanying the numerical modeling process, may provide an integrated perspective for instantaneous numerical modelling.

Prediction of Flyrocks, Airblasts and Ground Vibrations Using Neural Computing and Applications at ZCDC Mine

When an explosive detonates in a blasthole, approximately 20 to 30% of the energy is only utilized for fragmenting the rock mass whilst the bulk of the energy is lost in the form of ground vibrations, flyrocks and airblasts. Employees’ residences, mine offices, processing plants and engineering workshops built close to the mining area and are in danger of being damaged by blast induced hazards. In addition, due to flyrocks there was a high level of ore losses, ore dilution, equipment damages, mine roads and powerlines. To assess and reduce these negative impacts, monitoring of flyrocks, ground vibrations and airblast was carried out and generate a prediction model. A three-layer, feed-forward back-propagation of a 9-10-3 network architecture was trained, validated, and tested using the Bayesian regularization algorithm. A total of 100 monitored blast records obtained from the mine were used as input parameters for the ANN prediction model. Subsequently a Multivariate Regression analysis prediction model was run and used to compare with the results obtained from ANN model. Based on the study's findings, an ANN model proved to be the best for field predictions.To determine the relative impact of each input parameter on flyrocks, gound vibration, and Airblast, a sensitivity analysis was also carried out and lastly blast optimization which managed to reduce blast induced impacts by over 30%.

Modelling and Energy Optimization of a Thermal Power Plant Using a Multi-Layer Perception Regression Method

This research paper studies a thermal power plant model with an Artificial Neural Network that contributes to the accuracy improvement of actual measurement data. Neural Networks process the paradigm of algebraic expressions, and their training occurs via a Feed-Forward Back Propagation algorithm implemented in a MATLAB environment. The applied training case in a thermal power plant in Paracha includes three different algorithms, the Levenberg-Marquadt, the Scaled Conjugate Gradient, and the Bayesian Regularization, considering less number of samples to achieve more reliable results. The outcome highlights Bayesian Regularization Networks’ superiority in accuracy and performance compared to Levenberg-Marquadt and the Scaled Conjugate Gradient. The regression analysis estimates the relationship between input-independent and output-dependent variables, forecasts the energetic data, and highlights the benefits of the Bayesian Regularization method in the energy sector.

Harnessing solar power: Predicting photovoltaic potential in fiche, oromia, ethiopia with artificial neural networks

In developing nations like Ethiopia, especially in rural and semi-urban regions, reliable access to electricity through grid-based systems is often limited or unavailable due to inadequate energy infrastructure. To tackle this issue, photovoltaic (PV) technology plays a crucial role in converting solar energy into electricity, especially in remote regions. This study focuses on assessing the daily and monthly energy exposure of PV arrays located near Fiche town, Ethiopia. The study utilizes four network types - Feedforward Backpropagation (FFBP), Cascade Feedforward Backpropagation (CFBP), Error Backpropagation (EBP), and Levenberg-Marquardt (LR) - with eight input variables and a training-testing split of 70%-30% to estimate global solar radiation (GSR). Hidden and output layers are employed with a sigmoid and linear function as activation functions, respectively. The results reveal that the expected mean daily and monthly horizontal global solar radiation (HGSR) ranges between 3.282 kWh/m2/day ≤ HGSR ≤ 6.967 kWh/m2/day and 4.628 kWh/m2 ≤ HGSR ≤ 6.613 kWh/m2, respectively. Furthermore, the study determines that the daily energy delivered to the PV system and available from the PV array to load falls within the range of 0.656 kWh/m2 ≤ Ed ≤ 1.394 kWh/m2 and 0.567 kWh/m2 ≤ El ≤ 1.205 kWh/m2, respectively. The average results for the study period from 2016 to 2021 were 1.175 kWh/m2 and 1.015 kWh/m2 for the Ed and El, respectively. These results demonstrate that PV arrays near Fiche town have abundant solar energy resources, making them highly attractive for electricity generation. Furthermore, utilizing solar energy from PV arrays can help mitigate indoor pollution in the rural communities surrounding Fiche town.

Application of Artificial Neural Network Model for Improved Power System Protection in Port Harcourt 33 kV Network


 
 
 Electric power distribution networks are exposed to the environment due to their length, for this reason, they are mostly affected by faults. These faults disrupt the continuous flow of power supply. There is also an associated loss of power that is generated that also determines the state of the economy. In order to reduce system downtime, it is necessary to integrate a system that detects and classifies faults quickly in order to hasten its clearance. This will bring about improvement in the efficiency and integrity of the power network. The artificial neural network as proposed in this study is meant to detect, classify and locate fault on the Rukpokwu 33-kV
 feeder of Port Harcourt Electricity Distribution Company (PHEDC). Fault detector, classifier and locator, with feed-forward back propagation were employed in the research. Matlab Simulink software was used to model and simulate the distribution network. The model was trained using s values of voltages and currents. With the simulation results, the efficiency of the proposed network was demonstrated for fault detection, classification and location. Mean square error (MSE) and confusion matrix were used to evaluate the performance of the proposed model. Results showed that the acceptable MSE of 0.00000027736 and an accuracy of 100% were achieved, which is satisfactory.
 
 

Modeling of Dry EDM process parameters during machining of Inconel 718 using artificial neural network

Dry EDM is a modified technique with substituted gaseous medium instead of liquid dielectrics in the usual EDM phase. In this work, dry EDM attempt is made with copper electrode on Inconel 718 material using L27 orthogonal array with compressed air as a dielectric medium. This study takes into account process variables like gas dielectric pressure, pulse current, spark on time, and gap spark voltage. ANN models is developed using Feed-forward back propagation algorithms with trainlm, learngdm, MSE, transig as the training, learning, performance and transfer functions respectively for the Material Removal Rate (MRR), Tool Wear Rate (TWR) and Surface roughness (Ra) as a response characteristics. A good fitness is observed for the trained model. Experimental response values and the ANN predicted values is compared and found that the overall correlation coefficient was 0.94455 which shows good prediction accuracies and effectiveness of the model and ANN technique.

Artificial neural network model of non-Darcy MHD Sutterby hybrid nanofluid flow over a curved permeable surface: Solar energy applications

The conversion of solar radiation to thermal energy has recently much interest as the requirement for renewable heat and power grows. Due to their enhanced ability to promote heat transmission, nanofluids can significantly improve solar-thermal systems' efficiency. This section aims to study the heat transfer behavior of the Sutterby hybrid nanofluid flow of magnetohydrodynamics in the presence of a non-uniform heat source/sink and linear thermal radiation over a non-Darcy curved permeable surface. A novel implementation of an intelligent numerical computing solver based on multi-layer perceptron (MLP) feed-forward back-propagation artificial neural network (ANN) with the Levenberg-Marquard algorithm is provided in the current study. Data were gathered for the ANN model's testing, certification, and training. Established mathematical equations are nonlinear, which are resolved for velocity, the temperature in addition to the skin friction coefficient, and the rate of heat transfer by using the bvp4c with MATLAB solver. The ANN model selects data, constructs and trains a network, then evaluates its efficacy via mean square error. Graphs illustrate the impact of a wide range of physical factors on variables, including pressure, velocity, and temperature. In the entire study, the thermal energy improved by the SiO2 (silicon dioxide) - Au (gold) hybrid nanofluid than the SiO2-TiO2 (titanium dioxide) hybrid nanofluid. The higher internal heat generation/absorption parameter values increase the temperature.

Analysis of artificial intelligence approaches to predict the wall deflection induced by deep excavation

Abstract The geological condition of Ho Chi Minh (HCM) City is soft soil and high groundwater and includes two main structural layers such as Pleistocene and Holocene sediments. Therefore, deep excavation of all the high-rise buildings in the city is usually supported by concrete retaining walls such as the diaphragm or bored pile retaining walls. The system limits the excavation wall deflection during the construction process which could pose a potential risk to the construction and neighborhood areas. To estimate wall deformation at a highly accurate and efficient level, this study presents several machine learning models including feed-forward neural network back-propagation (FFNN-BP), long short-term memory (LSTM), bidirectional long short-term memory (Bi-LSTM), and support vector regression (SVR). The database for the experiment was obtained from a high building in HCM City, Vietnam. The database is deployed to implement the proposed algorithms in walk-forward validation technique. As a result, the Bi-LSTM model reduced prediction errors and improved robustness than the LSTM, FFNN-BP, and SVR models. Bi-LSTM, LSTM, and FFNN-PB could be used for predicting deep excavation wall deflection. In the meantime, not only could the estimated results support safety monitoring and early warning during the construction stages but also could contribute to legal guidelines for the architecture of deep excavations in the city’s soft ground.

Exploring the Influence of Induced Magnetic Fields and Double-Diffusive Convection on Carreau Nanofluid Flow through Diverse Geometries: A Comparative Study Using Numerical and ANN Approaches

This current investigation aims to explore the significance of induced magnetic fields and double-diffusive convection in the radiative flow of Carreau nanofluid through three distinct geometries. To simplify the fluid transport equations, appropriate self-similarity variables were employed, converting them into ordinary differential equations. These equations were subsequently solved using the Runge–Kutta–Fehlberg (RKF) method. Through graphical representations like graphs and tables, the study demonstrates how various dynamic factors influence the fluid’s transport characteristics. Additionally, the artificial neural network (ANN) approach is considered an alternative method to handle fluid flow issues, significantly reducing processing time. In this study, a novel intelligent numerical computing approach was adopted, implementing a Levenberg–Marquardt algorithm-based MLP feed-forward back-propagation ANN. Data collection was conducted to evaluate, validate, and guide the artificial neural network model. Throughout all the investigated geometries, both velocity and induced magnetic profiles exhibit a declining trend for higher values of the magnetic parameter. An increase in the Dufour number corresponds to a rise in the nanofluid temperature. The concentration of nanofluid increases with higher values of the Soret number. Similarly, the nanofluid velocity increases with higher velocity slip parameter values, while the fluid temperature exhibits opposite behavior, decreasing with increasing velocity slip parameter values.

Optimization of Monopole Consecutive Square-Shaped Patch Antenna with ANN for Wireless Communicaton Systems

In recent years, Artificial Neural Network (ANN) applications in antenna structures have gained significant traction due to their potential to reduce design and calculation times, offering optimization and prediction capabilities. This study introduces a novel monopole patch antenna design featuring a consecutive square-shaped broadband microstrip-line-fed antenna. The proposed antenna exhibits an impressive impedance bandwidth of 68% (1.55 - 2.82 GHz), a remarkable return loss of -47.25 dB, and a directivity gain of 2.77 dBi. Simulation studies were conducted using CST™ Studio Suite electromagnetic simulation software. The ANN model developed based on feed-forward backpropagation demonstrates exceptional agreement with the simulation results, showcasing an accuracy of 99.61805% and performing 2769.231 times faster. With advancing technology, ANNs present an effective solution for addressing complex antenna design challenges arising from escalating data rate requirements and uninterrupted data transmission. These results open promising avenues for further advancements in antenna design aided by ANN applications.

A new comprehensive model of thermal conductivity for hydrofluoroolefins refrigerants using feed-forward back-propagation neural networks

A new comprehensive model of thermal conductivity for hydrofluoroolefins refrigerants using feed-forward back-propagation neural networks

Rule-Based Non-Intrusive Load Monitoring Using Steady-State Current Waveform Features.

Monitoring electricity energy usage can help to reduce power consumption considerably. Among load monitoring techniques, non-intrusive load monitoring (NILM) provides a cost-efficient solution to identify individual load consumption details from the aggregate voltage and current measurements. Existing load monitoring techniques often require large datasets or use complex algorithms to obtain acceptable performance. In this paper, a NILM technique using six non-redundant current waveform features with rule-based set theory (CRuST) is proposed. The architecture consists of an event detection stage followed by preprocessing and framing of the current signal, feature extraction, and finally, the load identification stage. During the event detection stage, a change in connected loads is ascertained using current waveform features. Once an event is detected, the aggregate current is processed and framed to obtain the event-causing load current. From the obtained load current, the six features are extracted. Furthermore, the load identification stage determines the event-causing load, utilizing the features extracted and the appliance model. The results of the CRuST NILM are evaluated using performance metrics for different scenarios, and it is observed to provide more than 96% accuracy for all test cases. The CRuST NILM is also observed to have superior performance compared to the feed-forward back-propagation network model and a few other existing NILM techniques.