Feed-forward Back-propagation Neural Network Research Articles

Short Term Renewable Energy Forecasting Based on Feed Forward Back Propagation Neural Network Strategy

The fundamental inputs used as a renewable energy source are wind speed and solar radiation. Both parameters are very nonlinear and depending on their surroundings. As a result, reliable prediction of these characteristics is required for usage in a variety of agricultural, industrial, transportation, and environmental applications since they reduce greenhouse gas emissions and are environmentally benign. In this study, we used a Feed Forward Back Propagation Neural Network (FFBPN) technique to predict proper data such as temperature, relative moisture, sun radiations, rain, and wind speed. The FFBPN will be trained in such a way that it can conduct hybrid forecasting with little changes to the programming codes, ranging from hourly (short term forecasting) to daily forecasting (medium term forecasting). This feature is one of the significant improvements, showing the suggested hybrid renewable energy forecasting system's high robustness. Because the hybrid forecasting system is a unique approach, the system's accuracy will be determined by comparing the findings to the corresponding values of the persistent model, a stand-alone forecasting model. Finally, the completely created system package could be sold and/or used in future research initiatives to help researcher’s analyses, validate, and illustrate their models across a variety of areas.

Ultrasound assisted osmotic dehydration of dragon fruit slices: Modeling and optimization using integrated artificial neural networks and genetic algorithms

The effect of ultrasonication on osmotic dehydration of dragon fruit (Hylocereus undatus) slices was investigated using a feed forward backpropagation neural network in this study. The four process variables were ultrasonic power ( X UP $$ {X}_{UP} $$ : 100–300 W), ultrasound temperature ( X UT : $$ {X}_{UT}: $$ 30–50°C), osmotic solution concentration ( X SC $$ {X}_{SC} $$ : 40–60°Brix), and liquid to solid ratio ( X LS $$ {X}_{LS} $$ : 8:1–12:1 ml/g). Under different combinations of independent variables, the water loss ( Y WL $$ {Y}_{WL} $$ ), solid gain ( Y SG $$ {Y}_{SG} $$ ), and change in color ( Y ∆ E $$ {Y}_{\Delta E} $$ ) were evaluated. The ANN architecture consists of 4, 9, and 3 neurons in the input, hidden, and output layer, respectively. The three independent variables of ultrasound-assisted osmotic dehydration X UP $$ {X}_{UP} $$ , X UT $$ {X}_{UT} $$ , and X SC $$ {X}_{SC} $$ found to have a positive effect on the three responses water loss, solid gain, and change in color. The liquid to solid ratio was revealed to have a negative effect on color change and a positive impact on response water loss and solid gain. The output parameters of the ANN model (weights and bias) were integrated with genetic algorithm (GA) for the optimization purpose, and the best combination was picked based on the fitness value (FV). The optimal ultrasound-assisted osmotic dehydration condition obtained using integrated ANN-GA was 235 W, 38.74°C, 56°Brix, and 10.5 ml/g. The relative deviation between the experimental and integrated ANN-GA model predicted values for water loss, solid gain, and color change values under optimum conditions was less than 6.880%, suggesting that the predicted values for response were in good agreement with the experimental data. Practical applications The article describes the synergistic influence of sonication power and temperature on the osmotic dehydration of dragon fruit slices. Ultrasound-assisted osmotic dehydration (USOD) prevents enzymatic and oxidative browning, preserving the color. In dragon fruit slices flavor retention increases when sugar syrup is used as an osmotic agent. Furthermore, the fruit pieces absorb sugar and eliminate acid, resulting in a sweeter product than a conventionally dried product. The results showed that ultrasound assisted osmotic dehydration removed water from dragon fruit more rapidly at a lower temperature, keeping color, aroma, nutritional value, and flavor components. The combined impact of ultrasound and temperature also lowered the water activity of fruit and enhanced the product's storage life significantly.

Detection of transversal cracks in prismatic cantilever beams with weak clamping using machine learning

Because our infrastructure is aging and approaching the end of its intended functioning time, the detection of damage or loosening of joints is a topic of high importance in structural health monitoring. The most desired way to assess the health of engineering structures during operation is to use non-destructive vibration-based methods that can offer a global evaluation of the structure’s integrity. A comparison of using different modal data for training feedforward backpropagation neural networks for detecting transverse damages in beam-like structures that can also be affected by imperfect boundary conditions is presented in the current paper. The different RFS, RFSmin, and DLC training datasets are generated by applying an analytical method, previously developed by our research team, that uses a known relation, based on the modal curvature, severity estimation of the transverse crack, and the estimated severity for the weak clamping. The obtained dataset values are employed for training three feedforward backpropagation neural networks that will be used to locate transverse cracks in cantilever beams and detect if the structure is affected by weak clamping. The output from the three ANN models is compared by plotting the calculated error for each case.

Standalone Smart Irrigation System: An Analysis of Reliability Incorporating BPNN Technique under Various Failures

Observing the huge requirements of water and energy for traditional irrigation and contracting resources, the utility of a smart irrigation system has become necessary. A system that is capable of saving water depends on renewable energy resources for power requirements and runs automatically. Water-saving automatic irrigation systems powered by renewable energy could be a solution to India's rising demand for energy and water. In the current paper, the author has proposed an analysis of a standalone smart irrigation system for its reliability and functional behaviour under various operational events of failures and repairs. The authors have implemented an intelligent computational method because of the increased degradation in system function due to the complex structure of the proposed irrigation system. The system's mathematical formulation is stated in terms of neural networks, and a back propagation neural network technique is applied for fast and comparatively better outcomes. The author has determined state probabilities and other reliability parameters by means of neural networks in the proposed smart irrigation system. To improve the accuracy and consistency of reliability parameters, a Feed Forward Back Propagation Neural Network (FFBPNN) is applied. FFBPNN's learning mechanism can optimize the values of parameters by modifying neural weights. The MATLAB codes are used by the authors to demonstrate the numerical examples, and iterations are repeated until the precision in error tends to 0.0001. The sensitivity and cost of the system are also analyzed, which can help in managing the real-time operations of the system.

Performance Assessment of Solar-Transformer-Consumption System Using Neural Network Approach

Solar energy is one of the immeasurable renewable energy in power generation for a green, clean and healthier environment. The silicon-layer solar panels absorb sun energy and converts it into electricity by off-grid inverter. Electricity is transferred either from this inverter or from transformer, consumed by consumption unit(s) available for residential or economic purposes. The artificial neural network is the foundation of artificial intelligence and solves many complex problems which are difficult by statistical methods or by humans. In view of this, the purpose of this work is to assess the performance of the Solar - Transformer - Consumption (STC) system. The system may be in complete breakdown situation due to failure of both solar power automation subsystem and transformer simultaneously or consumption unit; otherwise it works with fully or lesser efficiency. Statistically independent failures and repairs are considered. Using the elementary probabilities phenomenon incorporated with differential equations is employed to examine the system reliability, for repairable and non-repairable system, and to analyze its cost function. The accuracy and consistency of the system can be improved by feed forward- back propagation neural network (FFBPNN) approach. Its gradient descent learning mechanism can update the neural weights and hence the results up to the desired accuracy in each iteration, and aside the problem of vanishing gradient in other neural networks, that increasing the efficiency of the system in real time. MATLAB code for FFBP algorithm is built to improve the values of reliability and cost function by minimizing the error up to 0.0001 precision. Numerical illustrations are considered with their data tables and graphs, to demonstrate and analyze the results in the form of reliability and cost function, which may be helpful for system analyzers.

The Synergy of Data From Profiling Floats, Machine Learning and Numerical Modeling: Case of the Black Sea Euphotic Zone

AbstractData from profiling floats in the Black Sea revealed complex temporal and spatial relationships between physical variables and oxygen, chlorophyll and the backscattering coefficient at 700 nm, as well as some limits in understanding the details of biogeochemistry dynamics. To account for different interdependences between physical and biogeochemical properties, a feedforward backpropagation neural network (NN) was used. This NN learns from data recorded by profiling floats and predicts biogeochemical states using physical measurements only. The performance was very high, particularly for oxygen, but it decreased when the NN was applied to older data because the interrelationships between the physical and biogeochemical properties have changed recently. The biogeochemical states reconstructed by the NN using physical data produced by a coupled physical–biogeochemical operational model were better than the biogeochemical outputs of the same coupled model. Therefore, the use of data from profiling floats, physical properties from numerical models and NNs appears to be a powerful approach for reconstructing the 4D dynamics of the euphotic zone. Basin‐wide patterns and temporal variabilities in oxygen, backscattering coefficient and chlorophyll were also analyzed. Of particular interest is the reconstruction of short‐lived biogeochemical features, particularly in coastal anticyclone areas, which are difficult to observe with available floats at the basin scale.

Prediction of long lead monthly three-dimensional ocean temperature using time series gridded Argo data and a deep learning method

• A deep learning model is proposed to predict long lead monthly 3D ocean temperature. • Nearly 17-year monthly time series gridded Argo data with 58 depth layers are used. • A case study is demonstrated in the South Pacific Ocean from multiple perspectives. • The model accurately predicts 1 year ahead ocean temperature from 0 to 1975dbar. • The model outperforms three baseline models on multiple metrics. Ocean temperature is a vital physical variable of the oceans. Accurately predicting the long lead dynamics of the three-dimensional ocean temperature (3D-OT) can help us identify in advance potential extreme events (e.g., droughts and floods) that may be caused by the changes of the 3D-OT, which however remains a challenge. To achieve this goal, a deep learning (DL) model was proposed to make predictions of the monthly 3D-OT for one year ahead using time series gridded Argo data. The DL model is comprised of a one-dimensional convolution (Conv1D) layer which is used for extracting latent features from the time series ocean temperature data, two long short-term memory (LSTM) layers which are used for capturing the long-term temporal dependencies hidden in the 3D-OT based on the features extracted by the Conv1D layer, and a fully-connected layer to output the predictions. The proposed DL model can well model the temporal dependencies and dynamic patterns of the ocean temperature at different spatial locations and in different depths by learning from simply the historical time series gridded Argo data. Experiments conducted in a sub-area of the South Pacific Ocean that predict the monthly 3D-OT with the lead time from 1 to 12 months show that the developed DL model surpasses the persistence model, the AdaBoost model, and the feedforward backpropagation neural network model (BPNN) when compared from multiple spatiotemporal perspectives using multiple statistics, indicating that the proposed DL model is a highly strong model for long lead monthly 3D-OT predictions.

An Efficient Estimation of Wind Turbine Output Power Using Neural Networks

Wind energy is a valuable source of electric power as its motion can be converted into mechanical energy, and ultimately electricity. The significant variability of wind speed calls for highly robust estimation methods. In this study, the mechanical power of wind turbines (WTs) is successfully estimated using input variables such as wind speed, angular speed of WT rotor, blade pitch, and power coefficient (Cp). The feed-forward backpropagation neural networks (FFBPNNs) and recurrent neural networks (RNNs) are incorporated to perform the estimations of wind turbine output power. The estimations are performed based on diverse parameters including the number of hidden layers, learning rates, and activation functions. The networks are trained using a scaled conjugate gradient (SCG) algorithm and evaluated in terms of the root mean square error (RMSE) and mean absolute percentage error (MAPE) indices. FFBPNN shows better results in terms of RMSE (0.49%) and MAPE (1.33%) using two and three hidden layers, respectively. The study indicates the significance of optimal selection of input parameters and effects of changing several hidden layers, activation functions, and learning rates to achieve the best performance of FFBPNN and RNN.

Welding sequence optimization to reduce welding distortion based on coupled artificial neural network and swarm intelligence algorithm

This study aims to develop a welding sequence optimization (WSO) framework based on coupled artificial neural network (ANN) and swarm intelligence algorithm for minimizing welding distortion of thin-walled squared Al–Mg–Si alloy tube components. This framework is mainly composed of two critical computer programs. Firstly, a multilayer feedforward backpropagation neural network (BPNN) system was established to rapidly estimate residual distortion for an arbitrary welding sequence so that welding sequence can be optimized for achieving desired welding quality. For this purpose, a series of nonlinear thermo-elastic–plastic finite element (FE) simulations were conducted and verified with experiments to generate the input database of the neural network. Subsequently, a reliable BPNN model was successfully created and trained within an acceptable error. Secondly, a novel swarm intelligence algorithm, namely, bees algorithm (BA) was proposed to solve the complicated WSO problems. In this optimization process, the trained BPNN model was implanted into this proposed BA for computing the fitness value of arbitrary welding sequences. Moreover, welding experiments were also performed to confirm the performance of the proposed optimization method. Comparing the results from experimental measurements, FE simulations, and proposed WSO framework, it is demonstrated that this proposed BPNN-and-BA-based WSO framework can be successfully applied in practical engineering to obtain an optimal welding sequence for minimizing final welding distortion.

2x2 Array Microstrip Patch Antenna Designing and Validation with Artificial Neural Networks

A novel and successful method for RF, microwave modelling, and antenna design has gained attention: computational modules based on neural networks. The behavior of active or passive components or circuits can be taught to neural networks. The design of a microstrip patch antenna utilizing an artificial neural network (ANN) is discussed in this research. Computer Simulation Technology (CST) is used to establish the antenna's size and characteristics in order to create a successful ANN model. The feed-forward back-propagation neural network and Levenberg Marquardt optimization algorithm are used to model the antenna design. High-Frequency Structure Simulator (HFSS) software, operating at 2.4 GHz, is used to build the initial microstrip patch (ISM band).A 2.4GHz operating frequency is used to create a 2*2 microstrip planar array. A variety of neural networks are evaluated and trained to produce the most ideal results after generating and modelling discoveries using HFSS simulation software and the Finite Difference Time Domain (FTDT) approach. In this research, neural networks are utilized. In order to identify the most ideal solution, optimization is performed using a radial basis function neural network (RBF NN) and a feed-forward backpropagation approach.

Artificial neural network model of co-gasification of petroleum coke with coal or biomass in bubbling fluidized bed

Co-gasification of petroleum coke (petcoke) with coal or biomass in fluidized bed is a promising way to avoid environmentally problems caused by its discharge. To give more accurately prediction in this process, feed-forward back propagation neural network (FFBPNN) with three optimization algorithms were conducted, including Levenberge Marquardt (LM), genetic algorithm (GA) and particle swarm optimization (PSO). The input parameters in the ANN model were petcoke ratio (W), equivalence ratio (ER), steam flow rate (S), particle diameter (Dp), volatiles (V), and fixed carbon (FC). And the output data were carbon conversion (X), ratio of H2/CO, LHV of syngas and gas yield (Q). The predicted data showed a good agreement with the experimental results. The PSO showed much better performances than those of LM and GA. With ER increased, the predicted X increased and the ratio of H2/CO decreased. But they were almost no changed with Dp increased. The contributive ratio of W was the largest (0.37) at petcoke ratio of 20%. The contributive ratio of ER increased not the same ratio as ER increased. The contributive ratio of particle size (Dp) almost not changed with Dp increased.

Artificial neural networks (ANNs) for modeling efficient factors in predicting pap smear screening behavior change stage.

Background and objectivesCervical cancer is ranked as the third most prevalent cancer that affects women all over the world and Pap smear seems to be the single most critical intervention to prevent cervical cancer. In the present study, the effects of demographic factors (age, education level, job, income level, marriage age, pregnancy, child number, breastfeeding, and menopause), insurance type, disease history and screening (sono and mammography, breast problem) and cancer information on Pap smear screening and behavior stage of change were investigated and modeled using an artificial neural network model (ANN).Materials and methodsData were collected from a descriptive-analytical cross-sectional study. This research was conducted on 1898 female employees of governmental agencies of Birjand, a city located in the east of Iran. The questionnaire consisted of four parts (socioeconomic, reproductive characteristics, information about cervical cancer screening, and stage of change for cervical cancer screening). Multilayer feed-forward back-propagation neural networks were used to detect the patterns between variables using a neural network with 14 inputs and one output. To find out the neural network with the minimum sum of squared errors, we evaluated the performance of all neural networks using varying algorithms and numbers of neurons in the hidden layer. For this purpose, the data collected from 1898 women were analyzed using SPSS-22 software.ResultsIn the optimal ANN model, the variables of marriage age, age, breastfeeding, and the number of children were identified as the most significant factors with 18.3, 16.3, 7.3, and 7.3 percent, respectively, whereas the history of cancer, job, pregnancy, and menopause had importance of lower than 5 percent.ConclusionOur findings showed that among many associated variables, the marriage age, age, breastfeeding, and the number of children were the most important predictors for the behavioral stage of change. Thus, it seems, focusing on these factors may lead to the adoption of effective programs and policies to improve cervical cancer screening practices in women.

Speaker Verification using CWT DWT Transform and Neural Network

In this Present study, the technique of wavelet transform and neural network were developed for speech based text-dependent and text0independent speaker identification. 390 feature were fed to feed-forward back propagation neural network for classification The function of feature extraction and classification are performed using wavelet and neural network system. The declared result shows that the proposed method can make an effective analysis with average identification rate reaching 98%. The best recognition rate selection obtained was for FFBPNN (Feed Forward Back Propagation Neural Network).

Predicting Effective Parameters in Cyclic Behavior of Reinforced Masonry Walls with Shotcrete Using Artificial Neural Networks

In Masonry buildings, walls are the main structural elements resistant to lateral loads. Reinforcing masonry walls by shotcreting one- or two-sides is one of the most common reinforcement methods. Considering the economic losses and casualties caused by damage to these elements in previous earthquakes, it is necessary to investigate the seismic behavior of these walls. The existing data can train artificial neural networks, and the behaviors could be generalized for future cases. In this study, the prediction of cyclic behavior parameters of reinforced masonry walls with one- and two-way shotcrete is investigated using different artificial neural network methods. Input parameters of the neural network include length, thickness, height, mortar shear strength, mortar compressive strength, mesh type, spring dimensions, rebar diameter, average thickness of shotcrete, and concrete compressive strength. Output parameters were the relative displacement of yield, relative displacement corresponding to maximum resistance, final relative displacement, yield strength, displacement resistance corresponding to maximum resistance, ultimate resistance, and initial stiffness. The results showed that the feed-forward back-propagation neural network could accurately predict the examined output parameters compared to other models, which can be considered as an alternative to some time-consuming and costly laboratory and analytical investigations.

Robust Process Parameter Design Methodology: A New Estimation Approach by Using Feed-Forward Neural Network Structures and Machine Learning Algorithms

In robust design (RD) modeling, the response surface methodology (RSM) based on the least-squares method (LSM) is a useful statistical tool for estimating functional relationships between input factors and their associated output responses. Neural network (NN)-based models provide an alternative means of executing input-output functions without the assumptions necessary with LSM-based RSM. However, current NN-based estimation methods do not always provide suitable response functions. Thus, there is room for improvement in the realm of RD modeling. In this study, a new NN-based RD modeling procedure is proposed to obtain the process mean and standard deviation response functions. Second, RD modeling methods based on the feed-forward back-propagation neural network (FFNN), cascade-forward back-propagation neural network (CFNN), and radial basis function network (RBFN) are proposed. Third, two simulation studies are conducted using a given true function to verify the proposed three methods. Fourth, a case study is examined to illustrate the potential of the proposed approach. In conclusion, a comparative analysis of the three feed-forward NN structure-based modeling methods and conventional LSM-based RSM proposed in this study showed that the proposed methods were significantly lower in the expected quality loss (EQL) and various variability indicators.

Modeling Industrial Energy Demand in Relation to Subsector Manufacturing Output and Climate Change: Artificial Neural Network Insights

The study aims to adopt an artificial neural network (ANN) for modeling industrial energy demand in Taiwan related to the subsector manufacturing output and climate change. This is the first study to use the ANN technique to measure the industrial energy demand–manufacturing output–climate change nexus. The ANN model adopted in this study is a multilayer perceptron (MLP) with a feedforward backpropagation neural network. This study compares the outcomes of three ANN activation functions with multiple linear regression (MLR). According to the estimation results, ANN with a hidden layer and hyperbolic tangent activation function outperforms other techniques and has statistical solid performance values. The estimation results indicate that industrial electricity demand in Taiwan is price inelastic or has a negative value of −0.17 to −0.23, with climate change positively influencing energy demand. The relationship between manufacturing output and energy consumption is relatively diverse at the disaggregated level.

Optimizing Feed Forward Backpropagation Neural Network Based on Teaching-Learning-Based Optimization Algorithm for Long-Term Electricity Forecasting

Optimizing Feed Forward Backpropagation Neural Network Based on Teaching-Learning-Based Optimization Algorithm for Long-Term Electricity Forecasting

Applying artificial neural-network model to predict psychiatric symptoms.

IntroductionMental disorders result in mental disabilities and discomfort in the affected person as they affect both thinking and behavior. Therefore, more vulnerable people must first be identified to improve the psychological level of society.AimThis study aims to determine the importance of gender, employment, education, place of residence, and age in predicting mental disorders using artificial neural network modeling.MethodsThe pattern held between variables in this study will be identified using multilayer feed-forward back-propagation neural networks with five inputs and 10 outputs. To determine the neural network with the least sum of square errors, we evaluated the performance of all neural networks with varying algorithms and different numbers of neurons in the hidden layer. Data were analyzed for 380 people aged 10–82 years using the SPSS software.ResultsThe optimal neural network model was effective in predicting mental disorders. In this model, the variables of the place of residence, education, age, gender, and employment were important in fitting the optimal model with 34.08, 20.11, 18.93, 14.55, and 12.33%, respectively. The accuracy rate for the neural network model was 99.2%.ConclusionTo achieve further results in improving mental health problems, it is better to focus more on employed, rural, and younger people with a non-tertiary education level.

A novel medium- and long-term runoff combined forecasting model based on different lag periods

Abstract The accuracy of medium- and long-term runoff forecasting plays a significant role in several applications involving the management of hydrological resources, such as power generation, water supply and flood mitigation. Numerous studies that adopted combined forecasting models to enhance runoff forecasting accuracy have been proposed. Nevertheless, some models do not take into account the effects of different lag periods on the selection of input factors. Based on this, this paper proposed a novel medium- and long-term runoff combined forecasting model based on different lag periods. In this approach, the factors are initially selected by the time-delay correlation analysis method of different lag periods and further screened with stepwise regression analysis. Next, an extreme learning machine (ELM) is adopted to integrate each result obtained from the three single models, including multiple linear regression (MLR), feed-forward back propagation-neural network (FFBP-NN) and support vector regression (SVR), which is optimized by particle swarm optimization (PSO). To verify the effectiveness and versatility of the proposed combined model, the Lianghekou and Jinping hydrological stations from the Yalong River basin, China, are utilized as case studies. The experimental results indicate that compared with MLR, FFBP-NN, SVR and ridge regression (RR), the proposed combined model can better improve the accuracy of medium- and long-term runoff forecasting in the statistical indices of MAE, MAPE, RMSE, DC, U95 and reliability.

Towards a better indoor positioning system: A location estimation process using artificial neural networks based on a semi-interpolated database

The Wi-Fi-fingerprinting positioning method is used widely in indoor positioning environments due to its simplicity and wide coverage. However, in the offline phase of the method, the collection process is a fundamental and critical step that requires time and effort. Moreover, the location estimation process, which is executed in the second Wi-Fi-fingerprinting phase (online phase), needs to be accurate enough to guarantee efficient indoor positioning. Hence, in this work, a novel indoor location-estimation process based on a semi-interpolated radio map and artificial neural network (ANN) is presented. A mobile application is built to gather the received signal strength indicator (RSSI) fingerprinting to construct a radio map, which is then expanded with the biharmonic spline interpolation (BSI) method through the estimation of more RSSI values. A feedforward back propagation (FFBP) neural network and generalised regression neural network (GRNN) were built in the online phase for the location-estimation process. They were trained using the expanded dataset by taking the reference point (X, Y) coordinates as their desired output and using two different forms of the data as their inputs. The first inputs are the RSSI values from the 17 access points (APs) – three of the APs have dual-band i.e, support both 2.4 and 5 GHz – and the second input is based on a selected set of APs, which produce a high level of acceptable RSSI and their coordinates. A comparison between these two models was done. The results show that FFBP outperforms GRNNs in terms of structure simplicity, while GRNNs achieved more accurate prediction results with an average distance error of up to 0.48 m. Hence, our proposed methodology leverages building a simple neural network topology that has good location estimation results for indoor positioning in a low-cost localisation process.