Back Propagation Neural Network Algorithm Research Articles

Bed density prediction of gas–solid separation fluidized bed based on genetic algorithm-back propagation neural network

The fluctuation of bed density in the process of mineral separation in gas–solid fluidized bed is particularly critical, and accurate prediction of bed density will play a crucial role in industrial production. In this paper, the axial and radial distribution of bed density in gas–solid fluidized bed under different ratios of dense medium and gas velocities was investigated. Based on the data of this study and other scholars, the corresponding data set was constructed. A 13–14-1 bed density prediction model was established by using BP neural network algorithm. The overall correlation of the model is 0.9771. The model is more accurate to predict the bed density of single dense medium. The mean absolute percentage error is 4.17 % and the root mean square error is 0.0298. The optimized GA-BP (Genetic Algorithm-Back Propagation) neural network by genetic algorithm results are closer to the actual experimental values. The overall correlation of the model is increased to 0.9793, and the prediction accuracy can reach more than 90 %. The GA-BP neural network prediction model is more accurate in predicting fluidized beds with density fluctuation range of 1.6–2.2 g/cm3, which can meet the needs of industrial applications.

A Pattern Recognition Method for Filter Bags in Bag Dust Collectors Based on Φ-Optical Time-Domain Reflectometry

The use of phase-sensitive optical time-domain reflectometry (Φ-OTDR)-distributed fiber vibration sensors to detect and identify damaged bags in bag dust collectors has the potential to overcome the inadequacy of traditional damaged bag detection methods. In our previous study, we verified the feasibility of applying this technique in the field of damaged bag detection in bag filters. However, many problems still occur in engineering applications when using this technology to detect and identify damaged filter bags in pulse-jet dust-cleaning bag dust collectors. Further studies are needed to characterize the fiber vibration signals inside different types of rectangular damaged filter bags. A filter bag damage identification and detection method based on empirical mode decomposition (EMD) and a backpropagation (BP) neural network is proposed. The signal feature differences between intact filter bags and damaged filter bags with different rectangular hole sizes and positions are comparatively analyzed, and optimal feature difference parameters are proposed. Support vector machine (SVM) and a BP neural network are used to recognize different types of filter bag signals, and the comparison results show that the BP neural network algorithm is better at recognizing different types of filter bags, obtaining the highest recognition rate of 97.3%.

Model for Technology Risk Assessment in Commercial Banks

As the complexity of banking technology systems increases, the prevention of technological risk becomes an endless battle. Currently, most banks rely on the experience and subjective judgement of experts and employees to allocate resources for technological risk management, which does not effectively reduce the frequency of technology-related incidents. Through an analysis of mainstream risk management models, this study proposes a technology-based risk assessment system based on machine learning. It first identifies risk factors in bank IT, preprocesses the sample data, and uses different regression prediction models to train the processed data to build an intelligent assessment model. The experimental results indicated that the Genetic Algorithm–Backpropagation Neural Network model achieved the best performance. Based on assessment indicators, indicator weight values, and risk levels, commercial banks can develop targeted prevention and control measures by applying limited resources to the most critical corrective actions, thereby effectively reducing the frequency of technology-related incidents.

Comparative Analysis of Convolutional Neural Network-Long Short-Term Memory, Sparrow Search Algorithm-Backpropagation Neural Network, and Particle Swarm Optimization-Extreme Learning Machine Models for the Water Discharge of the Buzău River, Romania

Modeling and forecasting the river flow is essential for the management of water resources. In this study, we conduct a comprehensive comparative analysis of different models built for the monthly water discharge of the Buzău River (Romania), measured in the upper part of the river’s basin from January 1955 to December 2010. They employ convolutional neural networks (CNNs) coupled with long short-term memory (LSTM) networks, named CNN-LSTM, sparrow search algorithm with backpropagation neural networks (SSA-BP), and particle swarm optimization with extreme learning machines (PSO-ELM). These models are evaluated based on various criteria, including computational efficiency, predictive accuracy, and adaptability to different training sets. The models obtained applying CNN-LSTM stand out as top performers, demonstrating a superior computational efficiency and a high predictive accuracy, especially when built with the training set containing the data series from January 1984 (putting the Siriu Dam in operation) to September 2006 (Model type S2). This research provides valuable guidance for selecting and assessing river flow prediction models, offering practical insights for the scientific community and real-world applications. The findings suggest that Model type S2 is the preferred choice for the discharge forecast predictions due to its high computational speed and accuracy. Model type S (considering the training set recorded from January 1955 to September 2006) is recommended as a secondary option. Model type S1 (with the training period January 1955–December 1983) is suitable when the other models are unavailable. This study advances the field of water discharge prediction by presenting a precise comparative analysis of these models and their respective strengths

Using Neural Networks (Back Propagation Network) to Determine the Variables Affecting Water Pollution in the Euphrates River

With the economy's steady growth and rapid development of modern industrial technology, environmental problems have increased. Water pollution is one of the complex problems that most countries of the world suffer from, especially developing countries, because it contains various pollutants, some of which can decompose as organic pollutants, and some of which are difficult to decompose, such as toxic heavy metals. In Iraq, according to the Iraqi Ministry of Environment - Water Pollution Division statistics for the year 2018, specifically the waters of the Euphrates River, a group of variables affecting the water quality of this river has been identified. The scarcity of water and the accumulation of various wastes in the water networks and on the banks of the Euphrates River also contributed greatly to the increase in the problem of pollution. The continued poor quality of the water has led to severe environmental health concerns. The research aims to shed light on the most important variables affecting the pollution of the waters of the Euphrates River and the water quality and quality in this river. The principal components analysis and the neural network (backpropagation algorithm) were used to determine the variables that affect the pollution of the Euphrates River water. They were applied to real data from the Ministry of Environment for the Euphrates water pollution. One of the most important conclusions we reached when using statistical methods was that the variable (DO2) is not significant and does not affect the pollution of the Euphrates River either. Still, When Using the algorithm, we get that all these variables (TH, TDS, EC) do not affect the water pollution of the Euphrates River.

Research on dynamic splitting tensile damage characteristics of steel fiber reinforced concrete under freeze-thaw cycle environment based on AE

This paper aims to investigate the dynamic splitting tensile damage characteristics of steel fiber reinforced concrete (SFRC) and achieve early warning of the dangerous stage under freeze-thaw cycle environment through acoustic emission (AE) technology, Mel-frequency cepstral coefficients (MFCCs) and genetic algorithm-back propagation (GA-BP) neural network. The effects of fiber content and loading rate on mechanical properties, wavelet energy spectrum coefficient and wavelet time-frequency-energy maximum, as well as MFCCs of AE signals were studied herein. Then a GA-BP neural network was established to realize the classification of the damage degree of SFRC under freeze-thaw cycling with MFCCs as the signal characteristic. The results indicate that the increasing loading rate leads to more serious brittle failure, but the appropriate steel fiber content can make the cracks generated develop fully during the dynamic splitting tensile failure process by analyzing the wavelet energy spectrum coefficient and wavelet time-frequency-energy maximum. It is also found that the damage degree of SFRC can be identified by analyzing the MFCCs. Furthermore, the GA-BP neural network performs well in the classification of the damage degree of SFRC with the recognition over 90%, which is more suitable for concrete conditions with a proper loading rate and fiber content.

Prediction of Converter Tapping Weight Based on Principal Component Analysis–Whale Optimization Algorithm–Backpropagation Algorithm Neural Network Model

Tapping weight is an important parameter in converter blowing process, wherein precisely predicting the quantity of steel required in an alloy baking converter can effectively guide the requirement of alloy ingredients. In practical production, the main approach is empirical estimation, despite its low accuracy. Employing a general neural network model for prediction requires to gather the converter blowing parameters and endpoint temperature measurement sampling parameters as the model inputs. However, this data cannot be obtained until the blowing process reaches its endpoint, rendering it impractical for alloy batching that requires advance preparation for baking. In this study, a principal component analysis–whale optimization algorithm–backpropagation algorithm (PCA–WOA–BP) neural network tapping prediction model is developed using raw material parameters available before alloy baking as input. This model is integrated into the intelligent alloy reduction model of a factory. The model achieves a regression coefficient R2 of 0.823, with 98.53% of furnaces having a prediction error of less than 2 t. The tapping weight of 20 consecutive heats in actual production is predicted; the error range is less than 80% within 1 t, and the converter tapping weight is predicted accurately.

Preschool Teachers Teaching Quality Evaluation Based on Neural Network Algorithms

Based on the current teaching situation of preschool teachers, in order to comprehensively evaluate the effectiveness of early childhood teaching, research have constructed a teaching quality evaluation model using fuzzy synthesis method and expert method. This model can handle the fuzzy relationship between evaluation indicators and achieve the evaluation of teaching quality. Considering that the teaching evaluation is influenced by many factors, the Genetic Algorithm back propagation (GA-BP) neural network algorithm is chosen for the solution model construction of the preschool teacher teaching quality. The entropy method chosen for the data calculation is to complete the preschool teaching quality evaluation. In the mean square error test for solving the model, the improved GA-BP model converged after 40 iterations with the model convergence speed increased by 34.65%. In the evaluation and prediction of preschool teacher indicators, the improved GA-BP model accurately evaluated the teacher classroom teaching indicators. In sample 3, sample 6, and sample 9, the improved GA-BP model scored 91, 89, and 88 points, respectively, close to the true scoring results. The improved model’s accuracy was high as 90% in teacher skills, personality charm, and academic research evaluation. The improved model also effected better in the application of the preschool teaching quality evaluation. The application effect of this model in teaching effectiveness evaluation and teacher quality evaluation is good, providing valuable reference for the establishment of early childhood education evaluation system.

Quantitative research on stress failure risk assessment for girth welds with unequal wall thickness of the X80 pipeline under lateral load

The cumulative length of the X80 steel pipeline extends to 17,000 km, with a predominant portion traversing mountainous terrains characterized by intricate geological formations. Recurrent geological hazards significantly impede the safe operation of oil and gas pipelines. Landslides can induce substantial axial forces on steel pipelines, significantly increasing the risk of pipeline failure. Moreover, as a result of sudden changes in mechanical structure, the pipeline's unequal wall thickness leads to local stress concentration, making it more vulnerable to external loads than the steel pipeline itself. Therefore, to analyze the mechanical response of steel pipelines under different working conditions, we establish a smooth particle hydrodynamics, finite element transverse landslide pipeline model for girth welds with unequal wall thickness and investigate a tearing accident caused by a lateral landslide in a section of the large-diameter natural gas pipeline. The landslide volume is found to be the main controlling factor for the failure of pipeline circumferential welds. According to the genetic algorithm backpropagation neural network, the von Mises stress of the most dangerous section of the pipeline is calculated based on pipeline and landslide parameters, and the error is 11.67 %. According to the API RP 581–2016 Risk-Based Inspection Methodology, a failure risk assessment matrix is established using the pipeline statute as an evaluation indicator to evaluate the landslide risk to pipelines in different regions. A variable-wall-thickness girth weld on the ZG pipeline was selected for application. This method has good applicability for evaluating the failure risk of circumferential welds with wall thicknesses of 12.8/15.3 mm under the action of small landslides.

Thermal management of methanol steam reforming reactor by back propagation neural network and particle swarm optimization algorithm

Methanol steam reforming (MSR) reactor is a device that converts the methanol into hydrogen. Since the MSR is an endothermic reaction, the methanol conversion is usually affected by the heating conditions. So, to improve the methanol conversion, the heating condition should be managed. This study aims to present a framework and methodology for studying the thermal management of a determined MSR reactor by controlling the heating conditions. Firstly, the characteristics of a three-dimensional MSR reactor are carried out, and the effects of the inlet condition parameters e.g. heating parameters and fuel supply parameters, on the methanol conversion are investigated. Then, the performance prediction of the MSR reactor is studied using several common intelligence algorithms and the corresponding surrogate models are obtained. The back propagation neural network (BPNN) surrogate model shows the best prediction accuracy than stepwise linear (SL), radial basis function neural network (RBFNN), linear support vector machines (L-SVM), and random forest algorithms (RFA). Finally, the thermal management of the MSR reactor is carried out by combining the surrogate model and particle swarm optimization (PSO) algorithm to obtain better performance. The results show that hydrogen production of the MSR reactor can be generally guaranteed by the characteristic curves of the heating parameters and fuel supply parameters, and the methanol conversion can be maintained steadily above 93% within the whole period of the hydrogen production requirements.

Optimisation of drag coefficient of a car based on back propagation neural network and genetic algorithm

Optimisation of drag coefficient of a car based on back propagation neural network and genetic algorithm

Fatigue Crack Growth Rate Prediction in Nickle based Super-alloys using Machine Learning Algorithm

The state-of-the-art research in any field is inclined towards the use of artificial intelligence (AI) and machine learning (ML) techniques due to their versatility and ease of application without compromise in the results. Predicting the service life of structural members subjected to fatigue loading, is a daunting task and is currently dependent on experimental, finite element (FE) and/or numerical approaches. Damage tolerant (DT) design incorporates fatigue crack growth rate (FCGR) test data to avoid catastrophic structural failure and is proven to be successful in predicting the life of a component. Such predictions also prove beneficial in applications such as structural health monitoring (SHM) and planning non-destructive inspection. Prediction methods and models, currently used, are application specific or inconsistent across different applications. Hence in the present work, an ML algorithm is proposed to predict the Paris region of the FCGR curve for nickel based super alloys, viz. GTM 720 and GTM 718. Experimental dataset for various stress ratios (R) of the alloys were used to train the back propagation neural network (BPNN) algorithm. The algorithm was appropriately tuned and evaluated so as to fit the training data well and achieve good generalization. Using this trained ML algorithm, the FCGR was predicted for a different stress ratio and compared with experimental results of GTM 720 and GTM 718, available in literature. The results obtained are quite comparable indicating promising applications for ML based algorithms in FCGR prediction.

Intelligent Prediction of FSW Physical Quantity and Joint Mechanical Properties

Friction stir welding (FSW) process parameters influence welding temperature field and axial force, which affect welding strength. At present, how the FSW process parameters of aluminum alloy 2219-T8 thick plates influence process physical quantity and how the process physical quantity changes the tensile strength about the welded joint are unknown. We focus on the intelligent prediction of FSW temperature, axial force, and mechanical properties, to provide a basis for FSW process control of aluminum alloy 2219-T8 thick plate. Firstly, we conducted the FSW experiment of aluminum alloy 2219-T8 thick plate. Then, we input the welding process parameters, set up a prediction model by particle swarm optimization-back propagation (PSO-BP) neural network to predict the peak temperature and axial force. Finally, we input the peak temperature and axial force, use genetic algorithm-back propagation (GA-BP) neural network to establish a weld tensile strength estimation model, and comply with the prediction of tensile strength.

Prediction of Transformer Oil Temperature Based on an Improved PSO Neural Network Algorithm

Abstract: In addressing the issue of power transformer oil temperature prediction, traditional back propagation (BP) neural network algorithms have been found to suffer from local optimization and slow convergence. This study proposes an oil temperature prediction model based on an improved particle swarm optimization (PSO) neural network algorithm, which introduces an asymmetric adjustment learning factor and a mutation operator. The BP neural network, genetic algorithm (GA) optimization neural network, and the improved PSO neural network are compared by considering various factors, such as ambient temperature, load changes, and the number of cooler groups under different working conditions. Results show that the proposed algorithm improves the actual change trend of oil surface temperature and makes the transformer operation more stable to a certain extent. Background: The mathematical model for predicting transformer oil temperature is clear, but the parameters in the model are uncertain and vary with time. When subjected to different operating conditions, such as ambient temperature, load changes, and the number of cooler groups acting independently or in combination, the prediction results of the oil temperature model vary with different system parameters. Objective: This paper aims to enhance the accuracy of transformer temperature prediction. In order to optimize the oil temperature prediction model, asymmetric adjustment learning factors and mutant operators are added to meet diverse system parameter requirements. Methods: The paper utilizes an oil temperature prediction model based on an improved PSO neural network algorithm, which introduces an asymmetric adjustment learning factor and a mutation operator to address the limitations of the standard PSO algorithm. Results: This paper has employed a fusion algorithm of the genetic algorithm of the BP neural network and the PSO algorithm, and conducted simulation and experimental analysis. The simulation and experimental results demonstrate the accuracy and effectiveness of the fusion algorithm. Conclusion: This study demonstrates enhanced prediction accuracy of transformer oil surface temperature using the improved particle swarm optimization neural network algorithm. This algorithm has less prediction error under different working conditions compared to other algorithms. By increasing population diversity and combining inertia weights, the algorithm not only greatly improves its search performance but also avoids local optimization.

Performance modeling of flame-assisted fuel cells based on a swirl burner

Aiming at the problems of a narrow operating range and complex modeling of Flame-assisted Fuel Cells (FFCs), an FFC system based on a swirl burner is proposed, and neural network algorithms are used to construct the prediction model for the polarization curve of the FFC system. First, the output voltage and power values of the FFC system are measured under different working conditions, and various experimental parameters are collected to form a dataset; second, the correlation analysis method is used to screen out the parameters that are highly correlated with the output voltage as the input variables of the neural network; finally, the prediction model of the polarization curve is constructed, and back propagation (BP), long short term memory, and 1D-CNN algorithms are chosen to examine the applicability of various neural networks for the FFC system. The experimental and polarization characteristic curve prediction results show that the FFC system can obtain a maximum output voltage of 10.6 V and power of 7.71 W. The average relative errors of the three algorithms are 5.23%, 4.08%, and 6.19%, respectively, with the BP neural network algorithm showing the best generalization ability. The study provides support for the application of the FFC system in aerospace and other fields.

A comparative study on the difference of color space conversion based on table look-up method and neural network

A comparative study on the difference of color space conversion based on table look-up method and neural network

Predicting Stock Market Trends Based on Macroeconomic Indicators through Machine Learning Approach: A Case Study of KSE 100 INDEX

The purpose of our research is to model the monthly price of the KSE 100 index based on Pakistan's macroeconomic indicators using a Machine Learning (ML) approach. The novelty of the study is forecasting the future value of the stock market using ML. Monthly data was collected for the period from Feb 2004 to Dec 2020. The output layer of our study is the closing price of the KSE 100 index, and the input layer consists of 16 macroeconomic variables of the Pakistan economy, which are the industrial production index (IPI), the exchange rate (EX-RATE), money supply (M2), consumer price index (CPI), foreign direct investment (FDI), Treasury bill on 3-months treasury, interest rate as KIBOR – Month Average (1 Month), Foreign Exchange reserves (FXR), Consumer Financing for house building (house financing), Balance of Trade (BOT), crude oil, Gold, Labor force participation rate, GDP growth (annual %), Households and NPISHs Final consumption expenditure (household consumption) (current US$), and Domestic savings. The prediction uses the Artificial Neural Network (ANN) Backpropagation algorithm. The model developed in this research achieved 99% accuracy using macroeconomic indicators. The accuracy level indicates that the model of the KSE 100 index can predict future trends. This study also forecasts the future monthly values of the KSE 100 index from Jan 21 to Jun 23 and daily future values from Oct 1, 2022, to Dec 31, 2020.

Application of the particle swarm optimization algorithm-back propagation neural network algorithm introducing new parameter terms in the application field of industrial design

With the development of the modern industry, to continuously use large precision instruments, the circuits protecting method through circuit design has gradually been promoted. And this method can prevent the excessive accumulation of electric heat from causing fires. The voltage at the starting point of power transmission is constant. However, during the transportation process, some lines have low resistance values, which can generate large instantaneous currents. To address this issue, this study conducted simulation experiments on self-collected Circuit dataset based on particle swarm optimization and backpropagation neural networks. This study first introduced new parameter term learning factors into backpropagation neural networks. Then they were imported into the support vector machine, and the nonlinear variables were mapped to high plane, and the optimal hyperplane was established. Then the traditional circuit design method was improved, 40 Resistor were connected in parallel and connected to the experimental circuit in series with the rheostat. Finally, the algorithm was introduced into Circuit dataset collected in this experiment, and its protective effect on the circuit was compared with the other three algorithms. Under the protection of this design, the working times of four algorithms were 0.28, 0.42, 0.38, and 0.43 s, respectively. Their phase displacements were 0.19, 0.26, 0.36, and 0.41, respectively. The circuit design method proposed in this study can effectively address circuit faults. And the fusion algorithm can disconnect the circuit at the fastest speed and significantly reduce excitation current intensity, and it is suitable for circuit design in the field of industrial design.

Spindle Thermal Error Prediction Based on Back Propagation Neural Network and Hybrid Taguchi Genetic Algorithm for a Computer Numerical Control Machine Tool

Spindle Thermal Error Prediction Based on Back Propagation Neural Network and Hybrid Taguchi Genetic Algorithm for a Computer Numerical Control Machine Tool

Research on T-shirt-style design based on Kansei image using back-propagation neural networks

Abstract Users’ Kansei image preferences have become one of the most important factors influencing purchase decisions. However, defining Kansei image can be complex. To address this issue, researchers have widely applied back-propagation neural networks due to their capacity to handle extensive data, adaptively adjust weights and biases, conduct multi-class classification and regression predictions, and offer interpretability analysis, among other features. In this study, a clothing-style design model based on users’ Kansei image cognition is proposed, using collarless T-shirts as an example. Furthermore, the attributes of T-shirt patterns are quantified using parametric graphics principles, and a semantic scale system for emotions is established through user research. The quantified sample data and corresponding semantic scale scores are then used as inputs for training a back-propagation neural network algorithm. Consequently, a design model grounded in users’ Kansei image cognition is developed, resulting in five optimal clothing design forms across various Kansei image categories. Additionally, the styles are showcased through the Style 3D platform, and the design evaluation is presented using radar charts. The results demonstrate that the five female T-shirt designs generated by the model align with users’ style preferences based on Kansei image.