BP Neural Network Algorithm Research Articles

Anomalous pattern recognition based on image recognition in food safety detection

In food processing, foreign matter inevitably contaminates packaged food. To ensure food safety, ray-based detection is used; however, the original images suffer from aberrations and noise that degrade quality and hinder further processing. Thus, images are preprocessed to enhance quality by highlighting key features and suppressing irrelevant ones before abnormal pattern recognition. Following image segmentation, a BP neural network algorithm is applied for foreign object detection. In tests with contaminants such as metal wires, stones, and glass, the algorithm identified distinct abnormal fluctuations at gray levels of 132, 108, and 34, respectively, allowing it to reliably detect foreign objects. Although the practical detection rate reached 100%, occasional misjudgments suggest that further optimization is needed. Overall, this method introduces a novel approach to detecting foreign objects in food and offers promising new strategies for improving food safety monitoring.

Stroke Risk Prediction and Assessment Based on Big Aata Analysis

Stroke is a common cardiovascular and cerebrovascular disease with high morbidity, high mortality and high disability rate. In this paper, a stroke risk prediction and evaluation model based on support vector machine, random forest, BP neural network and genetic algorithm optimization neural network algorithm was established by using a raw dataset including 10 characteristic variables such as gender, age, hypertension, heart disease, and 1 stroke target variable. The experimental results show that the average blood glucose level, body mass index, hypertension and other variables have a great impact on the risk of stroke, and the neural network algorithm optimized by the genetic algorithm performs slightly better than the other three models.

The research on traffic accident prediction based on simulated annealing optimized BP neural network algorithm

The research on traffic accident prediction based on simulated annealing optimized BP neural network algorithm

Optimization design of centrifugal pump cavitation performance based on the improved BP neural network algorithm

Optimization design of centrifugal pump cavitation performance based on the improved BP neural network algorithm

Application of Principal Component Analysis and BP Neural Network Algorithm in Stock Price Prediction

Application of Principal Component Analysis and BP Neural Network Algorithm in Stock Price Prediction

Еnеrgy Efficiеnt Nеural Nеtwork Control of a Brushlеss DC Motor

The paper considers the main trends in the development of electric motors for electric vehicles and mobile robots, as well as trends in the development of modern methods for calculating power electronics and electric drives based on an artificial neural network. Aspects of the efficiency development of modern synchronous and brushless DC motors are presented. Based on the mathematical model of a brushless DC motor, the architecture of a control unit with a neural network controller is built. A proactive calculation of the neural network was carried out, and the rules for adjusting the weighting coefficients were determined. Based on proactive calculation, a PID controller with self-adjusting parameters using a neural network was built, as well as a block diagram of the PID control system was built on the basis of the BP neural network; also, a speed controller was built using MATLAB modules. Besides, an S-activation function was built as a controller of the BP neural network; the function was based on the mathematical description of the neural network of the control unit of a brushless DC motor. The paper shows in detail the installation of a demultiplexer for better distribution of the S-function output. The resulting neural network encapsulates the S-function of the weight function. Based on the results of the neural network research and analysis of the BP neural network algorithm, a control algorithm has been established that is used to control the PID controller and is encapsulated in the simulation system. The theoretical possibilities of calculation based on a feedback neural network for constructing a simulation model of adaptive control of a brushless DC motor are demonstrated.

Non-parametric modeling of magnetorheological damper based on MEA-BP neural network

The process of parametric modeling of magnetorheological (MR) damper presents difficulties such as numerous parameters, complicated and time-consuming identification procedures, and difficulty in establishing the inverse model. In order to solve these problems, this paper adopts the mind evolution algorithm backpropagation (MEA-BP) neural network to successfully establish the forward model and the inverse model for prediction of electric current. These models can accurately capture the nonlinear hysteresis characteristics of MR damper. First, the mechanical properties of the MR damper were tested, and then the BP neural network structure of the forward and inverse models was determined according to its working principle. Aiming at the problems of traditional BP neural network which is easy to mature prematurely and converge slowly, the initial weights and thresholds of BP neural network were optimized by using MEA. Utilizing the Mechanical properties test results of the MR damper, the MEA-BP neural network was used to identify the forward and inverse non-parametric models of the MR damper. Compared with the traditional BP neural network model, the forward model and inverse model of the MR damper, based on the MEA-BP neural network, showed an improvement in prediction accuracy by 33.96% and 26.32%, respectively. The forward model based on the MEA-BP neural network demonstrated a 16.67% improvement in accuracy compared to the genetic algorithm (GA) BP neural network. The non-parametric model, developed using the MEA-BP neural network, more accurately captures the dynamic behavior of the MR damper. This advancement provides a robust foundation for research into the dynamics and semi-active control of MR adaptive damping systems.

Optimization design of the impeller structure for a micro centrifugal pump based on pressure pulsation and shear stress analysis

Micro centrifugal pumps are commonly used in the medical field for transporting drugs or blood, etc. Due to the particularity of fluid media, irregular flow is often generated inside the pump resulting in over-high pressure pulsation and shear stress, which will eventually cause damage to the medium. In this paper, the internal flow field of a micro centrifugal pump was investigated through both visualization experiments and numerical simulations. When the main structure parameters of the impeller were selected as optimization variables, with pressure pulsation and shear stress as optimization objectives, a combination numerical simulation solution of BP neural network and genetic algorithm was employed to optimize the objectives respectively. The results indicate that the BP neural network model and genetic algorithm have high accuracy and can be used to predict the performance of micro centrifugal pumps. After optimization, the pressure pulsation strength coefficient of the monitoring point near the volute tongue is reduced by 40%. The internal flow of the micro centrifugal pump becomes more stable. The maximum shear stress of the impeller is declined by 42% and 28% in the volute. The research is of great significance to improve the stability and efficiency of micro centrifugal pumps.

Research on partial discharge pattern recognition of transformer based on KPCA and JS-SVM algorithm

Abstract Transformers are essential for the safe operation of power grids, and their failures can lead to significant risks. To improve the accuracy of partial discharge (PD) diagnosis in transformer bushings, this paper proposes a PD pattern recognition algorithm combining Kernel Principal Component Analysis (KPCA) and Artificial Jellyfish Search (JS)-optimized Support Vector Machine (SVM). The proposed algorithm offers the advantages of efficient dimensionality reduction, high classification accuracy, and fast optimization. KPCA is first used to reduce the dimensionality of high-frequency signals, improving data processing efficiency and lowering computational complexity. The JS algorithm is then employed to optimize SVM parameters, significantly enhancing classification and recognition performance. Experimental results show that the proposed algorithm achieves a higher recognition rate in PD defect diagnosis, with overall recognition rates improved by 3.74% and 2.5% compared to traditional SVM and BP neural network algorithms. This study provides a novel approach for early defect diagnosis in transformers and holds significant engineering application value.

Multi-Objective Optimization of Injection Molding Process Parameters for Junction Boxes Based on BP Neural Network and NSGA-II Algorithm.

Many factors affect the quality of the injection molding of plastic products, including the process parameters, mold materials, type and geometry of plastic parts, cooling system, pouring system, etc. A multi-objective optimization method for injection molding process parameters based on the BP neural network and NSGA-II algorithm is proposed to address the problem of product quality defects caused by unreasonable process parameter settings. Taking the junction box shell as the object, numerical simulation was carried out using Moldflow2019 software and a six-factor five-level orthogonal experiment was designed to explore the influence of injection molding process parameters, such as the mold temperature, melt temperature, injection pressure, holding pressure, holding time, and cooling time, on the volume shrinkage rate and warpage deformation of the junction box. Based on a numerical simulation, the BP neural network and NSGA-II algorithm were used to optimize the optimal combination of injection molding process parameters, volume shrinkage rate, and warpage deformation. The research results indicate that the melt temperature has the most significant impact on the quality of the injection molding of junction boxes, followed by the holding time, holding pressure, cooling time, injection pressure, and mold temperature. After optimization using the BP neural network and the NSGA-II algorithm, the optimal process parameter combination was obtained with a melt temperature of 230.03 °C, a mold temperature of 51.27 °C, an injection pressure of 49.13 MPa, a holding pressure of 69.01 MPa, a holding time of 15.48 s, and a cooling time of 34.91 s. At this time, the volume shrinkage rate and warpage deformation of the junction box were 6.905% and 0.991 mm, respectively, which decreased by 33.2% and 3.8% compared to the average volume shrinkage rate (10.34884%) and warpage deformation (1.030764 mm) before optimization. The optimization effect was significant. In addition, the errors between the volume shrinkage rate and warpage deformation predicted by BP-NSGA-II and the simulated values using Moldflow software were 1.9% and 3.4%, respectively, indicating that the optimization method based on the BP neural network model and NSGA-II algorithm is reliable.

Exercise training load prediction based on improved genetic algorithm

With the advancement of digital information age, all kinds of sports also widely used digital mining technology, the technology can promote athletes’ professional level and physical quality, is in the field of sports improve athletes and coaches more effective way, but the bicycle sports for digital mining and research is still in the early state. Cyclists load ability in training is mainly influenced by the body function, the relationship between the two is complicated, this paper is the training load ability and physical function as a research object, explore the specific connection between the two aspects of cyclists, by constructing the cyclist training load prediction model, to apply in the “bicycle team training analysis system”. Strive to promote the overall development of cycling, to provide a basis for the scientific training of talents and training. The cyclist training load prediction model created in this paper can provide a scientific design basis for the training of athletes. When analyzing the body function, this paper mainly analyzes the influence of athletes on conforming sports through 25 aspects such as maximum oxygen intake, functional threshold power and blood oxygen saturation. Because the factors of athletes and the predicted results show non-linear correlation, this paper selects the BP neural network algorithm in the model algorithm, and the adaptive genetic algorithm of the selection operator is adjusted and optimized to clarify the initial weights and thresholds of BP neural network. Finally, the practice has proved that the improved algorithm has a more prominent global optimization ability, and the accuracy of the model has reached 93.28%.

Construction of intelligent coal blending system based on BP neural network and genetic algorithm

The quality of coke is crucial to the steel-making process, and advanced coal blending technology can effectively improve the stability of coke and reduce production costs. However, some of the current coal blending systems lack expertise and are out of touch with the actual situation of the factory, unable to meet production requirements. In the research process, this paper combined the conventional indicators measured in Chinese factories with actual data, and established a reliable database. In the back propagation neural network modeling process, several parameters were optimised, significantly improving the prediction performance compared to traditional regression methods. An evolutionary algorithm with a penalty matrix is used to optimise the coal blending ratio, resulting in a reduction in the ash and sulphur content of the coke, as well as a cost reduction of approximately 44 RMB per tonne. Based on the initial research results, an intelligent coal blending system was developed, which has been successfully applied in the production process of a coke plant, with simple operation and stable operation, providing important reference for the development of coal blending systems in other coke plants.

Machine Learning Model Construction and Practice for Personalized Training Programs in Physical Education and Sport Teaching

Abstract The implementation process of this paper to develop a personalized training program for physical education is mainly as follows: outlier processing, standardization, and correlation analysis of students’ physical education test scores. Then, the processed data were downscaled using principal component analysis and cluster analysis was performed. Finally, the BP neural network algorithm is used to predict the personalized exercise program, and the predicted program is adjusted with the help of NLP sentiment analysis. The practical analysis shows that the correlation coefficient between 50m running and standing long jump is −0.7483. The accuracy of the BP neural network model in predicting personalized exercise programs is 96%, and the adjusted and optimized personalized exercise program receives better feedback.

Data mining and neural network modeling for teaching and learning in vocational education: promoting innovation in academic management and teaching reforms

Abstract In order to provide students with dynamic and personalized academic early warning services, and at the same time provide university administrators with data-supported decision-making, this paper uses data cleaning, transformation and normalization methods to preprocess the data to normalize it into an analyzable dataset according to the characteristics of educational data. For this dataset, an academic early warning model (TabNet) combined with KNN neural network algorithm is proposed to train the early warning data and analyze and compare its performance with the BP neural network algorithm. Finally, it provides ideas for the construction of intelligent classrooms for vocational education teaching. The results show that the average accuracy and recall of the classifier on the test set are high, 93.11% and 74.35%, respectively; the classification of positive examples on the training set is not precise enough, and its average precision and recall are 81.36% and 47.07%, respectively. The recall, precision and F1 mean of the support vector machine on the training set are 87.36%, 91.14% and 81.57%, which are close to the test set. The Loss curve of the TabNet algorithm has a better generalization performance on both the test and training sets, with a low chance of overfitting and minimal differences. In addition, the TabNet neural network algorithm has higher accuracy on the ROC curve and is more valuable when using the classification results.

Research on aerodynamic attachments parameter optimisation by integrating BP neural network and genetic algorithm

Research on aerodynamic attachments parameter optimisation by integrating BP neural network and genetic algorithm

Research on aerodynamic attachments parameter optimisation by integrating BP neural network and genetic algorithm

Research on aerodynamic attachments parameter optimisation by integrating BP neural network and genetic algorithm

Evaluation of BP Neural Network Algorithms for Predicting Elastic Buckling Loads on Cold-Formed Steel Components

Evaluation of BP Neural Network Algorithms for Predicting Elastic Buckling Loads on Cold-Formed Steel Components

Integrating EPSOSA-BP neural network algorithm for enhanced accuracy and robustness in optimizing coronary artery disease prediction

Coronary artery disease represents a formidable health threat to middle-aged and elderly populations worldwide. This research introduces an advanced BP neural network algorithm, EPSOSA-BP, which integrates particle swarm optimization, simulated annealing, and a particle elimination mechanism to elevate the precision of heart disease prediction models. To address prior limitations in feature selection, the study employs single-hot encoding and Principal Component Analysis, thereby enhancing the model’s feature learning capability. The proposed method achieved remarkable accuracy rates of 93.22% and 95.20% on the UCI and Kaggle datasets, respectively, underscoring its exceptional performance even with small sample sizes. Ablation experiments further validated the efficacy of the data preprocessing and feature selection techniques employed. Notably, the EPSOSA algorithm surpassed classical optimization algorithms in terms of convergence speed, while also demonstrating improved sensitivity and specificity. This model holds significant potential for facilitating early identification of high-risk patients, which could ultimately save lives and optimize the utilization of medical resources. Despite implementation challenges, including technical integration and data standardization, the algorithm shows promise for use in emergency settings and community health services for regular cardiac risk monitoring.

Research on BP neural network model-based data mining technique in KOL identification

Abstract. Accompanied by the development of artificial intelligence industry, to play the data value of web crawler technology has been the focus of research in the field of computer network and data science, and the data mining technology based on artificial neural network intelligent algorithm is widely used. In view of this, this paper takes the BP neural network data mining technology, which has excellent nonlinear mapping capability, parallel processing capability and fault tolerance and is widely used, as the basis, and integrates the methods and ideas of data mining into the mining of data laws in the field of KOL identification of key opinion leaders, with a view to finding valuable intrinsic laws and relationships between the mining of web crawler technology and the identification of KOL features. The research content of this paper mainly includes two aspects of research work, the design of high-performance data mining technology and the actual work in the field of KOL recognition. On the one hand, this paper comprehensively describes the basic theory and methods of data mining, and focuses on the in-depth analysis and elaboration of BP neural network-based data mining technology on the basis of understanding and analyzing a variety of data mining technologies. On the other hand, this paper aims to solve the problem of bias in the prediction of traditional KOL model and design the experimental method of KOL recognition by BP neural network algorithm.

Investigation of failure analysis and failure delay strategy for H13 steel contact wire clip die

Investigation of failure analysis and failure delay strategy for H13 steel contact wire clip die