Neural Network Arithmetic Research Articles

Retraction Note: A fuzzy control model based on BP neural network arithmetic for optimal control of smart city facilities

Retraction Note: A fuzzy control model based on BP neural network arithmetic for optimal control of smart city facilities

Optimization of blasting parameters and prediction of vibration effects in open pit mines based on deep neural networks

Embedded systems in production equipment and Internet of Things (IoT) sensors on production lines are one of the elements that constitute an industrial cyber-physical system. In this paper, an in-depth study and analysis of the optimization of blasting parameters and prediction of vibration effects in open pit mines using deep neural network arithmetic are present. Based on the deep neural network research and analysis of the relationship between blasting parameters and rock fragmentation, a prediction model for blasting parameters and fragmentation for the East Open Pit Mine was established, and sensitivity analysis was performed on blasting parameters, and the unit consumption of explosives and the perforation rate were established. It was found that the average relative errors of both numerical simulation results and depth prediction results were no more than 10%, while the average relative errors of Sadowski's formula prediction results were more than 20%. The results show that the neural network optimized by a genetic algorithm and the numerical simulation has the highest accuracy in predicting the blasting result parameters. The research model and results obtained in this paper can be used as a reference guide for engineering practice.

RETRACTED ARTICLE: A fuzzy control model based on BP neural network arithmetic for optimal control of smart city facilities

In order to analyze and study the fuzzy control model of smart city optimization control, this paper designs a fully networked fuzzy controller based on BP neural network arithmetic, so that the realization process of fuzzy reasoning is networked and clear. The evaluation mode of wisdom town is constructed, and the index system is simplified according to the evaluation index system of the develop possible of wisdom town. The analog outcome shows that the above arithmetic can availably optimize the parameters and structure of the neuro vagueness control, and the designed neuro vagueness control has good performance. The build of wisdom cities is advantageous to further promoting the deep integration of industrialization, informationization, urbanization, and agricultural modernization, which is of great significance for solving urban development problems and promoting continuable town develop. To improve the development capacity of intelligent city, we should highlight the characteristics of smart cities; advocate the concept of sustainable development and low-carbon development, market-oriented, overall planning, top-level design, and micro-regulation; and take the road of development with local characteristics.

Optimization of impulse water turbine based on GA-BP neural network arithmetic

To develop an optimum design method for impulse water turbines with low specific speed, a representative impulse water turbine with low specific speed used in agricultural irrigation machinery was optimized with a combination of an orthogonal experimental design, a genetic algorithm, and a BP neural network in this study. Numerical calculation was applied to analyze interflow characteristics for optimized and original water turbines. Results showed that the internal flow characteristics of the optimized water turbine presented remarkable improvement compared with the original water turbine. Pressure distribution increased, the vortex strip in the draft tube was reduced remarkably, and impeller torque increased by 26 %. In addition, the optimized impeller was manufactured by 3D printing, and performance comparison was conducted between experiments of the optimized and original water turbines. The efficiency of the optimized water turbine reached 42.5 %, which exceeded the original water turbine’s of 8.5 %. With increasing rotating speed, maximum efficiency running point moved to a high flow rate, and highly efficient areas expanded. Internal characteristic analysis and a full-scale experiment for both water turbines showed that the performance of the optimized water turbine exhibited substantial improvement. The analysis and experiment also verified the theoretical correctness and feasibility of the proposed optimum design method.

Inversion of Density Interfaces Using the Pseudo-Backpropagation Neural Network Method

This paper presents a new pseudo-backpropagation (BP) neural network method that can invert multi-density interfaces at one time. The new method is based on the conventional forward modeling and inverse modeling theories in addition to conventional pseudo-BP neural network arithmetic. A 3D inversion model for gravity anomalies of multi-density interfaces using the pseudo-BP neural network method is constructed after analyzing the structure and function of the artificial neural network. The corresponding iterative inverse formula of the space field is presented at the same time. Based on trials of gravity anomalies and density noise, the influence of the two kinds of noise on the inverse result is discussed and the scale of noise requested for the stability of the arithmetic is analyzed. The effects of the initial model on the reduction of the ambiguity of the result and improvement of the precision of inversion are discussed. The correctness and validity of the method were verified by the 3D model of the three interfaces. 3D inversion was performed on the observed gravity anomaly data of the Okinawa trough using the program presented herein. The Tertiary basement and Moho depth were obtained from the inversion results, which also testify the adaptability of the method. This study has made a useful attempt for the inversion of gravity density interfaces.

Control Strategy Based on Wavelet Transform and Neural Network for Hybrid Power System

This paper deals with an energy management of a hybrid power generation system. The proposed control strategy for the energy management is based on the combination of wavelet transform and neural network arithmetic. The hybrid system in this paper consists of an emulated wind turbine generator, PV panels, DC and AC loads, lithium ion battery, and super capacitor, which are all connected on a DC bus with unified DC voltage. The control strategy is responsible for compensating the difference between the generated power from the wind and solar generators and the demanded power by the loads. Wavelet transform decomposes the power difference into smoothed component and fast fluctuated component. In consideration of battery protection, the neural network is introduced to calculate the reference power of battery. Super capacitor (SC) is controlled to regulate the DC bus voltage. The model of the hybrid system is developed in detail under Matlab/Simulink software environment.

Electric and Hydraulic Servo Actuator Fault Diagnosis Based on Genetic Optimization Neural Network

To resolve the complexity and uncertainty between the fault portent and the causes of electric and hydraulic actuator, a fault diagnosis model base on neural network arithmetic for actuator is designed. The genetic optimization arithmetic is used to initialize the value of weight and bias to improve the search efficiency and global convergence of the network. The simulation result indicate the feasibility and validity of the method based on genetic neural network model to diagnose the faults of electric and hydraulic servo actuator.

RBF Neural Network Arithmetic and Applications in Surface Interpolation Reconstruction

Aiming at problems such as: surface interpolation reconstruction of points cloud data，surface hole filling and two simple surface connection, a neural network arithmetic was employed. Based on radial basis function neural network, simulated annealing was employed to adjust the network weights. The new arithmetic can approach any nonlinear function by arbitrary precision, and also keep the network from getting into local minimum for global optimization feature of simulated annealing. MATLAB program was compiled, experiments on points cloud data have been done employing this arithmetic, the result shows that this arithmetic can efficiently approach the surface with 10-4 mm error precision, and also the learning speed is quick and reconstruction surface is smooth.

Laser Cladding Height Prediction Based on Neural Network

Real-time detection and closed-loop control of laser cladding height is necessary for forming high quality parts. Technological parameters are coupled and the forming process is a non-linear process. A large number of laser parameters affect the quality of the laser cladding surface. Based on the analysis of the influence of laser parameters on cladding height, the BP (Back propagation) neural network prediction model of cladding height is build. The neural network arithmetic is designed and the samples are acquired by laser forming experiment. The training samples are used to train the network to accomplish the mapping relation between input and output of the network. The test samples are used to verify the performance of the trained network. Simulation results indicate that the prediction model has sufficient accuracy. The BP neural network prediction model of cladding height is feasible and valid in theory and in practice. The laser cladding height BP neural network prediction model lays the foundation for real-time height prediction and closed-loop control in laser forming process, and it has great significance for improving the quality of formed parts.

Hole Filling Algorithm in Surface Reconstruction Based on Radial Basis Function Neural Network

Aiming at hole filling in points cloud data reconstruction, a novel neural network arithmetic was employed in abridged points cloud data surface reconstruction. Radial basis function neural network and simulated annealing arithmetic was combined. Global optimization feature of simulated annealing was employed to adjust the network weights, the arithmetic can keep the network from getting into local minimum. MATLAB program was compiled, experiments on abridged points cloud data have been done employing this arithmetic, the result shows that this arithmetic can efficiently approach the surface with 10-4 mm error precision, and also the learning speed is quick and hole filling algorithm is successful and the reconstruction surface is smooth. Different methods have been employed to do surface reconstruction in comparison, the results illustrate the error employed algorithmic proposed in the paper is little and converge speed is quick.