Research On Optimization Design Method Of Vertical Screw Conveyor Based On GA-BP Neural Network

Screw conveyor is widely used in agriculture,port,machinery,chemical industry,light industry and other fields because of its green and efficient conveying capacity.The design and manufacturing technology of screw conveyors in my country are backward,and the conveying efficiency is low.The main reason is that the design method of screw conveyors is not perfect.Screw conveyor design is a complex design process,including multiple input and output parameters,and the existing functional form cannot accurately represent the relationship between different input and output parameters.This paper,by using neural network theory and method,in view of the vertical screw conveying machinery with multi-dimensional complex mapping relation between the input and output multidimensional,and research on optimal matching design of design parameters to achieve collaborative optimization of performance parameters.,so as to improve the transmission efficiency of vertical screw conveyor,are of great theoretical significance and engineering value.The main research results are as follows:Based on discrete element theory,the simulation model of vertical screw conveyor and simulated particles,as well as the contact model between particles and between particles and between particles and components are established.In this paper,the research on the flow characteristics of the particles is carried out,and the above research provides a simulation method for the establishment of the neural network simulation sample set.Based on the periodic structure characteristics of the vertical screw conveyor,a multi-scale structural simulation model of the vertical screw conveyor is established,which is divided into macro-scale,meso-scale and micro-scale structural simulation models.And the multi-scale structure simulation model is used to simulate and analyze the particle axial velocity of the vertical screw conveyor,which proves that the macro-scale structure simulation model can be replaced by the micro-scale structure simulation model.The calculation amount of the simulation data is reduced,and the establishment efficiency of the neural network simulation sample set is improved.The influence degree of four influencing factors on the axial velocity of particles was studied,including screw speed,filling rate,helical pipe diameter and pitch.On this basis,the orthogonal test design was carried out,and a high-quality orthogonal test table was established.The multi-dimensional input design parameters of the neural network are established,and the simulation times of the neural network simulation sample set are reduced.In this paper,the performance characterization and performance parameter calculation method of vertical screw conveyor are studied.A calculation method of the average axial velocity of particles based on probability density,a calculation method of the conveying amount based on discrete element,and a calculation method of power consumption based on VDI are proposed.The performance parameters and calculation methods are established for the multidimensional output of the neural network.Through the establishment of BP neural network simulation training set,the construction of BP neural network model,and the optimization of genetic algorithm(GA),the GA-BP neural network model between design parameters and performance parameters is established.Both training regression and population regression studies on performance parameters of vertical screw conveyors were carried out.The reliability of the established GA-BP neural network model is proved.This paper uses the established GA-BP neural network model to predict the performance parameters of the vertical screw conveyor under different working conditions.And the best matching design of design parameters is carried out by using range analysis,variance analysis and comprehensive balance method.So as to achieve the synergistic optimization of performance parameters to improve the conveying efficiency of the vertical screw conveyor.In this paper,a vertical screw conveyor experimental system is built,and the experimental method and the experimental acquisition method of performance parameters are studied.The performance parameters such as average axial velocity,delivery volume,power and other performance parameters of particles are verified by experiments,which proves the correctness of the established GABP neural network model.This paper studies the optimization design method of vertical screw conveyor based on neural network through the methods of theoretical analysis,model establishment,simulation analysis and experimental verification.The research results can provide design ideas for the design and manufacture of my country’s conveyor field and ship unloader field,and accelerate the speed of my country’s screw conveyor and screw ship unloader field to the world’s advanced level,which has great theoretical significance and engineering value.