Welding process parameter optimization method based on a BP neural network

A BP neural network and process parameter optimization technology, applied in neural learning methods, biological neural network models, design optimization/simulation, etc., can solve the problems of complex welding process debugging, low welding production efficiency and process design efficiency, etc., to achieve The effect of simplifying the process of welding parameter adjustment, flexible debugging, and simplifying the adjustment process

Inactive Publication Date: 2020-04-21
上海智殷自动化科技有限公司
View PDF1 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The shortage of professional welding technicians, the complexity of the welding process debugging process, and t

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Welding process parameter optimization method based on a BP neural network
  • Welding process parameter optimization method based on a BP neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0020] The invention provides a process parameter optimization method based on BP neural network. The specific idea is: based on the finite element calculation method, a steel-aluminum laser welding model is established and a heat source is loaded, the heat source model is corrected through experimental temperature measurement, and then the calculated Analyze the temperature field and change law of the temperature field to obtain the training sample data of the neural network; establish a neural network model for steel-aluminum laser welding, use the training samples obtained to learn and train the neural network, and optimize the welding process parameters through the neural network; select The range of laser welding process parameters, design orthogonal experiments, obtain test results through numerical simulation, and obtain welding penetration data under different process parameters as neural network training samples; use training samples to train BP neural network to obtain...

Embodiment 2

[0022] The welding process is a complex process involving physical changes, heat transfer, mechanics, metallurgy and other aspects, including electromagnetic, heat transfer, melting and solidification, crystallization, transformation and other stages of process. It is impossible for finite element analysis software to analyze all processes All of these factors are taken into consideration, and some reasonable assumptions and simplifications should be made for the variables with weak influence when establishing the geometric model of laser welding. It avoids the problem that too much data leads to a large amount of calculation, too long time consumption or difficulty in calculation convergence. Considering that the experimental object is a steel-aluminum thin plate, the Gaussian surface heat source model is selected, and the laser power, defocus rate and welding speed are changed to obtain the penetration depth data under different process parameter combinations.

[0023] In th...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a process parameter optimization method based on a BP neural network. The specific thought of the method is as follows: based on a finite element calculation method, establishing a laser welding model and loading a heat source, correcting the heat source model through experimental temperature measurement, and then analyzing a temperature field and a change rule obtained through calculation to obtain training sample data of the neural network. The BP algorithm is a multi-layer feedforward network trained according to an error back propagation algorithm, the learning ruleis a steepest descent method, the weight and threshold of the network are continuously adjusted through back propagation, and the network error is reduced. The welding process parameter adjusting process is simplified, and high-quality process parameters can be quickly adapted according to welding workpiece parameters; and technological parameters are optimized, and the mechanical property of a weld joint is improved.

Description

technical field [0001] The invention relates to the field of welding, in particular to a method for optimizing welding process parameters based on a BP neural network. Background technique [0002] Due to the high complexity, uncertainty, variability and multi-variable coupling of the welding process, it is extremely difficult to establish a general mathematical model for the welding process. Since there is no general mathematical model, in the actual welding process, engineers with rich welding experience are required to continuously adjust parameters such as arc start, arc end, gas supply, and welding torch speed according to the current welding conditions, and continuously optimize the welding effect to meet product requirements. This method of process parameter adjustment has extremely high requirements on the technical level of welding commissioning personnel. The shortage of professional welding technicians, the complexity of the welding process debugging process, and...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F30/23G06F30/27G06F30/17G06N3/08G06F119/14
CPCG06N3/084
Inventor 李玉霞
Owner 上海智殷自动化科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap