Application of an ANN-GA approach to enhance the efficiency of resistance spot welding between SS400 steel and SUS304 stainless steel

Name: Apisit Kaewchalun

Organization : Nakhon Phanom University. Faculty of Industrial Technology

Name: Anantachin Khamsupa

Organization : Nakhon Phanom University. Faculty of Industrial Technology

Name: Kiattipong Onbut

Organization : Nakhon Phanom University. Faculty of Industrial Technology

Name: Yodprem Pookamnerd

Organization : Nakhon Phanom University. Faculty of Industrial Technology

Email : Yodprem.p@npu.ac.th

keyword: Dissimilar metal joints

LCSH: Electric welding

; Shear tensile strength

LCSH: Neural networks (Computer science)

LCSH: Genetic algorithms

Abstract: Resistance spot welding (RSW) of dissimilar metals, such as SS400 structural steel and SUS304 stainless steel, often faces challenges due to differences in material properties, leading to inconsistent weld quality. This study proposes a hybrid predictive and optimization approach by integrating an Artificial Neural Network (ANN) with a Genetic Algorithm (GA) to enhance weld performance. Experiments were conducted using varying electrode diameters (5, 8, and 10 mm) and welding times (10, 20, and 30 cycles) under a constant current of 8000 A. Tensile-shear force and Vickers hardness were measured and used to train a feedforward ANN with a 2–3–2 architecture. The ANN was then incorporated into a GA framework to identify optimal input parameters for improved mechanical properties. The ANN model achieved high prediction accuracy with R² values of 0.9565 for tensile-shear force and 0.9275 for hardness. The optimal welding parameters identified were a 9.59 mm electrode diameter and a welding time of 30 cycles, yielding predicted values of 7.85 kN and 445.02 HV. Convergence and regression analyses confirmed the robustness and generalization capability of the ANN-GA model. This hybrid approach reduces trial-and-error in welding parameter selection, enhances prediction accuracy, and optimizes weld quality. It demonstrates strong potential for real-world industrial implementation and paves the way for future adaptive control studies.

King Mongkut's University of Technology North Bangkok. Central Library

Address: BANGKOK

Email: library@kmutnb.ac.th

Created: 2025

Modified: 2025-11-25

Issued: 2025-11-25

บทความ/Article

application/pdf

BibliograpyCitation : ใน สมาคมวิศวกรเครื่องกลไทย และมหาวิทยาลัยเทคโนโลยีราชมงคลอีสาน วิทยาเขตขอนแก่น. คณะวิศวกรรมศาสตร์. สาขาวิชาวิศวกรรมเครื่องกล. การประชุมวิชาการเครือข่ายวิศวกรรมเครื่องกลแห่งประเทศไทย ครั้งที่ 39 (ME-NETT 2025) (pp.365-377). ขอนแก่น : มหาวิทยาลัยเทคโนโลยีราชมงคลอีสาน วิทยาเขตขอนแก่น, 2568

eng

©copyrights King Mongkut's University of Technology North Bangkok

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	ME-NETT 2025pp.365-377.pdf	559.8 KB

ใช้เวลา

0.041243 วินาที