Machine learning and ANFIS for welding strength

What is it about?

This study focuses on predicting welding strength in sheet metal joints with nuts using machine learning algorithms and the Adaptive Neuro-Fuzzy Inference System (ANFIS). The research aims to optimize industrial welding processes and improve quality control, specifically focusing on DD13 sheet metal joints with AISI 1010 nuts. The study investigates the impact of weld current, time, and hold time on joint integrity, considering them as critical input variables. Various machine learning algorithms, such as linear regression, random forest regression, ridge regression, Bayesian regression, K-Nearest Neighbors regression, decision tree regression, and ANFIS, are evaluated for their performance in predicting welding strength. Training and testing data include welding parameters and corresponding strength measurements. Performance metrics like R2 score, mean absolute error (MAE), mean squared error (MSE), and root mean square error (RMSE) are used to assess the predictive capabilities of these algorithms. The study finds that random forest regression performs most efficiently, with a high R2 score of 0.992 and minimal errors. ANFIS also demonstrates comparable performance, highlighting its effectiveness in this context. The findings from this study can be valuable for optimizing welding parameters in industrial settings, potentially leading to improved quality control and weld strength, especially in automotive applications. Industries can use machine learning and ANFIS to make informed decisions, optimize welding processes, and ensure joint integrity to meet the demanding requirements of various applications.

Featured Image

Photo by Rob Lambert on Unsplash

Photo by Rob Lambert on Unsplash

Why is it important?

Optimization of Welding Processes: By using machine learning algorithms and ANFIS, the study aims to optimize welding processes, which is crucial for improving the efficiency and effectiveness of industrial welding operations. Quality Control Improvement: The ability to predict welding strength accurately can help improve quality control measures, ensuring that welded joints meet the required standards and specifications. Cost Reduction: Optimized welding processes and improved quality control can lead to cost reductions for industries, as they can minimize material waste and rework. Enhanced Product Performance: Welding strength prediction can lead to stronger and more reliable joints, enhancing the overall performance and durability of welded components, particularly in demanding applications like automotive manufacturing. Decision-making Support: The findings of this study can provide valuable insights and guidance for industries looking to optimize their welding processes and improve quality control, enabling them to make informed decisions to meet the rigorous demands of their applications.

Perspectives