Statistical Method to Model the Quality Inconsistencies of the Welding Process

Mohammad Aminisharifabad,Qingyu Yang
DOI: https://doi.org/10.48550/arXiv.1902.08869
IF: 5.414
2019-02-24
Machine Learning
Abstract:Resistance Spot Welding (RSW) is an important manufacturing process that attracts increasing attention in automotive industry. However, due to the complexity of the manufacturing process, the corresponding product quality shows significant inconsistencies even under the same process setup. This paper develops a statistical method to capture the inconsistence of welding quality measurements (e.g., nugget width) based on process parameters to efficiently monitor product quality. The proposed method provides engineering efficiency and cost saving benefit through reduction of physical testing required for weldability and verification. The developed method is applied to the real-world welding process.
What problem does this paper attempt to address?