Altair > Case Studies > Reducing Weld Distortion in Automotive Manufacturing: A Case Study of Gestamp Tallent Ltd

Reducing Weld Distortion in Automotive Manufacturing: A Case Study of Gestamp Tallent Ltd

Applicable Industries

Automotive

Applicable Functions

Procurement
Product Research & Development

Use Cases

Manufacturing Process Simulation
Traffic Monitoring

About The Customer

Gestamp Tallent Ltd is a world-class designer, developer, and manufacturer of cutting-edge, chassis structural and suspension products, body in white structures, modules, and systems for the automotive industry. They specialize in developing innovatively designed products to achieve increasingly safer and lighter vehicles, thereby reducing energy consumption and environmental impact. With 96 manufacturing plants in 20 countries and a workforce numbering over 30,000 employees worldwide, Gestamp continues expanding in growth markets. They used the BMW MINI front subframe tower to demonstrate the weld distortion optimization approach.

The Challenge

Weld distortion is a significant issue in sheet steel product manufacture, especially in the automotive sector where tolerances are tight and complex high-performance components are the norm. The contraction caused by the welds cooling leads to distortion substantial enough to require additional processes to recapture the lost geometry. The sequence in which a part is welded largely affects the distortion of the part as the stiffness changes significantly depending upon which welds have already been executed. Gestamp Tallent Ltd, a world-class designer, developer, and manufacturer of cutting-edge automotive products, faced the challenge of accurately predicting weld distortion and optimizing the weld pattern to reduce it. They used the BMW MINI front subframe tower to demonstrate the weld distortion optimization approach. The tower is particularly susceptible to distortion due to its tall and thin dimensions.

The Solution

To tackle the challenge, Gestamp Tallent selected Altair’s HyperWorks CAE platform for use on their optimization projects. They chose Altair HyperStudy to further investigate weld removal optimization. HyperStudy enables users to run Design Of Experiments (DOE), optimization, and stochastic studies on models from numerous simulation codes. To prevent the optimization algorithm from removing too much weld from the tower and compromising the structural integrity, a constraint on the tower stiffness under loading was introduced. The initial optimization algorithm used was HyperStudy's Adaptive Response Surface Method (ARSM) managed by HyperStudy. After the ARSM algorithm gave a local solution, a genetic algorithm was run to investigate the possibility of a better global solution. A Hybrid Method Multi-Objective (HMMO) was then utilized for both gradient and global searching algorithms. HyperStudy was also used to optimize the sequence of welding to give minimal distortion of the tower.

Operational Impact

By using HyperStudy, Gestamp Tallent was able to balance the welds and suggest the most effective pattern to significantly reduce distortion without adding additional process costs or reducing the performance of the part. The use of HyperStudy to reduce weld distortion and define an optimum welding procedure produces high-performance, manufacture-friendly assemblies without expensive tool changes. The technology used in this project is generic and can be applied to any weld distortion project without any adaptation. This method is also best applied to areas in which welds can be balanced such that the distortional effects of one weld may be cancelled out or minimized by another.

Quantitative Benefit

Initial analysis work identified 20% of the total welds were critical for the stiffness and strength of the tower while inducing minimal distortion.
The weld removal optimization gave a 56% reduction in the distortion predicted at the top of the tower for an equivalent design.
The weld sequencing optimization resulted in a welding process with a predicted 93% reduction in distortion.