Machine-Learning Software Enables Ford to Select Most Productive Stamping Processes

September 28, 2020


Identifying the optimum stamping process for a given part design can be labor-intensive and time-consuming, relying heavily on a stamping engineer’s knowledge and skill level. By leveraging Knowledge Studio machine-learning software from Altair, Ford has been able to speed its stamping-process selection and increase first-time-through (FTT) rates.

At Ford Mexico, as with many production facilities, multiple sheet metal stamping processes are available to form nested and individual parts, including progressive, transfer and tandem press lines. For a given part design, many factors determine the best or most efficient stamping process,  including material type and thickness, part width, and desired surface finish. While process selection typically falls to the manufacturing-process engineer, growing design complexity, nonconventional material types and numerous process combinations prove challenging, often ultimately requiring a labor- and material-intensive trial-and-error prove-out process. 

Material utilization represents a particularly critical benchmark. Most automotive plants expect 60-precent material utilization in stamping operations, with the remaining 40 percent considered wasted. Ford sought to improve these numbers, while simultaneously improving the selection of the right stamping process the first time and increasing its FTT rates. 

To move toward achieving these goals, Ford Mexico began documenting and amassing its vast quantities of clean data associated with successful inhouse production runs. Spanning a 5-yr. period, process engineers recorded successful stamping processes for thousands of parts. The historical data provided valuable insight, but the question remained: How could Ford use this information to help automate and guide the selection of the best stamping process for a given part design?

More than 30 years ago, Ford began working with Altair to support the company’s product-development activities. Today, Ford employs Altair software globally to support the development of Ford cars, trucks and heavy equipment. That relationship would continue through this project. 

First learning of Knowledge Studio through an Altair technology briefing, Ford Mexico began exploring the possibility of applying Altair’s machine learning and predictive analytics product to support the project. Leveraging the data Ford collected for more than 3000 stamping processes identified as representative of future requirements, Ford’s stamping experts and Altair’s solution architects collaborated to develop an accurate, reliable machine-learning model with Knowledge Studio.

Knowledge Studio offers 15 different machine-learning models, allowing users to explore, select and train the model that best fits their data. Using subsets of the data, the team ran a series of tests to determine the most effective model. With an accuracy rate surpassing 90 percent, the decision-tree model produced the most consistent results. The process also revealed a surprising and valuable discovery: When selecting the optimal stamping process, the most important factors are the overall dimensions and thickness of the finished part. Alone, these factors do not enable a final decision, but when combined with all of the other data points, Knowledge Studio’s machine-learning algorithm provides Ford with results approaching 100-percent accuracy.

This further enhances production efficiency and business value, according to Ford officials. Using the predictive power of Knowledge Studio enables Ford to minimize manual trial-and-error process validations and rework, thus providing stamping-process engineers with more time to address the most difficult and complex part designs. Overall, the projected throughput has increased by a factor of three and has increased FTT rates, resulting in reduced rework time without increasing resources. In addition, the Knowledge Studio machine-learning model effectively captures Ford’s inhouse domain knowledge to support a faster learning curve when training new personnel.

Industry-Related Terms: Lines, Form, Material Utilization, Model, Surface, Thickness, Transfer
View Glossary of Metalforming Terms

 

See also: Altair Engineering, Inc.

Technologies: Software

Comments

Must be logged in to post a comment.
There are no comments posted.

Subscribe to the Newsletter

Start receiving newsletters.