GFT has deep knowledge in automotive and manufacturing, and it has worked closely for many years with Google Cloud, the U.S. auto manufacturer's preferred cloud provider. That enabled GFT to develop a cloud platform that could collect, aggregate, and synthetize manufacturing data from disparate sources onto a single application based on Google Cloud’s Manufacturing Data Engine platform.
GFT created data parameters, processed the collected information, and then used machine learning to develop dashboards and other data visualization interfaces for U.S. auto manufacturer and its suppliers.
For a test bed, GFT and the U.S. auto manufacturer selected a stamping plant that makes outer body panels for the F-150 and Transit commercial vans. The line produces about 900 parts per hour, or about 3 million panels a year. For each of those, the U.S. auto manufacturer sensors and other monitoring equipment track incoming materials, analyze assembly line operations, and collect images of each part produced by the line. The U.S. auto manufacturer hosted the image collection on Google Cloud, but the volume of data was overwhelming.
In a matter of months, the U.S. auto manufacturer had amassed 500 million data points that needed flexible and reliable storage. One of the first steps was working with the U.S. auto manufacturer to identify extraneous data that it didn’t need for the new application. GFT took the remaining data sets and defined operational and technical requirements, and then tested the system on the stamping line.
By processing and analyzing the stamping-line data, the U.S. auto manufacturer was able to spot variations in materials, control the stamping process parameters more effectively, and determine the output for better-quality parts.
Collecting all the data on a single platform allows the U.S. auto manufacturer to review data sets from different sensors. By comparing two sets of data from similar equipment, the U.S. auto manufacturer can monitor differences in performance, spot malfunctions early, and make repairs with less disruption to the production line.
GFT conducted the U.S. auto manufacturer pilot effort in about three months, and the entire rollout was completed in about six months.