The production capacity of the factory is limited by the long throughput time of the x-ray end-of-line test.
Requirements:
Develop algorithm to predict likelihood of manufacturing defects and therefor the need for quality testing by applying closed loop analytic approach
Increase throughput in production by reducing number of necessary x-ray tests
Save time, effort and costs
Our Solution:
Gather process data from the physical equipment and x-ray results of produced pieces (quality labels)
Model multivariate dependencies of process data and x-ray labels by means of supervised machine learning algorithms (binary classification)
Integrate predictive algorithm into Plant-IT system landscape
Continuously re-train the prediction model to increase the accuracy of the predictions
Minimize number of x-ray tests and test only when required to eliminate costly bottleneck
Continuously running algorithms integrated into the production process on-site
Re-trained, up-to-date prediction models
Siemens Amberg produces about 6 million SIMATIC-Products annually. 75% of the value chain is automated, making the factory one of Siemens’ leading manufacturing sites for digitalization.
The project addresses the surface mounted device line with X-ray end-of-line (EoL) testing which takes an X-ray of solder joints of BUS-Connector PINs within a SIMATIC ET200SP Base Unit.
