We have a problem
Our customer is an automotive manufacturer who is producing injection molding parts that are later used at the vehicle assembly line. Recently there have been several complaints from final assembly. Parts are defective several times per week and therefore negatively impacting vehicle output.
Looking into the problem
In order to tackle this problem a computer vision system is installed right before the parts are packaged for the assembly line. The responsible production engineer is training an artificial intelligence (AI) model within a day without the need of knowing all the data science going on behind the scenes (see chapter “Setting up Collaborative Vision in detail”). The images and results of the quality check are automatically transferred to the Intelligent Operations Platform (IOP). This makes all the data available for further analysis without the typical hurdles we see today when connecting new sensors. The logistics employees receive feedback once defective parts are detected and sort them out.
Creating objective measures
Having the data in IOP, the impact of the problem can be efficiently and objectively quantified. Gut feeling is replaced by evidence. Our scenario problem has worsened. Defects are now detected several times a day. While people are acting on the defects, the root cause is still undetected. Three things are making it complex for our production engineer: Parts cannot be traced back to the production situation in injection molding machine (IMM), the defects only happened sporadically, and the error image is vague.
In our next post we will show how our customer detected the root cause.
Find out how Capgemini can accelerate the effectiveness of your digital manufacturing operations at scale with our “Factory of the Future” approach here:
Stay tuned for part 2 of this blog series.