Issue link: https://iconnect007.uberflip.com/i/1372612
28 PCB007 MAGAZINE I MAY 2021 ing for a network of data flow throughout your entire enterprise, making it possible for anything to be stored digitally, and available for anyone involved. For example, in the case of a production floor, consider a quali- ty containment proce- dure in which there is a batch of product cur- rently in production. Some of these items roll off the line with some specs out of measure- ment, and collecting these measurements is critical. Aer it is identified and your quality engineer gets the information that something is off, the first thing is to contain the production as there may be many reasons and parameters involved. For instance, consider that the root cause of this quality defect is raw material. To identify the kind of raw material problem you're hav- ing, you will probably need your ERP system for this information. Consider you have a typ- ical 3.0 automation stack, or a pyramid if you will, in which at the bottom you have PLCs and control systems, followed by some SCADA and the MES system—if someone has this on paper it is probably completely impossible. It might take many hours, days, or even weeks for such an engineer to get the infor- mation and correlate everything in a stack like this. In some cases, we have heard about con- tainment procedures taking up to two weeks to collect all the necessary people who make the MRP meetings. But, if you can anticipate such a scenario and you have all your data available, you can establish either of the machine-learn- ing algorithms to analyze parameters and cor- relations in between so that you can stop any batch right away that has a specific feature you are looking for; then, the decision can take sec- onds. at is a huge advantage to preventing scrap manufacturing, and this is one of the use cases that provides huge benefits. Matties: Part of the challenge, though, is the skill set: to capture and interpret the data, and to make it meaningful to their process. at's a skill set oen missing in manufacturing right now, particularly for bare board fabricators. Zajac: It is considered uncommon to have this data available, but like the use of Microso Ex- cel is now a standard skill throughout the in- dustry, this will be the same skill that is re- quired in the future. Being able to analyze some data related to your scope of interest is going to be critical. Matties: I think the new role is being labeled as a data scientist. is is a new job title coming into the marketplace. Zajac: Right. at person will be involved at any step throughout the industry, from pro- duction floor through quality engineers, pro- cess engineers, and management as well. Alex Stepinski: Regard- ing the data scientist job description, I'm doing some mid-ca- reer executive educa- tion courses right now and the curriculum as- sociated with data sci- ence—if you compare it to the curriculum as- sociated with mechan- ical or chemical engineering—is a subset. It's taking what people with engineering degrees learn and separating it out into a separate cur- riculum. It's not as rigorous, so someone with a business background can take the course, as opposed to doing differential equations for fun, which Robert does in his spare time every night (laughs). Matties: I think there's the need for the skill set but there's also the need for a circuit board shop to hire this type of person. Robert Zajac Alex Stepinski