Week 7: Designing Experiments

This past week in my internship at Boston Scientific, I started working on my primary project. When one of Boston’s product lines first entered production, they used a particular technology to conduct one of the fabrication steps. However, in recent years, they began moving to a newer technology to conduct that step because the newer option had better process controls, and therefore should be able to more consistently produce products that meet their specifications. Currently, Boston has successfully validated (or shown that a process can consistently produce good results) this technology for products in a given product line, but not all.

For my assignment, I was added to a team that is beginning the validation process for a group of products with this new technology. The team’s next task is to conduct an experiment attempting to show the effect of all important factors on the process. The first task given to me was to fabricate some parts with the new technology under a wide range of settings and test them so that I could make recommendations for which parameters we should investigate, as well as the values those parameters should be set to in the experiment.

Got my own cubicle, with snacks from coworkers included!

Fabricating and evaluating the data took up the entire week, and I definitely learned some lessons along the way. When initially talking to the team, one of them said to just “play around with the settings to see what works”; I may have taken that a little too seriously at first. For the first product, I tended to investigate parameters as I thought of them and with little structure. This method was helpful in gaining some intuition as to how the different parameters affected the results, but I found later that it made it difficult to evaluate the data for trends because of the existence of confounding factors. I tended to change multiple variables at a time, and therefore could not get a clear picture of the data. As time went on and I started evaluating my data for recommendations, I saw the value of having a more structured approach to controlling the variables. I was also more constrained for time, having spent more time than I should have on the first of five products, so the structure also helped expedite the process. If I had the opportunity to do it again, I would have tried to be more methodical from the beginning.

I also learned a little about myself this week. My experience with this project thus far has been much more hands-on than theory-based. In manufacturing, you can propose all the ideas you want for why different parameters would be more or less effective at controlling a process, but at the end of the day you have to show it. Process validation requires evidence that using this machine at these settings with these restraints produces acceptable results by actually doing it that way. I have not had much time to dig into the theory behind my process beyond the basic physical principles I know from my classes at Texas A&M, but instead have needed to use a lot of trial and error. It may be that my limited time here makes it challenging to understand the theory behind the process, but I get the impression that manufacturing engineers tend to work far more in trial and error than in theory. Undoubtedly having a good understanding of a process would guide an engineer to good solutions, but it seems that there are often too many confounding factors that render a reliance on theory impractical to the manufacturing engineer. Nevertheless, it has only been one week, and I know better than to stop at the surface. I’m excited for the rest of the internship and for everything else I will learn in my time here!

This weekend, I made it back to the States for a friend’s wedding. While it was physically exhausting, it was great to see old friends, celebrate his marriage, and have a little taste of home. Being emotionally recharged, I’m ready to finish strong for the next 4 weeks!

This entry was posted in Ryan. Bookmark the permalink.

Comments are closed.