Perhaps the best things about travel and the conferences I attend are conversations with really intelligent people in the industry. The week of May 14 was no exception as I drove down to the headquarters building of the Procter & Gamble Co. in downtown Cincinnati for meetings of the Center for Intelligent Maintenance Systems (www.imscenter.net). There are so many people with Ph.D.s in engineering in attendance that I was totally intimidated. But the flow of ideas among academics, technology suppliers and end-users was tremendous. Two side conversations during the event really stuck with me. Jay Lee, professor of engineering at the University of Cincinnati and director of the Center, asked me several questions that I assembled as the lead question for this essay. So I was already pondering the future when I had an opportunity to chat with Rob Aleksa, who is machine control section head at P&G’s Corporate Engineering Technology Center. Rob and I had talked a few months ago about the needs of companies such as P&G for some sort of common method for original equipment manufacturers to program their machines. P&G’s problem (and it is not alone) is that it buys machines from many, probably hundreds, of machine builders. P&G must put the machines together into an assembly line and then maintain and optimize them. If every machine is different, that maintain-and- optimize task is huge. Research and implementation The Center’s research academics have studied many problems in manufacturing focused on machine component and sensor degradation. The idea was to determine signatures and then prognostics - algorithms that could take specific sensor input and predict component failure. Using this, maintenance personnel and engineers could optimize maintenance procedures and keep machines running for more hours per year. The algorithms have been defined, tested and packaged as a set of software tools known as the Watchdog Agent. The problem now is how can this be implemented? Let me tie those thoughts together by looking at a common denominator—technology suppliers. To expect these solutions to be implemented by either end-users or OEMs is similar to expecting the Cubs to win the World Series or the Browns to reach the Super Bowl. But if the technology suppliers actually did embed these ideas into their control platforms, then it would be easier to achieve adoption. In fact, it would be in everyone’s best competitive interest. Imagine being the first technology provider to implement the Watchdog Agent in its control platform easily configured in the engineering software? The company’s sales force could go to every manufacturer in the world and show them a tremendous value-add for improving machine uptime. The same thing for the first supplier who could package an easily configurable state model for machines into its engineering software. That would produce savings for OEMs by making it easier for them to move new technologies into machines, while at the same time endearing them to users who could then understand how their machines operate. This would be powerful. It would change the industry. What are the most important technologies coming in the next five years for manufacturing? Aleksa hit me with that question as I was pondering the technology implementation question. How about the next step in programming which is “non-programming.” Embed a machine state model that’s easy to configure into a machine program that is a “standard” within the company’s engineering tool. Next generation software engineering for the next generation of manufacturing engineers. And remember the point of all this, in the words of Jay Lee, “It’s not about maintenance. It’s about productivity.” The problem now is how can this be implemented?
Leaders relevant to this article: