Sam Hoff is the president and CEO of Patti Engineering, a Certified member of the Control System Integrators Association (CSIA). Under Sam’s leadership, Patti Engineering has become a leader in applications of Industry 4.0 and digital twining to improve manufacturing through AI and data science. Plant Service spoke with Sam early last year about how maintenance teams can get their digital programs off the ground, and Sam checked in recently with Editor in Chief Thomas Wilk about what teams can do to kick-start these programs in 2022.
PS: Sam, the last time we caught up, COVID was just a whisper coming from outside of the U.S., and your last published interview with us was in March 2020, the month everything sort of went down in the U.S. Our conversation then focused mostly on how plant teams could work with integrator partners to get their IIoT products off the ground. A lot of what we heard in the past 18 months has been that these products are only accelerating. Can you tell us some of what you've seen and heard in this regard?
SH: Tom, without a doubt, that projects have really accelerated. I think COVID has given Industry 4.0 a kick in the butt, to take it from the concept to practical applications. We're seeing all types of IIoT projects and IIoT requests. There's a couple of things driving this: (a) a shortage of workers and shortage of labor, and (b) wanting to keep people off the line and working remotely as much as possible. So, what kind of information can you glean from the production systems without being beside it to be able to track improvement.
PS: When it comes to IIoT projects, you also encourage siloed teams inside the plant to break down, especially the departments like controls, industrial engineering, and maintenance. Our readers have told us over the past 18 months that they want to do this, but they're also finding it a bigger challenge than they expected. It's one of the obstacles they see towards implementing things like predictive maintenance and digital projects. Can you share some of your thoughts or tips on how teams can do this successfully?
SH: You definitely need management buy-in on this. It has to be somebody pushing from above, because with these groups you're moving their cheese a little bit. They're going to be doing things differently. They're going to be interacting with people that maybe they're not used to interacting with. If you leave it up to all those groups individually, you're not really going to have success with it. It has to be driven by the operations management at that facility. We've seen successful implementations and unsuccessful implementations, and it all comes down to operations management.
PS: Is the skill to break down silos something that you find that people can learn on the fly, or is it something which you'd recommend, say, a training course or two could be helpful to help teams learn how to communicate across the boundaries?
SH: I think there has to be a regular cadence of meetings with all the stakeholders in it, but you can't have too many people in it. I would say the group has to be limited to about 6 to 10 people: a couple of representatives from each group, and maybe operations management. How do you train people to break down silos? I mean, it’s not really that easy, but once you get momentum, once these different groups that maybe do not communicate with each other often get those lines of communication going, it can be really successful.
PS: We've seen people doing it out of necessity too. The extent to which people are sharing their data outside the plant walls is increasing, especially on a quarterly and sometimes monthly basis, due to the shortage of people on site.
SH: Yep, and if you think about it, before if you had seven or eight plants across the USA, let's say, or even 15 plants across the world, it was easy to get a subject matter expert to go to this facility for a week, to go to that facility for a week. Well, with travel restrictions, and COVID, and all that stuff, now they can't go. So how do you give that subject matter expert eyes into the facility to be able to understand what is happening? A lot of that stems through digital technologies.
PS: Let's look forward a little bit into 2022. One issue that people that fought with this year and which we've heard they're looking forward to next year is cybersecurity, and I imagine as integrator, you get a lot of questions about whether people’s systems are secure, or what they can do to make them secure. What are some things that plants can do to be proactive in this area?
SH: Test, test, retest, and retest. There's been some pretty high-level cases recently, where you get the message, "You need to send Bitcoin to this encrypted address, and you get your system back running." There's been cases where even when you do that, you might be able to get back up on your backups or your systems quicker. So, you need to constantly make sure that your systems are backed up, you're prepared for it, you need to test – throw the kill switch some weekend, make sure you're going to see how long it will take you to recover, simulate what a cybersecurity attack will look like.
Listen to the entire interview
There's a lot of devices out there that can do packet sniffing as part of layers of defense, but you have to be testing. A lot of your industrial control components were not designed with cyber-security in mind, especially older ones. They are the most vulnerable part of your operations.
PS: If I can ask a question about that real quick, are there formal assessments or informal checklists people can use to look at their equipment?
SH: Absolutely. One of the organizations I really like is the (CS)2AI (www.cs2ai.org). They have monthly forums and they do webinars, so I would suggest going to that organization.
PS: Let's say the worst-case scenario happens, and a plant does get targeted. We've heard some talk of people putting response plans in place. Often that might be driven by finance to help limit the risk of exposure of a company. What are some things people can do to be prepared for when they get targeted?
SH: You definitely want a response plan in place but I think more than that, you want to test your response plan, so you want to run some type of simulations. You can even have some people try to hack into your systems to see what the vulnerabilities are. It's like anything, right? You have to test it in order to see how secure you are.
PS: Let me ask you one more question about 2022. Integrators are increasingly engaged with the maintenance teams, so given how many projects are digital, and how many systems are networked, what kind of projects do you anticipate working on the most in 2022?
SH: A lot of digital-type projects, a lot of getting information off the plant floor, and a lot of analyzing that information. We're seeing the need within integrators for doing some data science, you know, Power BI, trying to do some machine learning off the data that you're seeing, a lot of digital twinning and this kind of stuff. I really think digital technologies will drive a lot of projects in 2022.
PS: Your comment about data science is really interesting, and it makes sense that integrators will grow in that direction, but honestly, you're the first person I've talked to who has indicated that. Are you seeing more of that?
SH: Yeah, I'm starting to see some leading edge to that. And one of the things is, it's such a fascinating topic. I'm a huge baseball fan, and you'll see that data science has come to baseball big time, with the shifts that they employ and all this stuff. I saw Joey Gallo, for the Yankees, line out to left field yesterday while the Tampa Bay Rays have four outfielders on him! They're playing like a softball lineup against them because this guy's a fly ball or strikeout hitter. But the one thing about all the data science people in baseball is, you’ve got to be able to contextualize the data you're looking at.
Some companies have been burnt by bringing a data scientist who takes a look at the data coming out of the processes and says, "Well, this is what I'm seeing," but they have no ability to contextualize what they're looking at. And a lot of times, the conclusions they're driving are either inherently obvious, or they make no sense whatsoever, because they don't have the ability to contextualize the data. You still need some ability to contextualize the data when you're using data science strategies, and that's why integrators are uniquely suited for that.