Honeywell Process Solutions CTO Bruce Calder dives into how the Industrial Internet of Things, big data analytics and cloud computing can help transform and update a plant.
Bruce Calder is on proverbial fire for the Industrial Internet of Things.
After more than two decades with Honeywell, the last seven and change as engineering director for Honeywell Process Solutions, which focuses on automation control and control solutions, equipment and services, he received a promotion last November to that division’s CTO. For the last 14 months, his work life has focused on IoT, IIoT and a host of other related technologies that aim to make things easier.
Calder has worked in the industry since 1980, with previous stints at GE, Motorola and Canadian Pacific Railway before working his way up the tech ladder at Honeywell, and he speaks about it in full paragraphs. What follows is part of an hour-long conversation about what HPS is working on, what really has him excited right now, and what the future might hold (because you can’t talk tech without talking about the future at least a little bit).
IndustryWeek: Let’s start with the Internet of Things. What does it mean to you, what does it mean for you, and what does it mean for Honeywell?
Bruce Calder: We’re focused on what the Internet of Things — the Industrial Internet of Things — means for the industrial world. I think it’s a fantastic buzzword, there is a lot of hype, and no one really knows what it is. If you ask multiple people, you get slightly different answers. There is something substantial there. It’s a game changer. It’s just great because it gets people to wake up and go, We have to do something, or we’re not going to be competitive in the future.
In one of our industries, oil and gas, there’s a huge opportunity to deploy this technology because of the low oil prices. There are lots of struggles in the industry, customers don’t have much money, and they’re very competitive in driving down their costs — and this is the investment they should be making. We’ve been working on new technology in HPS, with a big push on all these emerging technologies like mobility, cloud, analytics, big data, collaboration, mobile devices like wearables and augmented reality. How do we leverage all that to offer more benefits to our customers? And how do you convince the customers to actually deploy it? This whole IIoT hype is waking them up. It’s a great driver for our industry.
IW: Data seems like a big part of that, at least from your perspective.
BC: The heart of what IIoT is all about is data, data analytics, and especially big data analytics. One thing we’ve been working really hard on is, in our advanced solutions portfolio, in HPS, we’ve always done a lot of sophisticated analytics-type applications, but they’ve been server-based. The IoT, cloud and mobility really super-charged that whole area of our portfolio, so now instead of deploying applications in a control system at a site, we can put them in the cloud, totally eliminate the whole maintenance aspect at the plant, get a lot more data available in those applications from multiple sites, and deliver new value. … Before this system, we had independent plants calculating their own KPIs, all trying to say they’re the best. Now, the corporation can see real data, unmassaged, with the exact same KPI calculation across all the plants, and now they really know what’s going on.
IW: What are some of your other favorite technologies since you took over as CTO 14 months ago?
BC: We’re working on is augmented reality, which is kind of a real fun technology, and an eye-opener for me. If you think of augmented reality, I’ve got my iPhone, and I open up a cabinet, and there are a bunch of flashing lights and all these terminals, and I’m supposed to do something. Well, which wire am I supposed to work with? What do those flashing lights actually mean? Which controller am I even supposed to be repairing or doing some maintenance on? I can just hold my iPhone up, basically with the camera and, overlaid on my phone, it tells me that’s the controller name, here’s what those flashing LEDs mean, this is what the code means, here are the names of the terminal screws, here’s the current value of each screw. That’s one simple example of what you can do with augmented reality. You can drive tremendous performance improvement.
Being able to collect all this data from tens of thousands of devices and then relate that data to other data you never thought to relate it to before. And then figure out how to find value in it that no one else has found before. This is where you can start to dream about process data, relate it to commodity pricing, relate it to weather information, relate it to things you never thought would have a connection. But if you have millions of data points, you can start to use statistical algorithms, data science techniques, to see if there are relationships between these kinds of things. Ultimately, the whole goal is to predict what’s going to happen, so you can prevent problems. You can go through huge amounts of data and look for these patterns, then figure out how to monitor and predict what will happen so you can repair the problem before it even causes an event, and maybe even no is productivity lost.
IW: If the manufacturing buzzwords for 2015 are IoT and IIoT, I feel like the buzzword for 2016, maybe for 2017, will be unplanned downtime. It keeps popping up in conversations and presentations.
BC: That has to be one of the big benefits of IoT and the whole data-driven analysis world, preventing downtime — and not just downtime in terms of loss of production, but incidents in our industry that can cause loss of life. We have a couple of projects with customers where we’re using big data analytics to do abnormal event detection, so we can detect an abnormal event just in its infancy and prevent it before it cascades into something that could ultimately be life-threatening.
We tend to have volatile processes that can explode, right? It’s fun to use data science techniques, doing analysis to try to recognize abnormal patterns without actually having any detail process knowledge. There are two different ways to do monitoring: One, you actually write equations and understand the processes that are happening, and modeling that, detecting things that seem off. Very difficult to keep that model up to date and working properly. And then there’s big data analytics, where you don’t try to understand the process, you just use statistics to go through, looking for patterns. When these events happened in the past, what were the things that correlated with them? Can you monitor and detect them happening in the future? It’s like machine learning: If you see events, you can predict them happening in the future. That’s pretty exciting.