#89 Predictive maintenance and data analytics

Buildings of Tomorrow ยท

Show notes

In this new episode with Tyson Soutter, the discussion is around how data analytics enables predictive maintenances.

Transcript

Hello and welcome to Buildings of Tomorrow. My name is John Lester, and in this episode we'll be talking about how data analytics enables predictive maintenance. I'm joined today by Tyson Suter, a global business development manager for digitalization at Siemens Smart-Infrastructure. Tyson, thanks for joining us. Thanks John. Give us an insight on how data analytics enables predictive maintenance. So we, we talked many times about data analytics and, and the benefits of it, but one of the biggest disruptors in our market is saying, We do maintenance through a task-based system, so there's a couple of options, it's either, uh, you know, reactive maintenance, meaning when something breaks, let's go fix it, or uh preventative, meaning task-based maintenance. Let's work through a list of items and, and check to make sure everything's working. So this is uh almost a schedule. Exactly, and this has been done for many years, it it's, it's, it's what's expected, um, and there's really important things that need to happen on as part of that maintenance. So, We're not talking about replacing everything or or getting rid it all, we're saying let's let's see what's really important and let's see what we can find and measure and and track with data analytics. So predictive maintenance is let's collect all the data available uh available and say OK let's check if the sensors are calibrated. Let's see if the actuators are driving open and closed. Let's make sure that the smoke detector systems are working. Let's see if doors are opening and closing as expected. This is all checks that we can do with data analytics, there's some really basic tasks that no longer required. And once you start doing that, you can say, OK, let's let's not check the things that are working. Let's create a list of the stuff that we know that's not working or we think could be near point of failure, going down in in performance, so you can really create a list of tasks that are driven through data where the individuals on site are saying, OK, I don't have to go and see if something's working. I know something's not right here, the data's telling me so. Let's go see what's happening. So changing this, this maintenance approach from reactive, like you mentioned, fixing something when it breaks, from having a schedule and changing things, which could mean we're changing things that have absolutely no issue and that we're spending money where we don't need to, and becoming proactive and understanding what could happen in the near future or what is not behaving how it should, and then directing our efforts there. Exactly, and and as more we get more data points every single day and, It it becomes more complex to maintain these systems, so it's very important that we have digital services that can really solve these problems. And when someone goes to a site, we wanna give them as much information as possible so they can fix the issue straight away, bring the right equipment, or even send the right resource to make sure that we're not wasting time or people's money. And this is an important point that you touched on very quickly is we know that uh devices are becoming more capable, we know that systems are becoming more complex, we know that more systems are talking and working on the same networks and all these sorts of things, so our buildings are getting more complex. If we took a traditional approach to maintenance, then this very quickly becomes untenable as far as the resources required to deliver this, so we need predictive maintenance, not just to improve what we do today, but to actually absorb uh absorb the stresses and the demands that we'll have in the near future. Exactly, so just because we have more sensors doesn't mean that they're going to be working, you know, even longer, they're still gonna fail, so we really need to make sure that we can, Scale our solution to enable these new technologies, you also mentioned earlier something about a workflow to make sure that when you go to fix something, you have the right parts already, so is this an aspect of, when you have to go and do maintenance, you're actually better prepared? Yeah, so you, when you, when you attend site, you know exactly what type of equipment you're gonna be working on. If you have the data, you can say, OK, I need, This looks like it's failed, clearly, let's bring a replacement. Like let's just do the work while I'm there. It really reduces the amount of visits required but also keeps everyone engaged. So that you can instantly deliver value, you're solving problems straight away and then you're showing the value to the customer, so they can really see the work you're doing. And we spoke in the past about the difference between data analytics and fault detection, you know, when you, when you receive a fault from a building that says something has failed, there could be multiple subsystems that are involved with that fault. So you don't know until you get there, so by taking this data analytics approach, you actually know. What's happened leading up to the fault so you can better determine what may have caused the issue. Exactly, and then you can start really pre-planning replacements and and and and keeping really critical parts on site so you reduce that downtime, which is, which is critical, cos you don't want to be in a position where, um, a missed issue or a a broken piece of equipment leads to, Uh, uh, a company shutting down or a, a plant shutting down, which, which is quite common, yeah, a loss of productivity or, or a, um, an interruption to service or whatever it might be. Wonderful, thank you, thanks, John. And thank you all for joining us here on Buildings of Tomorrow, please comment, share and like this episode and also ensure to subscribe to us here on this channel. We'll see you again soon.

Full transcript (WebVTT)