APM and AI myth busting and the critical role of new technologies
From food and beverage processing and pharmaceuticals to oil and gas production, most companies within the process industries are actively adopting Asset Performance Management (APM) technologies. As different sectors embrace an APM analytics methodology, each asset owner is taking a unique journey – but everyone is moving in a cohesive direction. This trend is not unique to any one industry, sector or part of the world.
There are many challenges right now, and they seem fairly consistent regardless of region. The biggest one? Data. That’s because APM technology, whatever your location, is very much based on data.
As more businesses gain a better understanding of how to use APM technology to their advantage, they’re also starting to better understand what type of data they need. And that it’s not necessarily cost-prohibitive. In fact, it’s more cost-effective than it’s ever been in our history.
Some companies have a great deal of data, and in some cases they have too much and don’t know what to do with it. Currently, this challenge of understanding data is a common thread across all industries and regions as we all aspire to understand how to use it to our advantage. As we get better at understanding the data this technology needs to thrive, we’ll be more effective at moving faster and faster.
Dispelling Myths About APM And AI
There are many myths circulating about digital transformation. Here are three which we at Stork have encountered, and how we’re dispelling them:
Myth #1: “APM technology and AI are not real, and do not apply to the operation of industrial plants.”
Obviously, this statement has no truth whatsoever. On the contrary, APM solutions and artificial intelligence are very real and very applicable to industrial operations. One reason we have joined forces with AspenTech is to help demystify these technologies by working on specific customer-focused digital initiatives. APM is helping numerous asset owners, and we’re supporting them to help them better understand how it functions, and why real-time and historical data are valuable. Sharing examples and case studies goes a long way toward dispelling this myth.
Myth #2: “APM worked in that industry over there, but it won’t work in mine.”
Whether you run a petrochemical plant or food processing operation, a lot of the equipment characteristics and causes of equipment failure over time are the same. They’re independent of the application or industry. There are some strong synergies between industries that are sometimes not well accepted, but in reality, they exist.
Myth #3: “Like many technologies, AI is just a flash in the pan. Wait long enough and it will disappear.”
The third myth is that you can just ride AI out, that it’s a new technology that will simply disappear in time. That line of thinking is entirely wrong. AI is here to stay and a game-changer in the world of maintenance (and will continue to be), not only in terms of its ability to help us better understand assets, but also to use information to make better decisions going forward – six months, twelve months out, even years in advance.
As AI becomes more commonplace, used by more and more people, we will quickly move past any discussion or attempt to not take it seriously. AI technology is an enabler which will be a strong value to those around the world who maintain assets. To take advantage of all the capabilities AI are and will open up for us, it will be a question of how to choose from the multitude of applications that are the right candidates for AI integration.
The reality is we’ve seen cases where we can bust that myth. We can reduce maintenance cost and not take down assets that didn’t really need to be taken down, and therefore get more runtime out of existing assets before a maintenance intervention is needed.
Another area where we’ve seen some really positive results is the ability to predict not days, not hours, but weeks in advance when critical assets are indicating failure. That gives precious time to be able to plan for that failure in advance, so it minimizes impact to the client, the environment, their cost structure, to their ability to meet customer demands. All those are essential business issues. These warnings help customers manage their business more effectively. We’re seeing demonstrations that two and three weeks longer of lead time in being able to predict failure of a critical asset are delivering strong benefits to our clients.
Educating clients to understand AI’s potential lasting impact
Some of our clients have a good understanding of the technology, and they’re eager to utilize it. Others are still trying to understand the previously mentioned point—is digital transformation real, and if it is, how do they implement it? What do they need to do to take advantage of the technology? The same applies to AI. There’s still a bit of education to be done with our clients throughout the process industries, but we’re doing it.
As a maintenance provider around the world, Stork is excited to be in the infancy stage of embracing technologies like predictive maintenance and use it to our clients’ advantage. Predictive analytics will become more understood by a broader range of people around the world. It’s now almost an expectation that plants need to move into that space. They may not be ready depending on their level of maintenance maturity, which is another education issue – are they ready to do something with the information when a failure is predicted?
Modelling capabilities and scenarios to play significant role, too
What will have the biggest impact moving forward? The modelling capabilities of products such as Aspen Fidelis Reliability™, specifically the ability to model what-if scenarios. It’s one thing to know that a failure is imminent. The next question is, “What am I going to do about it? I have several options—which one is best for my business?”
Analytics can model those scenarios using digital twins of plants, and frame them in terms that asset owners need: What’s the impact on safety, on the environment, on customer demand, on cost and operation of my facility? Those are critical decision points, and I’m excited about what these analytics bring to the decision makers we serve. We’ll look back 10 years from now and say, “Wow, prescriptive analytics really changed the game in terms of how we operate assets around the world.”