Everybody wants to join the connected world by using existing sensors and adding new ones to industrial equipment, but before diving in, it’s important to ask why. “What’s the real reason why we do asset monitoring—what is the challenge we’re trying to address?” asked Louise Pattison, product manager, APM Health, Meridium, of her Smart Industry 2015 session attendees. “It’s not just to gather lots of data.”
According to Pattison, there are four good reasons to monitor assets: to improve reliability, to raise productivity, to reduce risk, and to allow profitable growth by increasing production using existing assets. The reasons will vary in priority and importance at different facilities.
For example, “Manitoba Hydro sells electric power in a regulated market. Their strategy is based on availability and cost, not increasing capacity,” says Pattison. Nova Scotia Power has an aging workforce and equipment—much of it from the early 1900s. “They need to deal with knowledge gaps,” she said. Gladstone LNG in Australia is exploiting a brand-new field in the middle of the desert. They need to minimize and plan trips to equipment to reduce risk and minimize travel.
“Define your strategy, don’t just take all the data from all the equipment,” Pattison said. Use an intelligent asset strategy (IAS)—a logical framework for building an asset maintenance and monitoring program that addresses risk and delivers return on investment. The four steps in an IAS are to define operating context, assess risk of asset failure, define risk mitigation, and only then implement programs. Remember that asset monitoring is a key risk mitigation strategy, not the only one, she said. Consider it in the context of others, including planned maintenance and backup equipment.
“When monitoring is the right choice, how should you go about it?” said Pattison “The usual way is to start with the control system, sending sensor information to the historian. But also consider what we sometimes call the ‘Bubba sensor’—a mobile worker immediately sending information by mobile device.”
How will you use the data?
Data analysis methods range from simple to extremely sophisticated. “The easiest is simple trending, but this tends to have low returns,” said Pattison. Next up is using model-based analytics to correlate faults. This way, “I can tell when something’s wrong.”
The most sophisticated but still prevalent industrial approach is using smart sensors with built-in algorithms for common failure modes. “This puts model-based analytics in a smart sensor that tells you how well it’s feeling, and what’s going on in the equipment,” Pattison said. “Many signatures are known—you can buy off-the-shelf systems that will tell you what is failing and what to do.”
Beyond that are machine learning (“It’s interesting, but don’t start here,” Pattison said) and “diagnostics and prognostics” for applications that are not already known and defined. Pattison said, “This is expensive and hard to do, but with enough investment, you can predict the future.”
Monitoring can be used to meet objectives in several ways. The most common is by driving predictive maintenance—using temperature, pressure, vibration, etc. readings with limits. “You can know when a problem is developing so you can look into it,” Pattison said. “Diagnostic applications tell you when to do what.”
Another use is to do real-time risk monitoring: sensor and safety systems often are applied to mitigate risks, and monitoring can detect when degradation or failure is leading to increased risk.
A third use is to monitor operating integrity. When equipment is operated near or outside of its design limits—for example, a heat exchanger is designed to operate within assumed windows of temperature, pressure, acidity, humidity—it can shorten its life and lead to an unexpected failure, perhaps soon, perhaps further in the future due to cumulative damage. “This can affect planning years in advance,” Patttison said. “An excursion this month affects life 10 or 15 years down the road.” There’s no instant gratification, so this strategy is not always high on people’s lists, but it’s worth remembering.
In general, it’s good to monitor not just the asset, but also the things that affect the risk—to monitor the assumptions for changes that increase risk. If you are seeing unexpected failures, are the assumed actions being carried out, for example, oil changes? “Monitor work history as well as the sensors,” Pattison said. “This will also indicate whether you are receiving good return on the work—is it providing the best value for the cost?"
“A good IAS implementation can help you bring more revenue, lower maintenance costs and improve safety,” Pattison concluded, “by addressing the risks you know you have.”