How new technologies impact the role of Reliability Engineers – Case study at Total Refinery

Franky Oste works as a Reliability Engineer at Total Refinery. Total buys crude oil from all over the world. This means that the Antwerp plant benefits from a whole scale of different crudes, implying that every installation needs to be adaptable. This makes Franky’s job extremely demanding and confronts him with new challenges every day.

Reliability is key

“My role at the refinery is to guarantee that our installations get the longest possible life, and to avoid problems with the machines,” Franky explains. “We don’t want to start a shift with lots of problems and technical outages, whether it’s rotating or static equipment. That’s why we try to get the most out of the asset life time and go for maximum availability of our installations.”

Maximizing the lifetime of an asset also means gaining money. You don’t need to invest in new equipment when the old one is still working fine. “Some equipment at the refinery is older than I am – and I’m 59,” Franky laughs. “We have new equipment too, but it’s always a goal to maintain the oldest installation equally well. It’s a challenge to keep everything running and to avoid incidents. That’s where I come in as a Reliability Engineer.”

Franky uses different techniques in his job, from extremely early problem detection and problem solving, to handling them in a proactive way to further avoid problems. “We use a wide array of techniques, such as vibration measuring, ultrasound, flow measurements, infrared measuring… Oil analysis is of course of the utmost importance for us. In many cases, a simple oil sample can show you when there’s a failure. It’s like the blood in your body that can be analyzed in order to detect illnesses or abnormalities. In machinery, it’s exactly the same thing.”

“Another advantage of oil analysis is that you can predict when your machine is going to fail. It’s very interesting data, because you can easily let the machine run until 5 minutes before the failure,” Franky explains.

Detect, predict and prevent

“I started working for BASF Antwerp in 1984. After a couple of years of wrench work, I immersed myself in reliability. Since 1987, my main focus has been vibration measuring. IRD was the market leader, if I recall correctly. We’ve come a long way since then, with a wide range of available techniques and a lot of programs which make life much easier and quicker. We can detect, predict, prevent… in mere seconds. This is what makes reliability so interesting for me.”

In recent years, the job of a reliability engineer evolved quite a lot. “In the beginning, it was very difficult for the older generation to come to terms with the new technologies. But now it has become easier. Reliability is now even taught as a university course. When I started, there was no such schooling. Only the guy with his screw jack listening to the machines. And now, the young guns are already learning the different techniques from a young age. It’s beautiful to witness, really.”

And now the next leap is digitalization. “It’s still a big step to take, and that makes it a really nice challenge, even for the new generation,” Franky concludes.

The future of Reliability Engineers

With 35 years of experience, how does Franky look at the future? “In the past, I always needed to go to – or even in – the machines, which wasn’t always safe. With new technologies, like contactless sensors and the amount of data they generate, the analysis is much easier, much quicker, and a lot more safe, than what I grew up with.”

He gives an example: “When I did the first measurements on a pump, it took me two minutes. And then I still needed to go inside. Now, it doesn’t even take one second. And the system already gives you a suggestion for solutions: an imbalance, or bearing damage. You just need to look and select the right solution.”

The only requirement for good decision-making is the input of reliable data.If you put garbage in, you also get garbage out. “When you put the wrong bearing in the system, then the system might detect a bearing failure that’s not even there. So you need to start with the correct data to let the system give you correct suggestions.”

Rather a sound amount of data than a huge amount

So is Reliability Engineering evolving into managing data quality and checking predictions, rather than investing time in collecting data? “But there’s also a danger lurking. When you collect a huge amount of data, chances are that you can’t see the wood for the trees anymore. That adds to the risk of bad decision-making.”

“The quality and the quantity of your data are both equally important. And don’t collect a huge amount of data that you don’t need. You should be collecting information, not just data for the sake of data.”

Practical examples to inspire other reliability engineers

At Asset Performance 4.0, Franky is going to show some practical examples in remote monitoring and digitization at the refinery. They helped maximize reliability and avoided huge secondary damages. These cases can inspire all participants at Asset Performance 4.0.

“The first example we’ll treat in our presentation is about a centrifugal compressor. It’s a brand new case that just happened. We found extremely high vibration levels and a drop in the process. We needed to decide if we were going to restart the machine or not. It’s not an easy decision, even when you have vibration levels bigger than the clearance from your bearing. When you have a lot of data that you can combine, it makes the decision-making easier. We decided to remove the catalyst, so the next start-up would go easier. When you decide not to start and to empty the catalyst, you need a lot more time. You’ll see the whole decision-making process in my presentation.”

In this case, a lot of different techniques and tools were adopted. Among others, Franky sent a drone to go look in the stack. This and other state-of-the-art techniques were applied to find out the damage, the cause of the issues and to come to a good decision. “Thanks to this data, we were able to make the right call, and in this case, it really turned out great. We started the machine and the process again after only four days. It was amazing,” Franky shares happily.