Why digital photography and predictive modeling will remake the APM landscape

In this Big Picture Interview, Anne-Marie Walters explores picturing asset health.

Anne-Marie Walters, a chemical engineer, is global marketing director for Bentley Systems. In that role, she’s responsible for the global industrial process, offshore, and natural resources industries. Bentley is a leading player in the industrial 3D modeling world, and Walters spoke recently with Plant Services about the state of 3D predictive modeling, why it should be of interest to maintenance and operations teams – not just engineering teams – and how digital photography can provide at least as good of a view as sensors when it comes to asset performance management.

PS: Tell me a little more about the value you see in 3D modeling. How can predictive modeling enhance organizations’ asset performance management (APM) efforts? 

AW: Bentley acquired some technology which we now call ContextCapture which basically turns digital photographs into 3D models. ContextCapture software first of all generates a 3D mesh model so that you can read that into a CAD environment and start engineering on the side. After you’ve created the 3D mesh model and then you overlay back onto the model the actual photographs of the plant itself, you then can present to the engineers, the maintenance operatives, the operators, a 3D model of what they’re actually looking at as opposed to a simulated 3D or CAD model that doesn’t look real and they don’t necessarily recognize what they’re looking at because they’re not seeing familiar colors or shapes or sizes or wear and tear. The value of the 3D reality model, particularly with ContextCapture and with the high-resolution photographs that you’d put on the model, is that they can see every ding, every dent, piece of rust, and they can really recognize what they’re seeing and then they can also relate that information back to what they have.

One of the things we do show is if you combine the digital photography with something like optical character recognition, you can zoom in to say a nameplate on a pump or compressor or substation or something that’s got a serial number and you can get the software to pick up the text and automatically relate what you have in your 3D model back to the serial model number and the rest of the asset information you have. This is now. This is right now. ... You can capture all that information, come back to the office and play around with what you’ve got so you can make the best decisions and work it all out before you start to send people out there.

PS: How well is this new technology resonating? Are maintenance and operations teams making the connection to how this new software and these more hands-off approaches can aid in asset management?

AW: Today most asset information that operations and maintenance people are working with are old or out-of-date drawings. Being able to capture these reality models and bring that kind of information to their environment, they can start to make better connections between what engineering and design information they have with operations information and maintenance information and understand what options they have and make better decisions.

Visual software in the same way as sensors is flagging to the maintenance people visual information – there’s something changing about this view. Something’s beginning to shift a bit, move, shake a bit maybe. You can start to use digital photography almost instead of putting sensors on everything. You can imagine having a camera looking at a whole area of the plant and the software constantly comparing differences between photographs. It can direct your maintenance to those areas where they should be focusing their attention.

PS: What excites you about where all of this is going, about the future of predictive modeling?

AW: You’ve got literally unlimited computing capability now in the cloud. You can analyze all of this data, and if you then combine that with modeling technologies and you test against reality to make sure you’re models are OK, you can model into the future and predict asset life, how much longer something has got under the circumstances you have now, how much more life you’ve got in your assets. You can put all of this information together. Predictive modeling absolutely is something that is booming. It’s what people want. They want to know how much longer they’ve got before something is going to fail. That’s the whole thing about RCM, risk-based inspections – they want to direct their limited resources on the things that matter.

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.

Comments

No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments