UCD's Barry Brophy writes that it took 10 years to produce the first book in the field of underwater structural health monitoring - pulp fiction it is not
Elec

UCD’s Barry Brophy writes that it took 10 years to produce the first book in the field of underwater structural health monitoring – pulp fiction it is not.

Modern sensors are no use if you don’t know what to measure. This has been the case with underwater inspections of structures until very recently, as UCD professor Vikram Pakrashi explains.

“The state of the art was usually: the diver goes in to look at, say, marine growth – and most likely the diver is not an engineer – pokes it with a stick and comes up and says, ‘Oh there was a lot of marine growth there, here’s some pictures’. That’s all you have.

‘A completely different set of challenges’


“Underwater gives you a completely different set of challenges. It’s a different medium and your performance is dependent on two things: the amount of light and the amount of turbidity, which is floating particles.

This technology can aid the intelligent design of marine structures such as offshore wind turbines.

Below about 10 or 12 metres, there is absolute darkness, and you have to bring down this shining bright light, so if the water is really turbid, the light will scatter and you won’t see anything.

“Image processing has always been possible, but what was lacking was all of the benchmarking of how the test should be done. People didn’t know what size of damage you could detect, under what circumstances and how should you do the test?”

This ‘benchmarking’ is finally catalogued in a book that Dr Pakrashi has co-authored, published by Taylor and Francis, titled: ‘Image-Based Damage Assessment for Underwater Inspections‘. It is an interdisciplinary, inter-university work with co-authors Dr Michael O’Byrne and Prof Bidisha Ghosh, Trinity College Dublin, and Prof Franck Schoefs, University of Nantes.

The book brings together sensor technology, image processing and the vast experience of many practitioners to explain how best to carry out underwater tests on built infrastructure.

Marine growth alters geometry of structures on which it grows


Dr Pakrashi explains how this moves inspections on from just photographing underwater structures. “Let’s say you’re looking at the hydrodynamic force on the foundation of an offshore wind turbine or conventional oil rig platform.

“If you get the diameters at different depths (due to marine growth) you can estimate the extra hydrodynamic force because Morison’s equations says that when a wave hits a monopile the force is proportional to the square of the diameter as well as the surface roughness.

“So the marine growth could lead to an extra 10 to 15 per cent of stress on these structures that’s not usually unaccounted for in the design.

“We can know about growth dynamics from biologists and predict how they’re going to grow into the future, and then we can export the whole mesh into Fluent, or Ansys Aqwa, solve it, and get the extra hydrodynamic forces.

‘Structures in deeper waters where the wind is consistent’


“The sub-service foundation can easily cost about 30 per cent of the entire project, so it’s a big deal. And this will only become more significant as we get more structures in deeper waters where the wind is consistent.

“One of the interesting things about this work is that it doesn’t just enable engineers to design better structures, today, but also to monitor buildings that have been around for many years.

“This new kind of visual pattern-recognition technology is a really handy tool as we can use the most ubiquitous form or information available: pictures.

“You have pictures and underwater videos from 50 to 60 years back, so you can use these better and design better protocols for the future: when to take videos and when not to take them, under what conditions, how to reduce a diver’s diving time, what to do with the images, and so on.”

Video now a viable vibration sensor for larger structures


Video can also be used for over-water monitoring applications, as Vikram explained to me. “People always thought, ‘Come on, how much can a camera do?’ But for sub-10Hz vibrations it works.

“Because a video is 25 frames per second, which, when divided in half to allow for Nyquist, gives 13Hz, and throwing a little bit more away because of other factors is still practical for sub-10Hz, and definitely for sub-5Hz sampling. And a lot of structures vibrate at sub-5Hz frequencies because they’re so big, so camera systems can become a standard way to get things done.

“If you use a sensor, you have to have to go onto the structure, attach the sensor and then have a system to record your data. With a camera, you can focus anywhere and sometimes calibrate the image without the requirement of a sensor because you have the two depths of field, and we have a calibrated chequerboard and based on that, close and far away, we can figure the whole thing out.

“The only thing we need to have is a fair amount of vibration – on, say, a bridge that you are monitoring – and then separate out the walking frequency from the natural frequency. But it’s a much easier set-up with a camera.”

The book is the first in its field and I was interested to ask Vikram how it came about.

Video is now a viable tool for vibration measurement of large structures, such as on Daly’s suspension bridge in Cork.

“It was 10 years of sustained effort by the same group: ourselves, Trinity College Dublin, the University of Nantes and Capacites, a French company.

Image processing in structural health monitoring


“Ten years back, you wouldn’t see a lot around image processing in structural health monitoring, but I had done a couple of papers, and then in 2010, we ran a final year project, which was followed by an Irish Research Council funded PhD and then further SFT and MaREI Centre funded research until, only last year, we won our first Horizon 2020 EU funding.

“Interestingly one of the work packages in this project is around structural health monitoring of fish cages. Fish cage breaches lead to loss of salmon and there are associated health issues – due to worms – that increase the price and reduce the quality of the farmed salmon that you eat. Initially we never thought these kinds of applications would be there.”

So how hard was it to pull all of this expertise and research activity together into a single discipline-defining book?

“Extremely hard,” says Vikram. “Michael O’Byrne – he was the postdoc on this – spent, I’d say, more than 50 per cent of his time over two years writing the book. And then each author took responsibility for the two chapters in which they had expertise.

“This produced lots of queries and we’d read each other’s stuff maybe three times, and then do one last read before the publisher got it. You’re talking about four or five months at the end going back and forth. It was actually much harder to write the generic material than the original content we had researched.”

It was worth it in the end, though, as the book is both concise and insightful. It also includes snippets of MATLAB code so the reader can try some of the image processing routines out for themselves. It seems then that even in a digital, information age, there’s still nothing like a good book.

For the link to the book on the publisher’s website, please click here.

Author: Barry Brophy, research engineer, UCD School of Mechanical and Materials Engineering

http://www.engineersjournal.ie/wp-content/uploads/2019/01/a-aaaaamarine-1024x768.jpghttp://www.engineersjournal.ie/wp-content/uploads/2019/01/a-aaaaamarine-300x300.jpgDavid O'RiordanElecMaREI,marine,TCD,UCD
UCD's Barry Brophy writes that it took 10 years to produce the first book in the field of underwater structural health monitoring - pulp fiction it is not. Modern sensors are no use if you don’t know what to measure. This has been the case with underwater inspections of structures...