3D Models: Condition Monitoring & Inspection of Critical Infrastructure - Australian Mining Review


    Originally published in May 2022.

    The Australian Mining Review recently caught up with Daniel Lewis, the Digitalisation Manager at Applus+ in Australia, for insight into Reality Capture and its uses in Condition Monitoring.

    AMR: What is Reality Capture? 

    DL: Reality capture is the process of capturing the physical world into a digital environment. An example is when a client has many remote assets and they want to make them as photorealistic as possible. 


    AMR: What are the top reasons for using reality capture? 

    DL: Reality capture, or Digital Twin, is used for capturing and recreating a real-world asset in a digital environment, and using it for functions such as inspection, condition monitoring and assessment, engineering or consultancy. This allows the work to be completed remotely. 

    AMR: What are the cost benefits of using a 3D model for condition monitoring and assessments? 

    DL: There's the safety element of removing unnecessary personnel from site. Applus+’ capture team are generally quite small. Also of benefit is how fast it is to capture. It’s significantly quicker than conventional methods such as scaffolding, rope access or elevated work platforms. Where Applus+ have a lot of strengths is with the unmanned based capture using unmanned aerial and marine vehicles, we're able to capture the assets in a fraction of the time it takes to carry out a conventional assessment. All of this reduces overall cost, including removing the need for return visits to a site, significantly reducing travel costs. By capturing an asset we can then carry out a condition based assessment, marking up areas as required. Our clients can view those same areas without needing to go to site. 

    AMR: How have Applus+’ clients benefited from using 3D models? Can you give an example? 

    DL: An example is the reduction for multiple site visits. The client can see the results in the same environment that they've been captured in, enabling a richer understanding of what the issues are, compared to the conventional method of sharing photos and trying to explain them. Now the client can virtually see the asset and see that anomaly information themselves on the asset without needing to travel to the site. We can also use the data for reverse engineering into CAD to perform structural assessments. 

    AMR: How does the reality capture benefit condition monitoring? 

    DL: Some assets we may not be able to get access to for sustained periods of time, such as when processes are running or there are access restrictions, for example a wharf, chimney stack, pressure vessel or road bridge. By capturing the information and then reviewing it remotely we save on the costs of taking that asset offline. That time can be used for any remediation instead. 

    AMR: Is Reality Capture hard to use and interpret? 

    DL: No, they're very easy to interpret and control with a bit of practice. It’s controlled by dragging and dropping with the mouse. A benefit of using it for the condition assessments is that you can move up and down an asset, instantly. There's no need to climb around that asset or gain access to certain hard to reach locations. By moving and orbiting around the virtual asset, you can bring yourself to those locations as required. 

    AMR: Can you compare models of an asset taken at different times? 

    DL: Yes, you can do a side by side comparison to look at historic changes. There's overlay functionality currently in development. You can assess how rapidly degradation or deterioration of an asset is occurring. 

    AMR: Will the system automatically highlight any differences between the side by side comparison? 

    DL: Not currently. It's something on the road map but there are some challenges. It's about partnering with our clients, tailoring their computer vision models for  their specific assets that we are inspecting. 

    AMR: Is there an easy way to measure how long it takes to capture a particular asset and what kind of issues might arise if you're trying to capture too quickly? How will that reflect in the model? 

    DL: That is where our engineering and integrity knowledge at Applus+ comes in. Applus+ consult with clients when they are not able to provide a detailed or defined work scope. We work with them to find the most optimal way to capture an asset. The asset capture is then based on the resolution requirements for that capture. If we capture an asset at a high resolution, we need to take a larger number of photographs. If we capture it a lower resolution, a lower number of photographs, which correlates proportionally to the speed that we can capture. We can also create multi resolution models. We identify high risk locations on an asset and have our pilots really focus in on those areas with a variety of powerful inspection payloads to complement the reality models. 

    AMR: How is the data stored and accessed? Is the client in control of the access to the to the data or is that through Applus+? 

    DL: Two of the most important things in digitalisation are data security and data ownership. Applus+ utilise an ISO 27001 certified platform which goes through regular penetration testing, so the data is very secure. The client owns their data which is available in a number of formats. The cloud-based platform is specifically designed to make that data manageable and viewable inhouse or to third parties under the authorisation of the client. 

    AMR: How reliable is the data, for instance, how accurately could you see scale corrosion on a bolt head? 

    DL: It's precision digital twins, meaning it is photorealistic. Depending on how close we can get to an asset and the resolution requirements, we can see very small detail. We use a term which is Ground Sample Distance. GSD essentially means the ratio of pixels on a camera image to physical millimeters in the field. We can accurately go down to less than a millimetre if required. Resolutions are also dependant on the model size. The more data you capture the larger than model, so there's a trade-off to only capture the amount of data that you require. We can also complement the models with additional data, such as annotations or attaching additional photos. Applus+ have specific payloads for our UAS, which are designed specifically for high definition image capture. They can zoom in and see some really fine detail during inspections.

    AMR: Where do you see the future of 3D models? What's the next phase? Would it incorporate AI for example? 

    DL: I think some of the important next steps will be things like model segmentation or classification. Such as breaking an asset down into its components. For example separation of a vessel from associated access platforms or associated pipe work. Then there's the AI-type work, which we are currently working on to make the interpretation of those assets more efficient. 

    AMR: In this fast paced economy, where technology is constantly in development and we have a lot of innovation, is 3D capture money wasted down the line through advancement of tech? 

    DL: We're on the very beginning of the journey into the reality capture space, or metaverse of information. We've got a long time to go before we progress to the next technology leap. However, we do make sure that data is future compatible. The clients embracing inspection technology now will be much further ahead of their competitors who don’t have ‘big data’ through lack of adoption. When there is a technological evolution, information previously captured can be transformed into the new system or technology. This is certainly at the very foundation of what the team at Applus+ does. 

    Applus+ uses first-party and third-party cookies for analytical purposes and to show you personalized advertising based on a profile drawn up based on your browsing habits (eg. visited websites). You can accept all cookies by pressing the "Accept" button or configure or reject their use. Consult our Cookies Policy for more information.

    Cookie settings panel