Oct-2013

Laser surveying a revamp

Efficient project workflow in a revamp can be achieved using 3D design technology that integrates with laser scan and engineering data sources

GARY FARROW
AVEVA

Article Summary

Plant revamp projects are big business, but they are also particularly challenging. The new installation must arrive on time and fit right first time, or the contractor runs the risk of contractual penalties. Achieving this demands not only efficient business processes, but also detailed and reliable knowledge of the true condition of the existing facility.

This article describes current capabilities and best practice
in the use of 3D laser 
surveying for de-risking, planning and executing revamp projects.

In an ideal world, every operating plant would have a complete set of detailed and up-to-date engineering drawings or, better, a 3D CAD model that accurately describes it. In practice, not even the most capable of owner-operators would claim to have achieved this, although a few do get close to it. Design documentation rarely reflects the true as-built state of a plant, even at handover. Several years later, the divergence can be significant, as modifications and repairs accumulate without always being properly reflected in drawing updates. When tendering for a revamp project, therefore, any wise engineering contractor (EPC) must assume that no existing design information is trustworthy. Accurate surveying is essential.

Like the CAD tools used in plant design, surveying technologies have come a long way. Tape measures and theodolites gave way to photogrammetry, which, properly used, enabled a 3D model of an area to be compiled. However, recent rapid advances have made 3D laser scanning the tool of choice for surveying an operating plant. Today, there are a number of high-quality scanner systems and service providers to draw on. Equally important, there is also very sophisticated software for exploiting the rich data that laser scanners generate.

Equipment compactness and usability have advanced as rapidly as the now ubiquitous digital camera. Scanners are quick and easy to set up and use, and can capture a hemispherical 3D representation of their surroundings to an accuracy of a few millimetres over distances of tens of metres. However, just as a camera can take either a blurry snapshot or a studio portrait, according to the skill of the user, so the apparent simplicity of laser scanning can raise expectations of its capability that the unwary user, or their client, may be disappointed with. It is a tool and, as with many tools, there are right and wrong ways to use it.

Our team at LFM Software regularly encounters users of laser scanning who have experienced problems or sub-optimal results. Almost invariably, the causes lie in not appreciating a few basic principles. A business paper  summarises these.¹ Our top tips for a successful laser-surveyed revamp project follow.

Plan the project
Just like professional photography, success lies in preparation. When planning a laser survey, especially when only a tight time window is available, it is essential to coordinate the survey with other, unrelated plant activities to ensure that the work can be performed in the time available. For example, one cannot be expected to survey an area if a contractor’s crane is parked there for an essential maintenance task. Where areas to be scanned are normally off-limits, their being made accessible may be time-limited, so one should plan the survey sequence to accommodate this. A common pitfall is to attempt a more extensive scan than time or circumstances permit and be forced to take shortcuts.

Conversely, just as an owner-
operator tries to maximise the work performed during a shutdown, so it may be worth taking the opportunity to scan more than just the area of immediate interest, to anticipate future requirements. Forward-looking owner-
operators are increasingly commissioning whole-plant surveys for this reason.

Plan the scan
While currently available software can often work wonders with sub-optimal scan data, there are limits. High-quality survey data require a well-planned scan sequence. An experienced user will plan the number and positions of individual scans so as to achieve adequate overlaps between successive “point clouds”. This will ensure that the point clouds can be stitched together accurately on a common coordinate system and that areas shadowed in one view are scanned in at least one other (see Figure 1). While there are some guiding principles to optimum scan distribution, each situation is unique and a pre-survey site inspection can be time well spent to achieve a quality result.

Laser scanning is safe and non-
invasive, but one should nevertheless plan to avoid or restrict personnel movements in the area. Structural vibration can degrade scan accuracy, while it is not uncommon to find a scan including a detailed image of half a person standing in front of some important object, which may require re-scanning.

Use the right tools
There is little difficulty in selecting scanner hardware; all the leading suppliers’ systems are excellent. However, if scans are performed at different times using different scanners, there is a risk that these datasets may not all be compatible with the design software. At AVEVA, we have created applications and interfaces that can handle all the commonly used scan data formats, so that multiple individual surveys can be integrated into a single model. Users should either confirm that their own design solution can achieve this or, if not, restrict surveying to a known compatible scanner.

Much has been achieved in reverse engineering 3D point clouds into CAD model objects. Software can recognise simple geometric solids such as cylinders and create the corresponding 3D model. However, if the data represent a pipe, which has attributes and connectivity on a P&ID, there is relatively little value in modelling it as a solid cylinder. Current state-of-the-art software can infer that the data really do represent a pipe, can offer the designer a shortlist of its possible specifications from a design catalogue, and then reverse engineer an intelligent pipe object in the design system, accurately co-located with the original scan representation (see Figure 2). This can save considerable costly effort and brings reverse engineering of large plants into the realm of the practicable and affordable.