TL;DR
This laser scanning processing plant WA case study documents how Industrial Spatial Solutions captured a congested ore-processing facility in the Pilbara across a 36-hour shutdown window, registering 214 scan positions into a single point cloud georeferenced to GDA2020 / MGA2020 Zone 50. The 2-4 mm as-built dataset let the client's engineers detect a 63 mm clash in the proposed conveyor tie-in before fabrication, avoiding an estimated AUD 180,000 in rework and a second shutdown.
Key takeaways
- A 214-position terrestrial scan of a live Pilbara processing plant was completed inside a fixed 36-hour maintenance shutdown using two Leica RTC360 scanners working concurrent zones.
- The registered point cloud achieved a cloud-to-cloud registration accuracy of 3 mm and was georeferenced to GDA2020 / MGA2020 Zone 50 with AHD heights via a surveyed control network, so the data tied directly into the client's existing mine grid.
- Clash detection against the proposed conveyor and screen-house steel revealed a 63 mm interference with an existing cable ladder — caught at design stage rather than during installation.
- Bringing the as-built forward avoided an estimated AUD 180,000 in field rework and removed the risk of a second unplanned shutdown at roughly AUD 250,000 per day of lost throughput.
- Deliverables (registered E57 point cloud, Recap RCP, structural steel model and 2D sections) were issued 9 business days after demobilisation, ahead of the client's detailed-design lock.
The challenge: an undocumented brownfield plant on a fixed shutdown clock
The client operates an iron ore processing plant in the Pilbara region of Western Australia, roughly 200 km inland from Port Hedland. Like many facilities that have grown by accretion over fifteen-plus years, the plant's drawings no longer matched reality. Successive tie-ins, pipe reroutes and structural reinforcements had been made under time pressure without consistent as-built capture. When the operator commissioned a new screening circuit and a relocated transfer conveyor, the engineering contractor found that the available P&IDs and structural models were unreliable for clash-critical design.
Three constraints defined the brief:
- Access was only available during a planned 36-hour maintenance shutdown. The processing circuit runs continuously, and an unplanned stop costs the operation in the order of AUD 250,000 per day in deferred throughput. There was no scope to extend the window.
- The environment was congested and multi-level. The screen house, transfer towers and conveyor galleries are tightly packed with steelwork, chutes, cable ladders, hydraulic lines and dust-suppression pipework across five working levels — exactly the conditions where discrete total-station measurement is slow and incomplete.
- The data had to be survey-grade and tie into the mine grid. For the new steel to be fabricated off site and bolted up in a later shutdown, the as-built had to sit accurately within the client's GDA2020 / MGA2020 Zone 50 coordinate system with AHD levels, not just a local arbitrary datum.
The operator had been quoted a traditional measure-up by another firm that would have needed most of the shutdown for partial coverage of a single level. That was never going to resolve the clash risk across the full circuit.
The approach: control first, then concurrent scanning
ISS treated this as a control problem before a scanning problem. Reliable registration and georeferencing across a steel-dense, multi-level structure depends entirely on a sound control network — get that wrong and the point cloud drifts.
Establishing control (pre-shutdown, 1 day). A surveyor attended site the week before shutdown during a brief access window to install and observe a control network. Eleven control stations were placed on stable structure and concrete plinths and observed with a Leica TS16 total station, tied to two existing mine survey marks via GNSS observation on GDA2020. The network was least-squares adjusted to a 95% confidence misclose under 4 mm horizontal and 5 mm vertical (AHD), giving a rigid framework for the scan registration to lock onto.
Concurrent scanning (shutdown, ~30 hours of the 36-hour window). Two survey crews ran two Leica RTC360 scanners in parallel, each working a defined zone to avoid double-handling and cross-traffic. The RTC360 captures up to two million points per second and uses its Visual Inertial System to pre-register sequential setups in the field, which materially cuts office registration time on dense brownfield jobs. Black-and-white targets and spheres were distributed against the control stations so every zone could be tied back to the network rather than relying on cloud-to-cloud links alone.
Key field decisions on this job:
- Setup density was driven by occlusion, not a fixed grid. In the screen house, setups were spaced as close as 3-4 m to see behind chutes and around the vibrating screens; in open conveyor galleries they stretched to 12-15 m.
- Dust and residual moisture were managed actively. Surfaces near the screens were knocked down by the site cleaning crew ahead of the team, because airborne fines and wet steel scatter the return signal and inflate range noise. Where wash-down had left standing water, those bays were scanned last to allow drainage.
- Hot and live-adjacent areas were respected. Although the circuit was isolated, some thermal mass remained; scan positions and cable routes were planned with the site's permit-to-work and isolation regime, and crews held current site inductions and a job-specific SWMS.
A total of 214 scan positions were captured across the five levels inside the window, with overlap maintained at 20-30% between adjacent setups for robust registration.
Equipment and standards
| Item | Detail |
|---|---|
| Primary scanners | 2 x Leica RTC360 (terrestrial, time-of-flight + VIS pre-registration) |
| Control instrument | Leica TS16 total station; GNSS for grid connection |
| Registration software | Leica Cyclone REGISTER 360 |
| Modelling / extraction | Autodesk ReCap, AutoCAD, Navisworks (clash) |
| Coordinate system | GDA2020 / MGA2020 Zone 50, AHD heights |
| Scan positions | 214 across 5 levels |
| Registration accuracy | 3 mm cloud-to-cloud; control network <5 mm |
| Point accuracy (as-built) | 2-4 mm at typical working ranges |
Capturing to GDA2020 / MGA2020 is the practical standard for resources work in WA; tying the cloud to the client's existing mine grid meant the new steel model and the scan shared one coordinate truth, which is what makes off-site fabrication safe. Where drone capture supplements terrestrial scanning on these sites, any UAV work is flown under CASA Part 101 by a remote pilot operating within the company's RePL/ReOC arrangements — though on this congested indoor circuit the work was wholly terrestrial.
The result: a clash caught before steel was cut
Cyclone REGISTER 360 registered the 214 setups into a unified point cloud at 3 mm cloud-to-cloud, georeferenced to the control network. From the cleaned, colourised cloud, ISS extracted a structural steel model of the screen house and conveyor galleries plus 2D sections through the proposed tie-in corridors.
When the engineering contractor overlaid the proposed transfer conveyor and screen-house steel against the as-built model in Navisworks, the analysis flagged a 63 mm interference between a new support member and an existing cable ladder that did not appear on the legacy drawings. A second, smaller clearance issue was identified where a new walkway encroached on a hydraulic line bracket.
Because both were found at design stage, the contractor revised the member layout on paper. No steel was re-cut, no crew stood idle in a live shutdown, and the bolt-up proceeded in the later installation window as planned. The 2-4 mm point accuracy was well inside the tolerance the structural designers needed for connection design.
The outcome: time and money the client kept
The commercial case was straightforward once measured against the alternative of discovering the clash during installation:
- Avoided rework: approximately AUD 180,000 in re-fabrication, re-mobilisation and field modification, based on the contractor's own estimate for cutting and re-welding the affected steel on site.
- Avoided second shutdown risk: at roughly AUD 250,000 per day of lost processing throughput, even a half-day overrun to fix an installation-stage clash would have eclipsed the entire cost of the scan.
- Schedule held: the full deliverable set — registered E57 point cloud, Autodesk ReCap RCP, structural steel model and 2D sections — was issued 9 business days after demobilisation, ahead of the client's detailed-design lock.
- A reusable digital asset: the operator now holds an accurate, georeferenced as-built of the processing circuit that supports future maintenance planning, shutdown scoping and asset management, not just this one project.
For a brownfield WA processing plant, the lesson is consistent with what we see across the resources sector: the cost of a survey-grade as-built is small relative to a single clash found in the field, and the only practical way to capture a congested live circuit inside a fixed shutdown is dense terrestrial scanning anchored to a sound control network.
Frequently asked questions
How long does it take to laser scan a processing plant during a shutdown?
It depends on plant size, congestion and the number of levels, but a focused circuit like the one in this case study — five levels, screen house and conveyor galleries — was captured in roughly 30 hours of field time using two RTC360 scanners working concurrent zones. Planning the control network and setup strategy beforehand is what makes a fixed shutdown window achievable.
What accuracy can I expect from a laser scanning processing plant survey in WA?
For a well-executed terrestrial scan anchored to a surveyed control network, expect 2-4 mm point accuracy at typical working ranges and registration around 3 mm cloud-to-cloud. That is comfortably inside the tolerance most structural and mechanical retrofit designs require. Accuracy degrades with dust, dark or wet surfaces and grazing angles, which is why area cleaning before scanning matters.
Why georeference the point cloud to GDA2020 / MGA2020?
GDA2020 / MGA2020 (Zone 50 for most of WA) with AHD heights is the working datum for resources projects in the state. Tying the scan to the client's existing mine grid means the as-built, the new steel model and every other survey on site share one coordinate system — which is what lets steel be fabricated off site and bolted up accurately in a later shutdown.
Can you scan while the plant is operating?
Often, yes, with the right safety controls — scanning is non-contact and can capture data from safe stand-off distances. However, congested circuits with moving equipment introduce data noise and access restrictions, so clash-critical work like this is usually best done during a planned shutdown when the circuit is isolated and accessible across all levels.
What deliverables come out of a plant scan?
Typically a registered, colourised point cloud (E57 and Autodesk ReCap RCP), extracted 2D plans and sections, a 3D structural or piping model as required, and clash-detection models prepared for Navisworks. We scope deliverables to the downstream use — a tie-in design needs a different package to a maintenance asset record.
Request a quote
If you are planning a retrofit, tie-in or shutdown on a processing plant in WA and your drawings no longer match the steel, a survey-grade laser scan removes the guesswork before fabrication starts. Industrial Spatial Solutions captures congested, multi-level facilities inside fixed shutdown windows and delivers georeferenced as-built data your engineers can design against with confidence. Call us on 0407 057 015 to scope your laser scanning processing plant project and get a fixed-price quote, or request a quote through our contact page.
