Wall-Climbing Robot Rust Removal: The Strip-to-Grade Workflow

Rust removal is a workflow, not an event: assess, strip, verify the grade, recover the waste, hand over to coating. Magnetic climbing robots run that workflow on vertical steel where the alternative is scaffolding. This post follows the process with the machines in the climbing robot line.

Step 1: Assess the surface

Rust grade and coating condition set the process. Light-to-moderate rust goes to ultra-high-pressure water jetting; heavy scale and failed coating may need abrasive blasting (water sand or dry sand). The climbing wall robot carries both: a 3-DOF arm quick-changes between water rotating nozzles and sand spray guns, which is why it is the line's all-rounder. Record the target preparation grade in the work order — Wa 2.5 for water jetting, Sa2.5 for blasting are the published reference grades.

Step 2: Strip at a planned rate

Rates on record: 15–40 sqm/h on the climbing wall robot; 15–70 sqm/h on the tank rust removal robot; 50–70 sqm/h on the high-altitude and wind turbine models in water jet mode. Jet pressure is published per model: 2,800 bar (280 MPa) on the climbing wall, tank and wind turbine robots; the other models publish jetting rates without a pressure figure, so confirm the pressure specification for the model you plan around. Plan shutdowns on the conservative end of the band for your rust condition; the band exists because a polished tank wall and a scaled one are different jobs wearing the same name. Travel speeds (4 m/min working, 8–12 m/min transit) decide repositioning time between passes — on large walls it is a real share of the shift.

Step 3: Verify the grade

The deliverable is a grade, not an area. Sa2.5 (near-white metal) is the standard handover for recoating tank steel; verify by visual comparators and spot adhesion checks before the coating crew mobilises. The robot's consistency is the argument: a crawler holds standoff distance and traverse speed constant in a way a lance operator on a rope cannot at hour six.

Step 4: Recover the waste

Water jetting moves the waste stream; it does not eliminate it. Published recovery rates are 95% or more on the tank and hull models, and the high-altitude robot publishes a 98.7% crude oil recovery rate — the difference between a contained job and an environmental incident in a tank farm or dry dock. Confirm bunding and disposal routing before the first pass, not after.

Step 5: Hand over to coating

A prepared surface has a clock on it — flash rust waits for no schedule. The maintenance window closes properly when the coating robot (800 sqm/h dual-gun spraying) follows the stripping crawler on the same mobilisation. Strip and recoat planned as one job is the whole economic argument for the robot fleet; planned as two jobs, it is merely a faster way to prepare steel that then rusts again.

Safety note

The robots' stated purpose is replacing manual labour in high-risk scenarios — no scaffolders, no rope access at 30 m. Explosion-proof variants (Ex d IIC T4 and Ex d IIB T4 in this line) extend that argument into petrochemical zones. The risk assessment still belongs to the site: the robot changes who is exposed, it does not abolish the plan.

Jetting or blasting: choosing the process

Ultra-high-pressure water jetting — 2,800 bar (280 MPa) on the climbing wall, tank and wind turbine models that publish a pressure figure — strips without abrasive, contains no dust plume, and pairs with the recovery systems the tank and hull models publish (95%+ wastewater recovery). It is the default inside live petrochemical sites, where abrasive dust and sparks are unwelcome. Abrasive blasting (water sand or dry sand) cuts faster through heavy scale and failed thick coatings, and reaches Sa2.5 with a profile some coating systems demand. The climbing wall robot's quick-change arm exists for sites that need both in one mobilisation — jet the general field, blast the stubborn patches, and do not remobilise between them.

Crew, access and logistics

A robot crew is small — operator, support technician, and the site's permit holder — but the logistics are real: the high-pressure water supply and its hoses, power, waste recovery routing, and exclusion zones under the work face. Access decides the model: the 45 kg industrial crawler enters through DN600 manholes for tank interiors, while exterior work takes the larger frames. Plan hose and cable management up the structure as carefully as the robot's path down it; most stoppages on wall work are umbilical problems, not robot problems.

Records that protect the next cycle

Log what was stripped, to what grade, at what rate, per structure. The record prices the next shutdown honestly (rates achieved on your steel, not catalogue steel), supports coating warranty claims (preparation grade is a warranty condition), and builds the corrosion history that decides when a wall moves from maintenance to renewal. The robot makes the work faster; the records make the asset cheaper to own.

Measuring coverage and QA sampling

Robot work should be auditable in a way manual work is not, and owners should specify that audit trail before the work starts. The crawler's path is systematic, but the published specifications for these models do not state whether passes are logged, or which fields a log would carry. Ask the supplier to confirm — in the work order — whether the robot supports work logging, which fields it records (position, speed, process) and how the records are exported. Build QA sampling into the work order as well: grade verification points distributed across the wall (including the awkward zones: welds, curvature transitions, shadowed areas behind fittings), adhesion spot checks where the coating system requires them, and photographic records tied to coverage. When the coating crew later asks 'was this area really prepared?', the answer should be retrievable, not remembered — and that retrievability has to be contracted, not assumed, because coating warranties run for a decade.

Planning the fleet across a site

On a multi-tank site, sequence the fleet so no machine waits on another's output: stripping crawlers lead by a structure, the coating robot follows a structure behind, and the compact interior crawler works tank internals in parallel where entry permits. Do not size that sequence from the published rates alone: they are per-model figures published under stated conditions — the coating robot's 800 sqm/h is a dual-gun figure — and a stripping rate and a spraying rate are not interchangeable units. Build the fleet configuration from rates measured on site, the number of coating passes the system specifies and the waiting time between preparation and coating, and establish which machine actually sets the pace on your structure before the schedule is built around it.

Cost logic, stated plainly

The comparison that matters is not robot rate versus manual rate — though the published multiples (5–8x on the turbine model, 7–8x on the compact crawler's task class) are large. It is total job cost: scaffolding or rope-access provision and its inspection regime, the labour hours at height, the shutdown days, and the incident risk priced honestly. On tall or hazardous structures the access provision alone often exceeds the robot crew's cost, which is why these machines were adopted first exactly there. On low, easy walls the arithmetic can favour people — a credible supplier will tell you so, and the line's smaller crawlers exist for the middle ground.