The question
Our boat builder grinds fiberglass hulls by hand all day and the dust is making people quit, so could a cobot follow the curves of a hull without gouging the gel coat?
The short answer
A cobot can take over repetitive grinding on boat hull curves, but the realistic job is first-stage sanding with force control, a compliant tool and a path proven on your own hulls, not finish work close to the gel coat. Published evidence on gel coat protection is thin, so the depth tolerance has to be written down and tested in a pilot on one hull model. Enclosed extraction at the tool can cut dust sharply, yet the silica content of each product must be checked before deciding how much protection the crew still needs. If your hulls vary widely in shape or the grinding is mostly judgment-based fairing, keep that work with people.
Key takeaways
- A cobot with force control can grind boat hull curves, and first-stage sanding is the realistic starting point.
- Gel coat protection depends on the tool, the force setting and a path tested on each hull shape.
- Check the silica content of every product before assuming fiberglass dust is the only exposure.
- Pilot one hull model and measure dust, gel coat thickness and rework before committing.
- Crew turnover can justify automation, but the case must hold up on your own rework and dust records.
Can a cobot grind a fiberglass hull without gouging the gel coat?
Yes, a cobot can grind fiberglass hull surfaces, but the honest answer depends on how the robot is controlled and what the grinding is meant to achieve. Gel coat is a thin finish layer, and a rigid tool pressed against a curved hull on a fixed path will cut through it wherever the surface bulges or dips. A robot that holds steady contact force can follow those curves without digging in, so force control matters more than reach or payload when you compare machines.
Most published and vendor material on this topic describes first-stage sanding and roughening, where the goal is to scuff a surface before paint or fairing. One sanding system maker states that orbital end effectors are unsuitable for first-stage fiberglass sanding, and it brings the arm to the part rather than fixing the robot in one place. Those are vendor claims, so treat them as a reason to test, not as proof of performance on your laminate.
Gel coat removal and fairing on finished hulls are where the evidence runs thin. Our searches found no published results on gel coat removal or fairing on hulls, and no comparison of robots with skilled hand sanding. That gap is the main reason a pilot matters more than any brochure.
How does a cobot follow the curves of a hull?
A cobot holds a sander or grinder at a set angle and contact force while it moves along a path taught from the hull model or from a surface scan. Force control, sometimes called compliance, lets the tool give slightly when it meets a bump, so the disc stays flat against the surface instead of skipping over it. A mobile cobot can also move from one hull to the next, which helps yards that build several sizes on the same floor.
Technical data from one shipbuilding sanding supplier lists a sander running at 4000 to 10000 rpm, with 77, 125 and 150 mm discs, for roughening surfaces before painting. Disc diameter changes how tightly the tool follows a radius and how fast it removes material, so choose it from test panels rather than from a catalog. Speed and disc choice also interact with the resin and filler in your laminate, which is another reason to test before committing.
Research supports the basic approach. A peer-reviewed study on force-controlled surface treatment with industrial robots reported tests in the lab and in real conditions on refabricating fiberglass wind turbine blades, which are large curved composite parts. A University of Padua thesis on automating boat hull sanding in the antifouling region states that automation could reduce worker exposure to toxic dust and physical load. Both are research work, not production records from boatyards.
Has anyone shown a robot will spare the gel coat?
Evidence for gel coat protection on boats is thin, so you will have to build it yourself during a pilot. Published research covers fiberglass blade refabrication and hull sanding in general, not gel coat depth on finished production hulls. Vendor claims exist, but they are not controlled tests you can transfer directly to your own resin system.
Protection comes from three controls: a force setting low enough that the disc does not dig in, a path that stops short of the depth you cannot afford to lose, and an inspection after each pass. Your quality team should set the depth tolerance in writing before the robot runs on a hull, because the robot will do exactly what it was taught. A tolerance that exists only in the head of the most experienced finisher will not survive a shift change.
The comparison is less about speed than about consistency and judgment. The table below sets out where each option tends to win, including where a person is still the better choice.
- Consistency pass to pass
Cobot
Same force and path on every pass once the program is validated
Staff
Varies with fatigue and with the hour of the shift
- Following tight curves and corners
Cobot
Limited to the reach and path taught, and needs teaching for each hull shape
Staff
Adjusts by feel on the spot, including on shapes nobody planned for
- Gel coat depth control
Cobot
Holds a set force, but only as well as the program and tool setup allow
Staff
Feels the surface and stops when it looks thin, though skill varies by person
- Dust exposure at the operator
Cobot
Operator can stay outside the work zone while the tool runs, with extraction at the tool
Staff
Operator breathes dust unless extraction and respiratory protection are in place
- Setup and changeover
Cobot
New programming for each hull shape, plus a validated test pass
Staff
Changes shape by eye with little setup time
- Fairing and finish judgment
Cobot
Poor fit for final fairing that depends on judgment
Staff
Wins on finish quality and feel
- Crew availability
Cobot
Can run work that people would otherwise do, but still needs staff to load, check and service it
Staff
Needs a crew on every shift, so turnover directly disrupts output
How much dust do hull grinders breathe, and what is the legal limit?
Fiberglass grinding throws off fine dust, and the rule that matters most for crew health is the silica limit. According to a law firm summary of OSHA's respirable crystalline silica standard, the permissible exposure limit is 50 micrograms per cubic meter as an 8-hour average. Chemistry World reported that the 2016 rule halved the limit, so shop practices that were acceptable under older rules may now fall short.
Fiberglass resin is not automatically silica, but fillers and other materials in a laminate or gel coat can contain it. The safety data sheet for each product lists hazardous components, so pull those sheets before deciding whether the silica rule applies to your grinding. The answer shapes how much dust control, air monitoring and respiratory protection the job needs.
Dust capture at the tool is the main lever a robot offers. An enclosed sanding head with extraction keeps dust near its source, and the operator can stay outside the work zone while the tool runs. Monitor air at the crew position during the pilot, with and without extraction, and keep the records, because the rule is built around measured exposure rather than around how a robot looks.
Does a robot change the crew, staffing and turnover picture?
Hull grinding falls to fiberglass laminators and related finishing crews. The Bureau of Labor Statistics describes fiberglass laminators as workers who lay up layers on molds to form boat decks and hulls, which puts repetitive grinding close to the core of the job. A wage aggregator's summary of BLS data for May 2025 reports projected growth of 4.2% for the occupation from 2024 to 2034, so the skill is not disappearing.
Ship and boat building as a whole is a sizable employer. The FRED series for NAICS 3366 reports 108,362 hours worked in 2025, in a series updated in June 2026. A National Marine Manufacturers Association member survey from 2019 found that workforce challenges were a top concern for a large majority of members, though that survey is older and should be read with that in mind.
We did not find published turnover or separation rates for laminators or boat builders, so the turnover claim in your case has to come from your own records. Count how many grinding positions you filled over the past two years and how many of those hires left within the first year. A robot case rests on those numbers, not on an industry average.
How fast are US manufacturers adopting robots?
US adoption is rising. The International Federation of Robotics preliminary figures, released in June 2026, put US industrial robot installations at 38,000 in 2025, up 11%, with growth driven by food and non-manufacturing sectors. That trend suggests operators outside automotive and electronics are now testing robots for tasks once thought too variable.
Later coverage of the IFR World Robotics report gives 38,400 installations in 2025, up 12%, and says the US overtook Japan as the second-largest market. The figures differ between the preliminary release and the fuller report, so quote the one you use along with its date.
These totals cover all industrial robots, not boatyards, and the research did not produce a US-specific figure for the share of cobots in 2025. Adoption among boat builders is therefore a question for peers and for your site assessment, not a number to lean on in a budget meeting.
How does getting a cobot work with Service Robot Co.?
Service Robot Co. is an OEM-neutral integrator for US businesses. We select the right robot for the hull line across manufacturers, lease it through financing, deploy and integrate it, train staff and service every unit under our service plan, backed by regional service engineers in every state we service. One partner handles the whole lifecycle rather than a stack of separate vendors.
The process starts with a free site assessment, where we measure hull sizes, map the grinding zone and check dust extraction and power. The lease payment is quoted after that assessment, so the figure reflects your actual hulls and cell rather than a generic plan. Nothing on this page commits you to a lease, and a pilot on one hull model is a sensible first step.
Ask any integrator, including us, how they will handle gel coat depth limits, dust monitoring and programming for new hull shapes. Those three points decide whether a pilot succeeds, and the answers should be in writing before a lease is signed.
What should a pilot test before a lease is signed?
A pilot on one hull model tests what a quote cannot. Run the robot on test panels at several force settings, then on a real hull section, and measure gel coat thickness before and after each pass. Record the time needed to program a new hull shape, since that time decides how much value the robot adds on short runs.
Measure dust at the crew position with and without extraction at the tool, and compare the readings with the silica limit and your safety data sheets. Track rework on finished hulls for the whole pilot period, because rework is the cost your production team feels first and the one the owner will ask about.
Vendor training claims vary. One sanding cobot vendor says a technician can be trained in two hours, and a composites shop case study reported that staff learned to run a cobot sanding kit in one day. Both are vendor-reported, so use them to set a training plan that you check against your own operators.
Frequently asked questions
Want a cobot at your boatyards?
Tell us about your site. We will recommend the robot, quote the lease, and keep it serviced.
Sources
- Law firm summary of OSHA silica standard
- Chemistry World, news on the silica exposure limit
- Bureau of Labor Statistics, fiberglass laminators
- Wage aggregator summary of BLS data, fiberglass laminators
- FRED series, ship and boat building hours worked
- International Federation of Robotics preliminary US figures, June 2026
- The Robot Report, IFR World Robotics US coverage
- Vendor technical data for a mobile sanding cobot