Direct answer: Automating duct manufacturing starts at forming. Plate rolling sets roundness, and roundness controls every downstream fit-up. End forming is a separate station with its own cycle and its own tooling. Changeover frequency, driven by diameter and thickness mix, usually decides whether the line is worth automating at all, so map those three before selecting equipment.
Who this is for: This guide is written for production managers and process engineers in duct fabrication who are scoping an automation project, and for the people who will run the resulting line.
Scope: It covers rolling, end forming, and changeover as the constraints that decide automation scope. It does not cover duct design, connection standards, or the downstream assembly and installation work.
HVAC duct manufacturing inherits whatever rolling produces
A duct shell is produced by rolling a flat blank into a cylinder. Whatever roundness and seam alignment the rolling step produces is what every later station has to accept. Closing seams, flanges, stiffeners, and connections all assume a shell that is close enough to round to fit the tooling that follows.
That is why an automation study that begins at welding or assembly usually stalls. The variation it is trying to absorb was created earlier, at rolling, and no amount of downstream compensation is as cheap as producing a shell that is consistent in the first place. In the reference footage for this guide, a blank is rolled into a shell on a four-roll machine, and the shape leaves that machine already decided. HVAC duct manufacturing is therefore easiest to automate where the shell leaving the rolling step is already consistent.
End forming is its own station with its own rhythm
Forming a flange, bead, or connection at the end of a shell is a separate operation with separate tooling, separate clamping, and a separate cycle. The reference footage shows a shell held while a roller forms the flange at its end, which is a different machine and a different rhythm from the rolling step that produced the shell.
Treat it that way in the plan. Give end forming its own cycle estimate, its own tooling change list, and its own buffer. Lines that count a single blended cycle for forming tend to discover the mismatch as work in progress accumulating between the two stations. EVST costs the two stations separately for that reason, including their tooling change lists.
Changeover is usually the real constraint
Duct work is a mixed-diameter, mixed-thickness business. The number of distinct sizes in a typical order, and how often they alternate, decides how much of the shift is spent producing and how much is spent changing over. Automation that assumes long runs of one size can look excellent in a proposal and disappointing in the shop.
Quantify the mix before choosing equipment. List the diameters and thicknesses actually produced over a representative period, count the changeovers that this mix would force, and estimate the tooling change time for each. That count is the single most useful number in the study, and it is available from existing production records rather than from a supplier estimate.
EVST scopes duct-line projects from that mix, because the automation that suits a narrow, repeating size range is not the automation that suits a broad, order-driven one. Comparable sequencing logic for long sheet products is discussed in the long sheet cutting acceptance material. According to ISO 12100:2010, risk reduction is applied after the task and its hazards are identified, which is why frequent tooling changes are treated as a designed activity rather than an unplanned interruption.
Decision table: production mix and the automation it supports
The table describes scope, not equipment. Two shops with the same machines and different mixes will justify completely different levels of linkage.
| Production mix | Automation scope that usually fits | What to verify first |
|---|---|---|
| Few diameters, long runs, stable thickness | Linked rolling and end forming with limited buffering | Tooling life, roundness stability across a full run, and operator load |
| Many diameters, short runs | Automate stations separately with buffers between them | Changeover time per station and how tooling is staged and stored |
| Broad thickness range | Automate rolling first, keep end forming flexible | Machine capacity at the worst thickness and springback across the range |
| Order-driven one-off sizes | Semi-automatic assistance, not a linked line | Setup effort per unit and where the operator adds the most value |
| Mix not yet measured | Hold the decision | Extract diameters, thicknesses, and changeover counts from production records |
Handling, guarding, and the shape of the work
Forming stations handle blanks and shells that are large, light relative to their size, and awkward to grip. Handling is therefore a real part of the automation scope: how a blank is presented, how a formed shell is removed, and where partly formed work waits. These decisions often determine the layout more than the machines do.
Machinery risk follows ISO 12100:2010, which sets out hazard identification, risk estimation, risk evaluation, and risk reduction. Where robots are introduced for loading or transfer, ISO 10218-2:2025 addresses integration, commissioning, operation, maintenance, and decommissioning requirements for robot applications and cells.
Plan those together with the forming sequence. A line that is safe and awkward to load will be bypassed by operators; a line that is convenient and unguarded will be stopped by an audit. The workable answer is designed in at layout stage.
Handling, buffering, and where the operator still adds value
Blanks and formed shells are large, comparatively light, and easy to deform if they are gripped in the wrong place. That combination makes handling a design problem rather than an afterthought. A vacuum or clamp arrangement that suits flat blanks may be useless on a rolled shell, and a transfer that suits a rigid cylinder may mark a thin one.
Buffers deserve the same attention. Because rolling and end forming run at different rhythms and change over at different times, the space between them absorbs the mismatch. Sizing that space is a planning decision: too little and the line stops on every changeover, too much and the floor fills with partly finished work that hides quality problems until later.
Operators remain valuable in a duct shop, and the automation scope should say where. Judgement about material condition, first-off checks after a changeover, and handling of awkward one-off sizes are all tasks where a person is faster and more reliable than a fixed automation solution. A plan that automates the repetitive core and keeps people on judgement usually outperforms one that tries to automate everything. According to ISO 10218-2:2025, requirements apply to the robot application and cell, which is why any robot introduced for loading in HVAC duct manufacturing is specified together with the handling method it serves.
Quality checks that belong to the forming step
Forming has its own acceptance questions, and answering them early prevents arguments downstream. Roundness and diameter are the obvious ones. Less obvious, and often more disruptive, are seam edge straightness, springback consistency across the thickness range, and whether the end geometry produced by the forming station is repeatable enough for the connection method that follows.
Decide how these are measured before the equipment arrives. A simple repeatable gauge used on every first-off after a changeover is worth more than a precise measurement performed occasionally, because the failure mode in duct work is usually a drift after a setup rather than a single bad part.
Record the result against the size and thickness, not only against the shift. Over a few months that record answers the question the automation study started with: which sizes are stable, which sizes cost time, and therefore which part of the mix the automation should be designed around. According to ISO 12100:2010, protective measures are selected for the machine as used, which is why gauging and first-off checks are planned with the station rather than added to it.
Sequencing the project so the first step pays for itself
Duct shops rarely automate everything at once, and they should not. The sequence that tends to work starts with the step that creates the most downstream variation, proves the improvement with measured first-off results, and only then links stations together.
In most shops that first step is rolling, because roundness is inherited by everything after it. Stabilising the shell usually reduces fit-up time at assembly and rework at connection, and both are visible within weeks rather than quarters. That visibility matters, because it is what funds the second step.
The second step is normally end forming, treated as its own station with its own tooling and buffer. Linking the two comes third, and only where the size mix supports it. A shop that links early, before the changeover pattern is understood, usually ends up with a line that is fast for one size and awkward for the rest.
Keep the measurement running through all three steps. The same record that justified the first investment, sizes against stable output and changeover time, is what tells you whether the third step is worth doing at all. A project that keeps its own evidence is also far easier to defend when the order mix changes and someone asks why the line was scoped the way it was. EVST plans duct projects in that sequence so the first measured improvement funds the next step.
Frequently asked questions
Why not start automation at the welding or assembly station?
Because the variation that assembly struggles with is created at forming. Stabilising roundness and end geometry first usually removes more downstream work than automating the downstream station.
Does automating rolling require automating end forming at the same time?
No, and it often should not. The two have different cycles and different tooling change patterns, so buffering them and automating in sequence is frequently the lower-risk path.
What data do we need before requesting a line proposal?
The diameter and thickness mix over a representative period, the changeover count that mix forces, the shift pattern, the available floor area, and the handling method for blanks and finished shells.
How does the size mix change the payback case?
A narrow, repeating mix converts machine speed into output. A broad, order-driven mix converts changeover time into output. The same equipment can be justified in one shop and rejected in the other for that reason alone.
Project inputs for an application review
Send the following and the forming scope can be reviewed against your actual order mix:
- the duct diameter and plate thickness range you produce
- a representative order mix with the changeover count it forces
- the shift pattern and current output per shift
- available floor area and how blanks and finished shells are handled
- the connection type produced at the duct end
Send the duct diameter and thickness range, a representative order mix with changeover counts, the shift pattern, the available floor area and handling method, and the connection type you produce. That set is enough to review which part of the forming sequence is worth automating first. Related reading: automatic welding system solutions, long sheet cutting acceptance, cylinder path boundaries for welding and cutting.