Welding Automation by the Numbers
Chapter 10 of 10 · 10 min read
People, skills and the automation plan
Staff a welding automation project with operators, programmers, engineers and maintainers, train and redeploy welders, rank projects by payback and risk, stage the plan and measure what it delivers.
Every chapter of this book has ended up at people. The self-regulating arc needs someone who knows what parameters to set; fixtures need someone who understands locating; a robot cell needs programmers, operators, an engineer who owns its procedures and data, and maintainers who keep it running; guarding works only if people use it properly. Automation does not take skill out of welding. It moves skill from the welder’s hand into setting up, programming, checking and maintaining machines, and a fabricator that buys equipment without building those skills buys a cell that waits. The welders whose work changes are, for most fabricators, the best source of those skills.
This chapter sets out the roles an automation project needs and the training they take, considers what happens to welders whose work moves to machines, ranks the running case’s projects by payback and risk, stages them into a plan, and closes by bringing the book together.
The roles
| What they do | Background | Illustrative training | |
|---|---|---|---|
| Cell operator | loads, unloads, inspects, changes tips, clears faults, keeps records | experienced welder or fabricator | about 1 week, then coaching |
| Robot programmer | teaches and edits programs, sets parameters, runs first articles | skilled welder with an aptitude for machines | about 2–3 weeks plus months of practice |
| Welding engineer | owns procedures, monitoring limits, capability and program control | welding engineer or senior supervisor | robot and monitoring courses |
| Maintainer | maintains robot, positioner, power source, torch and safety system | electrician or fitter | maker's maintenance course |
| Fixture designer | designs and checks fixtures for new parts | toolmaker or design draftsperson | locating principles; offline software |
Some roles can be combined in a small fabricator: the programmer may also design fixtures, the welding engineer may also be the production supervisor. What cannot be combined away, however small the shop, is the knowledge each role brings. A cell with no one who can program it waits for the integrator for every new part; one with no one who understands welding makes programs that move well and weld poorly.
Welders as programmers
The fabricator chose two of its best welders to train as robot programmers, rather than hiring programmers who knew robots but not welding. Its reasoning was that robot handling can be taught in weeks, while the judgement of what makes a good weld, the angle that fuses both plates, the speed that avoids undercut, the sequence that controls distortion, takes years. Morton, in 1919, had observed that the automatic machines taught their inventor things that would benefit manual welders; the fabricator found the reverse equally true, that its welders’ experience made the robot’s programs better. The two programmers became the bridge between the shop floor and the cell.
Training
Training costs time and money, and is often left out of business cases. The fabricator planned about a week of operator training for four people, two to three weeks for each programmer plus months of supervised practice, a maintenance course for its electrician, and robot and monitoring courses for its welding engineer, together with cover for the welders while they trained. It also planned training in the cell’s safety procedures for everyone who would work near it. Welder Training by the Numbers covers how training is planned and how its effect fades without practice, which applied as much to programming as to welding.
What happens to the welders
The question every welder asks when a robot arrives is whether their job is going. In the running case, the cell did the work of about 2.4 welders, and positioners and carriages raised the output of the rest. The fabricator was growing steadily at the time, and its owner had said from the start that automation was to take on new work and make the existing work more competitive, not to reduce the number of welders. Two welders became programmers, one became the cell’s main operator, and the remaining output gain went into the outside work the cell had been bought to win and into growth on the frames. Over the following two years, as two older welders retired, the fabricator did not replace them on the bracket work.
Not every fabricator is growing, and automation can mean fewer welding jobs. Where it does, the obligations of employment law and the fabricator’s own values apply, and planning redeployment, retraining and the timing of changes with the people affected is part of the project. Either way, telling people early and honestly what is planned, and involving them in it, makes a cell far more likely to succeed: welders who help design the fixtures and trial the programs make sure the cell works; welders who fear it have every reason to let it fail.
Ranking the projects
The running case’s owner had three automation projects on the table, each with its own business case.
The smaller projects paid back faster and carried less risk: carriages and positioners use simple equipment, work with the welders the fabricator already had, and do not depend on winning new work. The robot cell carried more risk, in its dependence on utilisation, its learning period and its demands on part accuracy, but offered by far the largest saving in absolute terms. The owner ranked the projects by payback and by risk, and decided on the order in which to do them.
Staging the plan
- Surveymeasure products, volumes, welding hours, arc-on time and joint types; find the candidates
- Quick winspositioners for frames and carriages for rails; better fixtures for manual welding
- Prepare the partspress brake upgrade, bracket redesign with tabs and slots, part families
- Build the skillstrain programmers, operators, maintainers; trial on a borrowed robot
- Install the cellwith confirmed work, offline programming and monitoring from the start
- Measure and adjusttrack hours, arc-on time, capability, alarms and payback; refine
The order matters. Doing the positioners and carriages first gave the welders a taste of mechanisation that helped them, built confidence, and paid back quickly, funding part of what followed. Preparing the parts and the skills before the cell arrived meant the cell could start producing soon after installation, rather than spending a year waiting for the shop to catch up, as the failure in Chapter 1 described.
Measuring what it delivers
A plan is only as good as the measurement of its results. The fabricator tracked, for each project, the figures its business case had promised: for the cell, its hours of use, arc-on time, parts made, weld metal per part, alarms, capability and customer returns; for the positioners and carriages, the welding hours and over-welding on the frames and rails. Each quarter, the owner compared them with the business cases. Where results fell short, the reasons were found and fixed, and the owner reported them to the welders as openly as the successes; where results exceeded the plan, the reasons were noted for the next project, so that each business case was built on better evidence than the last.
Working with an integrator
Most fabricators buy their first cell through an integrator, a firm that selects the robot and welding equipment, designs the cell and its guarding, builds or specifies the fixtures, programs the first parts, installs and commissions the cell, and trains the fabricator’s people. A good integrator brings experience the fabricator does not have; a poor one sells a cell that works on the demonstration part and nowhere else. The fabricator’s owner asked three integrators to quote on the same specification, which listed the parts, volumes, weld sizes, tolerances and safety requirements, and asked each to weld sample parts from the fabricator’s own production on a demonstration cell. The samples were sectioned and gauged. The chosen integrator’s quote was not the cheapest, but its samples were the best, its training plan the most complete, and its references from other fabricators the strongest. The contract set out acceptance tests, on the fabricator’s parts, that the cell had to pass before final payment.
Keeping manual skills
Automation can erode the shop’s manual welding skills if the repeated work that once kept welders in practice moves to the robot, leaving them only the awkward, one-off jobs. Those jobs are exactly where skill matters most. The fabricator kept a share of repeated work on the shop floor for training, rotated welders through different products, and kept its welders’ qualifications current, so that the shop could still weld well by hand when the cell was down, a part was too large or a customer needed a repair. A shop that can only weld with its robot is a fragile shop.
Culture and resistance
People resist change for good reasons as well as bad: a new cell may threaten their job, their status as the shop’s best welder, or the way they have always worked, and early cells often do make work harder before they make it easier. The fabricator’s owner involved the welders from the survey onwards, showed them the business cases, asked for their ideas on fixtures and part design, and made the first projects ones that helped them, the positioners and carriages that took the worst of the overhead and repetitive welding off their shoulders. When the cell arrived, the welders who had helped design its fixtures were its strongest advocates, and its first programmers were people the rest of the shop already trusted.
After the first cell
The first cell teaches a fabricator more than any study. After a year, the running case’s fabricator knew its real changeover times, its real programming hours per part, the alarms that mattered, the parts that suited the robot and those that did not, and the real hours of work available. Those figures went into the business case for the next step, a second, smaller cell for the trailer crossmembers, which was approved on the strength of the first cell’s record rather than an integrator’s estimate. Each project built the evidence for the next, which is the strongest argument for starting small and measuring everything.
Safety and the people
The people plan includes safety from the start. Everyone who works in or near the cell is trained in its hazards, its safeguards and its procedures for loading, clearing faults, changing tips and isolating it, as Chapter 9 described, and the training is refreshed and recorded. Supervisors are trained to recognise the shortcuts that defeat safeguards, and to fix the reason for them rather than ignore them. A cell whose people understand why each safeguard is there is far safer than one whose people see the safeguards only as obstacles.
Bringing it together
This book began with the ladder of automation and the first lesson that each rung pays only where the work is ready. It went on to the self-regulating arc that made wire feeding and every later rung possible; to travel speed, the parameter machines hold best, and the over-welding its consistency removes; to fixtures and positioners, which decide whether the joint is where the robot expects it; to the robot cell’s payback, which depends above all on how many hours it runs; to programming and changeover, which set the smallest batch worth automating; to sensing, which widens the robot’s tolerance at a cost; to monitoring, which catches the mistakes a machine repeats; to guarding, which keeps people safe from the robot and the arc; and finally to the people and the plan.
The numbers in every chapter point the same way. Automation pays where the work repeats, the parts are consistent, the hours are there, the skills are in place and the results are measured. Where those conditions hold, it raises output, improves quality, uses less weld metal and spares welders the most repetitive and awkward work. Where they do not, it waits. The fabricator’s job is to make the conditions true before buying the machine.
A failure: the cell nobody owned
This is an illustration. A fabricator elsewhere installed a cell with the integrator’s programs for its first parts and no internal programmer, planning to have the integrator add new parts as needed. Within a year, the cell was welding only those first parts, because each new part meant a wait of weeks and an invoice from the integrator, and the production manager found it easier to keep new work on the shop floor. The cell ran about 600 hours a year.
The fabricator then trained one of its welders as a programmer and gave the welding engineer ownership of the cell’s work list and results. Within six months, the cell was running about 1,500 hours. The equipment had not changed at all; only the ownership of the cell had.
Exercise: ranking two projects
A fabricator can either buy a positioner for A$45,000 that saves about A$20,000 a year with little risk, or a small robot cell for A$280,000 that would save about A$85 for each hour it runs, with about 1,100 hours of confirmed work. Calculate each payback and the saving each would deliver over five years before discounting. Which would you do first, and why?
The positioner pays back in 45,000 ÷ 20,000 ≈ 2.3 years and saves about 5 × 20,000 = A$100,000 over five years, a net gain of about A$55,000.
The cell saves about 1,100 × 85 ≈ A$93,500 a year, paying back in 280,000 ÷ 93,500 ≈ 3.0 years, and saves about 5 × 93,500 ≈ A$467,500 over five years, a net gain of about A$187,500. The cell offers far more, but carries the risks of utilisation, learning and part accuracy. Doing the positioner first, while preparing the parts and training a programmer for the cell, captures a quick, safe gain and lowers the cell’s risk when it comes; both are worth doing, in that order.
