Optimizing laser cutting throughput on the shop floor

An operator reviews a cutting program at one of the company’s nine flatbed laser cutting machines at its South Bend, Ind., plant. Images: General Stamping & Metalworks

Most metal fabricators know all too well where their constraint processes are. Depending on the product mix at a given time, the constraint might be in bending, welding, or perhaps powder coating (oven cure time is what it is). Considering how far laser cutting has come in recent years, many might assume they have a healthy capacity buffer to ensure those downstream constraints are never starved of work.

General Stamping & Metalworks (GSM), a contract metal fabricator in South Bend, Ind., discovered that it’s not that simple. After rearranging what kind of work goes to which laser machine, managers found that they increased throughput dramatically. Manufacturing Engineer Harvey Ndalama described the result: “It’s almost as if we added another machine.”

GSM didn’t stop there, though. It also streamlined front-end activities, from the initial quote through order processing and order releases to the floor. This “quote-to-cut” cycle represents the headwaters of part flow. The more efficient and effective that cycle becomes, the more capacity the fabricator can sell—feeding the top line—and the more work it can send downstream using the same resources—feeding the bottom line. That’s not a bad combination.

GSM now has an estimating system that integrates bidirectionally with its ERP platform. The software at the core of the system is Paperless Parts. The effort has given customers the ability to initiate orders by simply clicking the link on a quote. The system also looks at a CAD model and recommends a manufacturing process.

The software gives process recommendations, which in turn give engineers a good starting point for exploring manufacturability options. GSM estimators have done manufacturability analysis for years, but new software is accelerating the process, from initial quote through order acceptance and beyond.

“In the past, administrative tasks took up most of our estimators’ time. The software has allowed them to spend more of their day leveraging their knowledge and experience to provide the best solutions for our customers.”

That was John Axelberg, GSM’s CEO, who explained that during the past year, company leaders have scrutinized front-end activities to determine which ones can happen concurrently. The chain of activities from a request for quote to the production of a first part needn’t happen sequentially.

“We’ve mirrored our ERP system in our estimating system,” Axelberg said. “That means all process routings are created in the estimate. And once we get an order, it is pushed directly into the ERP, and everything happens at once.”

Upon order acceptance, bend simulation and programming ensue (since, of course, the bend tooling determines the blank size). Once the job is released, the primary fabrication operation starts—usually at laser cutting.

Nesting occurs at the last possible moment, so what’s being produced reflects current demand. In the past, where each job landed depended mainly on machine availability. Over the past 18 months, though, that strategy has changed dramatically.

An operator loads sheet onto a flatbed laser at General Stamping & Metalworks. Compact storage racks, located near the point of use, are in the background.

For the lion’s share of GSM’s work, planners considered the soonest-available laser to be the quickest path toward completion. The laser is a “soft,” flexible tool, after all, and the fact that many machines can process a wide range of work is its primary selling point.

The South Bend plant has nine flatbed lasers, each with 60- by 120-in. beds. They don’t have storage and material handling towers. The plant has clearance limitations, so adding towers would require significant effort. Besides, not having towers actually benefited GSM’s flexible scheduling approach. Raw stock is stored in compact shelving along the plant wall, allowing fork truck drivers to retrieve and deliver an extraordinary variety of grades and thicknesses to any machine. (Incidentally, in its Tomah, Wis., plant, GSM does have a tower system connected to its 24-kW fiber laser dedicated to production work.)

Until recently, the thinking was that laser cutting isn’t a constraint process. Outside of the extremes (like the occasional 1-in. plate), a laser’s specific cutting time doesn’t hinder overall flow. If a lower-powered system happened to be available to cut a certain job calling for thick material, the job would be sent to that machine. Sure, the cutting speed would be slower, but waiting for available capacity at a higher-powered machine would be even slower.

As lean manufacturing practitioners continually preach, jobs spend most of their time between operations as work-in-process (WIP), not in a machine actually being worked upon. Using this logic, immediate availability matters more than cutting speed—right? As GSM found, not necessarily.

About three years ago, GSM began working with iNDustry Labs at the University of Notre Dame, which offers opportunities for students to work with local manufacturers. They analyzed six months’ worth of data from production control and ERP software, tracking cutting times, quality, machine performance (including instances of errors), and overall machine uptimes. They sent that data through specialized mathematical analysis software at the university, then came back with some surprising results.

“The students presented the data in a spreadsheet that showed an optimized allocation of materials to lasers,” Ndalama said, adding that those results ultimately led the fabricator to an entirely new laser cutting strategy.

Consider GSM’s newest TRUMPF laser at its South Bend plant. Installed in 2022, the 10-kW machine offers automatic nozzle changes and other features that help shorten job changeover times. The system has incredible speed and flexibility, and yet today, it rarely processes a low-quantity job. Instead, it cuts certain production orders involving 0.25-in.-thick mild steel—all day, every day.

Implemented after the analysis from iNDustry Labs, the strategy is a bit counterintuitive. GSM is a high-product-mix operation, and if a machine has all the latest technology to facilitate quick changeovers, why not use it? It’s also the highest-wattage laser on the floor. Why not use it to slice quickly through thick plate? Today, you’ll find thicker plate, all the way up to 1 in., being cut on GSM’s lower-powered machines.

As the students’ data analysis showed, the reasons behind all this go back to the mix of work—product type, quantity, the timing of orders (demand), and required lead times—and the types of lasers GSM has at its South Bend plant.

Again, schedulers used to send work to the first available machine, or multiple machines, and this included those few high-volume jobs involving 0.25-in. mild steel. They then sent other work to wherever open capacity existed. This meant that operators and material handlers spent a lot of time transporting a wide variety of material. Picture a tangled spaghetti diagram between the storage racks and the lasers.

Operators pull up a program to prep a job at a fiber laser at General Stamping & Metalworks.

GSM now sends all those high-volume 0.25-in.-thick jobs to the 10-kW machine, creating a high-quantity, low-product-mix “mini-factory.” Raw stock is stored near the point of use. Most important, segregating in this way ensures that the production work no longer “fights” for available laser cutting capacity with GSM’s long tail of lower-quantity orders.

The high-quantity work also makes the most out of the 10-kW machine’s cutting speed advantage. True, the material is only 0.25 in. thick, but GSM is cutting so much of it that the throughput difference is stark. Yes, the 10-kW system can cut much thicker stock much more quickly than lower-powered fiber lasers, but the quantities of thick plate (like 0.75 and 1 in.) are significantly lower.

The remaining work GSM cuts involves thicknesses from 0.015 in. all the way up to 1-in. plate. That huge range was partly why GSM previously focused on cutting flexibility, having lasers ready to cut any material, any thickness, at any time.

But the operation doesn’t cut similar quantities of all thicknesses, and the nature of demand varies between jobs. After the Notre Dame students gathered and analyzed the data, some patterns emerged. Outside of that 0.25-in.-thick production work, a significant portion of the plant’s long tail of lower-quantity orders involves mild steel 0.179 in. and thinner.

It’s true that modern fiber lasers can cut through thin sheet extraordinarily quickly. Even so, laser cutting has come a long way. The newer solid-state lasers at GSM, including 5- and 8-kW models, can cut thicker plate very quickly, and with good edge quality. A decade ago, GSM might have sent thick plate to its older CO2 lasers and thin sheet to the fiber lasers. Now, the shop is doing just the opposite.

Specifically, it’s sending 0.015- to 0.075-in. material to its oldest lasers, including a CO2 machine from 2005. Its 5-kW fiber lasers cut mild steel that’s 0.126 in. up to 0.135 in.—a narrow range, reflecting the higher volume of work in those material thicknesses—though those same machines also process aluminum and stainless steel between 0.09 in. and 0.135 in., a wide range that reflects the lower volume of work requiring those material types.

Several 8-kW fiber lasers cut material between 0.31 in. all the way up to 1 in. The wide range reflects the lower volumes GSM processes in each thickness. Meanwhile, one 8-kW laser is dedicated primarily to cutting 0.179-in. mild steel. Why? It again comes down to volume. So many parts require that thickness, and when all that volume takes full advantage of the 8-kW machine’s immense cutting power, overall throughput rises.

Ndalama added that each job is assigned a default and secondary laser. “The secondary laser is always a similar machine. If we have a bottleneck at the primary laser, we’ll route the extra demand to that secondary work center.”

When each laser works within a narrow band of process parameters, troubleshooting gets more efficient and proactive. “Because your processing parameters are almost constant, you’re not changing much,” Ndalama said. “This makes it a lot easier to troubleshoot a bad cut.”

An operator cutting similar-thickness plate can catch cut quality issues as soon as they become apparent and make the appropriate corrections. Conversely, if the same operator were cutting thin sheet for hours, certain cutting parameter problems, like beam centering, can “lie hidden” until the operator switches to thicker plate. “Thin material can tolerate a beam being slightly off center,” Ndalama said. “But if you jump to thicker material, that off-center beam will give you dross right away.”

Cutting data is fed directly into the Paperless Parts quoting system, which uses the cutting speeds the assigned laser can achieve. By segregating the laser cutting work, quoting has become more accurate.

Today, the spaghetti diagram between material storage and the lasers is far more orderly. Material is stored closer to its point of use—that is, its designated laser. And even after the increased throughput, the denesting teams still collaborate and keep up with the machines. As Ndalama explained, “Usually, someone who’s running a slower machine or a longer nest will help sort parts on another machine.”

How parts flow to the lasers isn’t set in stone, of course. If the product mix changes, so might the laser cutting optimization strategy. GSM’s exercise with iNDustry Labs proves an often-overlooked fact in custom and contract metal fabrication: The customer and product mix define the business, and that includes its operational strategy.

“We’re still keeping up with downstream requirements,” Ndalama said. “Again, it’s almost as if we added another machine, simply by rearranging the way we feed material.”

He added that this helps strengthen that excess capacity “buffer” needed in blanking to ensure those downstream constraint processes are never starved of work. Moreover, the increased quality (due to each laser operating within a narrow band of process parameters) has reduced the potential for rework—especially costly if it isn’t caught until it’s far downstream.

The basic tenets of lean manufacturing still apply. Quick changeover is important; so is material handling before laser cutting and part sorting after. But as sources explained, true improvement requires zooming out to see the big picture. Smart and efficient quoting really matters, along with order processing and, especially, effective job routings. As GSM has found, when the right job goes to the right laser, available capacity rises—with no additional machine purchase required.