Perfecting automated processes, flow of information in the fabrication shop

Automation software and the almighty algorithm are making metal fabrication technologies more intelligent and flexible than ever. And FABTECH 2023 in Chicago is poised to have even more digital transformation on full display. Butler Photography

The pandemic. Supply chain chaos. Long lead times. Unknown delivery dates. The level of orders tanking, then booming beyond everyone’s expectations. Talk about a few crazy years. Here’s hoping this year’s FABTECH ushers in a new era of steady, and at least somewhat predictable, growth.

The metal fabrication business continues its great digital transformation. We’re in an era of quick or even instantaneous quoting, of small batch sizes, of kit-based product flow, all with software offering new insights to improve throughput and for getting the best out of available talent—even in shop floors that process a multitude of disparate jobs over a given shift.

Some new machines give fabricators more capacity than they know what to do with, especially in cutting. Lights-out operation sounds great, but then what do you do with the pile of parts the next morning? Part-removal automation after blanking has come a long way in recent years and, with the perennial labor shortage, looks more attractive than ever. And as more automation reaches more shop floors, more experience the automation learning curve and the knowledge base grows, which in turn reshapes more metal fabrication careers.

Walk the FABTECH floor, and you’ll see two threads of automation, all driven by software and the almighty algorithm. One thread perfects the processes, the actual cutting, bending, and moving of metal. Another perfects the flow of information, from quoting and order processing through interconnected job routings, outside service providers, purchased components, and the ever-varying levels of capacity from the people and machines on the fab shop floor. The former (robots and other mechanical wizardry) gets all the attention, but the latter could well become the industry’s primary driver of change.

Cutting, Bending, Moving

Technology could (if it hasn’t already) drive some significant market shifts. Today, ultrahigh-powered fiber lasers are cutting plate well past 1 in. thick, with no noticeable dross. New approaches to assist gas mixing and delivery are driving the fiber laser’s cutting capabilities ever higher. Yes, the kilowatts keep increasing—15 kW, 20 kw, 40 kW—but it’s the machine and cutting head designs that are really making such ridiculously high-powered laser cutting a stable, repeatable process. When a laser can cut thick plate cleanly, no deburring required, the best fab shops start to take notice.

Laser beams today can be tweaked and shaped to meet a process’s cutting needs, and that includes ease of denesting and moving parts downstream. No matter how fast a laser cuts, a piece isn’t “finished” laser cutting until the next operation can do something with it. Automated cutting doesn’t look so automated when you look at the offload area, where an army of laborers shake, hammer, and lift cut pieces, then heave the skeletons into the scrap bin.

Here, algorithms again come into play, driving the laser beam (and specialized punch tools in punch and punch/laser combo machines) to produce specialized microjoints and cut geometries to ensure part stability during cutting, yet make part removal both reliable and easy. Other algorithms shape and manipulate laser beams to produce wider kerfs, to allow sufficient clearances for part-removal grippers to grasp cut blanks without issue.

Algorithms also drive the beam in a certain way to improve edge quality even in the most challenging geometries, ones where in years past, heat buildup in tight corners would have led to quality issues.

The goal is for the piece to flow immediately from blanking to bending, or whatever the next work center in the routing happens to be. No longer is being able to run something lights out a competitive advantage. Fabricators work to achieve optimal green-light-on time on their equipment, but customers don’t pay for green-light-on time; they pay for quality products shipped on time.

Hence the importance of flexibility, something that today applies to every process: plasma cutting, oxyfuel, multiprocess machines, waterjet. Nearly every player in this business has a renewed focus on agility, changeover, and the shrinking lot size.

In the forming department, air bending with supreme accuracy now is a given; material tolerances (range of thickness for each gauge, yield, and tensile strength variation) are one of the few wild cards remaining. The key now is quick changeover between jobs, achievable either with a strategically staged setup on a conventional press brake (row of punch and die sets across a brake bed, able to bend a variety of parts), or with automatic tool change on a press brake, panel bender, or folder. In many operations, the closer to kit-based flow a bending department can achieve, the greater the part flow velocity—and, ultimately, the greater the throughput.

The same goes for welding. The greater mix a welding operation can produce efficiently, the quicker small-lot jobs make it to downstream assembly and shipping. For several FABTECHs, attendees have witnessed robots “see” a welding job and, based on the weld data in CAD, commence welding—no manual programming whatsoever.

Welding cobots augment the capacity of a manual welding department, too, as do new innovations in laser welding, both automated and manual. When a laser can lay down a bead that requires no grinding, polishing, or any kind of postprocessing whatsoever, part flow velocity skyrockets.

Information Makes the Difference

Consider a modern precision fabricator who goes “all in” on the latest and greatest, complete with all the bells and whistles. It has laser cutting, punches, and punch/laser combos with multiple towers. It has part offloading automation, both from the punch (parts are sent down chutes or grasped with grippers directly off the punch bed) and laser (with grippers lifting pieces out of a skeleton). From here, parts go to flexible bending technologies (press brakes with automatic tool changes, panel benders, or folders), then on to the welding department.

How exactly should parts flow? How many pieces can actually be removed automatically from laser cutting? Could a gripper lift a “mini-nest” of tabbed-together parts? Or would it be more effective to send those small parts down a chute, or just shake the microtabbed parts out of a skeleton after cutting?

Which bending technologies make sense, and why? What kind of jobs flow well together (perhaps share common or similar setups), while still producing what’s needed downstream and not just adding to excessive work-in-process (WIP) inventory? Which works best for manual bending? Robotic bending? Brakes with ATC technology?

When it comes to overall production planning, what is the best time to release orders? Planners might look out into the schedule and nest some jobs early to optimize material yield, but not so early as to stifle part flow through the shop.

Moreover, how consistent is overall throughput? What technologies and training could make those throughput capabilities more consistent across shifts? Shops often automate to foster consistent production. A job programmed to be cut and automatically stacked will take a specific amount of time. A job programmed to be cut and then manually denested ... well, there’s a bit more variability to consider.

Is that “hard” automation—the machines, robots, and cartesian systems that move workpieces automatically—really the answer? Often, yes, mainly because labor simply isn’t available. Sometimes, though, people might just not have the right information. That information could have to do with levels of training, process standardization, or perhaps the fact that only certain people know how to do certain jobs well—and no work instructions, documentation, or use of technology has yet to unlock that tribal knowledge.

The scenario fosters “uneven productivity” in manual operation. It’s why many fab shops choose to automate. They see, say, denesting operations in action and don’t marvel at its speed—a team of engaged denesters sometimes outpace the automation, and they can work with any nest layout—no constraints on part orientation, kerf-width clearance, or other factors that help part-removal grippers to operate with repeatability and reliability.

But again, automation is consistent; manual denesters often aren’t. Let’s face it, spending hours working through a huge pile of cut parts is no one’s career aspiration.

Besides, no one can overcome changing demographics. More are comfortable in front of computer screens; fewer are comfortable with hands-on work. And they abhor tedium.

Certain manual jobs remain in metal fabrication, especially when it comes to moving jobs from one workstation to another. Someone has to manually lift, stack, drive a fork truck, or push a cart. Sure, mobile automation is coming—you’ll likely see a few examples at FABTECH this year—but its use hasn’t spread widely yet.

Mobile automation helps fill in the gaps created by the worker shortage, but the automation itself doesn’t solve the root of delivery problems. Emails get lost. Customers don’t respond to questions. Voicemails go unanswered. Approvals aren’t achieved. Jobs are released to the floor at the wrong time and in the wrong way. Such delays can last days or even weeks, and they can send ripples of disruption through the entire metal fabrication enterprise.

Here’s where the automation of information flow comes into play, from automated communications with customers, portals where customers can upload design files and other information, quick or even instantaneous quoting, all feeding off an interconnected web of data from the shop. When change happens (and in metal fabrication, it always does), the schedule immediately adapts, jobs are regrouped, renested, and re-optimized for material yield, part removal automation, press brake utilization, and other downstream processes—all in an instant. That’s the ideal, at least.

AI Steps in to Metal Fab

Richard Boyd is CEO of UltiSim, a North Carolina-based firm focused on digital twins, as well as Tanjo, which specializes in artificial intelligence (AI). He’ll be speaking at FABTECH this year about a manufacturing firm that’s on a quest to bring aspects popularized by ChatGPT’s large language models into the production control environment.

The goal is to ask a production system, “What should I pay attention to today?” From that, the system will look at all the data—the jobs on tap, historical information and trends, predictive maintenance information, and a multitude of other factors—then answer the question.

Although not complete yet, such a system would represent a new direction for metal fabrication, where more and more people work on the business rather than in the business. What type of work would suit our mix of technology and talent? Where can we handily underbid the competition? How can we perfect flow to make life easier for everyone on the floor?

The future of metal fab will focus on flow, not continually working overtime to muscle jobs through. Brake operators won’t put up with lifting large pieces to bend an edge flange. What about sheet followers? Special tooling? Maybe even a panel bender or folder? They’ll welcome part removal automation that eliminates tedious part stacking, even intelligent welding robotics that take over those lengthy, tiring welding jobs. And helping manage it all will be the almighty software algorithm.

It’s been a crazy few years, and this year’s FABTECH—the largest in the post-pandemic era—signals a shift in technological innovation. Metal fabricators aren’t in the parts-making business anymore. They’re in the business of information—gathered, analyzed, synthesized, and shared with intuitive software and good business practices. That’s what eliminates the unintended WIP and rework and allows for reduced batch sizes, getting closer to single-piece part flow and optimal job velocity, from initial quote to the final shipment. It also will help get more value out of every employee—a necessity, considering the chronic labor shortage. Along every step of the way, making best use of the right information at the right time will make all the difference in the world.