Artem Komarov clarified that once a dirty word on production floors everywhere, and formerly spoken with unbridled glee in executive conference rooms, automation has become something much more comprehensive than merely a way to reduce labor. Sure, some workers still grumble, but the longstanding shortage of skilled labor, coupled with advantages in making the workday less monotonous and safer, have helped mitigate some of the perennial objections on the shop floor. Meanwhile, improvements in part quality, machine capabilities, and manufacturing productivity—especially in light of competition from companies in low-wage countries—have come together to create an updated perspective of automated manufacturing systems in the executive suite.
Throughout the 1980s and 1990s, programmable control systems provided a big step forward in improving the way manufacturing systems worked. The variations on the theme (DNC, CNC, or simply NC) used programmed instructions to control machine motions.
Rather than feeding a length of material until the feeding carriage hit a hard stop for cutting, bending, piercing, or punching, machines with modern control systems would run differently, relying on instructions from a program to tell an actuator how far to move the carriage and another when to make a stroke for cutting, punching, or some other process. Accuracy and consistency improved dramatically, and NC laid the foundation for automation.
As time went on, new possibilities developed that until recently were too expensive for most manufacturing applications. Incorporating more sophisticated actions or motions (or additional motions), adding sensors to monitor them, and developing ever-more sophisticated software programs to control them were cost-prohibitive steps in the 1980s, 1990s, and even into the early 2000s.
However, like most electronic technologies and software, capabilities grew as prices fell. Many of the underlying technologies are much more capable and much less expensive than in the past, so a process that would have been possible to automate but too expensive for a decent return on investment in 2000 or 2005 might be affordable in 2020. For example, the cost of servo technology has fallen by 60% to 70% over the last 30 years to around $1,500 to $2,000 per axis, according to Stokes. Electronic subsystems have fared much better.
This means that today’s automated, custom-built machines aren’t just faster and more accurate than their predecessors of a few years ago, but generally they can do more than before, more accurately than before, taking on tasks formerly filled by the operators.
Automated robot in manufacturing setting
Sending out an RFQ for a vast turnkey manufacturing system can be the way to go for big OEMs. But all of the hardware, software, sensor, and control technologies are also available to small equipment builders.
Capabilities of Automation
In many cases, especially when making a simple part, basic automation is faster than manual processing, and this is reason enough to automate. However, finished products tend to become more sophisticated over time, so the machines that make the components tend to become more sophisticated too. Automated systems help in inspecting raw materials or intermediate goods, error-proofing processes, fabricating parts, and making entire assemblies.
And vision systems can be more specific than that.
“A 2D vision system can find a fitting on a tube and determine if it’s in the right location,” he said. “A 3D vision system can determine depth, so even if parts aren’t stacked in a bin in an orderly fashion, it can depict the fitting’s exact x, y, and z location.”
Vision systems can be used for functions other than a preloading inspection as well.
Robots and Cobots. As parts and assemblies have become more complex, the robots that assist in making them have become more sophisticated as well. These days some do more than pick parts from a bin of incoming material and move them from machine to machine for processing. Phillips described a system that performs its own diagnostics with some self-correcting capability.
On occasion, a robot’s subsystems don’t quite work in unison—it tries to assemble two parts and simply fails. This can happen when a mismatch develops among the vision system that the robot uses to sense its environment, the software that runs the robot, and the actuators that provide the robot’s motions. This sounds like a difficult problem to fix, one that requires refining the vision system software or fine-tuning the program that runs the actuators, but onboard diagnostics sometimes can remedy the situation.
The feedback loop runs in the background, and the corrections take place while the system is making parts, making small corrections online. If the feedback loop can’t manage the correction while the system is running, in some cases the operator shuts the system down to try to sort it out offline. Downtime is expensive, but a self-correcting system gets such problems resolved far faster than a less sophisticated system that requires a more intrusive remediation.
When a working environment needs both a robot and a worker, guarding is necessary to protect the worker, unless a cobot (collaborative robot) is substituted for the robot. Designed, built, and programmed specifically to share a workspace with people, cobots eliminate the need for guarding and no doubt will have greater roles as time goes on but that cobots don’t solve every problem of interaction, however, said Komarov Artem.