Skip to content

Computing Advances Impact Manufacturing Systems

Bruce Morey
By Bruce Morey Senior Technical Editor, SME Media

This is the first in a series of articles that will cover what I think is the most important trend in our industry—the accelerating improvement in manufacturing technology. We are all aware that equipment providers are constantly providing the industry with better solutions. This magazine devotes most of its pages to “new and improved” manufacturing devices of one sort or another. It is easy to fall into thinking this is normal, but it is a “new normal” in the long history of manufacturing. When seen in context, the improvements in recent years are so spectacular that to a rational mind it seems this rate of change can’t continue. Multi-axis machining to microns of accuracy, metrology devices that measures millions of points in a minute, lights-out manufacturing, collaborative robots, 3D printing. Where is the end?

Change Coming Sooner Rather Than Later

As Scott Walker, chairman of Mitsui Seiki USA and statesman emeritus for our industry pointed out to me, “CNC machines in the 1970s worked, though not really well. They weren’t very reliable, but they worked. They got reliable in the 1980s, but even then most machinists could make a part faster than I could punch a CNC tape.” Fast forward to the early 2000s and try and imagine a profitable machining operation without a CNC. And today? It is easy to project six-axis milling machines programmed almost instantly from a supplied CAD model, running lights-out to produce parts. The question is when. It’s probably sooner than you think.

It seems that we are getting closer to seeing “disruptive change,” in the language of Silicon Valley, those boys and girls in the high tech industry that are bringing us change in our everyday lives. Can that be true in manufacturing? Maybe not as fast as industries with a pure digital play, such as social networking or how we stream movies and shows. Manufacturing has some natural brakes. Why? Partly because of the physics of cutting metal and moving physical things. Also, think of the structure of the business (lots of small machine shops) and the level of investment (millions of dollars). “We have all of this crazy technology, but the reality is that manufacturing is an incremental improvement process,” said Walker. “We go step by step by step. And we don’t do it quickly, because the capital requirements to tool up are pretty intense.”

If we accept that computing technology is a key driver of all this, the only conclusion can be that we don’t know. That can be both exhilarating and disorienting.

The reason we don’t know is that computing, one of the key drivers in manufacturing technology, is advancing faster than most humans can comprehend. Most observers predict it will continue to advance faster and faster. A common plot of this is shown in Figure 1, which is a logarithmic-linear plot. A straight line in such a logarithmic-linear plot is exponential, but the curve shown in Figure 1 is doubly exponential.

chart-future-ventures.jpg
Over the last 120 years, advances in machine computation has improved at a doubly exponential rate, which according to Ray Kurzweil makes it difficult for humans to predict the implications on the future. (Provided by Ray Kurzweil, Steve Jurveston via Creative Commons)

Doubly exponential is what makes things interesting. As Ray Kurzweil, the principal author of this figure, has pointed out numerous times, human beings can predict the future with linear trends, but exponential and doubly exponential change is harder for us to grasp. He believes our brains weren’t built for it. His conclusion is that humans tend to under-predict the consequences from doubly exponential change. But for those making career decisions or investment decisions or any other prediction of the future, we must adapt to this “change in how things change.”

To illustrate this impact on manufacturing, I overlaid on the Kurzweil/Jurveston plot what I know of improvements in machining and metalworking equipment in Figure 2. The thing that should strike you, as it did me, is the acceleration of manufacturing improvements, especially after the NC and CNC revolutions that hit our industry. All are driven to one extent or another by computer advances. From the 15-30 years it took for full CNC machining to take hold, all the improvements I mention above are being introduced faster. CNC machining grew from a novelty in the 1970s to standard operating procedure by the early 2000s. Lights-out manufacturing is a reality.

chart-future-ventures-fig2.jpg
In this figure, the author has added in the implications of the double exponential improvement in computations to the world of machining and manufacturing. This should help predict what changes will be next.

Change Coming Sooner Rather Than Later

As Scott Walker, chairman of Mitsui Seiki USA and statesman emeritus for our industry pointed out to me, “CNC machines in the 1970s worked, though not really well. They weren’t very reliable, but they worked. They got reliable in the 1980s, but even then most machinists could make a part faster than I could punch a CNC tape.” Fast forward to the early 2000s and try and imagine a profitable machining operation without a CNC. And today? It is easy to project six-axis milling machines programmed almost instantly from a supplied CAD model, running lights-out to produce parts. The question is when. It’s probably sooner than you think.

It seems that we are getting closer to seeing “disruptive change,” in the language of Silicon Valley, those boys and girls in the high tech industry that are bringing us change in our everyday lives. Can that be true in manufacturing? Maybe not as fast as industries with a pure digital play, such as social networking or how we stream movies and shows. Manufacturing has some natural brakes. Why? Partly because of the physics of cutting metal and moving physical things. Also, think of the structure of the business (lots of small machine shops) and the level of investment (millions of dollars). “We have all of this crazy technology, but the reality is that manufacturing is an incremental improvement process,” said Walker. “We go step by step by step. And we don’t do it quickly, because the capital requirements to tool up are pretty intense.”

Computer-Intensive Manufacturing

However, the part of manufacturing that is computer intensive is changing rapidly. Let’s call it the data revolution. CAD, CAM, metrology devices that scan large areas, computerized statistical process control, robots, 3D printing—all of these are enabled by cheaper, faster sensors and computers. Now with the emergence of data transfer standards from MTConnect to WiFi to 5G to Model-Based Definition, the factory systems these machines and robots are embedded in is advancing into the age of Industry 4.0. You can see this in the lower right hand corner of Figure 2, with the emergence of digital twins, cloud computing and instances of practical artificial intelligence. Improving data collection, analysis, and presentation does not depend on the physics of a cutting tool gouging metal—or replacing a multi-million dollar machining center that still works—so its improvement will move at the pace of the double exponential curve of Moore’s Law.

And data matters. “The biggest advantage that manufacturing has is information,” agreed Walker. “I’ve run factories and there is tremendous pressure to get product out the door. It is traditional to have a weekly production meeting with all the department managers. You would think that production meeting would be based on information. It is not.”

As he describes it, what really happens is that such meetings are dominated by personalities, and the plant manager over time comes to know the personalities he can trust. “It is brutal,” he continued. “What useful information and data give you is the ability to eliminate the personalities and the struggle that people have in communicating with each other because it’s really hard to fight good data. And good data will make meetings faster, decisions smarter, and results better.”

And what will make useful data easy to use is visualization. Anyone with the right access can see on their smartphones what is happening and drill down and see causes and effects. Soon, practical augmented reality and virtual reality, enabled by faster computing and IoT devices, will enhance the ability to make use of this data for decision making. “Data instead of opinions,” concluded Walker.

Editor’s Note: Listen to a podcast interview with Scott Walker by Bruce Morey

  • View All Articles
  • Connect With Us
    TwitterFacebookLinkedInYouTube

Always Stay Informed

Receive the latest manufacturing news and technical information by subscribing to our monthly and quarterly magazines, weekly and monthly eNewsletters, and podcast channel.