Machine tools learn to improve thanks to AI

Share This Post

By Robin Cave, national applications engineer, Mazak Corporation

Artificial intelligence (AI) and machine learning can analyze performance data and deliver optimization results that exceed what human beings can accomplish.

For data giant Google, AI reduced server-farm cooling costs by an average of 30 percent after only a few months of use, with even greater optimization potential as the system continued to acquire performance data. Similarly, AI can reduce manufacturing setup time, enhance efficiency and workpiece quality, and boost productivity for better profits, helping shops progress past old-fashioned trial and error.

Until the advent of AI and machine learning in manufacturing technology, the old saying “If at first you don’t succeed, try, try again” described how shops attempted to make good production results repeatable. Because the machines could not provide first-hand insights into their performance, engineers identified good parts and doubled back to the production floor to ask operators how they achieved that level of quality. But even with operator notes, many shops failed to replicate success from day to day, and the large, growing volume of data generated through machine analytics provided too much information for timely processing.

Now, with smart CNCs and AI systems, machines can provide essential data about how they operate when they make a good part, and process all that information automatically to improve future results through machine learning. In an AI-equipped CNC, the controller can collect data about everything from axis thrust levels to spindle temperature and RPM, coolant behaviour and more, providing a far more detailed picture of each production task than any operator’s notes could capture. All this information can facilitate faster setups and shorter cycle times, improve surface finishes and boost machining accuracy, all while making overall machine operation easier.

But at the rapid pace of a busy production environment, humans cannot collect and analyze the exhaustive data required to recreate a successful production task at scale. To achieve these data-driven improvements, Mazak developed the Mazatrol SmoothAi CNC—the industry’s first dedicated CNC designed specifically for five axis machining and multitasking—and gave this new control the artificial intelligence to manage production project data, CAD/CAM software and machine behaviour. This enormous volume of information includes programs, tool and coordinate data, and all the parameters required for machining.

To process all this information automatically, the Mazatrol SmoothAi CNC includes the Mazak Smooth Project Manager, which provides an AI-managed set of analytics. Its advanced processing helps reduce the time required for data to flow into and out of a machine network, which in turn speeds setup. With the SMOOTH Project Manager in charge, the system can communicate the machine’s configuration, condition and behaviour to CAD/CAM software, virtually eliminating manual input and the errors it can cause. These advanced capabilities help flatten the learning curve for new machine operators just as they enable less experienced employees to succeed at more advanced production tasks.

Along with the information necessary to produce good parts, analytics also help eliminate machine crashes and production-quality hazards. On Mazatrol SmoothAi CNC-equipped Mazak machines, Ai Thermal Shield automatically adjusts temperature compensation to maintain precision as environmental temperatures change. As thermal compensation reports accumulate, artificial intelligence uses the long-term data to optimize compensation even further. The optional Mazatrol Smooth AI Spindle then monitors milling vibration and automatically adjusts spindle speed to minimize chatter and maintain excellent surface finishes.

AI enhances advanced processes as well as production of individual workpieces. Automated and lights-out machining have become essential in many shops, but these processes can fail unless tools, workpieces, fixturing and programming are available before production starts. On Mazatrol SmoothAi CNC-equipped systems with advanced robotic automation for high-mix, low-volume production, dedicated management software makes it easy to set up robots and verify all production resources before the cut begins. The same system also validates tool condition against the length of the automated production cycle, ensuring that sufficient tool life remains to complete the production process without an interruption. If any of these parameters fail the pre-check, alarms alert operators to correct problems before errors can endanger part production or even damage equipment.

Shops with stored Mazatrol programs from machines equipped with previous-generation Mazak CNCs can use Smooth CAM Ai File Manager to convert these programs for use on new machines with Mazatrol SmoothAi CNCs. That conversion process can work in both directions, making many programs backward compatible directly through a shop’s production network and ensuring that shops make the fullest use of all their machines.

In short, every aspect of production, planning and error prevention can benefit from AI’s ability to foresee how past experience predicts future performance, manage machine behaviour and share moment-by-moment updates on machine condition with CAD/CAM software and automation. Instead of “try, try again,” AI and machine learning make great results repeatable without guesswork. From start to finish, these developments can revolutionize metal cutting in all industries for shops of all sizes.

Mazak CNC controls.

Share This Post


Recent Articles

Wordpress Social Share Plugin powered by Ultimatelysocial

Enjoy this post? Share with your network