Going the Distance: Solving the “Last Meter” Connectivity on the Shop Floor

Magellan Aerospace Corp. of Mississauga, Ont., needed to get more production out of a three-machine cell at one of its aircraft component production plants. The problem was, these were three legacy machines with proprietary technologies.

Was the purchase of a fourth machine for the cell worth the large capital investment? What about getting those three machines to operate more efficiently? Was it possible? There was no way to know.

Separated from the company network by an unbrideable gap preventing connectivity and communications – a “last meter” so to speak – these machines were isolated from management visibility.

The concept of the “last mile” has often been used in telecommunications to describe the challenge of getting connectivity from the network to the customer’s premises. The term also has been used in logistics and supply chain management, referring to the last leg of the goods journey from the distribution hub to the destination.

But in recent years, a similar “last meter” concept has emerged in manufacturing, referring to the final connection from a machine or device to an enterprise’s network. Dave Edstrom, chief technology officer at Burlington, Ont.-based Memex Automation, says there are technical and business challenges involved in making that last-meter connection, in an interview with ThomasNet News.

“A device could range from a multimillion-dollar machine tool to an inexpensive sensor,” Edstrom noted. Solving the last-meter problem, Edstrom says, means “getting data from countless devices and putting it into an easy-to-read format, so data can become actionable intelligence anywhere, anytime, on any device.”

Today it’s not uncommon across a shop floor to find numerous stand-alone machines running several different proprietary controls in dozens of generational versions. It’s impossible to tie such a hodgepodge setup to the company network, limiting automation and giving management little visibility into factory operations and thus production optimization. Measures such as overall equipment effectiveness are simply a mystery.

 “Only 2 percent to 3 percent of all manufacturing plants are being monitored today,” said Edstrom. “And the average initial equipment utilization is 32 percent.”

Memex Automation is one of the companies working on the last-meter problem and often finds that installing a solution will “increase equipment utilization 20 to 50 percent with an ROI period of three to six months.”

One solution to this last-meter challenge began with the development of MTConnect, an open standard developed by industry players collaborating under the umbrella of the MTConnect Institute. Based on open protocols such as TCP/IP and XML, MTConnect fosters interoperability between devices and software applications. Edstrom, who was formerly chairman of the institute before moving to Memex, has referred to MTConnect as a “Bluetooth for manufacturing.”

“MTConnect has allowed us to link together devices, and especially the different ages of technologies on the shop floor,” said Jeff Estes of machine tool manufacturer Okuma America, speaking with ThomasNet News. Estes directs the Partners in THINC program at Okuma, a collaboration network working on the last-meter challenge that includes Memex, 5ME, Predator,and dataZen.

A factory might be employing a mix of machines of various vintages, “but now we can get at least basic information from all of those devices,” Estes commented. “We haven’t been able to do that with legacy technologies up until about two years ago.”

Previously, shop-floor analysis and troubleshooting required “a shotgun approach,” but attacking the networking challenge “has taken us that last step, that last meter,” Estes remarked. One reason the last-meter solution is important is that “rather than having to work on a whole bunch of things, we can narrow down on the real issue and focus our attention on that.”

Having visibility into machine outputs allows managers to diagnose problems using analytics and takes out the guesswork. “We might find out that the [cutting] tools are wearing out too fast,” said Estes, “so now we can focus on finding out what’s causing that. Memex’s software and others allow us to collect that data, bring it all together, and summarize it.”

Edstrom explains there is a “data dictionary” in MTConnect that lends meaning to the numbers. It standardizes many machine protocols into universal indicators such as “emergency stop,” “door open,” and “M00,” regardless of which control the manufacturer is using, easing the development of analytics apps. It is “the secret sauce,” Edstrom said, “a protocol pipe connecting manufacturing equipment to applications.”

Memex Automation has built a manufacturing execution system (MES), called MERLIN (Manufacturing Execution Real-time Lean Information Network), based on MTConnect and other protocols. MERLIN provides a real-time window into any machine of any vintage.

Memex Automation is able to combine MERLIN with custom-developed circuit boards and networking interfaces to convert legacy machines into networkable devices. Allowing managers to view basically any output that can be generated by an old machine, MERLIN can monitor its metrics, signals, and functions — number of fault events, quality, cycle time and count, part names and counts, alarm states, interrupted cycles, down time, feeds and speeds, idle times – and thus overall equipment effectiveness (OEE). It essentially gives the machine a second life.

In the case of Magellan Aerospace’s last-meter problem, the company decided to hold off spending money for a fourth machine in the problematic cell it was evaluating. Instead, the company installed MERLIN to monitor the three-machine cell and was able to identify an inordinate amount of optional stop time that was causing about 100 hours of idle time per month for each machine.

After making adjustments, Magellan Aerospace was able to go from an OEE rate of 36.9 percent to 85 percent. The company saved the capital cost of a new machine and monetized $40,000 per month of production time, which was enough to recover the cost of MERLIN in four months.

To see the full article, please click here.