March 18, 2014
With all the conversations you hear around the Internet of Things, manufacturing has had a pretty strong grounding in that area already, given the communication capabilities of sensors and controls that have been introduced to equipment over the years.
As Mike Hannah, manager of network and infrastructure product development for Rockwell Automation puts it: “For us, these things have been sitting out there on networks at the machine and plant level for over 10 years.”
That being said, the road to a truly smart manufacturing environment is not an easy one. Nor has anyone achieved the Utopian ideal of a world where every piece of information flows seamlessly from parts and receiving to production and shipping. But many players are getting closer to the mark with each passing day.
Smart manufacturing is perhaps best defined by the SMLC (Smart Manufacturing Leadership Coalition) a non-profit organization comprised of manufacturers, suppliers, technology companies, universities, government agencies and labs. It defines smart manufacturing as “the infusion of intelligence that transforms the way industries conceptualize, design, and operate the manufacturing enterprise.”
In more basic terms it is talking about what can be achieved by integrating network-based data and information that comprises all aspects of a manufacturing and supply chain enterprise.
Smart manufacturing could very well be a matter of competitive survival, says Rockwell’s chairman and CEO Keith Nosbusch. At the first Smart Manufacturing CEO Summit held last year he said that “the combination of automation and information is the next wave of productivity. Being able to get real-time feedback from the consumer and to apply that information quickly to the plant floor to change what you’re making – that seamless integration of the enterprise, the supply chain and the plant – is becoming the next wave of competitive differentiation.”
That was the thinking behind Moosehead Breweries’ $35 million capital-investment program to improve manufacturing efficiency at its Saint John, N.B. operations. According to Wayne Arsenault, vice president of operations, it is now in Phase 2 of its modernization project for its canning processes, with more phases on the drawing board.
Going smarter is a matter of survival in the brewing industry, Arsenault says. “In this business, consumers are always demanding more value for their money. They want consistent product at lower prices. In order to become more effective and efficient at reducing costs, we have to invest in controls and automation tools to be in the game. It’s an absolute necessity.”
The foundation for Moosehead’s transition to a leaner, smarter production facility was the bottling operation, where upgrades to systems and processes have already reduced labour requirements by 50%, while maintaining the same output.
“Right now we’re working on a controls and automation infrastructure upgrade as well as linking that architecture with the corporate IT infrastructure so we can make more real-time information available from the board room to the shop floor,” Arsenault explains.
It is a long way from the days of person-to-person and paper-based communications, when there was no platform linking machines with PC controls or any kind of HMI (human machine interface) for the operator, he adds. “Now the lines are linked end-to-end and information can be displayed in real time up to the boardroom level. The next stage is connecting handheld smartphone devices to PC-based controls on the shop floor to enable line diagnostics and troubleshooting.”
“We moved from having to hit start and stop to several HMI screens that show the interaction between each machine,” says Matt Walsh, manager, brewery maintenance and engineering. “At any point a technician can scroll in to view how the overall line performs, much like SCADA systems.”
With 12 main and 12 sub-machines (i.e. machinery that conveys product) fully linked, active monitoring can be performed on all sensors and variable speed drives and controls over Ethernet. Once wireless is added, operators will be able to use the HMI to zoom into an area and troubleshoot or trigger a stop signal if needed.
“Bringing it all together allows us to harvest business advantages out of the system and take advantage of real-time information and troubleshooting capabilities, as well as store programs and history of performance,” Arsenault notes. “Collecting and communicating feedback means we can look three stages ahead to see performance, and go three back to modulate things where necessary.”
Having worked on a number of infrastructure integration projects in manufacturing, including for Moosehead, Andy Bentley, president of McCrae Integration in Toronto says the biggest movement over the last few years on the production front is the breaking down of the isolation that used to exist around manufacturing systems.
“Now everything is being integrated with overall business systems from top to bottom,” he says. “Production machinery is talking through networking infrastructures to business systems such as SAP and Oracle more than it ever has in the past.”
While it all makes sense, it is not without its challenges. “There are different constraints that aren’t always understood by traditional IT specialists,” Bentley says. “They are used to working with different priorities and types of networks. Deploying systems on the shop floor that production personnel can service 24/7 gives the network a whole different look and feel.”
Security is also a challenge, because in many cases, third party vendors need to dial in to service equipment, and having VPN access opens up new vulnerabilities, he adds. “A lot are quite concerned about how to segregate the network for security to ensure they have protection from production to the main server environment. And IT’s drive for virtualization is not necessarily a fit with production requirements either.”
The biggest hurdle is getting process control IT on the same page, he says. “They never talked to each other in the past. But most projects now cross over and require a level of interdependency.”
Jim Simpson, associate partner, IBM Natural Resource Solutions Centre based in Calgary says integration of information between the front line machines and business system also relies on understanding the value of the usefulness of the data in hand. “There is a lot of instrumentation in all industries. While the extra data may be helpful, the problem is a lot of data does not mean a lot.”
The chemical and petroleum industry has recognized for some time the need to bring all the pieces of information into a common place to tie it all together, he says. “That’s the critical piece. The next part is developing the techniques and capabilities to get real, action-oriented intelligence out of it so you can take prescriptive [or preventative] action.”
The severe weather conditions in remote production locations puts a significant burden onus on the wireless side, Simpson adds. “Advanced wireless networking is a must do for anybody in this t his industry to have near real-time monitoring of operations. There are a lot of small facilities that need to be monitored. At the same time, there are facilities like oil sands mines that are famous for their size, where players have to invest in mesh type networks to get the coverage they need to move data around.”
The industrial world as a whole is going through a dramatic change on the wireless side, as operations increasingly look to technologies such as wireless cards, Wi-Fi and ZigBee, says Dilip Advani, director of product management and marketing for Fluke Networks in Santa Clara, Calif.
There are however, particular challenges that are unique to industrial settings. “There are constantly changing environmental conditions that have introduced multiple problems in terms of interference. Wireless communication is a completely different beast compared to a wired network where things are more controlled inside the cable.”
Testing for interference is a particularly critical issue in the manufacturing world, he says. “There are different types of invisible threats that exist that could cause problems. Video cameras, sensors, motion detectors … they can all cause problems for a Wi-Fi network. It’s important therefore for the organization to be able to detect sources of interference to ensure optimum network performance.”
Advani recommends using a spectrum analyzer at the Layer 1 physical layer to look at raw RF (radio frequency energy). “It can tell you if interference exists, the level of impact on your network and the root cause.”
Remote detection of devices can be done with a 24/7 intrusion detection and prevention monitoring device. “It helps detects threats that may exist within or from external sources in real time. This device will automatically alert you to the presence of a device [brought in from outside] and block it from operating in your environment.”
Whether it is a wireless or wired environment, the thinking part of a smart manufacturing operation is more than just cabling and access point decisions. Bob Speed, vice president of Lean Enterprise, for Belden says their Lean approach includes working on linking manufacturing capabilities and supply chain to meet customer consumption patterns and increase precision in production numbers.
“Doing it precisely allows us to get the output we’re looking for. We can actually go in to the machinery and tell it what to run and it will automatically know the processing parameters, set them and continually monitor them to ensure everything is staying where it is supposed to, regardless of where it is in the world. We can also gather data, transmit it for analysis, make decisions and transfer that back to the production machine to make adjustments.”
While it is ultimately about the integration of people, processes and technology, Speed notes that ultimately “machines are simply a lot smarter today.” C+
Denise Deveau is a Toronto-based freelance writer. She can be reached at firstname.lastname@example.org.