Manufacturers are aiming to exploit the benefits of full software compatibility between production, design and supply chain systems.
In recent years one of the holy grails of manufacturing has been the 'digital factory', a dream world of total seamless control throughout the whole production cycle. However, both software development and communications infrastructure have started to reach a point whereby building and operating a virtual factory is becoming a distinct possibility promising concrete benefits.
Visitors to this year's Hannover Messe, the industrial exhibition in Germany, had a glimpse of how this might work one day. At a huge stand measuring 160m in length, Siemens and Volkswagen replicated the entire factory line used to assemble the VW Tiguan car. But while many of the robots and automation tools where undoubtedly physically there doing their respective jobs, the 'action' from the press plant, paint shop and powertrain assembly line could only be followed virtually with the help of computer monitors.
These virtual paint-spraying robots weren't just an animation video but a faithful digital rendering of the real-life physical processes. Such representations enable companies to quickly customise products to individual buyer's wishes - indeed, some would argue that the process of 'mass customisation' is the ultimate promise of the digital factory. If every manufacturing process could be similarly modelled to the last detail, then ordering a robot to mix a certain unique paint colour - or altering any other part of the production process for that matter, even if this is already under way - should be as easy as changing the font colour in a word processor.
Manufacturing firms have, of course, long had experience of dealing with a vast range of software tools that help them visualise and manage real objects and processes. CAD (computer-aided design), CAM (computer-aided manufacturing), MES (manufacturing execution system), PLM (product lifecycle management), ERP (enterprise resource planning), MRP (manufacturing resource planning) or CNC (computer numerical control) are just a few examples.
What all of these systems have in common is that they generate, store and transmit data to connected hardware, be it to impart certain instructions, to obtain information from sensors or simply to provide information to users. One of the biggest challenges in industrial IT is making these multiple sources of information talk to each other.
If digital manufacturing is to provide the benefits of mass customisation, reduced time-to-market and cost savings that it promises, full compatibility will be needed between the software programmes running robots in the plants and those being used by the development, product design, supply chain and marketing departments.
Closing the loop
"We're at a juncture where PLM really is starting to become an integration of the virtual and the physical worlds," says Bill Carrelli, vice president of strategic marketing with Siemens PLM Software.
"All along, the PLM market for the most part has been primarily focused on supporting companies in the early stages of a product's lifecycle: product planning, product development and a lot of the virtual product development activities.
"And as much as there were still linkages to data, there wasn't an integration between them such that the manufacturing could leverage directly the PLM work that was being done and that the upfront PLM work could be influenced directly by the downstream manufacturing implications of what the design was actually doing."
What Siemens is trying to do, Carrelli explains, is to develop a closed-loop process where the manufacturing data that are generated on the shopfloor by work cells, machines tools, measurement and control systems are fed back directly, in real-time, into the product development tools upfront.
"Suddenly," he says, "we're immediately able to compare: is the product actually being manufactured as it was planned in design? Is it in spec with that? Are there any issue that are being raised on the manufacturing floor which weren't anticipated upfront where we can quickly make some corrections?
"At the same time, we're building a database of information, because this is a learning process: the more data we extract from manufacturing, the more accurate the virtual data upfront becomes in terms of our ability to predict manufacturing early on in the process."
One of the most obvious benefits of having full connectivity and software compatibility, particularly between design tools such as 3D CAD and shopfloor data systems, is the acceleration of time-to-market for new products.
Giorgio Tori, director of software for EMEA at Rockwell Automation, says: "You can imagine that, if the engineering data that represent the design of a product can be automatically transferred to the manufacturing site and can potentially automatically produce the code into the automation and the PLC [programmable logic controller] that are necessary to produce that particular component, then the engineering time for implementation of the software would be dramatically reduced."
A second advantage of such an approach would be an automatic reduction in the amount of programming errors. And that's not where the benefits end. "One thing you can do is to implement a virtual commission, so you can understand what would happen if you really started production at that particular point in time," Tori says.
"For example, if you were assembling a car, you could spot a potential interference between the physical movements of the robots that are supposed to assemble the body of the car and the body of the car itself. You do all that virtually, without actually doing the real commission, which reduces both time-to-market and costs.
"Are we there?" Tori asks. "Not yet, because nobody has a solution yet that could fully translate the conceptual design into PLC code. And, in any case, I doubt that this [fully automated process] will be possible for any type of production. I think there will always be something that will need to be partially manually adjusted. But, of course, times can be reduced."
Like Siemens PLM Software (a new Siemens division formed last year after the German firm acquired PLM specialist UGS), Rockwell is working on a solution that could unify design and manufacturing. The first Rockwell products to offer such functionality are expected to be announced at the 2009 Rockwell Automation Fair in California.
Even if robots and designers gain the capability to start talking to each other (which is essentially a software integration issue), there's still the communications challenge of linking up these two groups together online.
"Today with globalisation there are a lot of different development teams that are concurrently working on the manufacturing of products, and they could be spread right across the world," says Joel Shapiro, group manager of industrial measurements and control at National Instruments.
He says there are two types of measurements that are important to help coordinate this global production model. The first one is end-of-line measurements for real-time inspection of parts. The second one is measurements that allow companies to evaluate quality assurance, getting real-time feedback to, for example, pinpoint where the poor-performing parts of the process are.
"Connectivity allows you to take that data and pass it back to either R&D [research and development] teams or a centralised hub, where people can then access information all the way down to every sensor and analyse it," Shapiro says.
He accepts, though, that sometimes this won't be easy: "Even though the trend is going to be for people to want to have access to all that data, a lot of the existing infrastructure won't be able to support that.
"A lot of control systems out there are closed systems where you have connectivity to only one bus. Being able to connect to a variety of different busses as well as having different gateways that allow you to connect to those older sensors to take that data and publish it back on the central enterprise management system is going to be very important."
The automotive, aerospace and electronics industries are those particularly driving the demand for these types of applications.
Rockwell's Tori expects such demand to spill over into other manufacturing sectors wherever having global data visibility brings added value.
"You need to have the possibility to compare in a fair way the performance of different sites that may work in different time zones, operated by people with different cultural backgrounds and using different types of raw materials," Tori says.
"Despite all of these differences, the results of key performance indicators must be consistent throughout an organisation. They must be timely, and they must be received by the right people in order to facilitate the decision-making process."