The Digital Factory Decade

Simultaneous engineering has been the dream of automakers since Henry Ford introduced mass production. Computerization has brought it a step closer to reality, but completely joining every link in the design, assembly and supply chain has, until now, been a pipe-dream.

This decade will see a major step in the evolution of simultaneous engineering - if Audi has its way. The German automaker, together with software company Delmia, a division of CAD-CAM pioneer Dassault Systemes, is in the midst of a major project. Dubbed the 'digital factory' it could potentially shave months off the critical time to market of any new car program and boost the speed at which production ramps up to full volume. If Audi can manage this revolution, it will achieve an instant competitive advantage that will translate into millions of extra Euros on the company's bottom line.

It will mean a root and branch re-organization of the whole manufacturing process though, involving parts suppliers, the logistics supply chain and, of course, internal Audi departments.

"This is the dream we are working towards," said Werner Blom, the senior Audi engineer overseeing its plans. "We are striving towards the digital factory, one in which every step in the manufacturing process can be electronically simulated to improve planning, factory working and the quality of our cars."

Both Audi and Delmia agree that five years is a realistic time-frame to make a big dent in the project, although chasing it through to the hidden corners of the supply chain could take much longer. Blom said, "We think it will take five years to get into swing." Kilian Grefen, Audi account manager for Delmia, Dassault Systemes' manufacturing software, added, "The long-term goal is to get model cycle times down, but there's many major steps needed before that dream can become reality."

Efficient Environment
Today, Delmia is used in most factories to design and simulate critical processes, such as the body-framing station and engine-stuffing operations, dovetailing with design software such as CATIA, also from Dassault.

However, a sideways look at the process reveals that efficiencies are possible. For example, in a critical assembly process a robot may be needed. If design changes a component, manufacturing needs to know because maybe the operation of a robot needs changing, a simulation is at least needed and maybe even the whole operation redesigned.

But there is a time lag between supplier and manufacturing process, with design in the middle. Often design has to run a special computer run and delve into its archives to download information. If the component design was immediately available to manufacturing, the project could move closer to production faster. There are dozens of similar examples throughout any new car program.

"We want design to be able to work in the same environment as the product development environment, which means better communication, systems and working practices. That is the future," said Grefen.

Blom continued, "To make this all happen we need everyone in the chain working with the same processes and systems. Simultaneous engineering we have now. But it's the people as much as anything that we need to change. The next step will be not only working together, but giving the results together."

Database Discipline
So the starting point for the Audi digital factory is a huge database, which the company calls the IAPD or Integrated Audi Process Database. This will bring a discipline to every company and department that is a link in the chain - both internal to Audi and external - by setting standards for every detail of the process, such as components, factory fixtures, and assembly operations.

Structuring this database is currently absorbing the Audi and Delmia teams. "The goal is to reduce paperwork, because today there is still too much working on paper and we will link everyone to this database and align all departments to one status," said Blom.

Grefen added, "Audi has boiled this down to three essentials - product, process and resource or PPR. Resource, for example, means all parts bins, fixtures and so on through to the people and the factory floor space. It won't go as far as detailing whether the parts bins are full or not, but more which processes are needed at that bin. It's at the basic information level."

As might be imagined, structuring this database is not the work of a moment. At least 20 major functions - as diverse as the bill of materials, the quality plan, product design and the welding program - have to be accommodated in this automotive equivalent to Pandora's Box. Said Blom, "We have to link up the data of sequential production processes and apply control and monitoring to each stage." In practice that means applying the PPR rubric in detail, for example to the pressing of a door outer skin, the assembly of the door using that pressing plus all the other components in the door and then into the final assembly process.

Simultaneous Processing
The goal is to do this simultaneously, with design and manufacture overlapping their operations, constantly depositing real-time data on the IAPD, each department then accessing that data to update their own operations.

Blom dubs this 'simultaneous processing' and if achieved, it is easy to see how months could be squeezed out of both the product development cycle and the ramp-up to full production after SOP.

There is a massive technical computing project underway too, because each of the departments involved have their own software suites, for example CATIA, Enovia, and HLS (VW's own-brand factory layout software). They all have to be able to send the critical data to the database at the right time and in the right format and, of course receive data back. "It's a computing issue," conceded Grefen, "but it's also a communication problem combining humans with process and workflow. These are very different things coming together."

Workflow is one of the major potential savings in the digital factory. Grefen reckons that in today's simultaneous engineering process, up to 70 percent of engineers' time is wasted because they are either searching for information to answer another department's request or translating that information into format compatible between different software packages.

"That's a typical situation in the automotive industry with ineffective communications and different databases. In future there will be a single point of access." Nevertheless, he still believes there will be room for some departments to keep specialized databases "just for one or two persons".

Common Database
As an example of a time saving to made from the central database, there is no more persuasive an example as the addition of a humble part to the process. In today's simultaneous engineering environment, every individual department has to add the part to its system, defining it for their own part in the manufacturing process. Often, of course, that means several departments duplicating the same work. Take out that duplication by entering the data on a common database where all departments can access it and at a macro-level, there is an instant improvement in efficiency and reduction of time.

Although Audi may only be working on this initial step, Grefen reckons they have a "leading position" among European manufacturers. BMW, DaimlerChrysler, PSA Peugeot Citroën and Toyota are making similar plans, but at the moment VW's luxury brand is heading the field. "This whole process needs three to four years from ramp-up," added Grefen.

What is holding everyone back is the mammoth task of defining and re-organizing the workflows and communication methods in a process as complicated as car assembly. Blom isn't even sure how many people he has got involved in the project at Audi. But since every department is involved - and their numbers run into the dozens - it's reasonable to assume that thousands of manufacturing and design engineers around Europe are currently absorbed on digital factory projects.

Project Definition
It even seems impossible to define the exact scale of a project. Ask either Blom or Grefen how far they are into the project and how far it will go and the answer is a question. "Defining how much is done is very difficult," admitted Grefen, "because what is 100 percent of the project? There are a lot of different opinions about that. What is very clear is that the digital factory will be a continuous movement."

There are also different approaches to solving the same problem. Grefen said the Delmia and Audi approach is to re-define the whole work process - hence the enormous database program. "This is the solid foundation for the new methodologies," he added.

Companies from other countries - and Grefen singles out Japan as an example - are treating it as a huge simulation exercise. That means each link in the chain either doing its own simulations, allowing the results to be shared, or supplying data to allow one, wholesale simulation to be executed.

What is for sure is that the scale of the Audi/Delmia digital factory means that improvements will have be phased in. At the moment that means working back up the supply and process chain. Final assembly will be the first major operation to go live with the system - probably on the next generation Golf-based A3 hatchback, due for SOP late next year.

To make this happen, Audi's engineering teams have, in turn, broken the installation of a new production line down into four main segments: pre-planning, rough planning, final planning and implementation.

Planning the Process
Targeting efficiencies and time savings in the assembly process is entirely logical. According to Audi's production model, assembly planning is the longest process, so the gains to be made are biggest. It's also the part of the process that Audi organizes itself - there's not the added complexity of contractors and suppliers.

But it also requires a number of other computer systems to talk to the main database, such as ZENTO, which controls parts, MBB-2, the process database and PDVS, Audi's own software for 3-D engineering drawings. "We're starting by structuring the processes, adding more and more information through the planning process and saying: we need a four minute estimated cycle time on such and such an operation. We've got to bring a part to the assembly line, but what is the sequence time? We need to identify the critical paths, input the cycle times and organize the different processes on the factory floor. When we look at such things digitally, production becomes more efficient," added Delmia's Grefen.

The digital factory can also help with that other mysterious art - line balancing.

More significantly, it will also help cope with late changes, an ever-present feature of new car programs, for example when the design of a part needs to be changed to fit in with a change elsewhere. Said Grefen, "There will always be some mistakes and misfortunes or problems. But with better information and planning, they can be identified earlier."

The enormity of the task means that Audi won't go straight into reorganizing a whole factory production line, preferring instead to break its revolution down into bit-sized chunks. So lessons learned by the introduction of digital final assembly on the A3 will be followed on Audi's next new car program - tipped to be the next A6 and due in production around 2003.

During this next step, the digital factory will move back up the manufacturing process - probably to the press shop and BIW (BIW) assembly. "We've realized that we can't come to the 100 percent solution on a single model," said Grefen, "we'll learn on each model. It's a bit like a production ramp-up."

But moving the digital factory into the press shop and BIW will be quite a task, reckons the Delmia engineer. "A deep drawing simulation can be done today with CATIA, but we're looking much more at the end process. It's a much more organizational thing and they're running lots of pilots, looking for the methodology."

End-to-End Solution
The BIW section includes a huge number of operations and processes. "One the one hand, for every domain, you need an end-to-end solution, so everyone is collecting and feeding the same information into the database."

Just like in final assembly, the goal is the same, but Grefen reckons that in BIW a different approach is needed, because of the different techniques employed in the two departments: "Final assembly is dominated by organization of people, whereas in BIW it's process, technology and investment intensive."

Mistakes in BIW are particularly financially punishing, too, "Once the investment and installation is made, there's no chance to rectify the situation afterwards."

Line balancing is also a major problem in the press shop and BIW - a process where digital simulation can help enormously. "There's lots of different questions, like how many robots, how many fixtures and how to balance the output for each of these," added Grefen.

The digital factory team also has to decide when to open up the project to suppliers. In final assembly, that's not so important, but in BIW it becomes more significant, said Grefen, "That will be a critical point. Of course they're involved at the moment, but there are lots of questions about communications and security."

Another big question is cost. Reorganizing a factory, its workforce and supply base is not a matter of a few Euros, in fact surely it can be measured in dozens of millions of Euros? It's a question that Grefen prefers to answer with a question - "It's better to ask what are the benefits? If they are a better planning solution, with production to a better quality in a shorter time with less resources then it is surely worthwhile?"

If Audi's and Delmia's digital factory project can truly achieve those four goals, plus speed up the ramp-up of production, then the work of Werner Blom and his team will have been well worth the effort. And they must go down in history as the engineers that redefined the workings of a car factory. That is no pipedream.

Reproduced with kind permission from Automotive Manufacturing Solutions - Dec 2001.

Rand International Aggressively Enters the Dassault Systemes Marketplace

There was some exciting headline news coming out of the COE Spring 2002 Conference for anyone who follows this industry. On Monday of the COE Spring 2002 conference, Brian Semkiw president and CEO of Rand International, who also owns Trans Cat, announced Rand will be diverting most of their energy toward CATIA PLM activities worldwide.