So much owed by so many to so very few

It has long been a bone of contention among ‘ordinary’ CAD/CAM users that software developers spend a disproportionate amount of time and resource looking after the special needs of the automotive, aerospace and defence industries. Just after the initial deal with Ford, the next release of CAD/PDM (product data management) vendor SDRC’s I-Deas software became known as Master Series ‘Ford’ instead of 4. And IBM has been developing specific modules in its Catia 5 CAD/CAM specifically for automotive body-in-white and Class A surfacing. Why do they do that? Two points in defence of the developers. First, some recent stats from the massive EDS PLM Solutions IT house: the top 20 automotive suppliers all use its Unigraphics. 27 of the top 30 automotive manufacturers each have 100 or more seats of EDS software. And five companies in the automotive and aerospace sectors each have over 10,000 seats of EDS PLM Solutions software: Boeing (56,000 seats), General Motors (51,000 seats), Ford (22,000 seats), Bombardier (20,000 seats) and Delphi (12,000 seats). You can bet your life that IBM/Dassault, if the firms ever released any numbers, would have a similar picture to paint. These kinds of companies either use Unigraphics, Catia or both big time, so it should be no surprise that developers accommodate their needs. You could argue this should make you feel pretty secure, because with that kind of patronage, these companies are not going to go bust! Unfortunately, many users note the software module naming conventions and assume that because they indicate an automotive preference they are not more generally applicable. This is certainly not the case with Catia 5: there are some clever, tolerant ‘all terrain’ sweeping functions in the body-in-white modules that allow you to join moderately dissimilar sheet metal profiles in a semi-automatic fashion – a process that could take hours if you were to do it manually. Much wider usage These kinds of modelling technologies are certainly not just applicable to automotive body-in-white. Consumer product designers can take advantage of the vendors’ automotive Class A surfacing. As Phil Gray, managing director of industrial design firm Quadro Consulting, says: “When the surfacing tools arrived in Catia… it did all the surfacing we needed and it had all the integrated facilities of a robust 3D engineering design and manufacturing tool. The quality of surface that we can develop using ACA (Automotive Class A) and the economy and the ease with which you can develop it, is really impressive.” The point is Quadro designs phones (and many other things), not cars. And following the logic, the issue is that if the software is developed to meet the needs of the most demanding users in the world, its overall quality improves and everyone benefits. Time was when if you used software developed for the automotive/aerospace market you paid a very high price for it – not any more. There are developments from all the big boys that are very affordable even for smaller engineering and product design companies. Bennett Associates, a small mechanical engineering consultancy, is using Catia 5 to design the creep forming wing moulds for the new Airbus 380. The firm’s Richard Burgess says: “We could not actually do our job if it wasn’t for some of the high-level functionality. We need to design the substructure and the forming surface of the creep forming tool in parallel with the design and development of the final wing forming profile. Catia is the only software package which allows us to continue with the substructure assembly and at the very last minute substitute our dummy wing surface with the real thing.” Fact is, irrespective of where you stand on the issue, modern CAD/CAM would not exist if it hadn’t been for the lead taken by the automotive and aerospace industries. Just as Bezier replaced the Coons patch and both were overtaken by NURBs, so too have the design and manufacturing processes changed over the years, driven by competition in the automotive and aerospace industries – and the desire to reduce new product introduction times. Contrary to the way it is often presented, time-to-market has been the number one critical issue for every design and manufacturing organisation for the last 20 years: it’s not a recent invention of these youthful software sales and marketing types. Competition, itself driven by technology, is forcing vehicle makers to reduce their time-to-market from five years to two. The only way to bring about what is by any measure a drastic reduction in development time, has been to introduce high levels of standardisation and design and manufacturing automation. Here in particular, computer integration of design and manufacturing becomes essential. There is no time for extensive prototyping and testing – the majority of these processes just have to be simulated. And this time-to-market reduction not only affects the OEMs, but everyone in the manufacturing process. All auto makers have reduced the number of disparate computer applications and systems they use as part of the drive to streamlining their processes and gaining time-to-market advantage. Some are also trying to influence their suppliers to use the same systems so that assembler/supplier ‘extended enterprises’ can work together and compete more effectively. Of the ‘Big Three’, Chrysler (now Daimler-Chrysler) was probably the first to rationalise processes and, mostly from necessity, it adopted a single software architecture earlier than the rest. Ford and GM have both now long since adopted strategies to make CAD/CAM/CAE the centre of their business processes. Ford has C3P, CAD/CAM/CAE and PIM (product information management – engineering, geometric and process data, sometimes referred to as PDM) and GM has its C4 programme. C3P is supposed to unify CAD/CAM/CAE at all stages of component, powertrain and vehicle development. C4 at GM was started in 1985; by 1994 the firm had reduced the total number of systems from 26 to three, which has further been reduced now to a single Unigraphics system. Today, according to Jay Wetzel, vice president of GM’s Technical Centres, the company has 21 vehicles in development on 24-month cycles – more than any other auto-maker – “thanks to advanced computer-aided design tools that have already resulted in $1 billion worth of savings.” The objectives of all these programmes were to focus on total computer definition of all automotive components, and link to design with engineering and manufacturing at every stage. Critical to these is a single computer model passed to the various design applications and, using wide ranging computer simulation, tested in ways not possible before – with the intermediate objective of reducing the number of functional prototypes and hence the overall vehicle cycle time. For example, GM makes extensive use of WAVE large scale assembly-level parametric modelling. “What used to take days can now be reduced to minutes,” says Ron DeBrabant, director of design process and CAD/CAM integration for General Motors. “A control structure using parametric modelling is put in place for the total vehicle, including the vehicle architecture, assemblies and components. This control data – or interface parameters for the total vehicle – can then be shared from programme to programme in reusable digital-model libraries, and design teams can work concurrently within a common product framework. We can explore design alternatives quickly by changing the components’ design criteria, or parametrics, and all the associated assemblies and components are updated simultaneously.” That’s the scale of the opportunity from all this. But to achieve it, suppliers too have needed to embrace standardisation and invest in advanced technology just to stay in business. Many auto-makers have now suggested to their supply chains that the only way to do business in the future is by sharing information electronically. The implication is clear: in today’s design environment there’s simply no place for artificial walls between departments or functions, nor ‘electronic walls’ between computer systems. Integration has never been so critical: in a two year development cycle everything needs to be integrated at the outset and the design and manufacturing strategies defined. In this new environment there are whole processes, whole factories and supply chains, that need to be ‘controlled’ precisely. There is no room for duplication of data or sloppy manufacturing administration – and at the other end of the scale, that includes ensuring that truck loads of components arrive at a particular gate of a particular plant to within half an hour of the required time. At the front end, where information is shared, ‘digital mock-up’ effectively becomes possible for all stages of the process. Designs work first time because they have already been extensively simulated in the computer, and this has fundamental effects on new vehicle introduction times – and development costs. It sounds simple. The traditional styling approach was to make full scale ‘clay’ models to facilitate selection. Concepts would be passed from styling to engineering, tooling and manufacturing in a serial fashion, with each department having different computer systems. Drastic time savings of 30—50% are possible from concurrent engineering, executing product design, development and manufacturing in parallel, rather than sequentially. Simple yes, but its implementation requires collaboration and team working between functional groups both inside and outside companies from suppliers to customers. The first step to achieving this has nothing to do with IT: take down the existing physical walls between people and move them closer together, so they can communicate. But as the industry globalises and that ceases to be possible, you need ‘digital co-location’. Sharing is essential Then you absolutely need an IT and application infrastructure to enable data sharing, as described, so that all stages relevant to the design can and do start as early as possible. Remember, concurrent engineering also reduces the cost of change: it allows companies to identify changes early in the product development cycle, minimising cost while maximising opportunity. And don’t exclude shop floor simulation data; a large proportion of the cost of new model introduction involves reconfiguring and building the production lines. Simulation of the line structure with the CAD data is an important additional lesson from the automotive and aerospace sectors. These industries have had a profound influence over the development direction of modern design systems and beyond, not least of which is the notion of a single ‘math’ model: one digital copy of the data through all its processes from concept to maintenance. Add to that concurrent engineering, collaborative engineering and risk shared development and you paint a picture of engineering design development over the last 15 years. Perhaps the biggest beneficiary of all this is the wider industrial design community, a market too small to attract any dedicated attention from the developers. Efforts to accommodate concept design in the single math model have provided excellent platforms for industrial designers linked to the best surfacing technology available. There are technologies we simply wouldn’t have but for the deep financial pockets of the OEMs – visualisation centres, virtual reality caves, video conferencing for design review, crash analysis and sophisticated CFD (computational fluid dynamics) simulation, to name a few.