|
the analyst says
High-Fidelity Product Creation through Digital-Physical Convergence
By Bruce Jenkins, Principal and Senior Analyst, Spar Point Research LLC, Danvers, MA
Low-fidelity information. After decades of IT investment, product development and manufacturing organizations still find themselves plagued by inform-ation that's not complete enough, detailed enough, or timely enough to support confident, low-risk decision making.
The culprit? Gaps within product-creation IT solutions, gaps between solutions and gaps in manufacturers' product creation processes. Where are these gaps, and what can program managers and discipline leaders do about them?
Low Fidelity in Design and Production
Design systems are low-fidelity? How can that be?
It's not that solids models aren't dimensionally precise — reality isn't. More accurately, reality is stochastic. Contemporary 3D modeling systems define nominal, idealized geometry, but in the real world geometry is variable. Why this matters for manufacturers is that as-designed is never as-produced — parts coming off the production line will always vary from the digital model.
Understanding that variance is critical for managing quality and yield. Needed are information technologies to help determine what variances can be expected, the probable frequency and range of variances, what's acceptable given product requirements, how variances can be constrained — and at what cost.
Low Fidelity in Simulation and Test
In addition to production issues, design systems are low-fidelity in the insight they offer into product function and performance. Beyond simple mass properties, solids modelers "know" little or nothing of how a part or product will behave in use; 2D CAD systems, of course, even less. Meanwhile, the software tools that do know about product performance — those for simulation and analysis — traditionally were connected to design systems poorly, if at all.
Nor have simulation technologies in themselves been very high-fidelity. Any given analysis tool simulates only a narrow slice of a product's behavior under tightly bounded conditions. Analysis tools for one discipline have not been easy to use in concert with those for other disciplines. Exercising a solver over a range of conditions to get a more complete picture of product performance under a variety of conditions, or to explore alternatives, has been prohibitively difficult.
Physical test automation likewise offers only limited snapshots of a product's performance. And the domains of digital simulation and physical test are not well integrated. Because of this, manufacturers have a hard time capitalizing on the physical test's potential to boost confidence in simulation results and that of analysis to make physical testing more efficient and informative.
Higher Fidelity through Digital-Physical Convergence
On the horizon are technologies and approaches that we believe promise a leap forward in fidelity of IT-based product creation. Their essence is integration between computer-generated digital models and sensor-based digital capture of physical conditions. By bringing together computer-based abstraction and sensor-based measurement, these new approaches promise to help manufacturers bridge gaps between the nominal, idealized world of computer-aided design, analysis and manufacture, and the stochastic, probabilistic conditions of real-world production and use.
Bridging the gaps between design, simulation and test will yield greater understanding and control of the variables that affect performance, quality and cost. Bridging the gap between design and production will make it easier to optimize designs for production, then monitor production conditions to control quality, maintain yield and maximize the value of production assets.
Boosting Fidelity by Converging Processes — Today
Already we see program managers driving aspects of this convergence in their organizations and processes. While not experts in any one discipline, these individuals are collecting input from specialists in design, manufacture and test, then making decisions aimed at balancing and synthesizing these discipline-level viewpoints and needs. If you have responsibility for product development, what you can do today is look for efficiencies you can achieve by bringing these work processes together, then seek solution providers and partners who get it and will work with you to make it happen.
Why now? For many, it's the next challenge. With solids-based design affordable and well understood, practitioners are looking for the next constraint to overcome. Also, the considerable computing power needed to link multiple disciplines and run more work processes in parallel is now available and affordable, or well on the way. Of course the overriding driver is the pressure to deliver more new product with more features at higher quality, without massive new investment.
The good news is that there are solution providers who get it. Across the industry we see dawning realization of the potential in tighter integration between modeling and simulation, simulation and test, design and production. Some leading providers of test-and-measurement environments are reaching out to connect their solutions to product design systems. Analysis tool developers are working to devise more sophisticated ways to give designers functional insight than early tries at this. Software for control systems simulation is evolving to foster early involvement of this function. Of course not all of this may be available from a company's current solution providers, at least right away — some may have to look further afield.
The old way was to develop product with incomplete information and not much automation. Consequently, product development processes were carried out more or less serially. The promise of emerging tools and approaches is to help program managers and discipline leaders correlate and integrate processes so that more work can go on in parallel. Product performance analysis can be confirmed by physical test and measurement. Product design can be informed by manufacturing and producibility experience. Besides accelerating product delivery, the payoff is that information and decisions move upstream in product development, where their impact is greater.
The business objective of this new focus on integration is to get disciplines and functional specialties involved in the product development cycle early enough to make a difference. This isn't easy in a world where product creation is carved up into separate departments for design, analysis, test, manufacturing engineering, quality, value — but some are doing it today. If you're a program manager, of course you have to protect the edge your company has painstakingly built up in a particular discipline. The challenge is to do so while bringing that knowledge closer to all the other competencies spread throughout your organization. When that happens, everyone wins.
Bruce Jenkins is a principal and senior analyst at Spar Point Research LLC. He can be reached at bruce.jenkins@sparllc.com. Spar Point Research LLC identifies and analyzes industry best practices for creating business value by deploying and applying strategic information technologies. For more information, visit www.sparllc.com.
| 
