American Motors (now Chrysler) is largely attributed as the inspiration behind the emergence of the Product Lifecycle Management (PLM) software market in the mid-80s. AMC, looking for a way to speed up product development, first incorporated computer-aided design (CAD) to make engineering more productive, and then coupled that with a communications system to facilitate in design conflict resolution. This early adoption of PLM is said to have enabled Chrysler to reduce automotive development costs to half the industry average.
Many software solutions have since come together to organize and integrate the different phases of a product’s lifecycle. PLM is now a cohesive collection of tools. A handful of software providers covering the entire PLM range dominate the marketplace while others offer targeted niche applications. PLM enterprise software has emerged from nothing to a market valued at $8 billion in the short span of 15 years. Its forecasted growth rate of 9% will make it a $20 billion market by 2012. As dramatic as this is, PLM lacks one fundamental aspect before it can truly live up to its name – the means to accurately forecast the product in-service life from design to end of use.
PLM is used to design and analyze complex product concepts, pre-manufacture, with high fidelity, 3-dimensional analysis – even the capability to simulate how raw materials are processed into complex geometric shapes before the processing actually takes place. Yet, PLM today lacks an accurate and efficient means for computationally simulating how the fleet of finished products will behave in-service. PLM aftermarket currently consists of software comprised of cookbook assumptions or overly simplified models that grossly estimate aftermarket life and cost. The fact that this doesn’t work is illustrated by $30-40 billion in warranty payments annually in the US alone. Models, whether financial or engineering related, that are largely based on statistics and trending from past products simply don’t work even when supplemented with physical test data. These over-simplifications about aftermarket within PLM today were made because there was no alternative. But that’s no longer the case.
The historic barrier to understanding and predicting aftermarket behavior is now vanishing thanks to the emergence of sophisticated computational material science forecasting technology. The vexing issue of the past has been “variability.” The complexity starts within the material microstructure itself – variability within the sub-elements such as “grains,” their size, arrangement, density and so on, compounded by the variability inherently imperfect manufacturing processing (called “microstructural” variability). Variability is further complicated by manufacturing tolerances – for example, the reality that every drilled hole is not exactly the same size as those before it. Considering that the manufactured fleet encompasses hundreds or thousands of individual units, in the hands of an equal number of different customers, each one using the product in the different way and in a different environment, and you have an incredibly complex world made up of nothing but variability. The number of individual computational simulations required to evaluate all this variability was beyond the capabilities of computational processing just 15 years ago.
As a result, soon the standard PLM offering will be radically different from what is being offered today. Manufacturers will be able to create “digital twins” of their products that represent the actual physics of the real manufactured fleet, from concept to end of use. Future PLM software users will exercise this product simulator initially to optimize the product to be launched into the marketplace before the first one actually comes off the assembly line. But that’s just the beginning. The same simulator will be used, post-manufacture, to forecast fleet failure distribution, spare parts needs, and maintenance logistics requirements. Still further, because the simulator will predict failures in the product fleet, failure costs (such as warranty, maintenance, logistics, etc.) will also be accurately predicted to portray the product lifecycle. And for those of us that are in the PLM software market – this is all good – because the more PLM represents reality, the more it will be used and relied upon by the top executives that run the largest companies in the world. Now, that’s job security!
–by Loren Nasser, President & CEO