Vendor Accent

Quality Driven by CAE

Advanced CAE tools enable companies to design quality into their products, writes Massimo Fariello

Bringing a new product to market is a challenging endeavor, and consistently successful product launches are an indicator of a rare class of companies. Customer-focused product development teams involve cross-functional personnel from a variety of areas to support conceptual design efforts. Teams can include representatives from marketing, R&D, style, design, sales, supply management, production and finance.

Merging knowledge of the company’s products and processes with an understanding of customer needs, competitors and market trends, these teams develop a set of product functionalities and performances. If the “Concept Product” could be released to the market immediately, the chances of its success would be quite high.

The product development process is often complicated, however, and getting from Concept Product to “Real Product” requires time and money. Product development complications often result in the Real Product being close to –but not exactly the same as –the Concept Product. This is because design and manufacturing compromises are made during product development.

Successful product launches can be characterized by three factors—good interaction with users and the consumer market early in the concept phase, a real product that closely matches the Concept Product and an accurate feedback loop from users to the product design team to evaluate customer satisfaction. There is a fourth factor, however, that can lead to sustained enterprise success. That factor is “Quality.”

From concept to reality: the virtual master

The path from a Concept Product to a Real Product that is characterized by a statistical distribution includes two convergent processes: product development and production start-up. Product development that utilizes product lifecycle management (PLM) tools starts with a Concept Product and ends with a complete definition of the “Virtual Master.” The Virtual Master is the complete set of all design variables and their connected performance predictions in a digital environment.

Computer-aided engineering (CAE) tools are used to evaluate product design variables, with a goal of predicting and improving product performance.

Effective use of CAE technology will allow the connection of information between process and performance simulation, leading to a better estimate of the real nominal product performance.

The Virtual Master, in conjunction with CAE simulation, digitally represents the complete product. At this point in the design cycle, the Real Product may be different from the Concept Product, depending on the level of innovation and the compromises made. However, it is still regarded to be within the intent of the product development effort’s original scope. Following organizational approval, significant investment in capital equipment, facilities and tooling is made to support production start-up.

From virtual master to production reality

The reality of production start-up is the introduction of variances and statistical distributions inherent in a Real Product as compared to the perfect, clean and impeccably nominal Virtual Master.

This variability becomes evident as soon as the virtual product meets the “real” world. Every variable that was considered as nominal–thickness, material characteristics, process variables, tools and hundreds of others introduces variability by natural distribution.

To successfully pass quality gates, it is necessary for the product to perform satisfactorily, as related to a series of single variables and their production process distribution. This often results in an expensive convergence process to control input and output variability.

Different efforts to achieve the same quality

Given that the product distribution is a nonlinear function of the input variables’ distribution, the most natural way to improve product quality is to narrow the appropriate input distribution until the output is back within acceptable limits.

Cost-effective Quality is dependent on the methods chosen to pass the required quality gates, and making the convergence effort easier is an effective way to spend less money during product launch.

A key point is that the function is dependent on the product design itself serving as a function of design parameter variability in production.

One way to compensate for an offset mean value is to place tighter controls on the process variability than would normally be required. This can be expensive, however.

In contrast, CAE can be used to evaluate the natural behavior of processes and provide feedback to design prior to production. In this U-channel example, a stamping simulation with the new steel and the existing tool could be performed to discover that the mean value for spring back is now 4.8 degrees.

The Virtual Production

All the methodologies and tools that contribute to define the Virtual Master can be imagined as a connected “simulation engine.” To assess the effects of variability resulting from the real production world, this engine can be run several times on samples obtained through input distributions.

The result is a special kind of simulation – “Virtual Production” – whose main result is to evaluate the connection between input distribution and the quality gates in a pure digital environment.

Now, the design parameter variability function can be identified and is available in the design phase.

By starting from the target Quality, natural variances can be introduced and accounted for. Controllable input distributions can then be set to the minimum level necessary to reach Quality targets.

This efficiently places the Quality effort only where it is really needed and avoids setting impossible targets for the final product. It also avoids the expenditure of inordinate resources to control input distributions that have little effect on the global product.

Back to the Customers

The reliable “performance evaluation engine” methodologies developed by the CAE community over the past few decades can now be linked to Virtual Production, with the goal of assessing the performance of each product coming out of Virtual Production lines.

Having the ability to test every design and process variable choice as it relates to the effect on product performance (including cost) can help organizations anticipate problems, create positive cross-functional discussion, prevent costly manufacturing launch problems and enable product development organizations to focus on developing innovative and market-leading products.

The author is Vice President of Advanced Technologies, Altair