Process Integration and Design Optimization - PIDO

Nowadays, companies take advantages of virtual prototyping instead of physical testing because virtual prototyping is usually faster and less expensive. This means that product development is strictly linked with computer aided engineering (CAE) tools (e.g. computer aided design, finite element analysis, computational fluid dynamics).

A multi-disciplinary design process is characterized by subsequent transformations and additions of details. A complete multi-disciplinary design process can be separated into several phases receiving the contributions of many professional staff and different departments trying to achieve conflicting requirements. Each phase can be defined as a single module and each single module can be described as a finite group of tightly coupled relationships. However, most of the times, these phases are run separately from each other; even in the case where the results of one module are to feed another. Moreover, there is often a manual process of extracting the data of interest, and passing the results to the engineers and specialists who subsequently perform the next step of the process.

Companies are now recognizing that all these procedures should be automated in a way which allows all CAE tools to be run without human intervention. This is what a PIDO tries to solve. In few words, a PIDO can be described as a tool that allows the effective management of the design process, performs several simulations to explore the design space, orients the process to the optimum and/or determines its robustness.

Fig. 1 - left: Easy integration between a PIDO and a CAD. Right: Direct integration between modeFRONTIER and ANSYS Workbench

PIDO tools are becoming more and more important when designing a product because they assist engineers in overcoming the common error of giving priority directly to the product, rather than to the entire process. Most companies involved in product development are starting to take advantage of these tools, recognizing considerable time and cost savings. This is the reason why PIDO is gaining in popularity and is predicted to be one of the fastest-growing technologies in design process automation over the next years.

A good PIDO should be a complete tool and should face all the necessary aspects:

1. The formalization and manage-ment of the workflow to treat in a flexible and dynamic way, processes of any complexity grade.
2. A set of design of experiments and statistical tools to identify which parameters contribute most to the global performance.
3. The set of methods, technologies and strategies for multi-objective and multi-disciplinary optimiza-tion.
4. Meta-modeling and response surface methodologies (RSM).
5. Charts for visualizing the results for easy interpretation.
6. The set of the so-called techniques for decision support which allow the orientation and the documentation of choices. From all of the optimal options, the design team must formulate a number of plausible designs.
7. Advanced probability and statistical methods for reliability and uncertainties (design for six sigma). 

Fig. 2 - A PIDO tool such as modeFRONTIER can be coupled with queuing systems and a mixed architecture computer pool can be used to run in parallel several analyses and exploit distributed computing resources

The above list shows that a PIDO is not only a tool for CAE users, it can also be very useful for optimization experts, or managers who will be able to make decisions by exploiting detailed information coming from engineers through the integrated design process.

A commercially available PIDO tool is modeFRONTIER. A complete tool such as modeFRONTIER can eliminate much of the down time primarily due to its complete running cycle which encompasses the analysis, waiting for and checking the results, as well as re-executing the manually improved design.

For any questions on this article or to request further examples or information, please email the author:
Silvia Poles - Optimization Consulting
EnginSoft
info@enginsoft.it