Robust Design Optimization
Variability, uncertainty and tolerance have to be considered for design process of technical systems to assure the highly required quality and reliability. There is a class of variability and uncertainty, which is caused by environment influences (temperature, humidity, day light etc.), load variation (force, moment), human error etc.. They are uncontrollable, unpredictable and cause the uncertainty satisfaction of the required product functionalities. The design goal is assuring of the specified product functionalities in spite of unavoidable variability and uncertainty.
Winning customers and saving the product image, great efforts are done in the industry with extremely high effort and cost. Design of experiment with many prototypes is performed. Cost-intensive product changing during pre-series-production, even in the series-production are frequently the case. The new, innovative and cost-effective approach solving this problem is robust designing product parameters in the early design process. Thereby, optimal product parameters should be found. Within, the system behavior is robust in spite of unavoidable variability. E.g. the consistent variability und uncertainty leads only to the most small variability of the product properties. So, the required product specifications will be always satisfied in spite of variability and uncertainty. This process is so-called robust design optimization. There are different optimization goals:
Reliability based optimization: objective is the failure probability of a calculated distribution of the product properties being minimized
Variance based optimization: objective is the variance of an probability distribution being minimized.
Mean based optimization: objective is the mean value of a probability distribution being minimized or maximized
The great challenge of the robust design optimization is the long computing time of large deterministic product models. The optimization faces the technical feasibility and it is possible only for small product models. In OptiY®, brand-new optimization methods are available, which allow a fast optimization with considerably few number of model calculations. Robust design optimization is therefore also feasible for large product models in OptiY®.
Robust design optimization is also called design for six sigma (DFSS). It deals with an effective quality improvement tool: reduce variability and increase quality. The effort is concentrated to reduce the development cost and to improve the product quality.