Session: MSNDC-04-01

Paper Number: 148351

148351 - Cost-Driven Product Family Optimization Using Cost of Change 

The cost of reinventing the wheel can be tremendous. Therefore, most products are legacies of older designs. These products are not designed from scratch, termed greenfield design, but are created by improving existing designs conceptually or functionally, termed brownfield design. Brownfield design optimization problems are essentially different from greenfield optimization problems. Firstly, optimization has a definite starting point, which is the existing design. This limits the applicability of some optimization methods for such problems. Secondly, the cost of change (CoC) must be incorporated into the overall cost of product design. Depending on the maturity level of the existing model, additional costs may be incurred to redesign an existing product. CoC contribution to the overall cost depends on how much the new design differs from the existing one. In this study, a simplified CoC formula is introduced and integrated into a cost model as a penalty. The cost model and the CoC formula are parametric and their functions can fit the company cost data. A model-based cost optimization is carried out to demonstrate the effect of CoC on optimization results. The aim is to minimize the production cost of the product family while meeting the performance requirements of individual product variants. Therefore, the cost model is used as the objective function and the performance model as the constraint function. A two-stage nested optimization algorithm is chosen to optimize the commonality between components in the outer loop and the design attributes of the component variants in the inner loop. The approach is demonstrated on a four-frame shedding mechanism of a weaving loom as a numerical example. In this numerical example, product families that are optimized without considering CoC may incur cost overheads after manufacturing. Integrating CoC into the cost model yields more realistic optimization results and suggests a different standardization path.

Presenting Author: Emin Oguz Inci KU Leuven

Presenting Author Biography: EMIN OGUZ INCI is a post-doctoral researcher at Mechanical Engineering Department at KU Leuven. He holds a Ph.D. from KU Leuven, a M.Sc. from Computational Mechanics at TU Munich and an B.Sc. from Mechanical Engineering at METU. His research interests include mechatronic design, co-simulation of cyber-physical systems and design space exploration.

Authors:

Emin Oguz Inci KU Leuven
Pavel Eremeev KU Leuven
Martijn Vermaut KU Leuven
Klaas Gadeyne Flanders Make
Frank Naets KU Leuven