Modular product platform design

Abstract

Modular product platforms, sets of common modules that are shared among a product family, can bring cost savings and enable introduction of multiple product variants quicker than without platforms. This thesis describes the current state of modular platform design and identifies gaps in the current state. The gaps were identified through application of three existing methods and by testing their usability and reliability on engineers and engineering students. Existing platform or modular design methods either are meant for (a) single products, (b) identify only module "cores" leaving the final module boundary definition to the designer, and (c) use only a limited set of evaluation criteria.

I introduce a clustering algorithm for common module identification that takes into account possible degrees of commonality. This new algorithm can be applied both at physical and functional domains and at any, and even mixed, levels of hierarchy. Furthermore, the algorithm is not limited to a single measure for commonality analysis.

To select the candidate modules for the algorithm, a key discriminator is how difficult the interfaces become. I developed an interface complexity metric based on minimizing redesign in case of a design change. The metric is based on multiple expert interviews during two case studies. The new approach was to look at the interface complexity as described by the material, energy, and information flows flowing through the interface.

Finally, I introduce a multi criteria platform scorecard for improved evaluation of modular platforms. It helps a company focus on their strategy and benchmark one's own platform to the competitors'.

These tools add to the modular platform development process by filling in the gaps identified. The tools are described in the context of the entire platform design process, and the validity of the methods and applicability to platform design is shown through industrial case studies and examples.