| Nearly eighty percent of the product costs and quality problems are directly influenced by the decisions made early in the development cycle. Therefore, it is important to identify them as early, and accurately, as possible to avoid the expensive design rework including any potential warranty claims due to product failures in the field. Among others, the poor product architecture selection decision predominantly drives these costs and quality problems. It also influences the decisions in the domains of product, process, and the supply chain. Individual product architecture characteristics such as degree of commonality, nature of interactions, and interfaces between components may constrain strategic decisions like postponement and late customization. Poor product architecture raises costs with respect to manufacturing process, launching difficulties and other overhead costs. It is argued that such types of complex situations can be simplified with the modular architecture approach.; While researchers agree on advantages of product modularity, the modularization process still lacks a science based approach, especially in engineering design. Much of the previous product modularization methods are heuristics based. There are very few studies that provide a quantitative framework for product modularization. Even the few researchers who have studied the formal approach for optimization of modular product architecture have focused only on single-objective optimization, such as similarity index or costs of modularization. What is missing is the development of a quantitative framework that can optimally integrate the inherent effects of product architecture on product performance such as quality, reliability and so forth. In other words, how can we integrate the best practices such as "Design for X (DFX)" in modular design early in the conceptual stage so as to minimize the complexity and other problems in the downstream product development processes?; This research presents a framework for developing modular product architecture. It integrates DFX principles and practices in modular design during concept development phase. The methodology identifies optimal modules considering product quality, reliability, manufacturability and the modularization costs. The framework also establishes a structured process to analyze fuzzy product knowledge at an early stage of development hence enables design engineers to make decisions based on well-defined and measured objectives. This dissertation advances the current modularization methodologies by providing a scientific method to characterize the candidate modules and optimize them with respect to multiple design objectives. It provides a generic and repeatable modularization framework that can incorporate more than the above-mentioned four design objectives. The methodology is demonstrated through two case studies including an automotive climate control system. |
