On integrated modularization in heavy-duty truck architecting

Abstract

Road transports face increasing challenges with respect to safety, legislations on lower emissions and traffic congestion, as well as numerous business challenges related to paradigm shifts in technology, tightened delivery times and cost constraints. Combination of truck electrification and automation may be utilized to address some of these issues. Electrified and autonomous transport vehicles may be characterized as Cyber-Physical Systems (CPS). A drawback with CPS is the extensive increase of technical complexity, which introduce new challenges to Systems Engineering (SE). The added complexity is preferably targeted in the product architecting development stage of SE. 

Product architecting involves conceptual system design, module identification (clustering) and product layout design. A product architecture is the interrelation between physical components and their function, i.e. their purpose. Product architectures can be categorized as being modular, hybrid or integral. A modular architecture is a strategic means to deliver external variety and internal commonality. Modular subsystems enable concurrent development and modularization is, thus, a structured method to manage technical complexity. In this thesis, a new clustering-based methodology and process for heavy-duty truck modularization that integrates technical complexity, company business strategies and physical interference is proposed.  The main hypothesis behind the  presented research is that computer-based product architecture clustering analysis benefit from a quantitative complexity measure, as well as means to represent (model) and communicate product architecture related complexity. A variety of industrial cases of heavy-duty truck subsystems are used to describe the proposed methodology and to verify its performance, i.e. how well the proposed methodology and process supports the SE process. All investigated subsystems  contains synergistic configurations of mechanical, electrical and software technologies, i.e. they may be characterized as CPS. 

The presented research concludes that the proposed modularization methodology and process is capable of supporting the SE process by improving the quality of the module identification stage, by adding business strategies and physical interference to product architecture clustering. Moreover, it is confirmed that the new methodology is both scalable and flexible, allowing the consequences of different architectural trade-offs to be analyzed independently or combined depending on purpose. Furthermore, the newly developed architectural representations showed to make architectural discussions in general and modularity discussion in particular with and between domain experts efficient. Finally, the case studies clearly shows that the clustering results depend on the relative weights of the different types of component relations that are represented in the product architecture DSM (Design Structure Matrix). However, the importance of these weights are reduced when multiple business strategic and physical interference constraints are introduced.

urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-291106