Identification of architectural/multi-disciplinary design patterns for hyperspectral instruments

Commercial and institutional markets place high pressure (time and cost constraints) on companies designing and developing satellites and payloads. Those markets reward companies able to either inject disruptive solutions or identify reusable patterns (e.g. from past projects) and implement incremental improvement by exploiting heritage and available processes. Patterns usually can be identified in different domains of the realization of a space system: design, development, and verification. Yet it often is not trivial at all to identify patterns because of the lack of a taxonomy and ontology, as well as because of the complexity of the working framework of a project. Under such boundary conditions, the design and management of interfaces emerge as critical aspects of system engineering, given the complexities in system interactions. Recognizing the imperative to enhance project efficiency and limit escalating costs and development time due to project complexity, the reuse of architectural and design patterns from prior projects becomes a viable strategy, especially whilst dealing with interfaces. This work addresses the necessity for a systematic methodology to define, track and manage system patterns in space projects of a payload system, primarily focusing on interface patterns but with features applicable to general system engineering disciplines. The devised framework consists of methods and tools, comprising three key phases: general pattern mining, cataloguing of discovered patterns, and subsequent assessment of re-usability of interface patterns. The pattern mining phase employs databases, tables, matrices, technical and management documents, surveys, and interviews. Pattern cataloguing exploits the identification of recurring problems with patterns (and their description) as solutions, followed by classification based on purpose, system engineering discipline, and level of application. A connectivity map depicting interrelationships between patterns facilitates their application in different projects. The assessment phase aims at evaluating the re-usability of a pattern for a new project. Due to the vastness of disciplines, this thesis focuses on electro-functional (interface-related) patterns, evaluating them based on their purpose and applicability. Some evaluation tools, such as Excel and CAMEO Systems Modeler, are explored. Excel proves beneficial for defining physical connections and links among systems using dependency structure matrices (DSM), while CAMEO is promising for modelling the logic and behaviour of interfaces. This methodology is validated through two case studies involving electro-functional interfaces in two hyperspectral instruments for Earth Observation. The thesis concludes with a focus on improving the essential instruments for pattern mining and collecting necessary project information. Recommendations include modifications to the standard interface dependency matrix and an evaluation of the application of model-based system engineering (MBSE) to decompose physical, functional, and logical architecture, facilitating pattern identification.