Towards Whole Life-Cycle Management of Prefabricated Buildings based on Knowledge-Data Dual Driven Methods

As an efficient and environmentally friendly construction method, prefabricated buildings have garnered increasing attention compared to traditional buildings in the context of global urbanization and sustainable development. This research addresses the digital transformation challenges in prefabricated buildings by proposing a whole life cycle management framework based on knowledge-data dual-driven methods. First, country-specific, e.g., Australian, prefabricated building standards and construction cases will be collected to build a knowledge foundation model that integrates a knowledge graph with large language models. By aligning formats of building information modelling and digital twins, automated rule checking will be realized during the design phase, alongside safety, schedule, and quality management in construction. Finally, in the operation and maintenance phase, the knowledge base will support conventional deep learning methods in fault detection and diagnosis of photovoltaic systems to enhance the interpretability of the models and provide the repair or maintenance solutions for photovoltaic systems. This PhD research aims to develop a prefabricated building Web-based knowledge base and management framework, establish a standard dataset for knowledge extraction, and ultimately advance the digital transformation and sustainability of prefabricated buildings.