The integration of Generative Artificial Intelligence (GenAI) in engineering presents both opportunities and challenges, particularly in complex, safety-critical systems. This thesis explores two key aspects: (1) developing an AI-powered reverse engineering tool for Knowledge-Based Engineering (KBE) applications and (2) designing a communication framework to enhance interdisciplinary collaboration in AI-driven projects. The aerospace engineering component introduces the REProcess tool, leveraging Large Language Models (LLMs) to extract structured process representations from KBE applications, improving traceability and automation. The science communication component investigates collaboration barriers between AI experts and domain specialists, leading to the development of a structured ideation matrix for GenAI-powered automation systems. Together, these contributions demonstrate how Explainable AI principles can facilitate the adoption of GenAI in engineering workflows while bridging communication gaps. This interdisciplinary approach provides practical methodologies to enhance AI-driven engineering processes, ensuring both technical feasibility and effective human-AI collaboration.

Aviation has a significant contribution to climate change, which must be decreased. The A321 APPU could offer a substantial improvement, by introducing a hydrogen-powered Auxiliary Power and Propulsion Unit (APPU). This turboshaft engine is located in the tail cone and powers a boundary layer ingestion propulsor, taking over 10% of the thrust. The Steam Injection and Recovery (SIR) cycle is introduced to improve the efficiency of the APPU. This semi-closed water cycle can reduce fuel consumption and NOx emissions. Both the baseline and SIR APPU are modelled in pyCycle, an open-source gas turbine parametric analysis tool. Water properties and heat exchangers are added so pyCycle can model the SIR cycle. The baseline APPU has a thermal efficiency of 45% and a mass of 502 kg. The SIR cycle can increase the efficiency by 1.54% and decrease the NOx emissions by 33.4%, at the cost of a 15.7% mass increase.