Sustainability for future aviation is exceptionally challenging requiring development of disruptive technologies for a highly safety critical application.  This 3-year PhD project will develop a validated methodology for accelerating early stage design of integrated electrical propulsion systems for future aircraft which meet performance, functionality and regulatory safety requirements.  The project addresses the research challenge for a design methodology for disruptive, safety critical systems when starting from scratch.  This will be achieved by application of artificial intelligence (AI) methods to translate systems requirements from complex, interdependent and high numbers of regulatory documents and subsequently integrate these with a systems design methodology. This will enable confident, early stage selection of candidate solutions which are capable of pull through to high systems readiness and technology readiness levels. This research directly supports acceleration of certifiable electrical propulsion systems design, thus significantly contributing to achieving tight timeframes for decarbonisation of aviation.

The challenge of certification of disruptive technologies for aviation is of concern to the industry. Meeting regulatory requirements is significantly challenged by the complexity and volume of regulatory documentation.  Further challenges are introduced by human subjectivity in identifying and interpreting regulatory documentation.  Development of methods to efficiently assess the regulatory documentation, extract system requirements for solutions which are compliant, and combine with performance and functionality requirements are needed.   The disruptive nature of next generation aircraft technologies results in gaps in the regulatory documentation for these future aircraft.  Methods to conduct gap analysis of regulatory documentation are presented in [1], but does not pull through to translation and integration with systems design.

There is significant opportunity to streamline and automate the assessment of regulatory documentation using machine learning methods already developed for information retrieval and natural language processing. Large Language Models coupled with Retrieval Augmented Generation (RAG-LLM) offer the possibility of producing domain specific semantic models which can then be used to classify, query and segment regulatory document corpora, reducing the burden on compliance engineers in finding appropriate regulation and framing compliance appropriately.  Methodologies to extract systems requirements and integrate with systems level design is needed.

This PhD will address the research opportunity to: first develop a methodology to extract and pull through requirements for compliant electrical propulsion systems design; Second use this design methodology to assess capability of proposed solutions for pull through to high TRL/SRL; third apply the design method to propose options to overcome regulatory documentation gaps.  The project will be collaborative with Vertical Aerospace (VA), who are designing, building and testing new electrical vertical take-off and land (eVTOL) aircraft.  VA will provide real-world case studies and experience of the certification process for disruptive technologies.  Hence the project will provide strong academic research contributions, and new designs and methodologies to reduce time and cost associated with new technology development for industry.