Application of Digital Engineering in Project Bumblebee

Overview

Project Bumblebee aimed to develop a dynamic vehicle control system for submarines, ships, aircraft, and other complex platforms. Our objective was to create a system capable of determining proper control inputs (e.g. thrust, fin actuation) based on desired outcomes – effectively solving an inversion problem in control modelling.

Computational Fluid Dynamics (CFD) simulations, though detailed, are expensive and time-intensive – making high-fidelity simulation impractical for real-time or scalable solution development. We collaborated with the National Innovation Centre for Data (NICD), to apply digital engineering principles and techniques to the challenge instead.

Key Innovations

• Data Driven Modelling: Rather than relying solely on physics-based simulations, we focused on building a system model that predicts submarine state transitions based on prior states and inputs – decoupling the system dynamics from the control logic.
• Model-Based Systems Engineering: The team explored system-and-control approaches that offer better control granularity and system understanding than traditional data-driven methods.
• Open-Source Simulation Data: To train and validate models, we used open-source Autonomous Underwater Vehicle simulation datasets, allowing for scalable experimentation without re-simulating via CFD.
• Agile Sprint-Based Delivery: The project adopted an agile structure with weekly sprint skill sessions, rapid feedback loops, and progressive learning and deployment cycles.

Outcomes

The collaboration with NICD and subsequent in-house work yielded:

• A dynamic system model was successfully developed, capable of predicting control actions from state transitions – laying a solid foundation for further control model development.
• We gained enhanced data science capability, embedding modelling best practices and engineering rigour within the team.
• Ongoing validation includes testing against CFD simulations to further refine the model's fidelity and applicability.
• Critically, the initiative enabled us to advance our first demonstrator product and generated valuable Intellectual Property – positioning future control-system solutions for market readiness.

Digital Engineering Techniques Employed

• Model-Based Systems Engineering (MBSE) – structuring system and control models with clarity and traceability
• Data-Driven Modelling – using Autonomous Underwater Vehicle (AUV) data to train dynamic models, bypassing reliance on expensive simulations
• Agile and Sprint Framework – driving iterative progress with rapid feedback and knowledge transfer.
• Software Engineering Rigour – improving code quality, reproducibility, and dependency handling.
• CAD and Simulation Environment – supporting model development within robust digital frameworks

Strategic Impact

Project Bumblebee not only resolved an engineering hurdle – it catalysed a cultural and capability shift within Delkia. The team is now better equipped to take on similar digital engineering challenges, remaining open to future expert collaborations. The outcome provides a blueprint for embedding digital engineering in innovation: quick iteration, cross-disciplinary collaboration, and data-driven modelling.

Conclusion

Delkia's Project Bumblebee is a prime example of how digital engineering – when combined with agile delivery, data science, and model-based techniques – can transform complex system development. Despite initial obstacles, the project delivered a practical, accurate, dynamic model, enriched internal capability and paved the way for future product innovation.