To enable a UK-led R&D programme in Digital Twin technologies spanning sectors such as defence, maritime, aerospace, and healthcare, the customer required a highly specialised computing platform. This system needed to be capable of real-time modelling and analytics, demanding exceptional performance and reliability to support the requirements of the complex AI workloads and simulations being delivered. Due to the performance requirements, the solution needed to deliver superior thermal management while fitting within a 4U rackmount chassis, a compact design for the level of performance required. Energy efficiency and low acoustic emissions were also critical, particularly for prolonged usage in sensitive environments.
In addition to performance and thermal constraints, the platform would need to be maintainable and durable under challenging operational conditions. Rapid deployment was another essential factor, with a firm delivery window of six months. The customer was unable to find any standard commercial system able to meet these stringent demands, lacking the necessary customisation, thermal capabilities, scalability, and adaptability required for high-performance Digital Twin operations.
Drawing on its HPC engineering and cooling expertise, Captec designed and integrated a high-performance computing system for demanding AI and Digital Twin applications. Each server node combined dual AMD EPYC processors and dual NVIDIA L40S GPUs with Direct Liquid Cooling (DLC), enabling sustained operation at up to 2.2kW per node without thermal throttling. This maintained optimal processing performance under intensive workloads while reducing heat-related component degradation.
The integrated cooling approach also reduced operating noise, power consumption and maintenance requirements, supporting a lower total cost of ownership. Designed around a modular, industry-standard architecture, the system simplifies servicing, future upgrades and scalability. Captec’s engineering expertise and UK-based production capabilities supported delivery within the customer’s accelerated programme, providing a reliable foundation for evolving AI and simulation requirements.
Captec delivered a high-performance, direct liquid cooled server that aligns with the customer’s long-term Digital Twin R&D strategy. The system significantly enhances processing capabilities, enabling real-time Digital Twin modelling with greater speed and precision. It offers a quieter, energy-efficient, and more reliable performance profile, translating into reduced maintenance requirements and a lower total cost of ownership over its lifecycle.
The system now forms a foundational component within the UK’s advancement of virtual modelling technologies, reinforcing its leadership in the global Digital Twin ecosystem. The platforms modular design also provides the customer with an ideal foundation for further scalability and integration into broader infrastructure as the research programme evolves.
