GPU Accelerated Grid Optimisation

This project involves developing a new solver that leverages Graphics Processing Unit (GPU) acceleration to improve speed, accuracy, and scalability, addressing existing performance issues in parallel processing. The project includes research and development, implementation, thorough testing, and proof of concept validation. The learning from this project will be disseminated through workshops, conferences, public announcements, detailed reports, and publishing the output as part of HIGHS open-source solver which is developed and maintained by University of Edinburgh. 

Benefits

Optimisation modules are becoming increasingly critical for enabling net-zero operation of the GB grid. This transition will require solving large-scale optimisation problems, such as whole energy system optimisation, which current CPU-based solvers might struggle to handle efficiently within a reasonable timeframe due to limited parallel processing capabilities.  
Recent advancements in GPU technology offer a promising alternative. By exploring GPU-based optimisation, the project aims to significantly improve the speed and scalability of tools and processes, making them more effective and better suited to future system needs. 
These improvements are crucial for energy systems planning and operation, benefiting all related activities. Moreover, integrating advanced analytical tools would enable more effective and economical electricity system planning. 

ENA smarter networks portal