Rolls-Royce and Classiq have achieved a more than ten-fold reduction in quantum resource requirements for a complex engineering simulation, bringing practical quantum computational fluid dynamics (CFD) closer to reality. The collaboration focused on a publicly available Rolls-Royce application simulating transonic flow with shocks through a one-dimensional nozzle, a demanding task typically requiring significant high-performance computing power. Researchers successfully integrated a quantum linear solver into an existing classical CFD workflow, demonstrating that the simulation could still converge even with an approximate quantum component. “Quantum computing matters to enterprises if it can fit into the workflows that engineers and researchers already use,” said Nir Minerbi, co-founder and CEO of Classiq. This work, detailed in a new technical blog, suggests future quantum applications may not demand perfect quantum subroutines, opening the door to near-term applications in aerospace, energy, and beyond.
Hybrid Classical-Quantum Workflow for CFD Simulation
The companies’ recent work, detailed in a technical blog, explores integrating quantum computing into existing, demanding engineering workflows rather than treating it as a standalone solution. CFD, vital for designing systems from aircraft to turbines, traditionally requires substantial high-performance computing power; this hybrid approach seeks to alleviate that burden by offloading specific calculations to a quantum processor. Rather than replacing the entire CFD process, the team tested a quantum linear solver as a component within an established classical workflow, allowing the classical system to manage the overall simulation. Crucially, the research revealed that the workflow could still achieve convergence, meaning it could arrive at a stable solution, even when employing an approximate quantum solver. This challenges the conventional assumption that quantum components must be flawless to contribute meaningfully to engineering simulations.
The team achieved the resource reduction by utilizing a Chebyshev linear combination of unitaries (Cheb-LCU) approach, which significantly lowered the demands on the quantum hardware compared to a Quantum Singular Value Transformation-based solver, while still maintaining the simulation’s ability to converge. This focus on workflow integration, rather than isolated algorithm performance, is a key finding of the research. Classiq’s high-level quantum software platform facilitated the development and implementation of the quantum portion of the workflow, and the resulting quantum linear solver implementation is available in an open library to encourage further investigation. The work highlights the potential for near-term quantum applications that prioritize practicality and resource efficiency over absolute precision, potentially allowing fault-tolerant quantum computers to address real-world engineering challenges.
Quantum computing matters to enterprises if it can fit into the workflows that engineers and researchers already use.
