Introduction to the DLR Quantum Computing Initiative – Bringing quantum computing into applications with an example from the field of computational fluid dynamics
Quantum computing has the potential to become one of the key technologies of the 21st century. In order to achieve this, on the one hand, it is necessary to tackle the challenges that exist in the area of hardware in the construction of industrial-grade quantum computers, but on the other hand, questions about how exactly the possibilities of a quantum computer can be used to reach the expected acceleration also play a major role. In order to answer both questions together, DLR has founded the Quantum Computing Initiative to research the possibilities of this new technology in a network of external partners and internal projects. In addition to the presentation of the DLR quantum computing initiative, another focus of the talk will be to present methods for accelerating computationally expensive simulations, e.g. in the field of flow simulation, with the help of quantum computers.
Quantum computing has the potential to become one of the key technologies of the 21st century. In order to achieve this, on the one hand, it is necessary to tackle the challenges that exist in the area of hardware in the construction of industrial-grade quantum computers, but on the other hand, questions about how exactly the possibilities of a quantum computer can be used to reach the expected acceleration also play a major role. In order to answer both questions together, DLR has founded the Quantum Computing Initiative to research the possibilities of this new technology in a network of external partners and internal projects. In addition to the presentation of the DLR quantum computing initiative, another focus of the talk will be to present methods for accelerating computationally expensive simulations, e.g. in the field of flow simulation, with the help of quantum computers.
Accurately simulating turbulent, compressible fluids remains a major computational challenge for the aviation industry and a key to achieving future climate goals. Despite advances in computational fluid dynamics (CFD), limitations in turbulence models and the immense cost of high-fidelity simulations make scale-resolving full-aircraft simulations unrealistic. Quantum computing could overcome these barriers, as even few qubits can represent vast degrees of freedom. However, due to numerous obstacles such as the restriction to unitary operators, the read-out problem, and the limited availability of actual quantum hardware, the development of efficient algorithms for solving fluid mechanics equations is a real scientific challenge. This presentation highlights current efforts within the DLR QCI project ToQuaFlics to explore the potential of quantum computing for aerodynamic simulations.
High-performance computing (HPC) refers to the use of advanced computational models and technologies to significantly enhance the performance of processing highly complex problems. HPC has evolved dramatically since the first supercomputer was commissioned in the 1960s. Nowadays, HPC is characterized by the pursuit of exascale computing, i.e., the ability to perform 1018 calculations per second. In November 2024, Lawrence Livermore National Laboratory, California announced the new El Capitan System as the most powerful exascale supercomputer, capable of performing over 1.742 exaFlops. Over the past few years, HPC platforms have evolved from CPU-centric architecture to GPU-accelerated systems. Such computing capability empowers researchers to tackle a diverse array of scientific applications across numerous domains, including stock market prediction, task scheduling, weather forecasting, gene sequencing, drug discovery, and Computational Fluid Dynamics (CFD) (Netto et al., 2018; Jamshed, 2015). CFD is an applied science field that is concerned with the simulation and analysis of the physics of fluid flow using computational models and technologies. CFD finds extensive applications in diverse industries and scientific disciplines, including aircraft design optimization to facilitate the analysis of aerodynamic performance, drag reduction, and fuel efficiency enhancement (Moin and Apte, 2006). In addition, CFD plays a crucial role in oil reservoir simulation and recovery (Britto et al. 2020). Overall, CFD serves as a powerful tool for understanding and predicting fluid flow phenomena, enabling engineers and scientists to make informed decisions, optimize designs, and improve the performance of various systems and processes. Recently, Quantum Computing has emerged as a new computing approach that leverages quantum physics phenomena such as superposition and entanglement to provide computational speedup over classical computers. With the rapid progress of quantum computing, it is projected that future HPC platforms are venturing into Quantum Processing Unit (QPU)-based systems to tackle previously unsolvable problems. Several researchers have studied the potential use of quantum computing in solving complex CFD problems (Gaitan, 2020; Jaksch et al. 2023). Quantum computing can offer several advantages in CFD applications. First, quantum computing can address the exponential growth of computational complexity associated with high-resolution and turbulent flow simulations. Accordingly, more accurate and detailed simulations can be performed. For example, Berry et al. (2014) presented quantum algorithms that can effectively handle large-scale linear algebraic calculations, such as solving sparse matrix equations and eigenvalue problems, which are fundamental in CFD. Furthermore, quantum computers have the potential to accelerate optimization algorithms that facilitate faster design and optimization of fluid systems (Preskill, 2018). While quantum computing is still in its nascent stage, ongoing research in quantum hardware and algorithms demonstrates huge potential for future applications in solving complex CFD problems.
The computational fluid dynamics is a modern study that utilize computational methods such as the direct numerical simulation, and analyze or simulate the interested fluid mechanical models such as air foils and the turbulence model. This paper reports the acquired knowledge of a literature research and introduces the computational fluid dynamics andalgorithms of quantum computation as well as discusses examples of applications of the quantumcomputationalgorithms. Quantum computation or quantum information is an area that utilizes the properties of quanta and trytobuildvarious quantum circuits or even quantum computers. On the one hand, it has an analogy to the classical computation; onthe other hand, it also gains uniqueness and superiority at problem solving. Compared with the classical computingmethods, the quantum algorithms can be quite faster at computing with higher accuracy as well as security. Due tothedemand of improvement with respect to the computational methods in this area and the recent development in quantumcomputation field, there exists an opportunity to closely relate these two areas. By constructing new quantumalgorithms, the computing methods can be expected to be exponentially faster than classical computing methods. Aproposal israised to consider the features of the quanta and try to apply the advantages towards the computational fluid dynamics.
Abstract. This document is meant to review and discuss the possible applications of Quantum computing in the area of computational fluid dynamics (CFD). A review of the current state-of-the-art of quantum computing applied to computational fluid dynamics has been carried out, highlining how the technology is promising but still in an early stage of development. Furthermore, within the approaches developed to solve CFD problems with the use of quantum algorithms and / or quantum computers, this article discusses a quantum algorithm approach, based on the Lattice Boltzmann Method and developed to the study of 2D flow around a cylinder, a model which can be related to several industrial problems and, in the future, modified to simulate the refrigeration cycle used in aeronautical environmental control systems (ECS). This preliminary code helped to highlight the inherent difficulties to implement a quantum algorithm but helped also to demonstrate the applicability of quantum computing.
This chapter describes results of a recent investigation aiming to assess the potential of quantum computing and suitably designed algorithms for future computational fluid dynamics applications.For quantum computers becoming available in the near future, it can be expected that applications of quantum computing follow the quantum coprocessor model, where selected parts of the computational task for which efficient quantum algorithms exist are executed on the quantum hardware.For example, in computational fluid dynamics algorithm, this hybrid quantum/classical approach is discussed, and in particular it is shown how the approximate quantum Fourier transform (AQFT) can be used in the Poisson solvers of the considered method for the incompressible-flow Navier-Stokes equations.The analysis shows that despite the inevitable errors introduced by applying AQFT, the method produces meaningful results for three-dimensional example problems.A second example of a quantum algorithm for flow simulations is then described.This method based on kinetic modeling of the flow was developed to reduce the information transfer between quantum and classical hardware in the quantum coprocessor model.It is shown that this quantum algorithm can be executed fully on quantum hardware during a simulation.The conclusion summarizes further challenges for algorithm developments and future work.
The accurate simulation of turbulent compressible fluids is a major computational challenge for the aviation industry and possibly a key technology for improving the efficiency of future aircraft so that the ambitious climate targets can be achieved. Even though great progress has been made in the field of CFD in recent decades, and simulation software is available to a wide range of users from both commercial providers and research institutes, it still has two major shortcomings: a) The accuracy of the results is usually limited by the quality of the turbulence models. Even very advanced turbulence models such as Reynolds-Stress models are not able to accurately reproduce a wide range of flows, in particular at the border of the flight envelope. For unseen test cases involving complex physical phenomena or unconventional aircraft configurations, selecting an appropriate turbulence model is often challenging and not straightforward. b) In order to overcome the shortcomings resulting from turbulence modeling, scale-resolving methods such as large eddy simulations or even direct numerical simulations are increasingly being carried out. However, the cost of these simulations for flight Reynolds numbers is so high that even under very optimistic assumptions about the development of future HPC resources and advances in algorithms, full-aircraft simulations seem unrealistic. Even though the possibilities for the use of quantum computers in CFD are still limited due to their current stage of development, they may offer the potential to accelerate calculations in such a way that scale-resolving methods for complete aircraft seem possible. Since algorithms that can be efficiently implemented on quantum computers for solving compressible nonlinear transport equations are also largely not yet available, the time is now to deal with them.