# Analysis of Carleman Linearization of Lattice Boltzmann

^{1}

^{2}

^{3}

^{*}

## Abstract

**:**

## 1. Introduction

#### 1.1. Early Attempts for Quantum Simulation of Fluids

#### 1.2. Carleman Linearization

## 2. Lattice Boltzmann

#### Nonlinearity Ratio

## 3. Carleman Linearization for Lattice Boltzmann

#### 3.1. Number of Variables

#### 3.2. Carleman Linearization of Collision Step

#### 3.3. Carleman Linearization of Streaming Step

#### 3.4. Error Bound

## 4. Numerical Results

#### 4.1. Logistic Equation

#### 4.2. D1Q3

## 5. Conclusions

## Author Contributions

## Funding

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## Nomenclature

$\overrightarrow{v}$ | Continuum particle velocity |

${\overrightarrow{c}}_{i}$ | Discrete velocity in the ith direction |

Q | Number of discrete velocities number of modes at each lattice site, indexed by i |

D | Number of dimensions of the lattice |

x | Dimensions of the lattice, indexed by d, independent position vector variable |

N | Number of Carleman variables |

${N}_{{x}_{d}}$ | Number of sites across the dth dimension ${x}_{d}$ of the lattice |

G | Volume of the lattice in the units obtained by the product of the number of sites in each dimension ${\mathsf{\Pi}}_{d=1}^{D}{N}_{{x}_{d}}$ |

${f}_{i}$ | Discrete density distribution weight |

$\mathsf{\Omega}$ | Collision operator defined by $\frac{d\overrightarrow{f}}{dt}=\mathsf{\Omega}\left(\overrightarrow{f}\right)$ |

${w}_{i}$ | Weight of the ith discrete density |

${O}_{c}$ | Truncation order in Carleman linearization |

R | Ratio providing measure of nonlinearity parametrizing the error bound of the Carleman technique |

${F}_{0}$ | Coefficient vector of zero-order terms in a quadratic ODE |

${F}_{1}$ | Coefficient matrix of first-order terms in a quadratic ODE |

${F}_{2}$ | Coefficient matrix of second-order terms in a quadratic ODE |

t | Independent time variable |

$\overrightarrow{u}$ | Flow velocity |

$\rho $ | Local fluid density in lattice units |

c | Lattice speed |

C | Carleman linearization matrix |

$\epsilon $ | Norm of the solution error |

${f}_{C}$ | Approximated solution of the system |

$\Delta t$ | Discrete timestep |

V | Vector of Carleman variables |

$\overrightarrow{e}$ | Lattice vectors |

$Ma$ | Mach number |

K | Scaling factor of the logistic equation |

T | Total integration time |

p | Order of the polynomial describing the driving function $\mathsf{\Omega}$ |

${\overrightarrow{P}}^{\left(k\right)}$ | a vector of polynomial functions of kth order in Carlemann variables |

$Re$ | Reynold’s number |

L | Characteristic length scale defined in units of $\Delta x$ |

U | Characteristic velocity in lattice units |

$\nu $ | Kinematic viscosity in lattice units |

${c}_{s}$ | Speed of sound in lattice units |

## References

- Nitzberg, B. Weather Forecasting Gets Real, Thanks to High-Performance Computing. 2017. Available online: https://www.altair.com/c2r/ws2017/weather-forecasting-gets-real-thanks-high-performance-computing (accessed on 25 November 2021).
- Moss, S. China May Already Have Two Exascale Supercomputers. 2021. Available online: https://www.datacenterdynamics.com/en/news/china-may-already-have-two-exascale-supercomputers/ (accessed on 25 November 2021).
- Benzi, R.; Succi, S.; Vergassola, M. The lattice Boltzmann equation: Theory and applications. Phys. Rep.
**1992**, 222, 145–197. [Google Scholar] [CrossRef] - Succi, S.; Benzi, R. Lattice Boltzmann equation for quantum mechanics. Phys. D Nonlinear Phenom.
**1993**, 69, 327–332. [Google Scholar] [CrossRef] - Yepez, J. Quantum Computation of Fluid Dynamics. In Quantum Computing and Quantum Communications; Williams, C.P., Goos, G., Hartmanis, J., van Leeuwen, J., Eds.; Series Title: Lecture Notes in Computer Science; Springer: Berlin/Heidelberg, Germany, 1999; Volume 1509, pp. 34–60. [Google Scholar] [CrossRef]
- Vahala, G.; Yepez, J.; Vahala, L. Quantum Lattice Gas Algorithm for Quantum Turbulence and Vortex Reconnection in the Gross-Pitaevskii Equation. In Proceedings of the 2008 SPIE Quantum Information and Computation VI, Orlando, FL, USA, 23 April 2008; Volume 6976. [Google Scholar] [CrossRef]
- Yepez, J. Lattice-Gas Quantum Computation. Int. J. Mod. Phys. C
**1998**, 9, 1587–1596. [Google Scholar] [CrossRef] [Green Version] - Yepez, J. An efficient quantum algorithm for the one-dimensional Burgers equation. arXiv
**2002**, arXiv:0210092. [Google Scholar] - Yepez, J. Open quantum system model of the one-dimensional Burgers equation with tunable shear viscosity. Phys. Rev. A
**2006**, 74, 042322. [Google Scholar] [CrossRef] [Green Version] - Boghosian, B.M.; Yepez, J.; Coveney, P.V.; Wager, A. Entropic lattice Boltzmann methods. Proc. R. Soc. London Ser. A Math. Phys. Eng. Sci.
**2001**, 457, 717–766. [Google Scholar] [CrossRef] [Green Version] - Boghosian, B.M.; Taylor, W. Simulating quantum mechanics on a quantum computer. Phys. D Nonlinear Phenom.
**1998**, 120, 30–42. [Google Scholar] [CrossRef] [Green Version] - Steijl, R. Quantum Algorithms for Fluid Simulations. In Advances in Quantum Communication and Information; Bulnes, F.N., Stavrou, V., Morozov, O.V., Bourdine, A., Eds.; IntechOpen: Rijeka, Licko-Senjska, Croatia, 2020. [Google Scholar] [CrossRef] [Green Version]
- Mezzacapo, A.; Sanz, M.; Lamata, L.; Egusquiza, I.L.; Succi, S.; Solano, E. Quantum Simulator for Transport Phenomena in Fluid Flows. Sci. Rep.
**2015**, 5, 13153. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Budinski, L. Quantum algorithm for the advection—Diffusion equation simulated with the lattice Boltzmann method. Quantum Inf. Process.
**2021**, 20, 57. [Google Scholar] [CrossRef] - Lloyd, S.; De Palma, G.; Gokler, C.; Kiani, B.; Liu, Z.W.; Marvian, M.; Tennie, F.; Palmer, T. Quantum algorithm for nonlinear differential equations. arXiv
**2020**, arXiv:2011.06571. [Google Scholar] - Succi, S. Lattice Boltzmann 2038. EPL Europhys. Lett.
**2015**, 109, 50001. [Google Scholar] [CrossRef] - Steeb, W.H. Linearization Procedure and Nonlinear Systems of Differential and Difference Equations. In Nonlinear Phenomena in Chemical Dynamics; Vidal, C., Pacault, A., Eds.; Springer Series in Synergetics; Springer: Berlin/Heidelberg, Germany, 1981; p. 275. [Google Scholar] [CrossRef]
- Forets, M.; Pouly, A. Explicit Error Bounds for Carleman Linearization. arXiv
**2017**, arXiv:1711.02552. [Google Scholar] - Steeb, W.H. Linearization Procedure and Nonlinear Systems of DitJerential and DitJerence Equations. In Nonlinear Phenomena in Chemical Dynamics: Proceedings of an International Conference, Bordeaux, France, 7–11 September 1981; Vidal, C., Pacault, A., Haken, H., Eds.; Springer Series in Synergetics; Springer: Berlin/Heidelberg, Germany, 1981; Volume 12. [Google Scholar] [CrossRef]
- Liu, J.P.; Kolden, H.O.; Krovi, H.K.; Loureiro, N.F.; Trivisa, K.; Childs, A.M. Efficient quantum algorithm for dissipative nonlinear differential equations. arXiv
**2020**, arXiv:2011.03185. [Google Scholar] [CrossRef] [PubMed] - Succi, S. The Lattice Boltzmann Equation: For Complex States of Flowing Matter, 1st ed.; Oxford University Press: Oxford, UK, 2018. [Google Scholar]
- Bill, Y.; Meskas, J. Lattice Boltzmann Method for Fluid Simulations. 1997. Available online: https://www.researchgate.net/publication/265077140_Lattice_Boltzmann_Method_for_Fluid_Simulations/ (accessed on 25 November 2021).
- Randles, A.P.; Kale, V.; Hammond, J.; Gropp, W.; Kaxiras, E. Performance Analysis of the Lattice Boltzmann Model Beyond Navier-Stokes. In Proceedings of the 2013 IEEE 27th International Symposium on Parallel and Distributed Processing, Boston, MA, USA, 20–24 May 2013; IEEE: Cambridge, MA, USA, 2013; pp. 1063–1074. [Google Scholar] [CrossRef] [Green Version]
- He, X.; Luo, L.S. Lattice Boltzmann Model for the Incompressible Navier—Stokes Equation. J. Stat. Phys.
**1997**, 88, 927–944. [Google Scholar] [CrossRef] - Lallemand, P.; Luo, L.S.; Krafczyk, M.; Yong, W.A. The lattice Boltzmann method for nearly incompressible flows. J. Comput. Phys.
**2021**, 431, 109713. [Google Scholar] [CrossRef] - Itani, W.; Mezzacapo, A.; Succi, S. Quantum Carlemann Algorithm for Lattice Boltzmann Fluid Simulation. 2021; In preparation. [Google Scholar]
- Krüger, T.; Kusumaatmaja, H.; Kuzmin, A.; Shardt, O.; Silva, G.; Viggen, E.M. The Lattice Boltzmann Method: Principles and Practice, 1st ed.; Graduate Texts in Physics; Springer International Publishing: Cham, Switzerland, 2017. [Google Scholar] [CrossRef]

**Figure 1.**An illustration of the D1Q3 lattice Boltzmann scheme showing how the collision step is a relaxation of the discrete densities towards their equilibrium values whereas streaming involves assigning their values to neighboring cells in the respective directions.

**Figure 3.**Numbering convention used for the discrete densities of the D1Q3 lattice Boltzmann scheme. In 1D, the lattice vectors correspond to the respective scalars $-1,0,1$.

**Figure 4.**Illustration of how the streaming step preserves uniformity under periodic boundaries only, using D1Q2. The uniform initial flow with periodic boundaries coincides with the error-free case of collision without streaming while errors form at the boundaries when non-periodic boundary conditions are applied. We see that with the same initial conditions and periodic boundary, local and neighboring information of discrete densities are interchangeable. (

**a**) Uniform initial flow field, (

**b**) Periodic boundary conditions, (

**c**) Non-periodic boundary conditions.

**Figure 5.**Illustration of the propagation of Carleman linearization error with timestep in a two-dimensional domain with uniform initial flow and non-periodic boundaries. Lattice cells with no fill have discrete exact discrete densities whereas ones with red fill have discrete densities with error. (

**a**) Initial condition, at $t=0$, (

**b**) After one timestep, at $t=\Delta t$. (

**c**) After two timesteps, at $t=2\Delta t$.

**Figure 6.**The analytical (

**left**) (analytical integration) and numerical (

**right**) (discrete time-stepping) solutions of the Carleman-linearized logistic equation are shown with their corresponding errors (bottom) as a function of time, varying initial conditions and Carleman linearization orders. The predicted time of validity is shown as a vertical asymptote in each plot.

**Figure 7.**The solution of the discrete densities of the fluid in D1Q3 for successive collisions is shown for different $\frac{\Delta t}{\tau}$, for the exact and Carleman-linearized formulations as a function of time and Carleman linearization order. The bottom figures show the normalized errors for each discrete density. Note that the solution is exact beyond the first linearization order.

**Figure 8.**Visualization of the sparsity of the Carleman matrix for the collision term at various orders.

**Table 1.**Example of D1Q3 expanded to a second order truncation in Carleman linearization, with $N=9$ Carleman variables. Note that the dummy variable ${V}_{1}=1$ is defined to simplify the form of Equation (23).

Carleman Variables | Lattice Variables |
---|---|

${V}_{1}$ | 1 |

${V}_{2}$ | ${f}_{1}^{2}$ |

${V}_{3}$ | ${f}_{3}^{2}$ |

${V}_{4}$ | ${f}_{1}{f}_{2}$ |

${V}_{5}$ | ${f}_{1}{f}_{3}$ |

${V}_{6}$ | ${f}_{2}{f}_{3}$ |

${V}_{7}$ | ${f}_{1}$ |

${V}_{8}$ | ${f}_{2}$ |

${V}_{9}$ | ${f}_{3}$ |

**Table 2.**The maximum endtime validity for Carleman linearization of the logistic equation as predicted by Equation (41).

$\mathit{f}(\mathit{t}=0)$ | T |
---|---|

0 | ∞ |

0.1 | 2.40 |

0.2 | 1.79 |

0.3 | 1.47 |

0.4 | 1.25 |

0.5 | 1.10 |

0.6 | 0.98 |

0.7 | 0.89 |

0.8 | 0.81 |

0.9 | 0.75 |

1 | 0.69 |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Itani, W.; Succi, S.
Analysis of Carleman Linearization of Lattice Boltzmann. *Fluids* **2022**, *7*, 24.
https://doi.org/10.3390/fluids7010024

**AMA Style**

Itani W, Succi S.
Analysis of Carleman Linearization of Lattice Boltzmann. *Fluids*. 2022; 7(1):24.
https://doi.org/10.3390/fluids7010024

**Chicago/Turabian Style**

Itani, Wael, and Sauro Succi.
2022. "Analysis of Carleman Linearization of Lattice Boltzmann" *Fluids* 7, no. 1: 24.
https://doi.org/10.3390/fluids7010024