betaFeedback

WorksMathematics

Robust series linearization of nonlinear advection-diffusion equations

Abstract

We consider nonlinear partial differential equations (PDEs) for advection-diffusion processes which are augmented by an auxiliary parameter δδ such that δ=0δ=0 corresponds to linear advection-diffusion. We derive potentially non-perturbative series expansions in δδ that provide a process to obtain the solution of the nonlinear PDE through solving a hierarchical system of linear, forced PDEs with the forcing terms dependent on solutions at lower orders in the hierarchy. We rigorously detail our approach for a particular deformation that interpolates between linear advection-diffusion and the canonical Burgers' equation modeling nonlinear advection. In this case, we prove that the series has infinite radius of convergence for arbitrary integrable initial data, analyze the cases of a Dirac-delta initial condition (IC) (i.e., the fundamental solution) in an infinite domain and arbitrary IC in a periodic domain, and demonstrate the approach to turbulent behavior in a scenario with periodic forcing. We then treat models of nonlinear diffusion involving the pp-Laplacian operator, including generalizations of the Poisson equation in 11 and 22 dimensions, and the heat equation in 1+11+1 dimensions. We detail series expansions for two different deformations of these equations about their linear (ordinary Laplacian) counterparts, providing numerical evidence for the convergence of the series outside of a perturbative regime and demonstrating that the rate and radius of convergence are affected by choice of deformation. Our results provide a rigorous foundation for using series expansion techniques to study nonlinear advection-diffusion PDEs, opening new pathways for analysis and potential applications for quantum-assisted computational fluid dynamics.

Introduction

Advection-diffusion equations, in which the concentration of a conserved quantity (e.g., particles or energy) is moved throughout a domain through transport (advection) and diffusion, occur ubiquitously in the physical sciences and contain many paradigmatic examples of parabolic partial differential equations (PDEs) [1]. In a general sense, we can parameterize advection-diffusion equations in dd dimensions as the following equation for a density u:(0,∞)×Rd→Ru:(0,\infty)\times\mathbb{R}^{d}\to\mathbb{R}

∂tu=div⁡(j(u,∇u)) ,(1)\partial_t u = \operatorname{div}(\mathbf{j}(u,\nabla u)) \, , \tag*{(1)}

together with initial data u(0,x)=g(x)u(0,x)=g(x). We can interpret j\mathbf{j} as a flux of the conserved quantity, and several well-known instances of advection-diffusion follow from this interpretation. For example, if we consider the total flux to be the sum of an advective flux ja=vu\mathbf{j}_{a}=\mathbf{v}u with velocity v\mathbf{v} and a diffusive flux given by Fick’s law jd=−ν∇u\mathbf{j}_{d}=-\nu\nabla u with ν\nu the diffusivity (taken to be scalar for simplicity), we find the linear advection-diffusion equation

∂tu=v⋅∇u−νΔu ,(2)\partial_t u = \mathbf{v}\cdot\nabla u-\nu\Delta u \, , \tag*{(2)}

where Δ=∑j∂j2\Delta=\sum_j\partial_j^2 denotes the Laplace operator on Rd\mathbb{R}^{d}. The linearity of this equation as a PDE is inherited from the total flux j=ja+jd\mathbf{j}=\mathbf{j}_{a}+\mathbf{j}_{d} being a linear function of uu and its gradient, but nonlinear generalizations are possible. Specializing to 1 dimension and considering the case in which the advective velocity is proportional to the concentration, ja=12u2\mathbf{j}_{a}=\frac{1}{2}u^2, we obtain the canonical Burgers’ equation for nonlinear advection

∂tu=u∇u−νΔu .(3)\partial_t u = u\nabla u-\nu\Delta u \, . \tag*{(3)}

In a similar vein, if we consider a case in which the diffusivity depends on the concentration gradient to some power, we obtain the pp-Laplacian evolution equation

∂tu=Δp(u)(4)\partial_t u=\Delta_p(u) \tag*{(4)}

in which

Δp(u)=div⁡(∣∇u∣p−2∇u)(5)\Delta_p(u)=\operatorname{div}(|\nabla u|^{p-2}\nabla u) \tag*{(5)}

is the pp-Laplacian operator, defined for p∈(1,∞)p\in(1,\infty), and ∣⋅∣|\cdot| denotes the standard Euclidean norm of a vector in Rd\mathbb{R}^{d}. When p=2p=2, the pp-Laplacian is just the regular Laplace operator and (4) reduces to the heat equation for linear diffusion.

Nonlinear advection occurs even for simple, Newtonian, incompressible fluids through the celebrated Navier-Stokes (NS) equations. Burgers’ equation Eq. (3) arises as a simplification of NS through ignoring the pressure term, which additionally enables the equation to be consistently studied in one spatial dimension, in contrast to full NS. Nonlinear diffusion also arises in fluid dynamics through large-eddy simulation (LES) models [2], in which equations of fluid mechanics (e.g., NS) are filtered through a linear operator u→uˉu\to\bar{u} which commutes with constant-coefficient differential operators. Qualitatively, the action of the filtering operator is to separate out the large-distance and small-distance scales, and so should produce a set of equations for the behavior at large scales with a source term feeding in the effects of smaller scales, perhaps phenomenologically. Applying the filtering operator to the NS equations produces a system of equations that looks again like NS in uˉ\bar{u}, but with an additional diffusion term represented by the divergence of the so-called subgrid-scale tensor T(∇u)=∇u⊗∇u‾−∇u‾⊗∇u‾T(\nabla u)=\overline{\nabla u\otimes\nabla u}-\overline{\nabla u}\otimes\overline{\nabla u}. Ideally, the subgrid-scale tensor should be expressible as a function of uˉ\bar{u} alone (i.e., make no reference to uu) in order to simplify the problem, but such an exact closure does not actually reduce the number of degrees of freedom, a situation called the closure paradox [2]. Hence, practical LES approaches look for accurate models for T(∇u‾)T\left(\overline{\nabla u}\right) that are not expected to be exact but retain nice properties (e.g., that the filtered equations are provably well-posed). The LES model of Smagorinsky [3] amounts to taking

Tϵ(∇u)=−ϵ2∣D∣D,(6)T_{\epsilon}(\nabla u)=-\epsilon^{2}|D|D, \tag*{(6)}
D=[∇u+(∇u)T]/2,(7)D=\left[\nabla u+(\nabla u)^{T}\right]/2, \tag*{(7)}

with ϵ>0\epsilon>0 chosen to reproduce the physics of the underlying microscopic equations, such as the k−5/3k^{-5/3} scaling of energy-versus-wavenumber kk in NS. The Smagorinsky model falls into the more general class of Ladyzenskaja models whose subgrid-scale diffusion tensor takes the form T(∇u)=β(∣∇u∣2)∇uT\left(\nabla u\right)=\beta\left(|\nabla u|^{2}\right)\nabla u where β:[0,∞)→[0,∞)\beta:[0,\infty)\to[0,\infty) is monotone increasing and c1xμ≤β(x)≤c2xμc_{1}x^{\mu}\leq\beta(x)\leq c_{2}x^{\mu} for μ≥1/4\mu\geq1/4. Ladyzenskaja models enjoy the property that they produce a well-posed system of equations that converge to weak solutions of NS in the limit ϵ→0\epsilon\to0. We also see that Ladyzenskaja models with β\beta a pure power law produce the pp-Laplacian as a diffusion operator with p≥2.5p\geq2.5 (Smagorinsky is the case p=3p=3). While these two examples highlight important cases where nonlinear advection-diffusion arises in fluid dynamics, they are by no means exhaustive; Burgers’ equation appears in many other contexts to include condensed matter, nonlinear acoustics, and cosmology [4], and the pp-Laplacian appears in contexts of image processing, optimal transport, sandpile growth, and others [5].

Given the ubiquity of nonlinear advection-diffusion phenomena, there is considerable interest in solving nonlinear PDEs of the form Eq. (1) and a range of methods have been developed. Perhaps the simplest is direct numerical simulation (DNS), in which the solution is obtained in high spatial and temporal resolution without relying on additional models or uncontrolled approximations. Finite difference or finite element methods on fine grids applied directly to the PDE are examples of DNS approaches. However, as is well-known, phenomena such as turbulence introduce couplings between a very large range of scales, challenging the limits and the scalability of DNS applied to such problems. Even in the case of effective nonlinear models which arise from treating the large-scale dynamics at high resolution and phenomenologically treating the fine-scale dynamics (e.g., the LES models described above), the scales involved can be stressing for DNS, and care must be taken that the numerical scheme captures the properties of the solution. A relevant example arises for the finite-difference approximation to the pp-Laplacian; the only known finite-difference approximation to the pp-Laplacian which is monotone [6], i.e. converges to the proper viscosity solution, involves an operator whose support rr scales non-trivially with pp and with the grid spacing hh, requiring large stencils. This can be contrasted with the ordinary Laplacian, where low-order stencils with h∼O(r)h\sim\mathcal{O}(r) are accurate and monotone.

Given the difficulties of DNS applied to PDEs with conventional computing, there has been significant interest in applying alternative computing paradigms such as quantum computing to the solution of PDEs, see Refs. [7, 8, 9, 10, 11] for some recent reviews. A significant consideration here is that quantum computations are expressed in terms of unitary operations and so are inherently linear. A linear PDE can be related to a linear system of equations through discretization as in DNS approaches, and then known quantum algorithms for linear systems such as the Harlow-Hassidim-Lloyd (HHL) algorithm [12, 13] can be applied. Direct application of similar ideas to nonlinear PDEs requires multiple copies of a known fiducial state and incurs an exponential overhead [14], though this overhead can be reduced for certain types of polynomial nonlinearities together with variational means of representing quantum states [9]. Hence, significant effort has been devoted to means of linearizing the PDE in a means amenable to a quantum solution, with Carleman [15] and Koopman-von Neumann [16] linearizations being prominent approaches.

In this work, we take an approach in which the nonlinear PDE is converted into an infinite hierarchy of linear PDEs by introducing a deformation of the equation controlled by a parameter δ\delta such that δ=0\delta=0 is a linear equation and the nonlinear equation of interest is obtained at some value δ∗\delta_{*}, expanding the solution of our equation as a Taylor series in δ\delta about δ=0\delta=0, and then matching coefficients in the series expansion through the PDE order-by-order. As shown below, the resulting PDEs are all linear, and form a hierarchy in which functionals of the solutions from lower order appear as forcing terms in the equations for higher-order terms. This enables the collection of tools developed for linear PDEs to be brought to bear, including analytic techniques, local yet monotone finite difference schemes, and more straightforward adaptation for quantum computing.

The idea of using series expansions applied to nonlinear PDEs is not new; Bender et al. considered series solutions for a particular deformation of Burgers’ equation in Ref. [17] and the related homotopy analysis method [18] has a similar philosophy of using a homotopy to interpolate between a linear PDE and a nonlinear one. Our work expands significantly beyond this body of research, demonstrating that the series approach does not have to be interpreted perturbatively through the explicit identification of a deformation for the viscous Burgers’ equation proven to have infinite radius of convergence, and exploring the convergence behavior of nonlinear models of diffusion involving the pp-Laplacian through nonlinear deformations centered around the Laplace operator. These results provide a rigorous foundation for future studies and applications of series methods across the myriad application areas of nonlinear advection-diffusion equations.

Our work is organized as follows: in Sec. II we consider series methods applied to deformations of Burgers’ equation, which features linear diffusion but nonlinear advection. After presenting a methodology to obtain the hierarchy for a general homotopy connecting Burgers’ to linear advection-diffusion, we study a particular deformation in detail, and show that the associated series has infinite radius of convergence given mild constraints on the IC. We analyze the case of delta function IC to discuss practical matters of series convergence, and present an optimization we call refeeding – wherein a truncated series expansion is summed and used as a new IC for the series expansion technique – that improves the convergence behavior. We use these methodologies in a simulation of Burgers’ equation with forcing on a periodic domain, and see hallmark characteristics of Burgers’ turbulence [19, 20]. In Sec. III we apply similar techniques to nonlinear diffusive models involving the pp-Laplacian, studying both an evolution equation that generalizes the heat equation as well as a Dirichlet problem that generalizes Poisson’s equation. We again present a general method to derive a hierarchy of linear PDEs from a parameterized homotopy before studying two different homotopies – which we call the ordinary and the dual– in detail. Using exactly known solutions for the pp-Laplacian Dirichlet problem in 1D and 2D and the evolution equation in 1D, we study the convergence of these two series expansions for a range of pp and show validity outside of a naive perturbative range p=2+δp=2+\delta, ∣δ∣<1|\delta|<1. Finally, in Sec. IV, we provide a discussion and outlook. Some technical points of our exposition are presented as Appendices.

Notation

For m≥0m\ge0 and p∈[1,∞]p\in[1,\infty], Wm,p(Ω)W^{m,p}(\Omega) denotes the Sobolev space consisting of functions whose (distributional) derivatives up to order m≥0m\ge0 are ppth-power integrable over the (possibly unbounded) domain Ω⊆Rd\Omega\subseteq\mathbb{R}^{d}. When p=∞p=\infty, Wm,∞(Ω)W^{m,\infty}(\Omega) denotes the Sobolev space consisting of functions whose (distributional) derivatives up to order m≥0m\ge0 are essentially bounded on Ω⊆Rd\Omega\subseteq\mathbb{R}^{d}. The space W00,p(Ω)W_{0}^{0,p}(\Omega) coincides with the standard Lebesgue space Lp(Ω)={f:Ω→R:∫Ω∣f∣p<∞}L^{p}(\Omega)=\{f:\Omega\to\mathbb{R}:\int_{\Omega}|f|^{p}<\infty\}. Similarly, L∞(Ω)L^{\infty}(\Omega) denotes the space of essentially bounded functions on Ω\Omega. Most frequently, ∣⋅∣|\cdot| will denote the standard Euclidean norm of a vector in Rd\mathbb{R}^{d}. Occasionally we use ∣⋅∣C|\cdot|_{\mathbb{C}} to indicate the modulus of a complex number.

Nonlinear advection: deformations of Burgers’ equation

We begin with a statement of the most general form of nonlinear advection we consider, referred to as the homotopy Burgers’ equation (HPE). On a possibly unbounded spatial interval I⊂RI\subset\mathbb{R}, the HPE refers to the semilinear Cauchy problem

{∂tu+∂xh(u,δ)−ν∂x2u=fu(0,x)=g(x),(8)\begin{cases} \partial_t u+\partial_x h(u,\delta)-\nu\partial_x^2u=f\\ u(0,x)=g(x), \tag*{(8)} \end{cases}

where u:(0,∞)×I×[0,1]→Ru:(0,\infty)\times I\times[0,1]\to\mathbb{R} is the unknown field considered as a function of (t,x,δ)(t,x,\delta), ν>0\nu>0 is the viscosity, h:R×[0,1]→Rh:\mathbb{R}\times[0,1]\to\mathbb{R} is a smooth function that satisfies h(u,0)=vuh(u,0)=vu, with v∈Rv\in\mathbb{R} a constant, and h(u,1)=u2/2h(u,1)=u^2/2, f:(0,∞)×I→Rf:(0,\infty)\times I\to\mathbb{R} is an external force, and g:I→Rg:I\to\mathbb{R} is the given IC. If II is bounded, then uu may have Dirichlet or periodic boundary conditions (BCs), while if I=RI=\mathbb{R}, then we demand u(t,x;δ)→0u(t,x;\delta)\to0 as x→±∞x\to\pm\infty for any (t,δ)∈(0,∞)×[0,1](t,\delta)\in(0,\infty)\times[0,1] (and likewise for the initial data gg).

The conditions imposed on hh ensure that, as δ\delta varies, Eq. (8) defines a family of nonlinear advection-diffusion equations which “deforms” a linear advection-diffusion equation (δ=0\delta=0) into the classical Burgers’ equation, Eq. (3), (δ=1\delta=1). There are infinitely many such deformations; as an example, Ref. [17] considers h(u,δ)=u1+δ/(1+δ)h(u,\delta)=u^{1+\delta}/(1+\delta) with v=1v=1. The special case h(u,δ)=(1−δ)vu+δu2/2h(u,\delta)=(1-\delta)vu+\delta u^2/2, which we will call the linear homotopy, will be considered in detail momentarily. It is possible (and perhaps desired) to also have hh depend explicitly on the independent variables (t,x)(t,x). This does not appreciably change the following exposition and so we ignore this generalization.

Let u(t,x;δ)u(t,x;\delta) denote the solution to (8). Suppose δ↦u(t,x;δ)\delta\mapsto u(t,x;\delta) is analytic at δ=0\delta=0 for almost every (t,x)∈(0,∞)×I(t,x)\in(0,\infty)\times I. Then, there exists a sequence of functions un:(0,∞)×I→Ru_n:(0,\infty)\times I\to\mathbb{R} such that, for almost every (t,x)(t,x), δ↦u(t,x;δ)\delta\mapsto u(t,x;\delta) may be expanded as a convergent series with some radius of convergence r=r(t,x)>0r=r(t,x)>0:

u(t,x;δ)=∑n≥0δnun(t,x).(9)u(t,x;\delta)=\sum_{n\ge0}\delta^n u_n(t,x). \tag*{(9)}

The functions unu_n are then computed via

un(t,x):=1n!∂δnu(t,x;δ)∣δ=0.(10)u_n(t,x):=\left.\frac{1}{n!}\partial_\delta^n u(t,x;\delta)\right|_{\delta=0}. \tag*{(10)}

By expanding hh in a Taylor series about the point (u0,0)(u_0,0), plugging this and (9) into (8), and matching powers of δ\delta, we may derive a hierarchy of PDE each coefficient function unu_n must satisfy. The first few orders are derived here, with the general case treated in Appendix A.

We will suppress the dependence of uu on xx and tt to keep the notation uncluttered and write u(δ)u(\delta) for u(t,x;δ)u(t,x;\delta). First observe that h(u(δ),δ)∣δ=0=h(u0,0)=vu0h(u(\delta),\delta)|_{\delta=0}=h(u_0,0)=vu_0, which follows from the homotopy property of hh. The chain rule gives that

References

  1. [1]L. C. Evans, Partial differential equations, Vol. 19 (American mathematical society, 2022).
  2. [2]J.-L. Guermond, J. T. Oden, and S. Prudhomme, Journal of Mathematical Fluid Mechanics 6, 194 (2004).
  3. [3]J. Smagorinsky, Monthly weather review 91, 99 (1963).
  4. [4]U. Frisch and J. Bec, in New trends in turbulence Turbulence: nouveaux aspects: 31 July–1 September 2000 (Springer, 2002) pp. 341–383.
  5. [5]L. Diening, P. Lindqvist, and B. Kawohl, Oberwolfach Reports 10, 433 (2013).
  6. [6]F. del Teso and E. Lindgren, Journal of Scientific Computing 90, 67 (2022).
  7. [7]A. M. Dalzell, S. McArdle, M. Berta, P. Bienias, et al., arXiv (2023), 10.48550/arxiv.2310.03011, 2310.03011.
  8. [8]M. E. S. Morales, L. Pira, P. Schleich, K. Koor, et al., arXiv (2024), 10.48550/arxiv.2411.02522, 2411.02522.
  9. [9]D. J. Jaksch, P. Givi, A. J. Daley, and T. Rung, AIAA Journal 61, 1885 (2023), 2209.04915.
  10. [10]S. S. Bharadwaj and K. R. Sreenivasan, Indian Academy of Sciences Conference Series 3 (2020), 10.2991/isascs.03.01.0015.
  11. [11]I. Joseph, Y. Shi, M. Porter, A. Castell, V. Geyko, F. Graziani, S. Libby, and J. DuBois, Physics of Plasmas 30 (2023).
  12. [12]A. W. Harrow, A. Hassidim, and S. Lloyd, Physical Review Letters 103, 150502 (2009), 0811.3171.
  13. [13]H. Krovi, arXiv (2022), 10.48550/arxiv.2202.01054, 2202.01054.
  14. [14]S. K. Leyton and T. J. Osborne, arXiv (2008), 10.48550/arxiv.0812.4423, 0812.4423.
  15. [15]J.-P. Liu, H. Ø. Kolden, H. K. Krovi, N. F. Loureiro, K. Trivisa, and A. M. Childs, Proceedings of the National Academy of Sciences 118, e2026805118 (2021), 2011.03185.
  16. [16]I. Joseph, Phys. Rev. Res. 2, 043102 (2020).
  17. [17]C. M. Bender, S. Boettcher, and K. A. Milton, Journal of mathematical physics 32, 3031 (1991).
  18. [18]S. Liao, Communications in Nonlinear Science and Numerical Simulation 14, 983 (2009).
  19. [19]D.-T. Jeng, The Physics of Fluids 12, 2006 (1969).
  20. [20]J. Bec and K. Khanin, Physics reports 447, 1 (2007).
  21. [21]H. Fujita, J. Fac. Sci. Univ. Tokyo 13, 109 (1966).
  22. [22]P. Guidolin, L. Schütz, J. Ziebell, and J. Zingano, Journal of Mathematical Analysis and Applications 515, 126361 (2022).
  23. [23]S. Bianchini and G. M. Leccese, Journal of Mathematical Analysis and Applications 542, 128761 (2025).
  24. [24]A. Friedman, in Nonlinear Diffusion Equations and Their Equilibrium States I, edited by W.-M. Ni, L. A. Peletier, and J. Serrin (Springer New York, New York, NY, 1988) pp. 301–318.
  25. [25]C. Bandle and H. A. Levine, Differential and Integral Equations 7, 1169 (1994).
  26. [26]C. Bandle and H. Brunner, Journal of Computational and Applied Mathematics 97, 3 (1998).
  27. [27]J. D. Cole, Quarterly of applied mathematics 9, 225 (1951).
  28. [28]E. Hopf, Communications on pure and applied mathematics 3, 201 (1950).
  29. [29]The more general claim that, given f : ℝ → ℂ, |f|c > 0 implies |e^{∂²x}f|c > 0 is false. A counterexample we know of is a superposition of the form f(x) = (c₁e^{iξ₁x} + c₂e^{iξ₂x})e^{-x²/2} for appropriately chosen c₁, c₂, ξ₁, ξ₂ ∈ ℝ. However, this counterexample is not of the form e^{-δ₈R e²/2} T_g(x), and so doesn’t immediately imply that the assertion |(e^{δ₈R e²/2}e^{-δ₈R e²/2}T_g)(x)|c > 0 is false.
  30. [30]Hörmander, L., Grundlehren 257 (1983), https://doi.org/10.1007/b138375.
  31. [31]M. Okamura and T. Kawahara, Journal of the Physical Society of Japan 52, 3800 (1983).
  32. [32]P. Lindqvist, Notes on the stationary p-Laplace equation (Springer, 2019).DOI
  33. [33]E. DiBenedetto and M. A. Herrero, Transactions of the American Mathematical Society 314, 187 (1989).
  34. [34]E. A. Coddington and N. Levinson, Theory of ordinary differential equations (McGraw Hill Book Company, New York, 1955).
  35. [35]G. Aronsson and P. Lindqvist, Journal of Differential Equations 74, 157 (1988).
  36. [36]P. Lindqvist, Journal of Mathematical Analysis and Applications 127, 93 (1987).
  37. [37]J. Kinnunen and M. Parviainen, Advances in Calculus of Variations 3, 29 (2010).
  38. [38]The standard finite-difference scheme operates on the solution pointwise, and so would appear inappropriate to treat PDEs involving the p-Laplacian which require a weak formulation. However, properties of our series methodology make the Dirichlet problem amenable to finite difference approaches provided the grid does not encompass points where a singularity occurs. The zeroth-order equation is a Poisson equation, which will have a smooth solution due to elliptic regularity (assuming a regular enough source term). All higher-order equations take the form ∇ · ∇uₙ = ∇ · Fₙ, which becomes the solution of a linear system Aa = b in the finite-difference formulation. For properly defined Laplacian and derivative operators of the same order and stencil, the effects of non-differentiability are regularized on both sides similar to how regularization occurs in a finite element method (FEM) formulation. A FEM method can also be applied to the Dirichlet problem; however, we have kept the finite difference formulation here due to its simplicity in scaling to the 2D problem.
  39. [39]G. I. Barenblatt, Akad. Nauk SSSR. Prikl. Mat. Meh 16, 79 (1952).
  40. [40]K. ahm Lee, A. Petrosyan, and J. L. Vázquez, Journal of Differential Equations 229, 389 (2006).

Paper details

vlnc:W_g9j2ypvs
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents