Pith. sign in

REVIEW 4 major objections 6 minor 35 references

New highly efficient and accurate numerical scheme for the Cahn-Hilliard-Brinkman system

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A variable-time-step BDF-SAV scheme is proved unconditionally energy-stable for the Cahn-Hilliard-Brinkman system.

desk verdict Variable-step BDFk R-GSAV for Cahn-Hilliard-Brinkman is a useful, mostly sound contribution, but the manuscript under-specifies the chemical-potential update and has coefficient typos that a revision should fix. read the letter →

arxiv 2506.07128 v1 pith:DPCYUN6U submitted 2025-06-08 math.NA cs.NAphysics.comp-ph

Dawei Chen , Qinzhen Ren , Minghui Li This is my paper
classification math.NAcs.NAphysics.comp-ph MSC 74A5065M1265M70
keywords Cahn-Hilliard-Brinkmansystemgeneralizedscalarauxiliaryvariablerelaxationbackwarddifferentiationformulatimestepsunconditionalenergystabilityadaptivesteppinghybrid-orderalgorithm
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper constructs a family of high-order backward differentiation formula (BDF) schemes with variable time steps for the coupled Cahn-Hilliard-Brinkman (CHB) system of phase-field fluid flow, based on a generalized scalar auxiliary variable (SAV) approach with relaxation. The central theoretical result is that, with the four relaxation choices specified in Theorem 2.1, the auxiliary variable satisfies $r^{n+1} - r^n = -\tau_{n+1} \delta^{n+1} \kappa(\varphi^{n+1},u^{n+1}) \le 0$ for every step, so each scheme is unconditionally energy-stable under arbitrary time-step ratios. On top of this, the paper proposes adaptive time-step algorithms, including a hybrid-order algorithm that uses third-order BDF early and second-order BDF later, and reports large CPU-time savings at comparable or better accuracy than fixed-step alternatives and the prior IMEX-CNA adaptive scheme. A reader would care because unconditional stability under variable time steps makes long-time coarsening and buoyancy-driven flow simulations robust, while the hybrid switch addresses the known tension that high-order BDF schemes capture fast early dynamics but are inefficient or unstable for long runs.

What carries the argument

The load-bearing object is the relaxed generalized scalar auxiliary variable $r(t)=E_1(\varphi(t))$ together with the relaxation correction in (2.4g) and (2.5). The proof mechanism is the case analysis of Theorem 2.1: choosing $\sigma^{n+1}_0$ and $\delta^{n+1}$ according to whether the intermediate value $\tilde{r}^{n+1}$ is above, below, or equal to $E_1(\varphi^{n+1})$ makes the discrete energy identity $r^{n+1} - r^n = -\tau_{n+1}\delta^{n+1}\kappa(\varphi^{n+1},u^{n+1})$ hold, where $\kappa(\varphi,u)$ is the integral of the dissipative terms from the continuous energy law. This identity is what enforces both unconditional energy decay and the bound $r^{n+1} \le E_1(\varphi^{n+1})$, keeping the modified energy close to the physical free energy.

What would settle it

Run Scheme 1 on the Example 4.3 coarsening benchmark with time-step ratios above 4.8645, as produced by Algorithm 2 in Figure 4.6, and monitor $r^{n+1}-r^n$: any positive value would contradict Theorem 2.1; alternatively, benchmark Algorithm 2 against fixed-step R-IMEX-BDF2 at matched $L^2$ error and CPU time on a parameter set outside (4.3)-(4.8), where a cheaper fixed-step run would refute the reported efficiency advantage.

Watch

Extended reading notes

Core claim

The core claim is that the CHB system can be solved by R-IMEX-BDFk schemes, $k=1,2,3,4$, that combine an implicit-explicit BDF discretization with a relaxed scalar auxiliary variable. The auxiliary variable is defined as $r(t)=E_1(\varphi)=E(\varphi)+C_0$, and the scheme first computes an intermediate phase field, velocity, pressure, and auxiliary value, then corrects with a relaxation factor $\zeta^{n+1}_k = 1 - (1-\xi^{n+1}_k)^{k+1}$. Theorem 2.1 specifies four cases for choosing the relaxation parameter $\sigma^{n+1}_0$ and the nonnegative dissipation parameter $\delta^{n+1}$ so that $r^{n+1}$ is both nonincreasing and bounded above by the physical energy $E_1(\varphi^{n+1})$. This transfers the continuous dissipation law to the discrete level without restricting the ratio of adjacent time steps, which is the property that makes the subsequent adaptive algorithms possible.

Load-bearing premise

The adaptive efficiency claims rest on the untested assumption that $|1-\xi^{n+1}_k|^m$, with a manually chosen exponent $m$ in $(0,1)$, is a reliable proxy for the numerical error of the coupled CHB solution, and on a hand-picked switching time $T_c$; the paper offers no error analysis for either, and its own Table 4.4 shows that R-IMEX-BDF2A alone gives meaningless results for parameter sets (4.7) and (4.8).

Editorial extensions

If this is right

  • For any sequence of time steps, including ratios far above the previously noted bound 4.8645, the modified energy $r^n$ never increases, so no CFL-type step restriction is needed for stability.
  • The relaxation improves accuracy over plain IMEX-BDF at large time steps, as shown in Table 4.1, where the unrelaxed second-order scheme develops large errors at $\tau=2\times10^{-3}$ while the relaxed scheme remains accurate.
  • The hybrid-order Algorithm 2, using R-IMEX-BDF3A on $[0,T_c]$ and R-IMEX-BDF2A on $(T_c,T]$, is reported to cut CPU time by factors of roughly two to four against fixed-step R-IMEX-BDF2 at similar error and to beat IMEX-CNA in both accuracy and cost.
  • The same R-GSAV construction is claimed to extend directly to other phase-field equations, since the relaxation and adaptive error indicator do not rely on CHB-specific structure beyond the energy law.
  • The adaptive algorithms inherit the unconditional energy stability of the base scheme, so the modified energy remains monotone even when the step-size ratio fluctuates, as demonstrated in Figures 4.6 and 4.7.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The error indicator $|1-\xi^{n+1}_k|^m$, with a manually chosen exponent $m\in(0,1)$, is empirical rather than derived; a rigorous a posteriori bound linking it to the actual phase-field error would put the adaptive algorithms on the same footing as the stability theorem.
  • The switch time $T_c$ is hand-picked by observing where the fast transient ends; an automatic rule based on when $E'(t)$ or the step-size history crosses a threshold could make the hybrid algorithm self-tuning and more robust across different problems.
  • The stability proof guarantees decay of the modified energy $r^{n+1}$ but does not quantify the gap between $r^{n+1}$ and the original free energy $E(\varphi^{n+1})$; a step-size-dependent bound on that gap would tell users how much the modified energy can be trusted as a physical diagnostic.
  • Because the construction only uses the energy-dissipation structure, the hybrid-order adaptive strategy is a plausible candidate for other coupled phase-field systems, such as phase-field crystal or tumor-growth models, where high-order BDF schemes suffer the same early-transient versus long-time stability trade-off.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes a class of variable-time-step backward differentiation formula (BDF) schemes for the Cahn-Hilliard-Brinkman (CHB) system, built on a generalized scalar auxiliary variable (GSAV) approach with relaxation (R-GSAV). The schemes are claimed to be unconditionally energy-stable in a modified (SAV) energy, and the authors design adaptive time-stepping strategies, including a hybrid-order algorithm that switches between BDF3 and BDF2 according to a user-specified time Tc. Numerical experiments report temporal convergence rates for k=1,...,4, long-time coarsening simulations, and comparisons of CPU time and accuracy for the adaptive algorithms against fixed-step and existing adaptive schemes.

Significance. If substantiated, the paper would provide a useful extension of R-GSAV relaxation to the coupled CHB system, and the proposed hybrid adaptive strategy is a pragmatic idea for combining the short-time accuracy of BDF3 with the long-time robustness of BDF2. The central energy-decay identity r^{n+1}-r^n = -τ δ κ ≤ 0 is essentially algebraic and, under a consistent definition of the auxiliary variables, should hold for any nonnegative κ. However, the manuscript currently suffers from an under-specified scheme, coefficient typos, missing forcing terms in the accuracy test, and a strong dependence of the advertised adaptive gains on hand-tuned parameters. These issues prevent the main claims from being verified as written.

major comments (4)
  1. [Section 2.2, Scheme 1 (2.4a)-(2.4g)] The scheme does not specify how the chemical potential μ^{n+1} is updated. Step 1 defines an intermediate μ̃ and then (2.4g) rescales only φ, u, and p, leaving μ undefined. This is not a cosmetic omission: the next time step needs B_k(μ^n) in (2.4c), and Theorem 2.1's κ(φ^{n+1}, u^{n+1}) in (2.6)-(2.7) depends on μ^{n+1}. If μ^{n+1} is silently set to ζ μ̃, then κ(φ^{n+1}, u^{n+1}) is not the dissipation rate of the accepted state because M(ζφ̃)|∇(ζμ̃)|² ≠ ζ²M(φ̃)|∇μ̃|² for general mobility M; if μ^{n+1} is instead defined as the physical chemical potential -ε²Δφ^{n+1}+F'(φ^{n+1}), then μ^{n+1} ≠ ζ μ̃ because F' is cubic. The paper must state the update rule for μ and show that the relation between (2.4e) and (2.6) is meaningful under that rule; otherwise both the unconditional energy-stability theorem and the reported convergence rates rest on an unstated convention.
  2. [Section 2.2, coefficient formulas for A_k and B_k] The printed formulas for the second- and third-order coefficients contain sign and denominator errors. For k=2, A2(φ^n) is printed with a leading term (-t_{n+1}+t_{n-1})/(t_n-t_{n-1}) φ^n, which equals -(1+τ_{n+1}/τ_n)φ^n, together with a second term -τ_{n+1}²/(τ_n(τ_n+τ_{n+1}))φ^{n-1}; the standard nonuniform BDF2 coefficient A2 should be +(1+τ_{n+1}/τ_n)φ^n - τ_{n+1}²/(τ_n(τ_n+τ_{n+1}))φ^{n-1}. For k=3, B3(φ^n) contains denominators such as (t_n+t_{n-2}) and (t_{n-1}+t_{n-2}), which should be differences (t_n-t_{n-2}) and (t_{n-1}-t_{n-2}). As written, the R-IMEX-BDF2 and R-IMEX-BDF3 schemes are not the claimed high-order BDF discretizations and cannot be implemented. Corrected coefficient formulas for k=2,3,4 should be provided.
  3. [Section 4.1, Example 4.1] The accuracy test does not state that manufactured source terms are added to the CHB system. The exact solution (4.1) does not satisfy the homogeneous CHB equations (2.1) with the given parameters (ε=1, γ=2, M=ν=η=1, S=0); for instance, μ=φ³+φ and φ_t=-sin(t)cos(x)sin(y), so (2.1a) cannot hold without an external forcing term. The paper must specify the added forcing terms for the manufactured solution, or the convergence orders reported in Figure 4.1 cannot be reproduced.
  4. [Section 3 and Table 4.4] The robustness of the proposed adaptive algorithms is not established. The paper states in Section 3 that the error indicator |1-ξ_k|^m works when m is 'appropriately chosen between 0 and 1', but no selection rule or theory is given. Table 4.4 shows that R-IMEX-BDF2A with parameter sets (4.7) and (4.8) yields 'meaningless solutions', while Algorithm 2 succeeds under the same settings. Since the claimed efficiency of the hybrid algorithm depends on the hand-picked parameters (ρ, tol, r, m, τ_max) and the switch time Tc, the paper should provide a systematic parameter-selection strategy or at least a sensitivity study, and it should clarify why the hybrid algorithm is robust when the underlying single-order adaptive algorithms fail.
minor comments (6)
  1. [Section 2.1, equations (2.2a)-(2.3)] The derivation of the dissipation law appears to add (2.2b) to (2.2a) and (2.2c), although (2.2a) already contains (φ_t, μ); adding (2.2b) double-counts and would introduce a spurious factor 1/2. The correct step is to use (2.2b) to replace (φ_t, μ) with dE/dt. Please clarify the intended derivation.
  2. [Theorem 2.1 proof, equation (2.9)] The proof assumes E1(φ̃^{n+1}) ≥ 0 and then divides by E1 in (2.4e) and (2.9). With C0=0, E1 can be arbitrarily close to zero or exactly zero for constant states, making the division undefined. The theorem should assume E1(φ̃^{n+1}) > 0, or require C0 > 0, or discuss how zero-energy states are handled.
  3. [Remark 2.2, inequality (2.11)] The bound (Lφ^{n+1}, φ^{n+1}) ≤ (C_k M)^2 appears to require (Lφ̃, φ̃) ≤ E1(φ̃), but from the definition E1 = 1/2(Lφ̃, φ̃) + ∫F, only (Lφ̃, φ̃) ≤ 2E1(φ̃) holds under nonnegativity of F. The displayed square (Lφ̃, φ̃)^2 in (2.11) seems erroneous; please check the estimate.
  4. [Algorithms 1 and 2] The adaptive algorithms use E'(t) without defining its discrete counterpart; specify whether E'(t) is approximated by a finite difference of the discrete energy, and with which step size.
  5. [Algorithm 1, Step 4] In Step 4, the update formula Adp(e_n^{m_k}, τ_{n+1}) uses τ_{n+1} before τ_{n+1} has been computed; this should presumably be Adp(e_n^{m_k}, τ_n) or a clearly defined preliminary value.
  6. [Algorithm 2, Steps 1 and 6] The notation 'compute ξ_n by the R-IMEX-BDF3/BDF2' is inconsistent with the earlier definition ξ_k^{n+1}; this should be ξ^{n+1}_3 and ξ^{n+1}_2 respectively.

Circularity Check

1 steps flagged · score 6.0 of 10

The energy-stability theorem is self-definitional: the relaxation set Γ in (2.6) is defined by the very decay identity that Theorem 2.1 then 'proves', so the central stability claim reduces to a construction rather than an independent derivation.

  1. self definitional [Section 2.2, equations (2.5)-(2.7) and Theorem 2.1]
    "Step2:Compute scalar auxiliary variable rn+1 by using the following relaxation: rn+1 = σn+1 0 ˜rn+1 + (1 − σn+1 0 )E1(ϕn+1), σn+1 0 ∈ Γ, (2.5) where Γ = { σ ∈ [0, 1] s.t. rn+1 − ˜rn+1 / τn+1 = −δn+1κ(ϕn+1, un+1) + ˜rn+1 / E1( ˜ϕn+1) κ( ˜ϕn+1, ˜un+1) }, (2.6) with δn+1 ⩾ 0 to be determined. The following theorem summarizes the selection of σn+1 0 and δn+1. ... Given rn ⩾ 0, we have rn+1 ⩾ 0, ξn+1 k ⩾ 0 in the scheme (2.4)-(2.5) and rn+1 is unconditionally energy stable in the sense that rn+1 − rn = −τn+1δn+1κ(ϕn+1, un+1) ≤ 0"

    Equation (2.6) defines Γ as the set of relaxation parameters for which the target decay equality holds after combining with (2.4e). Theorem 2.1 then chooses σ0 and δ from that set and derives (2.7) by combining (2.4e) with (2.6). Thus the unconditional energy-stability inequality is not a consequence of the discretized CHB equations; it is the defining constraint of the relaxation step. The case analysis only checks that the proposed σ0,δ satisfy σ0∈[0,1], δ≥0, and r≤E1; the decay statement itself is a restatement of the definition of Γ. This is the standard SAV construction, but as a 'strict proof' of energy stability it is self-definitional.

full rationale

The central advertised result (Theorem 2.1) is a construction: the relaxation step (2.5)-(2.6) defines Γ by the very decay equality that the theorem asserts, so the proof that r^{n+1}-r^n = -τδκ ≤ 0 is an algebraic consequence of the definition of δ, not an independent derivation. This is a genuine self-definitional element in the main stability claim. I nevertheless keep the score at 6 rather than higher because the numerical content of the paper is not entirely circular: convergence rates in Example 4.1 are measured against the manufactured exact solution (4.1), and the adaptive/coarsening tests use an independent τ=1e-6 reference solution, so accuracy claims are externally benchmarked. The adaptive error indicator |1-ξ|^m and the switch time Tc are empirical/heuristic (Section 3), with the paper itself reporting failures for parameter sets (4.7)-(4.8) in Table 4.4; this is tuning, not a fitted parameter renamed as a prediction, so it does not add circularity. I find no load-bearing self-citation chain: the underlying SAV/relaxation framework is credited to [33] and [35], which are not invoked as an unverified uniqueness theorem, and the stability proof does not depend on the authors' own prior results. Separately, the scheme as written never updates μ^{n+1}: Step 1 defines φ̃, μ̃, ũ, p̃, r̃ and then rescales φ,u,p but not μ, while (2.4c) and κ(φ^{n+1},u^{n+1}) require μ; this is a formal well-definedness gap in Theorem 2.1 and the algorithm, but it is a correctness issue rather than a circularity and is not scored.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

No new physical entities are introduced; the auxiliary scalar r and relaxation parameter sigma are computational artifacts of the SAV framework, not physical postulates.

free parameters (5)
  • S (stabilization parameter) = S=0 (Example 4.1), S=1 (Examples 4.2-4.4)
    Added linear term S phi in (2.4b) to stabilize the IMEX treatment of the double-well potential; no theory fixes its value.
  • C0 (energy shift) = C0=0
    E1(phi)=E(phi)+C0; the positivity and regularity analysis in Remark 2.2 requires C0>0, but experiments set C0=0.
  • m (error indicator exponent) = 0.52, 0.70, 0.75, 0.85, 0.68, 0.70 in (4.3)-(4.8); m=1 in (4.9)-(4.10)
    The error indicator |1-xi|^{m}; the paper says it works for m in (0,1) based on experiments, not theory.
  • Adaptive controller parameters rho, tol, r, gamma*, tau_min, tau_max = six sets (4.3)-(4.8) and (4.9)-(4.10)
    Hand-tuned controller gains for Adp(e,tau); efficiency comparisons in Tables 4.2-4.6 depend on these choices.
  • Tc (switch time) = Tc=1.2 in Example 4.3; Tc=0.4 in Example 4.4
    The hybrid algorithm switches from BDF3 to BDF2 at this user-chosen time; no criterion for picking it is given.
assumptions (4)
  • domain assumption The continuous CHB system (2.1) has solutions and satisfies the energy law (2.3) for the data considered.
    Invoked in Section 2.1 to define the discrete dissipation target; existence follows from cited works [7,8] but is not reproved.
  • ad hoc to paper E1(phi~^{n+1}) > 0 at every step wherever the scheme divides by it in (2.4e), (2.4f), (2.9)-(2.11).
    The proof of r~>=0 and xi>=0 in Theorem 2.1 divides by E1; the paper does not establish a uniform positive lower bound.
  • ad hoc to paper |1-xi|^m is a reliable error indicator for the CHB system for suitably chosen m.
    Stated in Section 3 as an empirical finding; the entire adaptive strategy depends on it.
  • domain assumption The standard BDF variable-step coefficient formulas from [20] produce kth-order accuracy.
    Used in scheme (2.4a); the paper quotes them from Taylor expansion and cites [20].

how reviews work

0 comments
Cite this review

Pith. "Pith review of New highly efficient and accurate numerical scheme for the Cahn-Hilliard-Brinkman system." pith.science (2026). https://pith.science/paper/DPCYUN6U

@misc{pith2026250607128,
  author       = {Pith},
  title        = {Pith review of: New highly efficient and accurate numerical scheme for the Cahn-Hilliard-Brinkman system},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DPCYUN6U}},
  note         = {Machine review of arXiv:2506.07128}
}
read the original abstract

In this paper, based on a generalized scalar auxiliary variable approach with relaxation (R-GSAV), we construct a class of high-order backward differentiation formula (BDF) schemes with variable time steps for the Cahn-Hilliard-Brinkman(CHB) system. In theory, it is strictly proved that the designed schemes are unconditionally energy-stable. With the delicate treatment of adaptive strategies, we propose several adaptive time-step algorithms to enhance the robustness of the schemes. More importantly, a novel hybrid-order adaptive time steps algorithm performs outstanding for the coupled system. The hybrid-order algorithm inherits the advantages of some traditional high-order BDF adaptive strategies. A comprehensive comparison with some adaptive time-step algorithms is given, and the advantages of the new adaptive time-step algorithms are emphasized. Finally, the effectiveness and accuracy of the new methods are validated through a series of numerical experiments.

Figures

Figures reproduced from arXiv: 2506.07128 by the authors.

Figure 4.1
Figure 4.1. (Example 4.1) Temporal convergence rates: L 2 errors of phase field function ϕ, the velocity (x-component of u1, y-component of u2), and the derivative of pressure (px and py) as functions of the time step size τ for Scheme 1. where the following parameters are used in this test: ε = 0.05, γ = 4, M = ν = η = 1, S = 1. We use the results from the semi-implicit/BDF2 scheme with τn = 1e−6 as the reference solution. The… view at source ↗
Figure 4.2
Figure 4.2. (Example 4.2)Time histories of the modified energy computed by (a) the R-IMEX-BDF2 scheme, (b) the R-IMEX-BDF3 scheme,using large time step sizes τn=1,4. In addition, we simulate the spinodal decomposition of a homogeneous mixture into two coexisting phases governed by the CHB system as a further test of the algorithms developed here [PITH_FULL_IMAGE:figures/full_fig_p011_4_2.png] view at source ↗
Figure 4.3
Figure 4.3. (Example 4.2)Time evolutions of modified energy (left), error (mid￾dle), and en+1 (right). (a) T = 1.8 (b) T = 4 (c) T = 8 (d) T = 15 (e) T = 80 (f) T = 150 (g) T = 200 (h) T = 500 [PITH_FULL_IMAGE:figures/full_fig_p012_4_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4.4
Figure 4.4. Figure 4.4: (Example 4.2)Time evolution of the coarsening process with ϵ = 0.05 by using the R-IMEX-BDF2 scheme with τn = 0.001. S = 1 and γ ∗ = 1. Similar to the Example 4.2, we use the results from the R-IMEX-BDF2 scheme with τn = 1e − 6 as the reference solution. To present t…
Figure 4.5
Figure 4.5. Figure 4.5: (Example 4.3)Modified energy and original energy history curves (a)(b)(c) of schemes with different parameters; The snapshots (d)(e)(f) of the inter￾faces between two phases are depicted at T = 500. not exceed 4.8645. Our approach, similar to the adaptive algorithm d…
Figure 4.6
Figure 4.6. Figure 4.6: (Example 4.3)Histories of time steps and time step ratios obtained by Algorithm 2 with (4.3)(4.7).The average time step is 2.9e-3 [PITH_FULL_IMAGE:figures/full_fig_p016_4_6.png]
Figure 4.7
Figure 4.7. Figure 4.7: (Example 4.3)Histories of time steps and time step ratios obtained by R-IMEX-BDF2A with (4.5).The average time step is 1.5e-3. Example 4.4. (Buoyancy-driven flow) We consider in this last example the buoyancy-driven flow that two heavy fluid layers sandwich a light f…
Figure 4.8
Figure 4.8. Figure 4.8: (Example 4.4)Snapshots of buoyancy-driven flow function ϕ by R￾IMEX-BDF2A with (4.10) [PITH_FULL_IMAGE:figures/full_fig_p018_4_8.png]
Figure 4.9
Figure 4.9. Figure 4.9: (Example 4.4)The snapshots of buoyancy-driven flow function ϕ by different algorithms at T=7.8. 5. Concluding remarks For the Cahn-Hilliard-Brinkman system, we proposed unconditionally energy-stable R-IMEX￾BDFk schemes. These schemes only require solving elliptic and…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages

  1. [35]

    Zheng and X

    N. Zheng and X. Li, New efficient and unconditionally energy stable schemes for the cahn– hilliard–brinkman system , Appl. Math. Lett., 128 (2022), p. 107918

    work page 2022

  2. [21]

    Jiang, H

    P. Jiang, H. Jia, L. Liu, C. Zhang, and D. W ang, Highly efficient, robust and uncon- ditionally energy stable second order schemes for approximating the cahn-hilliard-brinkman system, Appl. Numer. Math., 201 (2024), pp. 175–186. NEW SCHEMES FOR THE CAHN-HILLIARD-BRINKMAN SYSTEM 21

    work page 2024

  3. [1]

    D. M. Anderson, G. B. McF adden, and A. A. Wheeler, Diffuse-interface methods in fluid mechanics, Annu. Rev. Fluid Mech., 30 (1998), pp. 139–165

    work page 1998

  4. [2]

    Chen and C

    L. Chen and C. Xu, A time splitting space spectral element method for the cahn-hilliard equation, East Asian J. Appl. Math., 3 (2013), pp. 333–351

    work page 2013

  5. [3]

    Chen and J

    L. Chen and J. Zhao, A novel second-order linear scheme for the cahn-hilliard-navier-stokes equations, J. Comput. Phys., 423 (2020), p. 109782

    work page 2020

  6. [4]

    Y. Chen, Y. Huang, N. Yi, and P. Yin, Recovery type a posteriori error estimation of an adaptive finite element method for cahn–hilliard equation , J. Sci. Comput., 98 (2024), p. 35. 20 D.W. CHEN, Q.Z. REN, AND M.H. LI

    work page 2024

  7. [5]

    Cheng and J

    Q. Cheng and J. Shen, Multiple scalar auxiliary variable (MSA V) approach and its applica- tion to the phase-field vesicle membrane model , SIAM J. Sci. Comput., 40 (2018), pp. A3982– A4006

    work page 2018

  8. [6]

    Colli, G

    P. Colli, G. Gilardi, A. Signori, and J. Sprekels, Cahn–Hilliard–Brinkman model for tumor growth with possibly singular potentials , Nonlinearity., 36 (2023), pp. 4470–4500

    work page 2023

Show all 35 references
  1. [7]

    Collins, J

    C. Collins, J. Shen, and S. M. Wise, An efficient, energy stable scheme for the cahn- hilliard-brinkman system , Commun. Comput. Phys., 13 (2013), pp. 929–957

    2013

  2. [8]

    Conti and A

    M. Conti and A. Giorgini, Well-posedness for the brinkman–cahn–hilliard system with unmatched viscosities, J. Differ. Equ., 268 (2020), pp. 6350–6384

    2020

  3. [9]

    Ebenbeck and H

    M. Ebenbeck and H. Garcke, On a cahn–hilliard–brinkman model for tumor growth and its singular limits , SIAM J. Math. Anal., 51 (2019), pp. 1868–1912

    2019

  4. [10]

    Ebenbeck, H

    M. Ebenbeck, H. Garcke, and R. N¨urnberg, Cahn–hilliard–brinkman systems for tumour growth, Discrete Contin. Dyn. Syst. - S., 14 (2021), pp. 3989–4033

    2021

  5. [11]

    Feng and H

    X. Feng and H. Wu, A posteriori error estimates for finite element approximations of the cahn-hilliard equation and the hele-shaw flow , J. Comput. Math., (2008), pp. 767–796

    2008

  6. [12]

    U. S. Fjordholm, S. Mishra, and E. Tadmor, Well-balanced and energy stable schemes for the shallow water equations with discontinuous topography , J. Comput. Phys., 230 (2011), pp. 5587–5609

    2011

  7. [13]

    Garcke, K

    H. Garcke, K. F. Lam, E. Sitka, and V. Styles, A cahn–hilliard–darcy model for tumour growth with chemotaxis and active transport , Math. Models Methods Appl. Sci., (2016)

    2016

  8. [14]

    R. Guo, Y. Xia, and Y. Xu, An efficient fully-discrete local discontinuous galerkin method for the cahn–hilliard–hele–shaw system , J. Comput. Phys., 264 (2014), pp. 23–40

    2014

  9. [15]

    Han and X

    D. Han and X. W ang, A second order in time, uniquely solvable, unconditionally stable numerical scheme for cahn–hilliard–navier–stokes equation , J. Comput. Phys., 290 (2015), pp. 139–156

    2015

  10. [16]

    , A second order in time, decoupled, unconditionally stable numerical scheme for the cahn– hilliard–darcy system, J. Sci. Comput., 77 (2018), pp. 1210–1233

    2018

  11. [17]

    Hou and Z

    D. Hou and Z. Qiao, An implicit–explicit second-order bdf numerical scheme with variable steps for gradient flows , J. Sci. Comput., 94 (2023), p. 39

    2023

  12. [18]

    Methods Partial Differ

    , A linear adaptive second-order backward differentiation formulation scheme for the phase field crystal equation , Numer. Methods Partial Differ. Equ., 39 (2023), pp. 4174–4195

    2023

  13. [19]

    Huang and J

    F. Huang and J. Shen, A new class of implicit–explicit bdfk sav schemes for general dissipa- tive systems and their error analysis , Comput. Meth. Appl. Mech. Eng., 392 (2022), p. 114718

    2022

  14. [20]

    Huang, J

    F. Huang, J. Shen, and Z. Yang, A highly efficient and accurate new scalar auxiliary variable approach for gradient flows , SIAM J. Sci. Comput., 42 (2020), pp. A2514–A2536

    2020

  15. [22]

    Knopf and A

    P. Knopf and A. Signori, Existence of weak solutions to multiphase cahn–hilliard–darcy and cahn–hilliard–brinkman models for stratified tumor growth with chemotaxis and general source terms, Commun. Partial Differ. Equ., 47 (2022), pp. 233–278

    2022

  16. [23]

    Le Bars and M

    M. Le Bars and M. G. Worster, Interfacial conditions between a pure fluid and a porous medium: implications for binary alloy solidification , J. Fluid Mech., 550 (2006), pp. 149–173

    2006

  17. [24]

    Li and C

    M. Li and C. Xu, New efficient time-stepping schemes for the navier–stokes–cahn–hilliard equations, Comput. Fluids., 231 (2021), p. 105174

    2021

  18. [25]

    Li and J

    X. Li and J. Shen, On a sav-mac scheme for the cahn–hilliard–navier–stokes phase-field model and its error analysis for the corresponding cahn–hilliard–stokes case , Math. Models Methods Appl. Sci., 30 (2020), pp. 2263–2297

    2020

  19. [26]

    H.-l. Liao, B. Ji, and L. Zhang, An adaptive bdf2 implicit time-stepping method for the phase field crystal model , IMA J. Numer. Anal., 42 (2022), pp. 649–679

    2022

  20. [27]

    C. Liu, J. Shen, and X. Yang, Decoupled energy stable schemes for a phase-field model of two-phase incompressible flows with variable density , J. Sci. Comput., 62 (2015), pp. 601–622

    2015

  21. [28]

    Ngamsaad, J

    W. Ngamsaad, J. Yojina, and W. Triampo, Theoretical studies of phase-separation ki- netics in a brinkman porous medium , J. Phys. A-Math. Theor., 43 (2010), p. 202001

    2010

  22. [29]

    J. Shen, J. Xu, and J. Yang, The scalar auxiliary variable (SA V) approach for gradient flows, J. Comput. Phys., 353 (2018), pp. 407–416

    2018

  23. [30]

    Shen and X

    J. Shen and X. Yang, Numerical approximations of Allen-Cahn and Cahn-Hilliard equations, Discrete Contin. Dyn. Syst., 28 (2010), pp. 1669–1691

    2010

  24. [31]

    W ang and S

    C. W ang and S. M. Wise, An energy stable and convergent finite-difference scheme for the modified phase field crystal equation , SIAM J. Numer. Anal., 49 (2011), pp. 945–969

    2011

  25. [32]

    N. K. Yamaleev and M. H. Carpenter, A systematic methodology for constructing high- order energy stable weno schemes , J. Comput. Phys., 228 (2009), pp. 4248–4272

    2009

  26. [33]

    Zhang and J

    Y. Zhang and J. Shen, A generalized sav approach with relaxation for dissipative systems , J. Comput. Phys., 464 (2022), p. 111311

    2022

  27. [34]

    Zhang, Y

    Z. Zhang, Y. Ma, and Z. Qiao, An adaptive time-stepping strategy for solving the phase field crystal model , J. Comput. Phys., 249 (2013), pp. 204–215

    2013

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.