A Conservative and Positivity-Preserving Discontinuous Galerkin Method for the Population Balance Equation

Guanyang Liu; Yong-Tao Zhang; Ziyao Xu

REVIEW 3 major objections 5 minor 64 references

A Conservative and Positivity-Preserving Discontinuous Galerkin Method for the Population Balance Equation

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read For the population balance equation, a discontinuous Galerkin scheme conserves mass to round-off and keeps number densities nonnegative.

desk verdict Strong numerics and a genuinely useful scheme, but the positivity theorem is not proven as written—several steps are dimensionally inconsistent, including a division by zero on the first cell. read the letter →

arxiv 2507.16631 v1 pith:KQ4WD6W7 submitted 2025-07-22 math.NA cs.NA

Ziyao Xu , Guanyang Liu , Yong-Tao Zhang This is my paper · ORCID

classification math.NAcs.NA MSC 65M6045K0565M70

keywords populationbalanceequationdiscontinuousGalerkinmethodmassconservationpositivity-preservingaggregationandbreakagemoment-conservinglimitercommonrefinementquadraturenumberdensity

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

The paper targets the population balance equation (PBE), the integro-differential model describing how particles of different sizes grow, nucleate, aggregate, and break in settings such as aerosols and crystallization. It aims to show that a high-order discontinuous Galerkin discretization can preserve the two physical properties that matter most in simulations: total particle mass and nonnegative number densities. To obtain exact mass conservation, the aggregation and breakage birth and death terms are rewritten as symmetric double integrals evaluated on a common refinement of the integration domain, so that quadrature errors cancel when the terms are summed. To obtain positivity, the authors prove that the first moment (local particle mass) in each cell stays nonnegative under a time-step condition, then apply a new scaling limiter that restores nonnegativity while preserving that moment. In the reported tests the scheme is third-order accurate and keeps total mass deviation at round-off level.

What carries the argument

The load-bearing mechanism is a symmetric common refinement of the integration regions for the birth and death integrals. On the aggregation region $\Omega_a$, the birth regions $A_i$ and death regions $B_i$ are subdivided into a single triangular partition $T$; on the breakage region $\Omega_b$, a common refinement $E_0$ is built from the regions $C_i$ and $D_i$. The quadrature operators $Q_T$ and $Q_E$ are chosen so that, after the change of variables, the same quadrature points evaluate the birth mass term and the death mass term, and the rules are invariant under exchanging $u$ and $w$; with test function $v$, the two sums cancel exactly, independent of polynomial degree and mesh. The companion mechanism is a moment-conserving scaling limiter: given a cell polynomial with nonnegative $s$-th moment, the limiter shifts the polynomial up by a constant and rescales it, preserving $\int_{I_i} v^s n_h(v)\,dv$ while clipping negative values to zero.

What would settle it

Implement the aggregation-only scheme on a highly nonuniform mesh where the common refinement is triangulated without enforcing $u$-$w$ symmetry, or with a triangle quadrature rule that is not symmetric, and track the total mass over a long run: a mass drift away from its initial value at a rate proportional to the asymmetry would refute the central conservation claim. A positive control uses the symmetric refinement and shows the drift staying at round-off.

Watch

Extended reading notes

Core claim

The paper claims that both goals can be met simultaneously in a DG framework that works directly with the number density $n(v,t)$, not with a mass-density reformulation. For aggregation and breakage, the birth and death integrals are evaluated on the same triangulated common refinement of their integration regions, using quadrature rules that are symmetric under swapping the two particle sizes $u$ and $w$; as a result, when the test function is $v$, the discrete birth and death contributions to the total mass cancel exactly, so the numerical mass $M_h(t)$ satisfies $dM_h/dt = 0$ up to rounding. For positivity, the paper proves that after one forward Euler step the first moment on each cell is nonnegative under the CFL bounds $\Delta t \leq 1/(v_{\max}\|\beta(u,w)n_h(w)\|_\infty)$ for aggregation and $\Delta t \leq 1/\|u\,p(u,w)\gamma(w)\|_\infty$ for breakage. A new scaling limiter then replaces a cell polynomial $n_h$ by $\theta(n_h + m)$, choosing $m$ and $\theta\in[0,1]$ so that the $s$-th moment is conserved while all point values become nonnegative; taking $s=1$ conserves mass, and the same construction is valid for any moment. Numerical benchmarks with analytical solutions and with physical aggregation kernels show third-order convergence and mass deviations around $10^{-16}$.

Load-bearing premise

The exact mass cancellation requires the triangulated common refinement and the quadrature rules to remain symmetric when the two particle sizes $u$ and $w$ are swapped, including after the geometric transformation that places quadrature points on physical cells; if that symmetry is lost, the birth and death quadrature errors no longer cancel.

Editorial extensions

If this is right

Aggregation and breakage can be simulated on arbitrary nonuniform meshes with high-order polynomials while total particle mass stays constant to round-off, without reformulating the equation in mass-density variables.
The positivity limiter is not tied to the cell average: any physically meaningful moment can be conserved during limiting, including particle mass or, in length coordinates, the third moment that represents mass.
The proven CFL conditions give a computable guarantee that local mass on each cell remains nonnegative per forward Euler stage, and the fallback strategy of halving the time step terminates after finitely many halvings.
Because the time integration uses SSP-RK stages, which are convex combinations of forward Euler steps, the conservation and positivity properties carry over to high-order fully discrete time stepping.
In positivity-challenging examples, the limiter allows time steps about ten times larger than the unlimited scheme before the computation becomes unstable, while preserving the accurate solution.

Reading between the lines

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

The moment-conserving limiter should transfer to other kinetic and population models in which a weighted moment, rather than the cell average, is the conserved physical quantity, such as coagulation models formulated in particle-length coordinates.
The common-refinement cancellation suggests a general recipe for conservative discretization of integral birth-death operators: refine all integration domains into one symmetric partition before applying quadrature; this may extend to two-dimensional internal coordinates or multi-particle collision integrals.
The swap-symmetry assumption can be probed directly: running the aggregation scheme with an intentionally non-symmetric triangulation or non-symmetric quadrature rule should produce mass drift at the level of the asymmetry, whereas the symmetric construction should stay at round-off.
The fallback time-step halving could be replaced by an adaptive controller that monitors the proven moment-positivity condition and adjusts the step before negative moments appear.

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.

Desk Editor's Note

read the letter

The headline is: the scheme is real and the numerics are strong, but the positivity theorem is not proven as written—several steps don't parse dimensionally, so the central theoretical claim is currently unsupported.

What's new: the common-refinement double-integral quadrature for aggregation/breakage, and the scaling limiter that preserves the first moment. The mass conservation in the tests is at round-off, third-order accuracy is clean across many benchmarks, and the positivity-preserving limiter demonstrably rescues a computation that blows up without it. That is a useful engineering contribution.

Where it's soft: Theorem 4.1's proof. The aggregation step asserts an equality that mixes 1/v_max and v_{i+1/2}, which have different units, and the bracket can be negative anyway. The breakage step divides by v_{1/2}=0 on the first cell. Earlier in the same proof, the lower bound drops the death terms on A_i∩B_i, which is the wrong direction for a lower bound. So the main guarantee—the positive first moment before limiting—is not established by the text. A corrected proof may well exist, but it's not here.

Also, the conservation proof (Prop 3.1) rests on the u-w symmetry of the common refinement and of the quadrature rule on each triangle. That is plausible for a u-w symmetric partition with D3-symmetric rules, but it is asserted, not demonstrated; a referee should ask for a precise argument covering the affine mapping.

The novelty claim is a bit strong: the mass-density DG method in [31] already preserves mass and positivity, and the aggregation-map idea in [34,42,52] is the direct ancestor. The specific new combination is legitimate, but the "first work" phrasing should be toned down.

Bottom line: this is a well-executed numerical paper with a broken (as typeset) theory section. A serious referee could reasonably recommend major revision: fix the positivity proof, clarify the symmetry argument, and soften the novelty claim. The experiments justify the referee time. I'd bring it to the reading group to see if the proof can be repaired.

Referee Report

3 major / 5 minor

Summary. This paper develops a discontinuous Galerkin method for the population balance equation (PBE) in the number-density formulation. The aggregation and breakage operators are rewritten as double integrals over symmetric regions A_i, B_i and C_i, D_i, respectively, and are evaluated by quadrature on common refinements of the integration domains. The authors prove exact mass conservation for the semi-discrete scheme, derive CFL-type conditions under which the first cell moment remains nonnegative after forward Euler steps, and introduce a scaling limiter that preserves an arbitrary prescribed moment while enforcing pointwise nonnegativity. Numerical tests for pure aggregation, pure breakage, aggregation--breakage, and aggregation--growth report third-order accuracy, mass conservation to round-off, and improved robustness with the limiter.

Significance. If the theoretical results hold, the paper's main contribution is a positivity-preserving DG framework in which the conserved quantity is a general moment rather than the cell average; this is exactly what is needed for the number-density form of the PBE, where mass, not particle number, is conserved. The numerical verification is unusually complete: order-of-accuracy studies, round-off-level mass conservation over long times, physical kernels, and a clear demonstration that the limiter prevents blow-up. The construction via common refinements is elegant and likely to be reusable. However, the proof of the first-moment positivity theorem (Theorem 4.1) as typeset contains unjustified and dimensionally inconsistent steps, and the conservation proof leaves the global symmetry of the quadrature rules implicit; these issues must be repaired before the central claim is fully supported.

major comments (3)

[§4.1, Theorem 4.1 (aggregation case)] The aggregation part of the proof contains the displayed equality Q_Ii[v n_h] - Δt Σ_{T⊆B_i\A_i} Q_T[uβ(u,w)n_h(u)n_h(w)] = Σ_{T⊆B_i\A_i} Q_T[u n_h(u)(1/vmax - v_{i+1/2} - Δtβ(u,w)n_h(w))]. This equality is not derived and is dimensionally inconsistent: 1/vmax and v_{i+1/2} have different units, and a one-dimensional cell integral Q_Ii[v n_h] cannot be replaced by a sum of two-dimensional triangle quadratures without an argument. The claimed nonnegativity therefore does not follow from the written proof. A viable repair is to use the cancellation of birth and death over A_i, obtaining Q_Ii[v n_h] - Δt Σ_{T⊆B_i\A_i} Q_T[uβ(u,w)n_h(u)n_h(w)], and then bound the death term by Δt (vmax - v_{i+1/2}) ||β n_h||_{L∞(Ωa)} Q_Ii[u n_h] ≤ Δt vmax ||β n_h||_{L∞(Ωa)} Q_Ii[u n_h]; the stated CFL condition then yields the result. Please replace the displayed equality with this (or an equivalent) argument.
[§4.1, Theorem 4.1 (breakage case)] The breakage proof uses the factor w/v_{i-1/2} and the inequality w/v_{i-1/2} ≥ 1. For i=1, v_{1/2}=0, so this step is undefined; this is not a harmless edge case because the theorem is stated for every cell. The displayed equality immediately before this step is also not justified. The first cell can be handled separately: parents in the first cell fragment only into daughters in the first cell, so the birth term cancels the death term and the first moment only increases. For i>1, a valid argument can use the mass-conservation identity (2.3) to bound ∫_0^{v_{i-1/2}} u p(u,w) du ≤ w. Please rewrite the proof accordingly.
[§3.3, Proposition 3.1] The proof of Proposition 3.1 concludes by relying on the 'u–w symmetry of the integrand, the partition T, and the quadrature rule'. The quadrature rules of [57] are D3-symmetric on an equilateral reference triangle, but the triangles in T are arbitrary non-equilateral physical triangles after the affine mapping, and a D3-symmetric rule does not automatically remain invariant under (u,w)↦(w,u) on such a triangle. Please state explicitly how the global u–w symmetry of the full quadrature set is guaranteed (for example, by pairing each triangle with its reflection and using compatible rules on each pair), because the exact mass cancellation in (3.27) depends on this property.

minor comments (5)

[Eqs. (3.6)–(3.7)] The displayed quadrature approximations contain a spurious minus sign; the same integrals appear with a plus sign in (3.4) and in the component form (3.30). Please remove the minus signs.
[Figure 3.4 caption] The caption reads 'breakge region' and should read 'breakage region'.
[§4, limiter definition] The symbol m denotes both the time level (dropped at the start of the proof of Theorem 4.1) and the limiter shift in (4.2)–(4.3); please use different symbols to avoid confusion.
[Reference [57]] The first author of reference [57] is listed as 'S. Wandzurat'; the standard spelling is 'S. Wandzura'. Please verify.
[§5.1, time-step choice] The practical time step Δt = 10Δv is much larger than the theoretical CFL restriction in Theorem 4.1; since Remark 4.1 introduces a fallback halving strategy, a sentence connecting the practical choice to the theory would improve clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the conservation and positivity results are derived from the discretization and stated CFL conditions, not from fitted data or load-bearing self-citation.

full rationale

I walked the derivation chain and found no step in which a claimed prediction or first-principles result reduces by construction to its own inputs. The exact-conservation results in Propositions 3.1 and 3.2 are design consequences: the common-refinement quadrature is deliberately constructed so that birth and death contributions cancel when summed over the full domain. The proof rests on the stated u–w symmetry of the partition and quadrature, and the paper explicitly describes this as a construction rather than as an empirical prediction. No fitted parameter is renamed as a prediction. The positivity-preserving limiter (4.2)–(4.3) is likewise explicit about its construction: the scaling factor θ is defined precisely so that the s-th moment is preserved, so Theorem 4.2's conservation property is an identity following from the definition of θ, not a circular derivation of an independent claim. The positivity property follows directly from θ ≥ 0 and the definition of m. Self-citations [59, 60, 61] appear only for auxiliary extensions (Lax–Wendroff, implicit time marching, exponential time differencing) and are not load-bearing for the central conservation or positivity claims. The central external ingredients are the symmetric triangle quadrature rules of [57], the aggregation-map idea of [34, 42, 52], and the Zhang–Shu framework [63], none of which are self-citations. I did note two non-circular correctness concerns. First, the proof of Theorem 4.1 contains an asserted equality in which “1/v_max − v_{i+1/2}” is dimensionally inconsistent, and the breakage case divides by v_{1/2} = 0 on the first cell; as typeset, the proof does not establish the advertised first-moment positivity guarantee. Second, Proposition 3.1 assumes that the D3-symmetric quadrature rules of [57] preserve the u–w swap after affine mapping to non-equilateral physical triangles, which is asserted but not demonstrated. These are proof-gap and missing-support issues, not circularity, because the claimed results are not equivalent to their inputs by construction. The paper is otherwise self-contained with respect to its theoretical claims, and the numerical benchmarks are external analytical solutions rather than data fitted from the scheme itself.

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

The method introduces no fitted parameters and no new physical entities. It relies on standard mathematical assumptions about the PBE and on the stated symmetry and exactness properties of the quadrature rules, plus the SSP-RK stage structure.

assumptions (4)

domain assumption Kernels are bounded or mildly growing so all PBE integrals remain finite and no gelation occurs.
Section 2, in the remark after Proposition 2.2, restricts attention to bounded or mildly growing kernels; the conservation proof requires finite integrals.
domain assumption The common refinement T can be constructed by constrained Delaunay triangulation enforcing u-w symmetry, and the D3-symmetric quadrature rules of [57] can be mapped so that QT[f(u,w)] = QT[f(w,u)].
Needed for exact cancellation in Proposition 3.1; asserted in Section 3.3 but not demonstrated for arbitrary non-equilateral triangles.
standard math The Gauss-Lobatto and D3-symmetric triangle quadrature rules are exact for the polynomial degrees appearing in the moment identities.
Used in the positivity proofs of Theorem 4.1 and in the accuracy proof of Theorem 4.2.
standard math SSP-RK time integration is a convex combination of forward Euler stages, so positivity and conservation proven for forward Euler transfer to the full time discretization.
Stated in Section 3.6 and invoked in Section 4 before Theorem 4.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Conservative and Positivity-Preserving Discontinuous Galerkin Method for the Population Balance Equation." pith.science (2026). https://pith.science/paper/KQ4WD6W7

@misc{pith2026250716631,
  author       = {Pith},
  title        = {Pith review of: A Conservative and Positivity-Preserving Discontinuous Galerkin Method for the Population Balance Equation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KQ4WD6W7}},
  note         = {Machine review of arXiv:2507.16631}
}

read the original abstract

We develop a conservative, positivity-preserving discontinuous Galerkin (DG) method for the population balance equation (PBE), which models the distribution of particle numbers across particle sizes due to growth, nucleation, aggregation, and breakage. To ensure number conservation in growth and mass conservation in aggregation and breakage, we design a DG scheme that applies standard treatment for growth and nucleation, and introduces a novel discretization for aggregation and breakage. The birth and death terms are discretized in a symmetric double-integral form, evaluated using a common refinement of the integration domain and carefully selected quadrature rules. Beyond conservation, we focus on preserving the positivity of the number density in aggregation-breakage. Since local mass corresponds to the first moment, the classical Zhang-Shu limiter, which preserves the zeroth moment (cell average), is not directly applicable. We address this by proving the positivity of the first moment on each cell and constructing a moment-conserving limiter that enforces nonnegativity across the domain. To our knowledge, this is the first work to develop a positivity-preserving algorithm that conserves a prescribed moment. Numerical results verify the accuracy, conservation, and robustness of the proposed method.

Figures

Figures reproduced from arXiv: 2507.16631 by the authors.

**Figure 3.1.** Integration regions on Ωa used in the DG formulation (3.9) for aggregation. Since the aggregation kernel β(w, u) is generally non-polynomial (see (5.19)–(5.21)), the integrals on the right-hand side of (3.9) are evaluated using numerical quadrature in practice, which introduces truncation errors. This gives rise to an issue of mass conservation, as the birth and death terms may not cancel exactly when summed over t… view at source ↗

**Figure 3.2.** Example of common refinements for the partitions (3.10), (3.11), and (3.13) of the aggregation region. T 0 is the coarsest common refinement, and T is a refinement obtained using constrained Delaunay triangulation (CDT) applied to T 0 with u–w symmetry, which is quadraturefriendly. adopt the quadrature rules developed in [57], which exhibit D3-symmetry, have positive weights, and include no points outside the trian… view at source ↗

**Figure 3.3.** Integration regions on Ωb used in the DG formulation (3.19) for breakage. are the integration regions for the birth and (partial) death terms of breakage, respectively (see an illustration in [PITH_FULL_IMAGE:figures/full_fig_p011_3_3.png] view at source ↗

Figures from the paper (8 more)

**Figure 3.4.** Figure 3.4: The coarsest common refinement E 0 for the partitions (3.20) and (3.21) of the breakge region Ωb . We denote by E 0 = {Ei,j = Ci ∩ Dj} L i,j=1 the coarsest common refinement of the partitions (3.20) and (3.21) for the breakage region Ωb ; see [PITH_FULL_IMAGE:figure…

**Figure 5.1.** Figure 5.1: Each finer mesh at level i is obtained by equally splitting each cell of the level i − 1 mesh into two, for i = 1, 2, 3. The results are presented in [PITH_FULL_IMAGE:figures/full_fig_p020_5_1.png]

**Figure 5.2.** Figure 5.2: Example 2: Aggregation–breakage. Left: numerical and exact number densities; right: deviation of the total mass over time. A nonuniform mesh with L = 15 cells is used. Level ||n − nh||L1 Order ||n − nh||L2 Order ||n − nh||L∞ Order 0 6.08 × 10−4 – 5.19 × 10−4 – 8.18 ×…

**Figure 5.3.** Figure 5.3: Example 3: Aggregation–growth. Left: numerical and exact number densities; middle: evolution of the total number over time; right: evolution of the total mass over time. A nonuniform mesh with L = 15 cells is used. 25 [PITH_FULL_IMAGE:figures/full_fig_p025_5_3.png]

**Figure 5.4.** Figure 5.4: Example 4: Physical kernels. Left: numerical and reference number densities; right: deviation of the total mass over time. A nonuniform mesh with L = 15 cells is used. 28 [PITH_FULL_IMAGE:figures/full_fig_p028_5_4.png]

**Figure 5.5.** Figure 5.5: Example 5: Positivity test, Case I. Top row: DG solution (∆t = 0.003) without limiter; Bottom row: DG solution (∆t = 0.02) with limiter. 0 1 2 3 4 5 6 7 8 9 10-10 100 1010 1020 1030 1040 1050 1060 With limiter No Limiter (a) Case II: L 1 -error over time 0 1 2 3 4 5 …

**Figure 5.6.** Figure 5.6: Example 5: Positivity test, Cases II and III. [PITH_FULL_IMAGE:figures/full_fig_p030_5_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 64 canonical work pages

[31]

H. Liu, R. Gr¨ opler, and G. Warnecke, A high-order posit ivity-preserving DG method for coagulation–fragmentation equations, SIAM Journal on Scientiﬁc Computing , 41 (2019), B448–B465

work page 2019
[57]

Wandzurat and H

S. Wandzurat and H. Xiao, Symmetric quadrature rules on a triangle, Computers & Mathe- matics with Applications , 45 (2003), 1829–1840

work page 2003
[1]

Ascher, S.J

U.M. Ascher, S.J. Ruuth and R.J. Spiteri, Implicit-expl icit Runge-Kutta methods for time- dependent partial diﬀerential equations, Applied Numerica l Mathematics, 25 (1997), pp.151- 167. 29 10-3 10-2 10-1 100 101 102 103 -0.5 0 0.5 1 1.5 2 Numerical Solution Exact Solution (a) nh(v, t) at t = 5.04 10-4 10-3 10-2 10-1 100 101 102 103 -400 -300 -200 -100 0...

work page 1997
[2]

Attarakih, C

M.M. Attarakih, C. Drumm and H.J. Bart, Solution of the po pulation balance equation using the sectional quadrature method of moments (SQMOM), Chemical Engineering Science , 64 (2009), pp.742-752

work page 2009
[3]

Bleck, A fast, approximative method for integrating t he stochastic coalescence equation, Journal of Geophysical Research , 75 (1970), 5165-5171

R. Bleck, A fast, approximative method for integrating t he stochastic coalescence equation, Journal of Geophysical Research , 75 (1970), 5165-5171

work page 1970
[4]

Bourgade and F

J.P. Bourgade and F. Filbet, Convergence of a ﬁnite volum e scheme for coagulation- fragmentation equations, Mathematics of Computation , 77 (2008), 851-882

work page 2008
[5]

Cheng and C.-W

Y. Cheng and C.-W. Shu, A discontinuous Galerkin ﬁnite el ement method for directly solving the Hamilton–Jacobi equations, Journal of Computational Physics , 223 (2007), 398–415

work page 2007
[6]

Cheng and C.-W

Y. Cheng and C.-W. Shu, A discontinuous Galerkin ﬁnite el ement method for time-dependent partial diﬀerential equations with higher-order derivativ es, Mathematics of Computation , 77 (2008), 699–730

work page 2008

Show all 64 references

[7]

Chew, Constrained Delaunay triangulations, Proce edings of the Third Annual Symposium on Computational Geometry, (1987), 215–222

L.P. Chew, Constrained Delaunay triangulations, Proce edings of the Third Annual Symposium on Computational Geometry, (1987), 215–222

1987
[8]

Cockburn, S

B. Cockburn, S. Hou, and C.-W. Shu, The Runge–Kutta local projection discontinuous Galerkin ﬁnite element method for conservation laws IV: the multidimensional case, Math- ematics of Computation , 54 (1990), 545–581

1990
[9]

Cockburn, F

B. Cockburn, F. Li, and C.-W. Shu, Locally divergence-fr ee discontinuous Galerkin methods for the Maxwell equations, Journal of Computational Physics , 194 (2004), 588–610

2004
[10]

Cockburn and C.-W

B. Cockburn and C.-W. Shu, TVB Runge–Kutta local projec tion discontinuous Galerkin ﬁnite element method for conservation laws II: general framework , Mathematics of Computation , 52 (1989), 411–435

1989
[11]

Cockburn and C.-W

B. Cockburn and C.-W. Shu, The Runge–Kutta local projec tion-discontinuous-Galerkin ﬁnite element method for scalar conservation laws, ESAIM: Mathematical Modelling and Numerical Analysis, 25 (1991), 337–361

1991
[12]

Cockburn and C.-W

B. Cockburn and C.-W. Shu, The local discontinuous Gale rkin method for time-dependent convection–diﬀusion systems, SIAM Journal on Numerical Analysis , 35 (1998), 2440–2463

1998
[13]

Diemer and J.H

R.B. Diemer and J.H. Olson, A moment methodology for coa gulation and breakage problems: Part 2—moment models and distribution reconstruction, Chemical Engineering Science , 57 (2002), pp.2211-2228

2002
[14]

R. L. Drake, A general mathematical survey of the coagul ation equation, Topics in Current Aerosol Research, Pergamon Press, Oxford, 1972, 201–376

1972
[15]

Filbet and P

F. Filbet and P. Lauren¸ cot, Numerical simulation of th e Smoluchowski coagulation equation, SIAM Journal on Scientiﬁc Computing , 25 (2004), 2004-2028

2004
[16]

S. K. Friedlander, Smoke, Dust, and Haze: Fundamentals of Aerosol Dynamics , 2nd ed., Oxford University Press, Oxford, 2000

2000
[17]

Gabriel and L.M

P. Gabriel and L.M. Tine, High-order WENO scheme for pol ymerization-type equations, ESAIM: Proceedings, 30 (2010), 53-69. 31

2010
[18]

Gelbard and J.H

F. Gelbard and J.H. Seinfeld, Numerical solution of the dynamic equation for particulate systems, Journal of Computational Physics , 28 (1978), 357-375

1978
[19]

Gottlieb, D.I

S. Gottlieb, D.I. Ketcheson, and C.-W. Shu, High-order strong-stability-preserving time dis- cretizations, Journal of Scientiﬁc Computing , 38 (2009), 251–289

2009
[20]

Gottlieb, C.-W

S. Gottlieb, C.-W. Shu, and E. Tadmor, Strong stability preserving high-order time discretiza- tion methods, SIAM Review, 43 (2001), 89–112

2001
[21]

Hasseine and H.J

A. Hasseine and H.J. Bart, Adomian decomposition metho d solution of population balance equations for aggregation, nucleation, growth and breakup processes, Applied Mathematical Modelling, 39 (2015), 1975-1984

2015
[22]

Hidy and J.R

G.M. Hidy and J.R. Brock, The dynamics of aerocolloidal systems, Pergamon (1970)

1970
[23]

Ketcheson, C.B

D.I. Ketcheson, C.B. Macdonald and S. Gottlieb, Optima l implicit strong stability preserving Runge–Kutta methods, Applied Numerical Mathematics , 59 (2009), 373-392

2009
[24]

Kolewe, S.C

M.E. Kolewe, S.C. Roberts and M.A. Henson, A population balance equation model of ag- gregation dynamics in Taxus suspension cell cultures, Biotechnology and bioengineering , 109 (2012), 472-482

2012
[25]

Kumar and D

S. Kumar and D. Ramkrishna, On the solution of populatio n balance equations by discretiza- tion—I. A ﬁxed pivot technique, Chemical Engineering Science , 51 (1996), 1311-1332

1996
[26]

Kumar and D

S. Kumar and D. Ramkrishna, On the solution of populatio n balance equations by discretiza- tion—II. A moving pivot technique, Chemical Engineering Science , 51 (1996), 1333-1342

1996
[27]

Kumar and D

S. Kumar and D. Ramkrishna, On the solution of populatio n balance equations by discretiza- tion—III. Nucleation, growth and aggregation of particles , Chemical Engineering Science , 52 (1997), 4659-4679

1997
[28]

P.L.C. Lage, Comments on the” An analytical solution to the population balance equation with coalescence and breakage-the special case with constant nu mber of particles” by DP Patil and JRG Andrews [Chemical Engineering Science 53 (3) 599-601], Chemical Engineering Science , ...

2002
[29]

Leyvraz and H.R

F. Leyvraz and H.R. Tschudi, Singularities in the kinet ics of coagulation processes, Journal of Physics A: Mathematical and General , 14 (1981), 3389

1981
[30]

Lim, J.M

Y.I. Lim, J.M. Le Lann, X.M. Meyer, X. Joulia, G. Lee and E .S. Yoon, On the solution of population balance equations (PBE) with accurate front t racking methods in practical crystallization processes, Chemical Engineering Science , 57 (2002), 3715-3732

2002
[32]

Liu and X

C. Liu and X. Zhang, A positivity-preserving implicit- explicit scheme with high order polyno- mial basis for compressible Navier–Stokes equations, Journal of Computational Physics , 493 (2023), 112496. 32

2023
[33]

Liu and S

X.D. Liu and S. Osher, Non-oscillatory high-order accu rate self-similar maximum-principle- satisfying shock-capturing schemes I, SIAM Journal on Numerical Analysis , 33 (1996), 760–779

1996
[34]

Liu and S

A. Liu and S. Rigopoulos, A conservative method for nume rical solution of the population balance equation, and application to soot formation, Combustion and Flame , 205 (2019), 506- 521

2019
[35]

Lombart and G

M. Lombart and G. Laibe, Grain growth for astrophysics w ith discontinuous Galerkin schemes, Monthly Notices of the Royal Astronomical Society , 501 (2021), 4298-4316

2021
[36]

Lombart, Grain growth by the Galerkin method for the f ormation of planets, Doctoral dissertation, Universit´ e de Lyon, (2020)

M. Lombart, Grain growth by the Galerkin method for the f ormation of planets, Doctoral dissertation, Universit´ e de Lyon, (2020)

2020
[37]

Maisels, F.E

A. Maisels, F.E. Kruis and H. Fissan, Direct simulation Monte Carlo for simultaneous nucle- ation, coagulation, and surface growth in dispersed system s, Chemical Engineering Science , 59 (2004), 2231-2239

2004
[38]

Marchisio, J.T

D.L. Marchisio, J.T. Pikturna, R.O. Fox, R.D. Vigil and A.A. Barresi, Quadrature method of moments for population-balance equations, AIChE Journal , 49 (2003), 1266-1276

2003
[39]

McLeod, On an inﬁnite set of non-linear diﬀerential equations, The Quarterly Journal of Mathematics, 13 (1962), pp.119-128

J.B. McLeod, On an inﬁnite set of non-linear diﬀerential equations, The Quarterly Journal of Mathematics, 13 (1962), pp.119-128

1962
[40]

Melzak, A scalar transport equation, Transactions of the American Mathematical Society , 85 (1957), 547-560

Z. Melzak, A scalar transport equation, Transactions of the American Mathematical Society , 85 (1957), 547-560

1957
[41]

Myerson, Handbook of industrial crystallization, B utterworth-Heinemann, (2002)

A. Myerson, Handbook of industrial crystallization, B utterworth-Heinemann, (2002)

2002
[42]

O’Sullivan and S

D. O’Sullivan and S. Rigopoulos, A conservative ﬁnite v olume method for the population bal- ance equation with aggregation, fragmentation, nucleatio n and growth, Chemical Engineering Science, 263 (2022), 117925

2022
[43]

Patil and J.R.G

D.P. Patil and J.R.G. Andrews, An analytical solution t o continuous population balance model describing ﬂoc coalescence and breakage—a special case, Chemical Engineering Science , 53 (1998), 599-601

1998
[44]

Patterson, W

R.I. Patterson, W. Wagner and M. Kraft, Stochastic weig hted particle methods for population balance equations, Journal of Computational Physics , 230 (2011), 7456-7472

2011
[45]

Qamar and G

S. Qamar and G. Warnecke, Numerical solution of populat ion balance equations for nucleation, growth and aggregation processes, Computers & Chemical Engineering , 31 (2007), 1576-1589

2007
[46]

Ramabhadran, T.W

T.E. Ramabhadran, T.W. Peterson and J.H. Seinfeld, Dyn amics of aerosol coagulation and condensation, AIChE Journal , 22 (1976), 840-851

1976
[47]

Reed and T

W.H. Reed and T. Hill, Triangular mesh methods for the ne utron transport equation, Technical Report, Los Alamos Scientiﬁc Laboratory, 1973

1973
[48]

Rigopoulos, Population balance of particles in ﬂows : From aerosols to crystallisation, Cam- bridge University Press, 2024

S. Rigopoulos, Population balance of particles in ﬂows : From aerosols to crystallisation, Cam- bridge University Press, 2024

2024
[49]

Rigopoulos and A.G

S. Rigopoulos and A.G. Jones, Finite-element scheme fo r solution of the dynamic population balance equation, AIChE Journal , 49 (2003), 1127-1139. 33

2003
[50]

Rivi` ere, Discontinuous Galerkin Methods for Solvi ng Elliptic and Parabolic Equations: Theory and Implementation, SIAM, 2008

B. Rivi` ere, Discontinuous Galerkin Methods for Solvi ng Elliptic and Parabolic Equations: Theory and Implementation, SIAM, 2008

2008
[51]

Scott, Analytic studies of cloud droplet coalesce nce I, Journal of the Atmospheric Sci- ences, 25 (1968), 54–65

W.T. Scott, Analytic studies of cloud droplet coalesce nce I, Journal of the Atmospheric Sci- ences, 25 (1968), 54–65

1968
[52]

F. Sewerin, An eﬃcient implementation of a conservativ e ﬁnite volume scheme with constant and linear reconstructions for solving the coagulation equ ation, Chemical Engineering Science, 280 (2023), 119020

2023
[53]

Shu, Total-Variation-Diminishing time discret izations, SIAM Journal on Scientiﬁc and Statistical Computing, 9 (1988), 1073–1084

C.-W. Shu, Total-Variation-Diminishing time discret izations, SIAM Journal on Scientiﬁc and Statistical Computing, 9 (1988), 1073–1084

1988
[54]

Shu, Discontinuous Galerkin methods: general ap proach and stability, in Numerical Solutions of Partial Diﬀerential Equations , 201 (2009), 1–44

C.-W. Shu, Discontinuous Galerkin methods: general ap proach and stability, in Numerical Solutions of Partial Diﬀerential Equations , 201 (2009), 1–44

2009
[55]

Shu and S

C.-W. Shu and S. Osher, Eﬃcient implementation of essen tially non-oscillatory shock-capturing schemes, Journal of Computational Physics , 77 (1988), 439–471

1988
[56]

von Smoluchowski, Versuch einer mathematischen The orie der Koagulationskinetik kolloi- der L¨ osungen,Zeitschrift f¨ ur physikalische Chemie, 92 (1918), 129–168

M. von Smoluchowski, Versuch einer mathematischen The orie der Koagulationskinetik kolloi- der L¨ osungen,Zeitschrift f¨ ur physikalische Chemie, 92 (1918), 129–168

1918
[58]

Xu and C.-W

Y. Xu and C.-W. Shu, Local discontinuous Galerkin metho ds for high-order time-dependent partial diﬀerential equations, Communications in Computational Physics , 7 (2010), 1–46

2010
[59]

Xu and C.-W

Z. Xu and C.-W. Shu, Third-order maximum-principle-sa tisfying and positivity-preserving Lax–Wendroﬀ discontinuous Galerkin methods for hyperboli c conservation laws, Journal of Computational Physics , 470 (2022), 111591

2022
[60]

Xu and C.-W

Z. Xu and C.-W. Shu, On the conservation property of posi tivity-preserving discontinuous Galerkin methods for stationary hyperbolic equations, Journal of Computational Physics , 490 (2023), 112304

2023
[61]

Z. Xu, Z. Sun, and Y.-T. Zhang, Stability and time-step c onstraints of exponential time- diﬀerencing Runge–Kutta discontinuous Galerkin methods fo r advection–diﬀusion equations, arXiv:2503.03019 (2025)

2025 arXiv
[62]

Z. Xu, H. Zhao and C. Zheng, Accelerating population bal ance-Monte Carlo simulation for coagulation dynamics from the Markov jump model, stochasti c algorithm and GPU parallel computing, Journal of Computational Physics , 281 (2015), 844-863

2015
[63]

Zhang and C.-W

X. Zhang and C.-W. Shu, On maximum-principle-satisfyi ng high-order schemes for scalar conservation laws, Journal of Computational Physics , 229 (2010), 3091–3120

2010
[64]

Ziﬀ and E.D

R.M. Ziﬀ and E.D. McGrady, The kinetics of cluster fragm entation and depolymerisation, Journal of Physics A: Mathematical and General , 18 (1985), 3027. 34

1985

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [31]

H. Liu, R. Gr¨ opler, and G. Warnecke, A high-order posit ivity-preserving DG method for coagulation–fragmentation equations, SIAM Journal on Scientiﬁc Computing , 41 (2019), B448–B465

work page 2019

[2] [57]

Wandzurat and H

S. Wandzurat and H. Xiao, Symmetric quadrature rules on a triangle, Computers & Mathe- matics with Applications , 45 (2003), 1829–1840

work page 2003

[3] [1]

Ascher, S.J

U.M. Ascher, S.J. Ruuth and R.J. Spiteri, Implicit-expl icit Runge-Kutta methods for time- dependent partial diﬀerential equations, Applied Numerica l Mathematics, 25 (1997), pp.151- 167. 29 10-3 10-2 10-1 100 101 102 103 -0.5 0 0.5 1 1.5 2 Numerical Solution Exact Solution (a) nh(v, t) at t = 5.04 10-4 10-3 10-2 10-1 100 101 102 103 -400 -300 -200 -100 0...

work page 1997

[4] [2]

Attarakih, C

M.M. Attarakih, C. Drumm and H.J. Bart, Solution of the po pulation balance equation using the sectional quadrature method of moments (SQMOM), Chemical Engineering Science , 64 (2009), pp.742-752

work page 2009

[5] [3]

Bleck, A fast, approximative method for integrating t he stochastic coalescence equation, Journal of Geophysical Research , 75 (1970), 5165-5171

R. Bleck, A fast, approximative method for integrating t he stochastic coalescence equation, Journal of Geophysical Research , 75 (1970), 5165-5171

work page 1970

[6] [4]

Bourgade and F

J.P. Bourgade and F. Filbet, Convergence of a ﬁnite volum e scheme for coagulation- fragmentation equations, Mathematics of Computation , 77 (2008), 851-882

work page 2008

[7] [5]

Cheng and C.-W

Y. Cheng and C.-W. Shu, A discontinuous Galerkin ﬁnite el ement method for directly solving the Hamilton–Jacobi equations, Journal of Computational Physics , 223 (2007), 398–415

work page 2007

[8] [6]

Cheng and C.-W

Y. Cheng and C.-W. Shu, A discontinuous Galerkin ﬁnite el ement method for time-dependent partial diﬀerential equations with higher-order derivativ es, Mathematics of Computation , 77 (2008), 699–730

work page 2008