Pith. sign in

REVIEW 3 major objections 6 minor 39 references

High order-accurate solution of scattering integral equations with unbounded solutions at corners

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper shows that scattering integral equations whose density blows up at corners can be solved to high order by folding a corner-grading change of variables into a new unknown, without singular basis functions or a priori singularity…

desk verdict Strong 2D corner-regularized BIE method with credible high-order convergence; the 3D extrapolation is speculative and the regularity argument is asserted rather than proved, but the numerical evidence carries the paper. read the letter →

arxiv 2501.18065 v1 pith:ZNQWYC47 submitted 2025-01-30 math.NA cs.NAphysics.comp-ph

Constantine Sideris , Davit Aslanyan , Oscar P. Bruno This is my paper · ORCID
classification math.NAcs.NAphysics.comp-ph MSC 65R2065N3878A45
keywords electromagneticscatteringacousticintegralequationsdomainswithcornersunboundedsolutionsatchangeofvariablesNyströmmethodhigh-orderaccuracy
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 addresses a known failure of high-order boundary integral solvers: at a corner, the integral-equation density can blow up, and standard high-order discretizations drop to first-order accuracy. The authors claim that in the simplest setting where this happens (2D TM scattering, a Neumann-Helmholtz problem whose density diverges at corners), a regularized combined-field integral equation can be made high-order again by a change of variables whose Jacobian vanishes at the corner, with that vanishing line element folded into the unknown. The new unknown is smooth where the old density is singular, so the Chebyshev-based Nyström discretization converges at observed 8th to 10th order, giving field errors below $10^{-11}$ at distance $10^{-8}$ from the corner for interior angles as small as $0.01$ radians. This matters because the method needs neither the singular exponent nor singular basis functions, and the authors expect the same strategy to extend to 3D Maxwell problems, where unbounded edge currents cannot be avoided by formulation choice.

What carries the argument

Two mechanisms carry the argument. First, the corner-regularized change of unknown $\psi_q(\theta)=\phi(s_q(\theta))\widetilde{L}_q(\theta)$ in equation (44): the change of variables $s(\theta)$ (based on the function $w(\theta)$ of equation (32), with $s(\theta)\sim\theta^p$ as $\theta\to 0$) grades the mesh toward the corner, and multiplying the unbounded density $\phi$ by the vanishing line element $\widetilde{L}_q$ transfers the singularity into the known geometry factor, leaving a smooth unknown for Chebyshev approximation. Second, the precision-preserving quadrature layer: near-corner source-target differences are evaluated through Taylor expansions of the parametrization around the corner to avoid catastrophic cancellation, and the $1/|r-r'|$ peak of the normal-derivative kernel is integrated with adaptive Gauss-Kronrod quadrature rather than fixed Fejér rules. Together these make the regularized equation (44) numerically tractable at the extremely fine mesh spacings the grading produces.

What would settle it

Run the sharpest tested case (the $0.01$-radian parallelogram) with the change-of-variables order reduced to $p=2$ and measure the convergence slope of the field error at distance $d=10^{-8}$ from the corner: if the slope remains high, the smoothness of $\psi$ is not the limiting premise, and if it drops to first or second order, the vanishing Jacobian is what carries the high-order convergence.

Watch

Extended reading notes

Core claim

The paper's central claim is that the infinite-density corner difficulty can be removed by a change of unknown rather than by adding singular basis functions. Starting from the operator-regularized combined field integral equation (CFIE-R), the authors apply a change of variables $s(\theta)$ whose Jacobian and its first $p-1$ derivatives vanish at the corner, and they define a new unknown $\psi_q(\theta)=\phi(s_q(\theta))\widetilde{L}_q(\theta)$, where $\widetilde{L}_q$ is the line element in the new variable. Because $\widetilde{L}_q$ vanishes at the corner, $\psi$ is smooth where the original density $\phi$ blows up, so the Chebyshev-based rectangular-polar Nyström discretization can approximate it accurately. With CoV order $p=4$ and $p=6$, plus precision-preserving treatments for near-corner and self interactions, the computed field converges at observed rates between 8th and 10th order and reaches relative errors below $10^{-11}$ at distance $10^{-8}$ from the corner, for square, parallelogram, and teardrop cross-sections including interior angles down to $0.01$ radians. The method also reproduces the theoretical corner exponent $\nu=-1/3$ for a $90^\circ$ corner to about eight digits, and it stays resonance-free and well-conditioned where the underlying MFIE becomes singular.

Load-bearing premise

The method's accuracy rests on the assumption that the chosen change-of-variables order $p$ makes the product of the density and the Jacobian, $\psi(\theta)=\phi(s(\theta))\widetilde{L}(\theta)$, smooth at the corner; the paper verifies this empirically for the tested angles rather than proving it or giving an automatic rule for $p$.

Editorial extensions

If this is right

  • Field evaluations at distance $d=10^{-8}$ from a corner reach relative errors below $10^{-11}$, with observed convergence between 8th and 10th order, for interior angles as small as $0.01$ radians.
  • The corner-regularized CFIE-R system remains resonance-free and well-conditioned at wavenumbers where the MFIE matrix is singular, and GMRES reaches a $10^{-5}$ residual in fewer than about 15 iterations across the tested wavenumber range.
  • Because the method needs no a priori singularity analysis or singular bases, the same solver handles corners of different angles and curved boundaries (square, parallelogram, teardrop) with essentially the same accuracy.
  • Each ingredient is necessary: omitting the change of unknown makes the scheme worse than the unregularized formulation, and omitting the change of variables reduces convergence to first order with errors above $10^{-2}$ even past 1000 unknowns.

Reading between the lines

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

  • If the same smoothness premise holds for edge singularities, the change-of-unknown mechanism should transfer to 3D Maxwell problems, where the tangential current component diverges at edges in every formulation; the paper presents this as expected future work, and the mechanism does not require knowing the singular exponent.
  • The reported 12-digit error floor is attributed to the Gauss-Kronrod quadrature's accuracy limit in the implementation; replacing it with higher-precision adaptive integration is a direct, testable route to machine-precision corner fields.
  • The text notes that smooth patches dominate the error at coarse resolutions; an error-balancing strategy that concentrates unknowns near corners only as the mesh refines could reduce cost for a given tolerance.
  • If the central claim is right, the practical consequence is that singular-basis expansions—whose exponents for 3D corners are not known in closed form and would require their own numerical solution—can be bypassed entirely.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The manuscript develops a high-order Nyström solver for 2D TM electromagnetic scattering by PEC scatterers with corners, targeting integral equations whose density blows up at corners. The method combines an operator-regularized combined-field integral equation (CFIE-R), a graded change of variables s(θ) of order p, a change of unknown ψ(θ)=φ(s(θ))L̃(θ), precision-preserving treatment of near-corner source-target differences, and adaptive Gauss-Kronrod quadrature for precomputed weights. Numerical results for square, parallelogram, and teardrop cross-sections report relative errors near 10^-11 or below at distance 10^-8 from corners, observed convergence orders 8-10 for interior angles down to 0.01 radians, favorable conditioning and GMRES iteration counts versus MFIE, recovery of a theoretical corner exponent, and agreement with a commercial FDTD solver.

Significance. If the claims hold, the work is significant: it would remove a long-standing barrier for high-order integral-equation solvers on non-smooth geometries, avoiding singular basis functions or geometry-dependent exponents, and it offers a plausible route to 3D edge singularities. The paper has genuine strengths: an honest ablation study (Section 4.1), an exact-solution test for a monopole source (Fig. 9), recovery of the corner exponent as an independent post-hoc check (Fig. 16 right), and comparisons with a commercial solver (Fig. 14). No parameters are fit to the reported errors; the CoV order p is the only user parameter. The main caveat is that the central regularity assumption is asserted rather than proved, and the fixed-p strategy is not justified for arbitrary corner angles.

major comments (3)
  1. [Section 3.2, Eq. (42); Section 3.4] The claim that ψ(θ)=φ(s(θ))L̃(θ) has 'a number of bounded derivatives at the corner point' is not proved, and no rule is given for choosing p for a given geometry. The paper's own estimate ψ_q(θ′)∼(θ′)^{(p−1)−νp} implies that the number of bounded derivatives is controlled by β=p(1−ν)−1; for a near-crack geometry (ν→1/2) and p=6, β≈2, which is not by itself compatible with the 8th–10th order rates reported in Fig. 13. Because this regularity is the mechanism by which the Chebyshev/RP discretization is claimed to converge at high order, the central assertion 'without requiring a priori analysis of the geometry' is not yet established. Please provide a theorem or rigorous a priori bound connecting p, ν, and achievable order, or an automatic p-selection rule, or explicitly narrow the claim to the tested angles and p values.
  2. [Eq. (44)] The main regularized equation contains index errors that make it ambiguous as printed: in the first source integral the unknown is written ψ_q(θ′) although the integration is over patch q′, so the displayed equation appears to couple every target point to the single unknown ψ_q(θ). The regularization terms contain similar notational inconsistencies between the target and source patch indices. Please correct the indices and verify that the discretized system uses ψ_{q′} in all source-patch integrals; as written, Eq. (44) cannot be implemented unambiguously.
  3. [Section 4.3, Section 4.4] The plane-wave convergence studies for the parallelogram, needle, and teardrop geometries use reference solutions computed by the same method on finer grids. The only fully independent accuracy tests are the monopole exact solution for the square cross-section (Fig. 9) and the FDTD field comparisons for the square at k=10 (Fig. 14). The reported 8th–10th order rates for the small-angle cases and the density-error results near corners therefore rest on self-convergence. Please provide an independent reference for at least the small-angle cases, or clearly state that the reported rates for those cases are self-convergence estimates.
minor comments (6)
  1. [Eq. (21)] Equation (21) gives the same expression for the tangent vector τ and the normal vector n; the normal should be a 90-degree rotation of the tangent, e.g., n=(−dy/ds, dx/ds)/L.
  2. [Section 3.4 versus Section 4.4] The sign convention for the singularity exponent ν is inconsistent: Section 3.4 uses ν as a positive exponent in ψ_q(θ′)∼(θ′)^{(p−1)−νp}, while Section 4.4 states that φ(d)∼d^ν with ν=−1/3 for the same 90-degree corner. Please clarify the sign convention so the two statements are compatible.
  3. [References] References [24] and [25] are the same paper and should be merged.
  4. [Abstract and Remark 1] The abstract's phrase 'near machine precision accuracy' is overstated given Remark 1, which states that the adaptive Gauss-Kronrod integration enforces a 10^-12 accuracy floor in the Matlab implementation; the reported errors near 10^-11 are not machine precision in double arithmetic.
  5. [Section 4.1 versus Section 4.3] Section 4.1 describes the original formulation as achieving second-order convergence, while Section 4.3 (Fig. 10) describes the no-CoV method as exhibiting first-order convergence; these statements should be reconciled.
  6. [Section 4.3 discussion of CoV (32)] The text attributes the superiority of CoV (32) over (34) to an 'evenly split distribution' in Section 3.2 but later attributes it to an 'uneven distribution' in Section 4.3; the wording should be harmonized.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the corner-regularized R-CFIE solver is validated against manufactured exact solutions, fine-mesh references, and an independent FDTD solver, not fitted to its own outputs.

full rationale

The paper's central numerical claim is not circular. The algorithm contains no parameters fitted to the reported error values; the high-order convergence is measured, not imposed. The key construction, ψ_q(θ)=φ(s_q(θ))L̃_q(θ) (Eq. 42), is a change of unknown, and the assertion that ψ is smooth at corners is a regularity assumption whose validity depends on the CoV order p relative to the corner singularity exponent—an unproved point that is a correctness risk, not a circular reduction. The corner-exponent extraction in Section 4.4 is a post-hoc consistency check against the known exponent ν=−1/3, not an input to the solver. The self-cited references ([11] for the operator-regularized CFIE-R and [14] for the corner exponent) are independent prior results and are not load-bearing for the novel corner-regularization contribution; the method is also benchmarked against a manufactured point-source solution (Section 4.3, Figure 9) and a commercial FDTD solver (Section 4.3, Figure 14). No equation or fitted parameter reduces by construction to the claimed outcomes, so the score is 0.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard integral identities, the prior CFIE-R operator regularization [11], and an empirically validated regularity assumption about the corner singularity being cancelled by the chosen CoV order p. No new physical entities are posited.

free parameters (1)
  • CoV order p (Eqs. 32, 35) = 4 and 6
    Hand-chosen polynomial order in the grading change of variables. The method assumes p is large enough that ψ = φ L̃ is smooth at the corner; no automatic selection rule is provided, so this parameter is tuned by the user per experiment.
assumptions (5)
  • domain assumption The CFIE-R regularized combined field integral equation (31) is uniquely solvable and well-conditioned for all wavenumbers.
    Section 3.1 adopts the operator-regularized formulation from [11]; the resonance-free and conditioning claims in Section 4.2 inherit this prior result, which is cited but not re-derived here.
  • standard math Identity (8): the hypersingular normal-derivative operator equals k^2 ∫ G (n·n) φ + ∂τ ∫ G ∂τ φ for closed curves.
    Section 2.2 relies on this tangential form to avoid normal derivatives; standard vector calculus identity cited to [11].
  • domain assumption The corner singularity of φ is a power law d^{-ν} and the CoV order p is large enough that ψ = φ L̃ has bounded derivatives at the corner.
    Section 3.2, after Eq. (44): the method's high order depends on the regularized unknown being smooth. The paper validates p=4 and p=6 for the tested angles and recovers ν=-1/3 for the 90-degree corner, but provides no general proof.
  • domain assumption Adaptive Gauss-Kronrod quadrature computes the precomputed weights (48) accurately enough to support the claimed convergence.
    Section 3.4 and Remark 1: the overall accuracy floor is set by Matlab's GK error limit of 10^-12; the near-corner kernel peaks are handled by GK rather than by a dedicated singularity-cancellation change of variables.
  • ad hoc to paper The proposed 2D corner-regularization strategy generalizes directly to 3D edge singularities.
    Stated in the Abstract and Introduction, but no 3D formulation, algorithm, or experiment is provided. This is an extrapolation that the paper does not substantiate.

how reviews work

0 comments
Cite this review

Pith. "Pith review of High order-accurate solution of scattering integral equations with unbounded solutions at corners." pith.science (2026). https://pith.science/paper/ZNQWYC47

@misc{pith2026250118065,
  author       = {Pith},
  title        = {Pith review of: High order-accurate solution of scattering integral equations with unbounded solutions at corners},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZNQWYC47}},
  note         = {Machine review of arXiv:2501.18065}
}
read the original abstract

Although high-order Maxwell integral equation solvers provide significant advantages in terms of speed and accuracy over corresponding low-order integral methods, their performance significantly degrades in presence of non-smooth geometries--owing to field enhancement and singularities that arise at sharp edges and corners which, if left untreated, give rise to significant accuracy losses. The problem is particularly challenging in cases in which the "density" (i.e., the solution of the integral equation) tends to infinity at corners and edges--a difficulty that can be bypassed for 2D configurations, but which is unavoidable in 3D Maxwell integral formulations, wherein the component tangential to an edge of the electrical-current integral density vector tends to infinity at the edge. In order to tackle the problem this paper restricts attention to the simplest context in which the unbounded-density difficulty arises, namely, integral formulations in 2D space whose integral density blows up at corners; the strategies proposed, however, generalize directly to the 3D context. The novel methodologies presented in this paper yield high-order convergence for such challenging equations and achieve highly accurate solutions (even near edges and corners) without requiring a priori analysis of the geometry or use of singular bases.

Figures

Figures reproduced from arXiv: 2501.18065 by the authors.

Figure 1
Figure 1. Partitioning of the domain boundary Γ as the union (15) of nonoverlapping patches Γ [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the graded-mesh Changes of Variables (CoV) in equations (32) and (34) of order [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. The blue crosses display the magnitude of the kernel [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: The real and imaginary parts of the quantities [PITH_FULL_IMAGE:figures/full_fig_p015_4.png]
Figure 5
Figure 5. Figure 5: Plane wave incident upon a PEC cylinder of square cross-section. [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]
Figure 6
Figure 6. Figure 6: Error ε (equation (51)) as a function of N (equation (19)) in the solution u at distances of d = 10−8 (left) and d = 1 (right) from a corner point, obtained by means of implementations with key improvements omitted, as described in Section 4.1, compared to results prod…
Figure 7
Figure 7. Figure 7: Eigenvalue distributions for the left hand-side operators in the “change-of-unknown” regularized [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]
Figure 8
Figure 8. Figure 8: Condition number κ2 (left) and number of GMRES iterations required to reach 10−5 residual (right) as functions of the wavenumber k for a PEC cylinder with square cross-section of side 2. 4.3 Field evaluation accuracy In this section, the accuracy of the proposed method…
Figure 9
Figure 9. Figure 9: Convergence of the field u at a distance 10−8 away from a corner of a PEC cylinder of square cross-section with point-source excitation. For the next example, the scattering from the same square cross-section cylinder is considered with the planewave incident excitatio…
Figure 10
Figure 10. Figure 10: Convergence of the field u for PEC cylinder of square cross-section for the non-resonant frequency k = 10 (left) and the resonant frequency k = 9.5548 (right) at a distance 10−8 away from a corner [PITH_FULL_IMAGE:figures/full_fig_p021_10.png]
Figure 11
Figure 11. Figure 11: Convergence of the field u for PEC cylinder with a parellogram-shaped cross-section at a point 10−8 away from a corner. Left: Acute angle. Right: Obtuse angle. is considered next where the parameter α controls the angle of the corner; α values resulting interior angle…
Figure 12
Figure 12. Figure 12: Left: Convergence of the field u for PEC cylinder with a teardrop-shaped cross-section at a point 10−8 away from the corner. Right: comparison of the algorithm’s performance for the two CoVs considered [PITH_FULL_IMAGE:figures/full_fig_p022_12.png]
Figure 13
Figure 13. Figure 13: Convergence of the field u for PEC cylinder with teardrop-shaped (left) and parellogram-shaped “needle” (right) at points 10−8 away from a corner. (For the right image the observation point lies at a distance d = 10−8 of the needle corner point with acute angle 0.57◦ …
Figure 14
Figure 14. Figure 14: Total field magnitude plots for Ex, Ey and Hz field components for a cylinder with square cross-section. Upper row: Fine-mesh results produced by the commercial solver Lumerical. Lower row: Results produced by the proposed solver. 4.4 Density errors and density corner…
Figure 15
Figure 15. Figure 15: Total field magnitude plots for Ex, Ey and Hz field components for cylinders with a parallelogram and teardrop shaped cross-sections −1/3 as d → 0: the precision of the numerical solution ψ suffices to produce the singular exponent ν = νe(0) in the density ϕ with an a…
Figure 16
Figure 16. Figure 16: Left: Absolute value εabs of the error in the evaluation of the unknown ψ for a PEC scatterer with a square cross-section. Right: Error in the numerical asymptotics of the corner exponent vs. an exact result near a corner point. 24 [PITH_FULL_IMAGE:figures/full_fig_p…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [1]

    S., and Turc, C

    Anand, A., Ovall, J. S., and Turc, C. Well-conditioned boundary integral equations for two- dimensional sound-hard scattering problems in domains with corners. The Journal of Integral Equa- tions and Applications (2012), 321–358

    work page 2012

  2. [2]

    Atkinson, K. E. The numerical solution of integral equations of the second kind , vol. 4. Cambridge university press, 1997

    work page 1997

  3. [3]

    E., and Graham, I

    Atkinson, K. E., and Graham, I. G. An iterative variant of the Nystr¨ om method for boundary integral equations on nonsmooth boundaries. The mathematics of finite elements and applications, VI (Uxbridge, 1987), London (1988), 297–303

    work page 1988

  4. [4]

    B., and Pitk ¨aranta, J

    Babuˇska, I., Kellogg, R. B., and Pitk ¨aranta, J. Direct and inverse error estimates for finite elements with mesh refinements. Numerische Mathematik 33 , 4 (1979), 447–471

    work page 1979

  5. [5]

    The post-processing approach in the finite element method—part 2: the calculation of stress intensity factors

    Babuˇska, I., and Miller, A. The post-processing approach in the finite element method—part 2: the calculation of stress intensity factors. International Journal for Numerical Methods in Engineering 20, 6 (1984), 1111–1129. 25

    work page 1984

  6. [6]

    Interpolated Factored Green Function

    Bauinger, C., and Bruno, O. P. “Interpolated Factored Green Function” method for accelerated solution of scattering problems. Journal of Computational Physics 430 (2021), 110095

    work page 2021

  7. [7]

    AIM: Adaptive integral method for solving large-scale electromagnetic scattering and radiation problems

    Bleszynski, E., Bleszynski, M., and Jaroszewicz, T. AIM: Adaptive integral method for solving large-scale electromagnetic scattering and radiation problems. Radio Science 31 , 5 (1996), 1225–1251

    work page 1996

  8. [8]

    Boyd, J. P. Chebyshev and Fourier spectral methods . Courier Corporation, 2001

    work page 2001

Show all 39 references
  1. [9]

    A fast direct solver for the integral equations of scattering theory on planar curves with corners

    Bremer, J. A fast direct solver for the integral equations of scattering theory on planar curves with corners. Journal of Computational Physics 231 , 4 (2012), 1879–1899

    2012

  2. [10]

    P., Elling, T., Paffenroth, R., and Turc, C

    Bruno, O. P., Elling, T., Paffenroth, R., and Turc, C. Electromagnetic integral equations requiring small numbers of krylov-subspace iterations. Journal of Computational Physics 228 , 17 (2009), 6169–6183

    2009

  3. [11]

    P., Elling, T., and Turc, C

    Bruno, O. P., Elling, T., and Turc, C. Regularized integral equations and fast high-order solvers for sound-hard acoustic scattering problems. International Journal for Numerical Methods in Engineering 91, 10 (2012), 1045–1072

    2012

  4. [12]

    P., and Garza, E

    Bruno, O. P., and Garza, E. A Chebyshev-based rectangular-polar integral solver for scattering by geometries described by non-overlapping patches. Journal of Computational Physics 421 (2020), 109740

    2020

  5. [13]

    P., and Kunyansky, L

    Bruno, O. P., and Kunyansky, L. A. A fast, high-order algorithm for the solution of surface scattering problems: basic implementation, tests, and applications. Journal of Computational Physics 169, 1 (2001), 80–110

    2001

  6. [14]

    P., Ovall, J

    Bruno, O. P., Ovall, J. S., and Turc, C. A high-order integral algorithm for highly singular pde solutions in lipschitz domains. Computing 84 (2009), 149–181

    2009

  7. [15]

    Y., Gimbutas, Z., Greengard, L

    Cheng, H., Crutchfield, W. Y., Gimbutas, Z., Greengard, L. F., Ethridge, J. F., Huang, J., Rokhlin, V., Yarvin, N., and Zhao, J. A wideband fast multipole method for the Helmholtz equation in three dimensions. Journal of Computational Physics 216 , 1 (2006), 300–325

    2006

  8. [16]

    Inverse acoustic and electromagnetic scattering theory, 1998

    Colton, D. Inverse acoustic and electromagnetic scattering theory, 1998

    1998

  9. [17]

    L., and Fix, G

    Cox, C. L., and Fix, G. J. On the accuracy of least squares methods in the presence of corner singularities. Computers & mathematics with applications 10 , 6 (1984), 463–475

    1984

  10. [18]

    J., and Rabinowitz, P

    Davis, P. J., and Rabinowitz, P. Methods of numerical integration . Courier Corporation, 2007

    2007

  11. [19]

    Demkowicz, L., Devloo, P., and Oden, J. T. On an h-type mesh-refinement strategy based on minimization of interpolation errors. Computer Methods in Applied Mechanics and Engineering 53 , 1 (1985), 67–89

    1985

  12. [20]

    Recursive elements, an inexpensive solution process for resolving point singularities in elliptic problems

    Devloo, P. Recursive elements, an inexpensive solution process for resolving point singularities in elliptic problems. In Proceedings of 2nd World Congress on Computational Mechanics. IACM, Stuttgart (1990), pp. 609–612

    1990

  13. [21]

    J., Gulati, S., and W akoff, G

    Fix, G. J., Gulati, S., and W akoff, G. On the use of singular functions with finite element approximations. Journal of Computational Physics 13 , 2 (1973), 209–228

    1973

  14. [22]

    Ganesh, M., and Graham, I. G. A high-order algorithm for obstacle scattering in three dimensions. Journal of Computational Physics 198 , 1 (2004), 211–242. 26

    2004

  15. [23]

    Givoli, D., Rivkin, L., and Keller, J. B. A finite element method for domains with corners. International journal for numerical methods in engineering 35 , 6 (1992), 1329–1345

    1992

  16. [25]

    High-order methods for linear functionals of solutions of second kind integral equations

    Graham, I., and Chandler, G. High-order methods for linear functionals of solutions of second kind integral equations. SIAM journal on numerical analysis 25 , 5 (1988), 1118–1137

    1988

  17. [26]

    Solving integral equations on piecewise smooth boundaries using the RCIP method: a tutorial

    Helsing, J. Solving integral equations on piecewise smooth boundaries using the RCIP method: a tutorial. In Abstract and applied analysis (2013), vol. 2013, Wiley Online Library, p. 938167

    2013

  18. [27]

    E., and Stephan, E

    Heuer, N., Mellado, M. E., and Stephan, E. P. A p-adaptive algorithm for the bem with the hypersingular operator on the plane screen. International journal for numerical methods in engineering 53, 1 (2002), 85–104

    2002

  19. [28]

    A Chebyshev-based high-order-accurate integral equation solver for Maxwell’s equations

    Hu, J., Garza, E., and Sideris, C. A Chebyshev-based high-order-accurate integral equation solver for Maxwell’s equations. IEEE Transactions on Antennas and Propagation 69 , 9 (2021), 5790–5800

    2021

  20. [29]

    J., and Akin, J

    Hughes, T. J., and Akin, J. Techniques for developing “special” finite element shape functions with particular reference to singularities. International Journal for Numerical Methods in Engineering 15, 5 (1980), 733–751

    1980

  21. [30]

    Linear integral equations, 1989

    Kress, R. Linear integral equations, 1989

    1989

  22. [31]

    A Nystr¨ om method for boundary integral equations in domains with corners.Numerische Mathematik 58 , 1 (1990), 145–161

    Kress, R. A Nystr¨ om method for boundary integral equations in domains with corners.Numerische Mathematik 58 , 1 (1990), 145–161

    1990

  23. [32]

    Singular finite elements for the fracture analysis of v-notched plate

    Lin, K., and Tong, P. Singular finite elements for the fracture analysis of v-notched plate. Inter- national Journal for Numerical Methods in Engineering 15 , 9 (1980), 1343–1354

    1980

  24. [33]

    The hp-version of the boundary element method in R3: The basic approximation results

    Maischak, M., and Stephen, E. The hp-version of the boundary element method in R3: The basic approximation results. Mathematical methods in the applied sciences 20 , 5 (1997), 461–476

    1997

  25. [34]

    Diagonal forms of translation operators for the helmholtz equation in three dimensions

    Rokhlin, V. Diagonal forms of translation operators for the helmholtz equation in three dimensions. Applied and computational harmonic analysis 1 , 1 (1993), 82–93

    1993

  26. [35]

    Families of consistent conforming elements with singular derivative fields

    Stern, M. Families of consistent conforming elements with singular derivative fields. International Journal for Numerical Methods in Engineering 14 , 3 (1979), 409–421

    1979

  27. [36]

    Locally corrected multidimensional quadrature rules for singular functions

    Strain, J. Locally corrected multidimensional quadrature rules for singular functions. SIAM Journal on Scientific Computing 16 , 4 (1995), 992–1017

    1995

  28. [37]

    Numerical treatment of vertex singularities and intensity factors for mixed boundary value problems for the laplace equation in R3

    von Petersdorff, T., Andersson, B., et al. Numerical treatment of vertex singularities and intensity factors for mixed boundary value problems for the laplace equation in R3. SIAM Journal on Numerical Analysis 31 , 5 (1994), 1265

    1994

  29. [38]

    A finite-element method for Laplace-and Helmholtz-type boundary value problems with singularities

    Wu, X., and Han, H. A finite-element method for Laplace-and Helmholtz-type boundary value problems with singularities. SIAM journal on numerical analysis 34 , 3 (1997), 1037–1050

    1997

  30. [39]

    Discrete boundary conditions for elasticity problems with singularities

    Wu, X., and Xue, W. Discrete boundary conditions for elasticity problems with singularities. Computer methods in applied mechanics and engineering 192 , 33-34 (2003), 3777–3795

    2003

  31. [40]

    Yaghjian, A. D. Augmented electric-and magnetic-field integral equations. Radio Science 16 , 06 (1981), 987–1001. 27

    1981

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.