chunk_uid
stringlengths
40
40
chunk_type
stringclasses
2 values
chunk_index
int64
0
6.71k
total_chunks
int64
1
6.71k
section_title
stringlengths
1
157
embed_text
stringlengths
1
83.3k
spans
dict
paper_doi
stringlengths
0
63
paper_id_arxiv
stringlengths
9
16
title
stringlengths
7
245
authors
listlengths
1
768
categories
listlengths
1
7
year
int64
2k
2.02k
language
stringclasses
2 values
discipline
stringclasses
8 values
dense_vector
listlengths
1.02k
1.02k
d4281958079281c89405631f9ca4db59ed09d74d
subsection
54
74
Error analysis
With regularities (REF ), it is straightforward to check that, for any 1 \le n \le N , the weak solution (c,\mu ,{{v}},p) to model problem (REF ) satisfies\big ({\partial _t c}(t^n),\chi \big ) + a_{\mathcal {D}}(\mu ^n,\chi ) + a_{\mathcal {A}}(c^n,{{v}}^n,\chi ) = 0, \quad \forall \chi \in S_h,\\ \big (\Phi _{+}\,\!^...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.047665707767009735, -0.014121194370090961, -0.019438575953245163, -0.014464497566223145, 0.0017699174350127578, -0.011801992543041706, 0.05492847040295601, 0.02894425205886364, 0.02345140650868416, 0.04064706712961197, -0.060177192091941833, -0.02894425205886364, 0.0377175509929657, 0.0...
b6f83d4ab09d40949d438b20f92c4c70eb44028e
subsection
55
74
Error analysis
In addition, there exists a constant C_1>0 independent of mesh size h, such that|(\lambda , \phi )| \le C_1 \Vert \phi \Vert _{\mathrm {DG}} \Vert \mathcal {J}(\lambda )\Vert _{\mathrm {DG}}, && \forall \phi \in H^1(\mathcal {T}_h)\,,\quad \forall \lambda \in M_h.The linearity of the operator \mathcal {J} is easy to ch...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1016/j.cam.2017.04.002", "end": 1720, "openalex_id": "https://openalex.org/W2606223826", "raw": "N. Chaabane, V. Girault, C. Puelz, and B. Riviere, Convergence of IPDG for coupled time-dependent Navier–Stokes and Darcy equations, Journal...
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07306291908025742, 0.009308350272476673, -0.035676922649145126, -0.022355299443006516, 0.004352798219770193, -0.005592640023678541, 0.0015660154167562723, 0.027864012867212296, 0.03564640134572983, 0.031190602108836174, -0.04010220244526863, -0.02021895721554756, 0.04007168486714363, 0....
1dde68fdad71c544e195ffb2b2cdef8010ffc610
subsection
56
74
Error analysis
There is an approximation operator \mathcal {R}_h: H_0^1(\mathcal {T}_h)^d \rightarrow \mathbf {X}_h satisfyingb_\mathcal {P}(\phi ,\mathcal {R}_h({{v}})-{{v}}) = 0,\quad \forall {{v}}\in H_0^1(\mathcal {T}_h)^d, \quad \forall \phi \in Q_h,and for all E in \mathcal {T}_h, for all {{v}} in H_0^1(\mathcal {T}_h)^d \cap W...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03324412181973457, 0.016728905960917473, -0.027764489874243736, 0.02242222987115383, -0.0014042513212189078, -0.006105440203100443, 0.044294968247413635, 0.0014385944232344627, 0.01471411157399416, 0.02942822314798832, -0.04163910448551178, -0.03165670856833458, 0.043073881417512894, 0....
f72bb88905c8011d8fb4b1c4738a55a031a4f03e
subsection
57
74
Error analysis
We also have for all s, 1 \le s \le q+1,\forall {{v}}\in H_0^1(\mathcal {T}_h)^d \cap H^s(\Omega )^d,\quad \Vert \mathcal {R}_h({{v}})-{{v}}\Vert _{\mathrm {DG}} \le C h^{s-1} \vert {{v}}\vert _{H^s(\Omega )}.With the operator \mathcal {R}_h, we have a bound for the form a_\mathcal {C} (see Proposition 6.2 in ). [Bound...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1016/j.cam.2017.04.002", "end": 313, "openalex_id": "https://openalex.org/W2606223826", "raw": "N. Chaabane, V. Girault, C. Puelz, and B. Riviere, Convergence of IPDG for coupled time-dependent Navier–Stokes and Darcy equations, Journal ...
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.029936591163277626, 0.01704341173171997, -0.018050452694296837, -0.02703753300011158, -0.014548694714903831, -0.026686593890190125, -0.00537470169365406, 0.021834487095475197, 0.022887302562594414, -0.014716535806655884, -0.03457508608698845, -0.007430744357407093, 0.045622020959854126, ...
de7a2861e0d0940706f587f4e29c0c9514c9299e
subsection
58
74
Error analysis
There exist a constant C, independent of mesh size h and time step size \tau , such that for all 0 \le n \le N\Vert c^n - \mathcal {P}_hc^n\Vert _{\mathrm {DG}} &\le Ch^q \Vert c\Vert _{L^\infty (0,T;\,H^{q+1}(\Omega ))},\\ \Vert \mu ^n - \mathcal {P}_h\mu ^n\Vert _{\mathrm {DG}} &\le Ch^q \Vert \mu \Vert _{L^\infty (0...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.030302420258522034, -0.01171815674751997, -0.019728614017367363, -0.010932368226349354, -0.0012511572567746043, 0.015227500349283218, 0.008491085842251778, 0.00785787869244814, 0.021223900839686394, 0.009734613820910454, -0.0032251630909740925, -0.05215189978480339, 0.01383139193058014, ...
9aaba2d1fb2620f77698bceff42a8b79959854bb
subsection
59
74
Error analysis
We note that for all n \ge 1a_{\mathcal {D}}(\zeta _\mu ^n,\chi )=0, \quad b_{\mathcal {P}}(\phi ,{\zeta }_{{{v}}}^n)=0, && \forall \chi \in S_h, \quad \forall \phi \in Q_h.Therefore, from (REF ) and (REF ), the error equation becomes, for any \chi \in S_h , \varphi \in S_h , {{\theta }} \in \mathbf {X}_h , and \phi \i...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.012212013825774193, 0.02614789828658104, -0.05891667306423187, -0.05540791526436806, -0.019038842990994453, 0.02839045412838459, 0.012082342058420181, 0.017833659425377846, 0.037162356078624725, 0.013905372470617294, -0.022867968305945396, -0.014607124961912632, -0.0026620817370712757, 0...
5e7b57d4ac485712accf639780468a4ca06a70f9
subsection
60
74
Error analysis
Suppose (c,\mu ,{{v}},p) is a weak solution of (REF ) with regularity (REF ). Then, under as:CHNS:assumptionA and sufficiently small time step size \tau , there exists a constant C independent of mesh size h and time step size \tau such that for any m \ge 1\max _{1\le n\le m}\Big (\Vert \xi _c^n \Vert _{\mathrm {DG}}^2...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03013644739985466, -0.015716729685664177, -0.05358947068452835, -0.011177188716828823, -0.015426809899508953, 0.008224579505622387, 0.03237951546907425, 0.06951982527971268, 0.01898214779794216, 0.011154300533235073, -0.0215456523001194, -0.035736486315727234, 0.012138503603637218, 0.00...
c9a10ceea40b362866b8297bfaee11856f84a53b
subsection
61
74
Error analysis
\end{aligned}Choosing \varphi = \delta _\tau \xi _c^n in () and adding and subtracting the appropriate terms, we obtain\begin{aligned}\kappa a_{\mathcal {D}}(\xi _c^n,\delta _\tau \xi _c^n) - (\xi _\mu ^n,\delta _\tau \xi _c^n) = (\zeta _\mu ^n,\delta _\tau \xi _c^n) + \big (\Phi _+\,\!^{\prime }(c_h^n)-\Phi _+\,\!^{\p...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.025488130748271942, 0.004864875227212906, -0.046611227095127106, -0.024709749966859818, -0.02406873181462288, 0.002999052405357361, 0.052533019334077835, 0.05195305123925209, 0.029608964920043945, 0.02249670960009098, -0.015269984491169453, -0.007944054901599884, -0.0331803560256958, 0.0...
f209df7808cdcb3712018ce93124d9b35e5cf178
subsection
62
74
Error analysis
\end{aligned}Summing (REF ) – (REF ), we obtain the following equation\begin{aligned}a_{\mathcal {D}}\big (\mathcal {J}(\delta _\tau \xi _c^n),\mathcal {J}(\delta _\tau \xi _c^n)\big ) + \kappa a_{\mathcal {D}}(\xi _c^n,\delta _\tau \xi _c^n) + \mu _\mathrm {s} a_{{\varepsilon }}({\xi }_{{{v}}}^n, {\xi }_{{{v}}}^n) + (...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0404781773686409, 0.02635965868830681, -0.06086990237236023, -0.04502663016319275, -0.024757016450166702, 0.030526529997587204, 0.03846342861652374, 0.04701085761189461, 0.025657549500465393, 0.01593484915792942, -0.019842244684696198, 0.013271409086883068, -0.021582257002592087, 0.0122...
61fe588accd229050f006345e7d18648541751ab
subsection
63
74
Error analysis
\end{aligned}The remainder of the proof consists of finding lower bounds for the terms in the left-hand side and upper bounds for the terms in the right-hand side of the equation above. We will then utilize Gronwall's lemma. For the left-hand side of (REF ), since a_{\mathcal {D}} and the inner product are both symmetr...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06691314280033112, 0.03266289085149765, -0.02315402217209339, -0.030785540118813515, -0.027320830151438713, 0.00624639680609107, 0.011119119822978973, 0.04374385625123978, 0.01820879988372326, 0.02071193791925907, -0.06575314700603485, 0.026603467762470245, -0.02622189186513424, -0.0111...
efed813cef8f138bd134049d35ae2047219af3b2
subsection
64
74
Error analysis
By Cauchy–Schwarz's inequality, Poincaré's inequality, Young's inequality, and using a Taylor expansion, we haveT_1 &\le \Vert \delta _\tau c^n - (\partial _t c)^n\Vert _{L^2(\Omega )}\Vert \mathcal {J}(\delta _\tau \xi _c^n)\Vert _{L^2(\Omega )} \\ &\le C_P \Vert \delta _\tau c^n - (\partial _t c)^n\Vert _{L^2(\Omega ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 5043, "openalex_id": "", "raw": "B. Rivière, Discontinuous Galerkin Methods for Solving Elliptic and Parabolic Equations: Theory and Implementation, Frontiers in Applied Mathematics, Society for Industrial and Applied Mathematics,...
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06751617789268494, 0.048714205622673035, -0.03198777511715889, 0.015749705955386162, 0.00941624864935875, 0.014032804407179356, -0.00941624864935875, 0.0434337817132473, 0.024784419685602188, 0.020007619634270668, -0.024524977430701256, -0.022251037880778313, -0.012796635739505291, 0.01...
ecc545677266154b505ffe97c0cf34dc87b3ec57
subsection
65
74
Error analysis
We give an outline of the proof for completeness.\begin{split} T_{11} + T_{12} =& - a_{\mathcal {C}}({{v}}_h^{n-1},{{v}}^n,{{v}}^n,{\xi }_{{{v}}}^n) + a_{\mathcal {C}}({{v}}_h^{n-1},{{v}}_h^{n-1},{{v}}_h^n,{\xi }_{{{v}}}^n)\\ =& -a_{\mathcal {C}}({{v}}_h^{n-1},{{v}}_h^{n-1},{\xi }_{{{v}}}^n,{\xi }_{{{v}}}^n) - a_{\math...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.007792978547513485, 0.015013786032795906, -0.03274348005652428, 0.017851756885647774, 0.025557003915309906, -0.03704621270298958, 0.005721717141568661, 0.040738627314567566, 0.007098743692040443, 0.014372953213751316, -0.019118163734674454, -0.03637486323714256, -0.0273421797901392, -0....
6dbb5e80f23499290c08189d443966f2f837f946
subsection
66
74
Error analysis
We rewrite the second term as:T_{\mathcal {C}}^2 = -a_{\mathcal {C}}({{v}}_h^{n-1},{\xi }_{{{v}}}^{n-1},{{v}}^n, {\xi }_{{{v}}}^n) + a_{\mathcal {C}}({{v}}_h^{n-1},{\xi }_{{{v}}}^{n-1},{\zeta }_{{{v}}}^n, {\xi }_{{{v}}}^n).Note that {\xi }_{{{v}}}^{n-1} belongs to V_h and we apply lem:CHNS:boundconvection to the first ...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07529406249523163, 0.02721325121819973, -0.022698048502206802, -0.004572405945509672, 0.005544852465391159, -0.013507474213838577, 0.002286202972754836, 0.021920090541243553, 0.01893029361963272, -0.004885114263743162, -0.05018587410449982, -0.02956237830221653, 0.0015320801176130772, -...
db02ec2c81da4126c8cd19fd300d86712c3bb978
subsection
67
74
Error analysis
The term simplifiesT_\mathcal {C}^4 = \sum _{E\in \mathcal {T}_h} \int _E \left(({{v}}^{n-1}-{{v}}^n)\cdot \nabla \mathcal {R}_h{{v}}^n\right) \cdot {\xi }_{{{v}}}^n \\ + \sum _{E\in \mathcal {T}_h} \int _{\partial E_-\setminus \partial \Omega } \vert ({{v}}^{n-1}-{{v}}^n)\cdot {{n}}_E\vert \left((\mathcal {R}_h{{v}}^n...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.01915125921368599, 0.04953385144472122, -0.04935073480010033, -0.006798315327614546, 0.0015364879509434104, -0.01367293018847704, 0.005314283538609743, 0.04254478961229324, 0.0345485657453537, 0.01632816530764103, -0.03674599900841713, -0.027178000658750534, 0.014054429717361927, -0.005...
466a5bc975f42a31297ccae89ba3022b384f7322
subsection
68
74
Error analysis
Using Poincaré's inequality, we obtain:|a_\mathcal {A}(\zeta _c^{n-1},{\xi }_{{{v}}}^n,\mu ^n)| \le \Vert \zeta _c^{n-1}\Vert _{L^2(\Omega )} \Vert {\xi }_{{{v}}}^n\Vert _{L^4(\Omega )} \Vert \mu ^n \Vert _{W^{1,4}(\Omega )} \le C h^q \Vert c^{n-1}\Vert _{H^{q+1}(\Omega )} \Vert {\xi }_{{{v}}}^n\Vert _{\mathrm {DG}},|a...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.048311639577150345, 0.03625661879777908, -0.04165848717093468, -0.01924225315451622, -0.012528065592050552, -0.0019236531807109714, -0.0088581508025527, 0.021058136597275734, 0.03128201514482498, 0.002288928022608161, -0.03396768867969513, -0.0017290943069383502, -0.00680956058204174, -...
804cc382ca2444b0f0534db9b3cb89610677ace1
subsection
69
74
Error analysis
We note that \delta _\tau c^n - (\partial _t c)^n - \delta _\tau \zeta _c^n belongs to M_h by taking \chi = 1 in (REF ). We choose \chi = \xi _\mu ^n in (REF ), use coercivity of a_\mathcal {D}, lem:CHNS:errorpropertyofJ, Cauchy–Schwarz's inequality, triangular inequality, and Poincaré's inequality to obtain:K_\alpha \...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03793569654226303, 0.017136679962277412, -0.043642837554216385, -0.03164868801832199, -0.022904859855771065, 0.028688298538327217, 0.029466545209288597, 0.04727465286850929, 0.0313740111887455, 0.0254379753023386, -0.03979738429188728, -0.012467201799154282, -0.004005679860711098, 0.017...
57d8c35636603915fbc38987db6c9311b2a9bff3
subsection
70
74
Error analysis
Using Young's inequality, we have\Vert \xi _\mu ^n\Vert _{\mathrm {DG}}^2 \le C\Big ( \Vert {\xi }_{{v}}^n\Vert _{L^2(\Omega )}^2 + \mu _\mathrm {s}K_{{\varepsilon }}\Vert {\xi }_{{v}}^n\Vert _{\mathrm {DG}}^2 + \Vert \xi _c^{n-1}\Vert _{\mathrm {DG}}^2 + \tau ^2 + h^{2q}\\ + \Vert J(\delta _\tau \xi _c^n)\Vert _{\math...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06587234139442444, 0.003746909089386463, -0.02474028430879116, -0.02557971328496933, -0.025228679180145264, 0.00015882392472121865, 0.004571076482534409, 0.015674442052841187, 0.03705700859427452, -0.009172677993774414, -0.015674442052841187, -0.02060418389737606, -0.016849644482135773, ...
cfe743bd52cc1a6c15b81ea881b3f2304cc82036
subsection
71
74
Error analysis
\Vert \partial _{tt} c\Vert _{L^2(\Omega )}^2.To this end, combining (REF ) with all the bounds for T_1 to T_{16}, and choosing the values r_1 = K_\alpha /9, r_2 = 1/18, and r_3 = \min {\lbrace 1,K_\alpha \rbrace }/18C yields&\frac{K_\alpha }{2}\Vert \mathcal {J}(\delta _\tau \xi _c^n)\Vert _{\mathrm {DG}}^2 + \frac{\k...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05956832319498062, 0.04403539374470711, -0.046049486845731735, -0.05129833519458771, -0.03799311816692352, 0.03475836291909218, -0.00804111361503601, 0.032927367836236954, 0.014617431908845901, 0.035887472331523895, -0.025588134303689003, -0.008376795798540115, -0.004947497043758631, 0....
b35f426651e010796460a2cd873a65b706099b18
subsection
72
74
Error analysis
Then, under as:CHNS:assumptionA and sufficiently small time step size \tau , there exists a constant C independent of mesh size h and time step size \tau such that for any m \ge 1\max _{1\le n\le m}\Big (\Vert c(t^n)-c_h^n \Vert _{\mathrm {DG}}^2 + \Vert {{v}}(t^n)-{{v}}_h^n\Vert _{L^2(\Omega )}^2\Big ) + \tau \sum _{n...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.010459101758897305, 0.002441219985485077, -0.054469719529151917, 0.00290276319719851, -0.0119314631447196, -0.014204848557710648, -0.02248973958194256, 0.04101249575614929, -0.014838039875030518, 0.00013898742327000946, 0.0052867671474814415, -0.03326162323355675, 0.002500343369320035, ...
667b4fd5c20c27bf1a82802acb5012e2608ef7dc
subsection
73
74
Conclusions
In this paper, we have formulated an interior penalty discontinuous Galerkin method for solving the Cahn–Hilliard–Navier–Stokes equations. The time discretization utilizes a convex-concave splitting of the chemical energy density and a Picard's linearization for the convection term. Existence and uniqueness of the nume...
{ "cite_spans": [] }
1807.02725
Numerical analysis of a discontinuous Galerkin method for Cahn-Hilliard-Navier-Stokes equations
[ "Chen Liu", "Beatrice Riviere" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.007711320649832487, -0.011835980229079723, -0.011622306890785694, -0.009996863082051277, -0.005837862845510244, -0.034981388598680496, 0.029242731630802155, -0.030189000070095062, 0.017856990918517113, 0.04078109562397003, -0.028128577396273613, -0.0016025505028665066, -0.0222678221762180...
8b3696458ef8778513ffd55254d00281165496ee
abstract
0
84
Abstract
We introduce an approximation technique for nonlinear hyperbolic systems with sources that is invariant domain preserving. The method is discretization-independent provided elementary symmetry and skew-symmetry properties are satisfied by the scheme. The method is formally first-order accurate in space. A series of hig...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.021026192232966423, -0.009490777738392353, -0.015136417001485825, 0.019774997606873512, -0.01580779068171978, -0.04150307923555374, 0.04260168969631195, -0.00735077541321516, 0.025008657947182655, 0.02662605606019497, -0.04177773371338844, -0.012130496092140675, 0.0030993798281997442, 0...
4114c4c7c0098a4a2c7fb482915737c46863abd6
subsection
1
84
Introduction
The present paper is concerned with the approximation of hyperbolic systems in conservation form with a source term:{\left\lbrace \begin{array}{ll} \partial _t + ()=(), \quad \mbox{for}\, (,t)\in _+,\\ (,0) = _0(), \quad \mbox{for}\, \in ^d. \end{array}\right.}The space dimension d is arbitrary. The dependent variable ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03167073056101799, 0.009946683421730995, -0.02425648458302021, 0.02248682826757431, -0.011289181187748909, -0.04582797363400459, 0.029153548181056976, -0.0011985220480710268, 0.013302926905453205, 0.05626283586025238, -0.04015287011861801, 0.006662905216217041, 0.0006421675207093358, 0....
90a0e27c3b055271e682e7b2d618c427157c5795
subsection
2
84
Introduction
A brief overview of explicit Runge Kutta Strong Stability methods is made in §. The key result of this section is a reformulation of the Shu-Osher Theorem REF which does not involve any norm. We show therein that only convexity matters. It seems that the result, as reformulated, is not well known in the literature. We ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04854240268468857, 0.00290804379619658, -0.010221059434115887, -0.006761916913092136, -0.005163922905921936, -0.010693973861634731, -0.003146408125758171, 0.024728860706090927, 0.032554835081100464, 0.017513098195195198, -0.025339072570204735, 0.025781476870179176, -0.018977608531713486, ...
e3bd04e20455588f02297f54a8f2077c11ee43db
subsection
3
84
Preliminaries
We recall in this section key properties about the system (REF ) that will be used repeatedly in the paper. The reader who is familiar with hyperbolic systems with source terms, Riemann problems, and invariant sets is invited to jump to §.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.021718360483646393, 0.020283697172999382, -0.02959374524652958, 0.03568343445658684, -0.014934234321117401, -0.053448837250471115, 0.06526191532611847, 0.029135873541235924, 0.004498585127294064, 0.008394306525588036, -0.04731336236000061, 0.013652194291353226, -0.006299545522779226, 0....
20b08131382a63359edee0cabc5582973fdd0152
subsection
4
84
Riemann problem space average and maximum wave speed
We consider (REF ) without source term in this subsection, () = 0. Instead of trying to give a precise meaning to the solutions of (REF ), which is either a very technical task or a completely open problem, we instead assume that there is a clear notion of solution for the Riemann problem. That is to say we assume that...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03821749985218048, 0.02855628728866577, -0.036324940621852875, 0.00031502917408943176, -0.013408176600933075, 0.030174121260643005, 0.03342505171895027, 0.02516799420118332, 0.022191790863871574, -0.01131720282137394, -0.011149314232170582, 0.03205141797661781, 0.006620144005864859, 0.0...
82c637eb966182fe90d643d6608b30d034fdf168
subsection
5
84
Invariant sets and invariant domains
We introduce in this section the notions of invariant sets and invariant domains. Our definitions are slightly different from those in ChuehConleySmoller1977,Hoff1985,Smoller1994,Frid2001. We associate invariant sets with solutions of Riemann problems and define invariant domains only for an approximation process; our ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05022554472088814, -0.0006121810292825103, -0.04897448420524597, 0.02340400591492653, 0.0022904344368726015, -0.029384689405560493, 0.04015602543950081, 0.043359965085983276, 0.046533387154340744, 0.017667431384325027, -0.0343279093503952, 0.041986845433712006, -0.024700837209820747, 0....
bc301790fb5626d3038ff8ef3c733e0fa777af4b
subsection
6
84
Examples
We briefly go over some examples of systems with source terms and show that the proposed definition for invariant sets is meaningful/useful.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.003936609718948603, 0.01930159516632557, -0.04885668680071831, 0.05669938772916794, -0.01505982130765915, -0.08355378359556198, 0.04537782073020935, 0.014953014440834522, -0.018859107047319412, 0.018355587497353554, 0.006026980467140675, 0.005515831522643566, -0.0009464838658459485, -0....
f2edeb91e0820cd54294808bce7ebff70fa50a2e
subsection
7
84
Euler + co-volume EOS
For the compressible Euler equations with covolume of state the dependent variable is =(\rho ,,E), where \rho is the density, is the momentum, and E is the total energy. The flux is ()=(\rho , \otimes + p , (E + p)) where := /\rho and the pressure is given by the equation of state p ( 1 - b \rho ) = ( \gamma - 1 ) e \r...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.008857128210365772, -0.007045269478112459, -0.011550122871994972, 0.007323723752051592, -0.008613004349172115, -0.04400337487459183, 0.03301778808236122, 0.017775287851691246, 0.01002434641122818, 0.02220003679394722, -0.011771360412240028, 0.0023783030919730663, -0.008727437816560268, ...
dc0fa9684409e890fdaf8a4188b982b9a5d3b2c1
subsection
8
84
Shallow water
Saint-Venant's shallow water model describes the time and space evolution of a body of water evolving in time under the action of gravity assuming that the deformations of the free surface are small compared to the water elevation and the bottom topography z varies slowly. The dependent variable is =(,), where is the ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.074335016310215, 0.014059876091778278, -0.028791086748242378, 0.00003236279371776618, 0.001068031881004572, -0.05794835463166237, 0.0019663230050355196, 0.028119752183556557, 0.019758589565753937, 0.00768982945010066, -0.015326256863772869, -0.012984215281903744, -0.004214911255985498, ...
9e23fd2a271e44de166333f42857e5947682358a
subsection
9
84
ZND model
We now consider the Zel'dovich–von Neumann–Döring model for compressible reacting flows. The dependent variable is =(\rho _1,\rho _2,, E), where \rho _1 is the density of the burned gas (fuel), \rho _2 is the density of the unburned gas, is the momentum of the mixture, and E is the total energy. The flux is ()=(\rho _1...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.011744186282157898, -0.005646977107971907, -0.009073318913578987, 0.03369871899485588, -0.008798601105809212, -0.06623751670122147, 0.0446569062769413, 0.03201988711953163, 0.007215158082544804, 0.030661558732390404, -0.02086329087615013, -0.004929658491164446, -0.015071324072778225, -0...
fa54c1c76c700acb6c6237d1108f540fc70850b6
subsection
10
84
ZND model
Moreover, \rho _2 -\tau \kappa (T)\rho _2 =\rho _2(1 -\tau \kappa (T)) \ge \rho _2(1 -\tau \kappa _0); hence \rho _2 -\tau \kappa (T)\rho _2\ge 0 provided \tau \le \tau _0:=\kappa _0^{-1}. Finally, observing that \rho :=\rho _1 +\tau \kappa (T)\rho _2 + \rho _2 -\tau \kappa (T)\rho _2 >0, we have \rho e(+ \tau ()) = E ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.0333264134824276, 0.02517792396247387, 0.011391103267669678, 0.059297826141119, 0.026032445952296257, -0.02055434137582779, 0.014366676099598408, 0.09161712229251862, -0.053224604576826096, 0.022324426099658012, 0.012733925133943558, -0.024231843650341034, -0.033356934785842896, 0.017975...
6f9f9beebf157d86e79dd28bc5245909d0a0394f
subsection
11
84
Euler equations with sources
In some astrophysical applications one may want to solve the compressible Euler equations with Coriolis effects, gravitation effects and some heat transfer effects due to the emission and/or absorption of radiation. The dependent variables and the flux are the same as those of Euler's equations, but the source term is ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04857566952705383, 0.00672416016459465, -0.030939528718590736, 0.03569945693016052, -0.005755392834544182, -0.05388481914997101, 0.01513412781059742, 0.038567617535591125, 0.023708099499344826, 0.010107217356562614, -0.017880240455269814, -0.0006626902031712234, -0.02109929360449314, -0...
e2b701ebad7ed504e4c5758006d5977f43ed4e27
subsection
12
84
Abstract low-order approximation
In this section we describe a generic invariant domain preserving technique for approximating solutions to (REF ). In order to stay general we present the method without referring to any particular discretization technique, we are going to use instead the graph theoretic language to describe the method. The method is i...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05253618210554123, -0.00040781707502901554, 0.02152397111058235, -0.023659586906433105, -0.02185956761240959, -0.0539700984954834, 0.03911230340600014, 0.006300070788711309, 0.012287425808608532, 0.023812131956219673, -0.02739691734313965, 0.039325863122940063, -0.0172832440584898, 0.04...
ad20e86b00fb11495d4e2e09d3a2d9630ec5d936
subsection
13
84
The low-order scheme
To identify properly the time stepping technique, we denote by t^n the current time, n\in , and we denote by the current time step size; that is t^{n+1}:=t^n+. We now address the approximation in space by assuming that we have at hand some finite-dimensional vector space X_h with some basis \lbrace \varphi \rbrace _{i\...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.00962909311056137, -0.013856127858161926, -0.04922892898321152, 0.017716920003294945, 0.006912803743034601, 0.00444830534979701, -0.002243227791041136, 0.0014249073574319482, 0.03460979834198952, 0.04068329930305481, -0.015099821612238884, 0.03305327519774437, -0.002605654066428542, 0.0...
31d976ed0126835f2c85b5c4cdb5ac381c2a3729
subsection
14
84
The low-order scheme
We assume that the graph viscosity \lbrace d_{ij}\rbrace _{(i,j)\in } is scalar and has the following properties:d_{ij}= d_{ji}> 0, \quad \text{if} \quad i\ne j.Although the diagonal value d_{ii} is not needed, we adopt the convention d_{ii}:=-\sum _{j\in (i)\backslash \lbrace i\rbrace } d_{ij}. This convention will he...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.043514594435691833, -0.004886237438768148, -0.03216296061873436, -0.022276055067777634, -0.0026071376632899046, -0.007548683788627386, -0.010054740123450756, 0.012312861159443855, 0.025937872007489204, 0.045101381838321686, -0.04806135222315788, -0.001804207917302847, -0.00661034323275089...
492aa35e2d34aebd96e879b13f25ea82560fe9cb
subsection
15
84
The low-order scheme
Note that if all the values \lbrace _j\rbrace _{j\in (i)} are constant, the graph viscosity term \sum _{j\in (i)} d_{ij}(^n_j - ^n_i) vanishes; which in some sense implies that (REF ) is a first-order consistent perturbation of (REF ). The scalars m_i and the vectors \lbrace _{ij}\rbrace _{j\in (i)} are not uniquely d...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.034181077033281326, 0.0007324516773223877, -0.0475178025662899, -0.023011188954114914, 0.011696337722241879, 0.015854526311159134, -0.0030213631689548492, 0.010574771091341972, 0.011749745346605778, 0.03387589007616043, -0.04660223796963692, 0.0016823498299345374, -0.01716683618724346, ...
92089f938e620f5f8fe33a54a3c0ad9e815d2e5c
subsection
16
84
Invariant domain preserving graph viscosity
Now we propose a definition of the graph viscosity that makes the algorithm (REF ) invariant domain preserving. Recall that the discretization setting is still unspecified. Most of the arguments presented in this subsection are generalizations of those in §3.2, §4.1 and §4.2 of GuePo2016.Since \sum _{j\in (i)} (_i^n)_{...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07281192392110825, -0.007518456317484379, 0.000054893600463401526, -0.022627845406532288, 0.01690603606402874, 0.009696558117866516, -0.004100629594177008, 0.029494015499949455, 0.017302749678492546, 0.03622286394238472, -0.04723925143480301, 0.013213562779128551, -0.00958212185651064, ...
4697e5d53f877fc402f9a23759108dfd67cc5347
subsection
17
84
Invariant domain preserving graph viscosity
The state \overline{}_{ij}^{n} defined in (REF ), with d_{ij} as defined in (REF ), belongs to .Let us set t_{ij}:= \Vert _{ij}\Vert _{\ell ^2}/(2d_{ij}), then according to Lemma REF , we have \overline{}_{ij}^{n}:=\overline{}(t_{ij},_{ij},_{i}^n,_{j}^n) \in if \lambda _{\max }(_{ij},_{i}^n,_{j}^n) t_{ij}\le \frac{1}{...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03510455787181854, 0.025168441236019135, -0.05308419466018677, -0.018895410001277924, 0.0005485087167471647, 0.06666813045740128, -0.013042106293141842, 0.05925038829445839, 0.004033208359032869, 0.011340298689901829, -0.029609931632876396, 0.037821345031261444, -0.01819331757724285, 0....
1df10d699473ecc49672db015adc733bc6887f74
subsection
18
84
Invariant domain preserving graph viscosity
But we have established in Lemma REF that \overline{}_{ij}^{n}\in . Then, the convexity of implies _1 is in . Since is an invariant set according to Definition REF and _i^{n}\in by assumption, the condition 2\le \tau _0 implies that _2:= _{i}^{n} + 2(_i^n) is a member of . In conclusion, the convexity of implies that _...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.026389308273792267, 0.022863611578941345, -0.03214336931705475, 0.016789034008979797, 0.008638720959424973, -0.004731454886496067, -0.014133313670754433, 0.04127049818634987, 0.02829715423285961, -0.003630628576502204, -0.019978949800133705, 0.0015052895760163665, -0.015163550153374672, ...
aebe4320bf7cfcb5eea19147a2fed3f761f0b3fe
subsection
19
84
Invariant domain preserving graph viscosity
Then recalling (REF ), the CFL condition and the convexity of \eta imply that\eta (_i- (_i^n)) \le \Big (1 - \sum _{j\in (i)\backslash \lbrace i\rbrace } \frac{2d_{ij}}{m_i} \Big )\eta (_i^n) + \sum _{j\in (i)\backslash \lbrace i\rbrace }\frac{2d_{ij}}{m_i}\eta (\overline{}_{ij}^{n}).Lemma REF implies that \eta (\overl...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05344251170754433, 0.028140489012002945, -0.034794095903635025, -0.024142220616340637, 0.0033897538669407368, -0.0043912287801504135, -0.018236380070447922, 0.02930029295384884, 0.0010739625431597233, 0.019060451537370682, -0.01950300671160221, 0.061133839190006256, -0.021639486774802208,...
700fe6433459d11ae5b31f8700bcc0d1891f9809
subsection
20
84
Invariant domain preserving graph viscosity
In this case one has to come up with some informed guess. We now give a lower bound on \lambda _{\max }(,_L,_R) that guaranties positivity if it happens that some components of , say , has to be positive (think of the density and the total energy in the Euler equations or the water height in the shallow water equations...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.057575713843107224, 0.04936804994940758, -0.026636777445673943, 0.010068885050714016, -0.007490640506148338, -0.0006908056093379855, -0.005751469172537327, 0.026285892352461815, 0.03173224255442619, 0.028360692784190178, -0.017986690625548363, 0.02469928003847599, -0.003060712246224284, ...
cf8a0154579c037f6f125e7f94c66c539c147154
subsection
21
84
Examples of discretizations
In this section we illustrate the GMS-GV scheme described in § in the following three space discretization settings: finite volumes, continuous finite elements, and discontinuous elements.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.003536369651556015, -0.007732708938419819, 0.007232963107526302, -0.040895234793424606, -0.019806722179055214, 0.006710328161716461, 0.02832147479057312, 0.006679809186607599, 0.036256372928619385, 0.05087489262223244, 0.0015011450741440058, 0.00046278946683742106, 0.010803665965795517, ...
022c3d1a425ee444764a1de85eeaeaa3069dcc2f
subsection
22
84
Finite Volumes
We now illustrate the construction of the abstract low-order scheme (REF )–(REF ) in the context of finite volumes (FV).
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.011213981546461582, -0.0034595513716340065, 0.0023801103234291077, 0.015104546211659908, 0.006415618117898703, 0.03228406235575676, 0.004363535437732935, 0.01727105677127838, 0.0054506054148077965, 0.046747807413339615, -0.012419293634593487, 0.0022523319348692894, -0.0062897466123104095,...
3b2b40ea5f066c9314ebad360dc78458e33322db
subsection
23
84
Technical preliminaries
We unify our presentation by putting into a single framework the so-called cell-centered and vertex-centered finite volume techniques, see Figure REF . We refer the reader to Barth2004,Eymard2000 for comprehensive reviews on the finite volume techniques. For any manifold E\subset ^d of dimension l we denote by |E| the ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.016294747591018677, 0.015325911343097687, -0.046107422560453415, 0.004439861048012972, -0.01205323077738285, -0.01901053823530674, 0.03246743604540825, -0.0005597504787147045, 0.009329810738563538, 0.05031842365860939, 0.01193117257207632, 0.023526685312390327, 0.011099652387201786, 0.05...
0f58b9b1101f6cc07d2b7e5d8eab88f3bb776908
subsection
24
84
Technical preliminaries
Similarly, letting \partial K to be the boundary of the cell K, we denote by (\partial K) the set of indices of the shape functions with non-vanishing trace on \partial K:(K) := \big \lbrace i \in \varphi _{i|K} \lnot \equiv 0 \big \rbrace ,\qquad (\partial K) := \big \lbrace i \in \varphi _{i|\partial K} \lnot \equiv ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.008743672631680965, 0.027497554197907448, -0.06824032217264175, -0.030305296182632446, -0.019577892497181892, -0.01861654594540596, 0.02517811581492424, 0.02087494730949402, 0.009933911263942719, 0.04214663803577423, 0.02075287140905857, -0.0012493686517700553, -0.003177783451974392, 0....
0b1f3194f3cfd7bc44c6ac1878a21248afe58ea8
subsection
25
84
Definitions of
We define the connectivity graph (,) by identifying the vertices of this graph with the cells in _h, and we say that a pair of cells K_i,K_j form an edge of the graph, (i,j)\in , iff the cells K_i and K_j share an interface, \partial K_i\cap \partial K_j is a (d-1)-manifold of positive measure. For any i\in we define t...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.008697551675140858, 0.0244752150028944, -0.03503434732556343, -0.03048720583319664, -0.0246735792607069, -0.017990197986364365, 0.018280116841197014, -0.0055275228805840015, 0.018890470266342163, 0.07348667830228806, -0.01570136845111847, 0.009880113415420055, 0.00040316773811355233, 0....
a5c4858a14563b9540d0fc5cf1995bb8e67968a9
subsection
26
84
Definitions of
Let us mention in passing that while any family of vectors of the form _{ij} =\alpha _{ij} |\Gamma _{ij}| satisfies the conservation constraint (REF ), only the factor \alpha = \frac{1}{2} leads to a consistent discretization of the divergence operator.We define the connectivity graph (,) by identifying the shape funct...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.013225079514086246, -0.0018334249034523964, -0.03299783170223236, 0.003125026123598218, -0.03087632544338703, -0.012461947277188301, 0.0028903628699481487, 0.007493957411497831, 0.004296434111893177, 0.048107851296663284, -0.03348623961210251, 0.008524185977876186, -0.00220735976472497, ...
c51647be43d55831a93f64f056f225f7c3eb33e9
subsection
27
84
Definitions of
On the other hand, the skew-symmetry property _{ij} = -_{ji} follows using integration by parts if is the d-torus (which is the case for periodic boundary conditions) or if either \varphi _i or \varphi _j vanish at the boundary of D (which is the case when we solve the Cauchy problem).We start by defining the undirecte...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.019181106239557266, 0.014877946116030216, -0.019944077357649803, -0.004093342460691929, -0.014046307653188705, -0.010200930759310722, 0.07068169116973877, -0.014496460556983948, 0.0030366270802915096, 0.07306215912103653, -0.03558499738574028, -0.00308812758885324, 0.019928818568587303, ...
22e3f01e0dea04658c3a56d0882cfa4e2013c803
subsection
28
84
Definitions of
Inserting (REF ) into (REF ) and integrating by parts, we obtain\int _K \nabla ((_h)) \varphi _i\approx \int _K ((_h)) \varphi _i\\ +\int _{\partial K} \tfrac{1}{2}((_h^\mathsf {e})-(_h^\mathsf {i}))_K \varphi _i s + \int _{\partial K} \alpha _{\partial K}^n (_h^\mathsf {i} -_h^\mathsf {e}) \varphi _i s.We now consider...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.025043733417987823, 0.027073480188846588, -0.03201812878251076, -0.020266957581043243, -0.022434057667851448, 0.04004555568099022, 0.03217074275016785, 0.02792811207473278, -0.003086590440943837, 0.05179672688245773, -0.014185343869030476, -0.00039631593972444534, 0.010225046426057816, ...
8f70b55ccaef6200c0f597595e23050506ffe2e9
subsection
29
84
Definitions of
\end{array}\right.}Therefore, (REF ) can be rewritten as follows:\begin{aligned}\int _K \nabla ((_h)) \varphi _i\approx \sum _{j\in (i)} (_j^n)_{ij} + \int _{\partial K} \alpha _{\partial K}^n (_h^i -_h^e) \varphi _i s. \end{aligned}The set of coefficients \lbrace _{ij}\rbrace _{j \in (i)} defined in (REF ) satisfy the...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.007724211551249027, 0.04644444212317467, -0.04003620892763138, -0.0005745477974414825, -0.02578551508486271, 0.018919549882411957, 0.014525331556797028, 0.020475834608078003, -0.007506789173930883, 0.022856036201119423, -0.03286508843302727, 0.017454810440540314, -0.031186740845441818, ...
5bbf66db264706299deb02843d83d66c46696cc1
subsection
30
84
Definitions of
The partition of unity property on \partial K (see (REF )) implies that \sum _{j \in (\partial K)} _{ij}^{\partial } = \int _{\partial K} \varphi _i _K s and \sum _{j \in (\partial K)} _{ij}^{\partial } = \int _{\partial K} \varphi _i _K s; hence, the last two summations cancel each other. This completes the proof.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0007147453143261373, 0.059018224477767944, -0.019942300394177437, 0.00014697815640829504, -0.047086410224437714, -0.012603167444467545, 0.035276662558317184, 0.014121346175670624, -0.010070327669382095, 0.008002858608961105, -0.020003333687782288, 0.009559182450175285, -0.0023554649669677...
ffe0f69f316ad2afec3f6beb81d39c4d56f4401a
subsection
31
84
Continuous finite elements
We describe in this section one possible implementation of the abstract low-order scheme (REF )–(REF ) in the context of continuous finite elements (cG). The set of the d-variate polynomials of degree at most k\in is denoted _{k,d}. The reader who is familiar with GuePo2016,guermondpopovsecondorder2018,GuerNazPopTom201...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0557868592441082, -0.012191930785775185, -0.02059963345527649, -0.03360028937458992, -0.0040016695857048035, -0.04049735143780708, 0.02044704370200634, 0.007137391250580549, -0.003929189406335354, 0.031174110248684883, 0.0015888420166447759, 0.0184175968170166, -0.004901949781924486, 0....
c81d559675e4df12394e9b71748a5f9909d342d9
subsection
32
84
Discontinuous finite elements
We finally describe in this section one possible implementation of the abstract low-order scheme (REF )–(REF ) in the context of discontinuous finite elements (dG). This section builds on top of the definitions and notation already introduced in §REF .
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.016400648280978203, -0.003211475908756256, -0.0028109948616474867, -0.023311804980039597, -0.014074045233428478, -0.020489368587732315, 0.05501464009284973, 0.026530910283327103, -0.0015552010154351592, 0.04030745476484299, -0.01853654719889164, -0.002420049160718918, 0.001790722017176449...
dbeb5972808e19c0fea5f9fc24721d7e7e536f9b
subsection
33
84
Graph viscosity for dG
It is important to notice at this stage, that the formulation of the viscous fluxes \int _{\partial K} \alpha _{\partial K}^n (_h^{n,\textup {i}} -_h^{n,\textup {e}}) \varphi _i s in (REF ) is not compatible with our pursuit of a purely algebraic formulation. Note that the dissipation in (REF ) is active only on \parti...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.0105596287176013, 0.01423718687146902, -0.023286117240786552, -0.04059047996997833, -0.01033836416900158, -0.03146525099873543, 0.048952728509902954, 0.01306982897222042, 0.005142477806657553, 0.04852546006441116, -0.00025726694730110466, -0.017151765525341034, 0.0001257723051821813, 0.0...
5e8abe1a921ec38ed5cf3358269d0376051908e3
subsection
34
84
Runge Kutta SSP time integration
Increasing the time accuracy while keeping the invariant domain property can be done by using so-called Strong Stability Preserving (SSP) time discretization methods. The key idea is to achieve higher-order accuracy in time by making convex combinations of forward Euler steps. More precisely each time step of a SSP met...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03182481974363327, 0.00892498530447483, -0.03658480942249298, 0.009664920158684254, -0.017941508442163467, 0.019436635076999664, 0.02045881189405918, 0.007979089394211769, 0.03277071565389633, 0.050040941685438156, -0.008062999695539474, 0.0353948138654232, -0.028865080326795578, -0.002...
e59d2035c355bd385f37b4c5a76f607306c20a0f
subsection
35
84
SSPRK methods
We are going to illustrate the SSP concept with explicit Runge Kutta methods. Let us consider a finite-dimensional vector space E, a subset A\subset E and a (nonlinear) operator L:[0,T]A\longrightarrow E. We are interested in approximating in time the following problem \partial _t u +L(t,u)=0 with appropriate initial c...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06327501684427261, 0.04667828977108002, -0.04637320339679718, 0.021417098119854927, -0.02799171954393387, 0.03246124088764191, 0.01571197248995304, 0.009213623590767384, 0.03142394497990608, 0.039691802114248276, 0.02253066375851631, 0.051468152552843094, -0.006319111678749323, 0.007169...
f6cdcf56c47fb4e863161f08200e93d8f11a0996
subsection
36
84
SSPRK methods
Any Runge Kutta method that admits an (\alpha -\beta ) representation as defined above is said to be SSP for a reason that will be stated in Theorem REF .[Midpoint rule] The midpoint rule, defined by the Butcher tableau\begin{array}{c|ccccc} 0 \\ \frac{1}{2} & \frac{1}{2} \\[2pt] \hline & 0 & 1 \end{array}does not have...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06208077445626259, -0.0009437826811335981, -0.028781795874238014, 0.0017883704276755452, -0.03174238279461861, 0.017534613609313965, 0.003839988959953189, 0.04398151487112045, 0.011827089823782444, 0.01191102433949709, -0.00505894236266613, 0.04907860606908798, -0.010453620925545692, 0....
b4aacc04a83f2252592c8eaf5dfa2b154b718391
subsection
37
84
SSPRK methods
\end{aligned}[SSPRK(3,3), SSPRK(4,3)] The following Runge–Kutta methods, which are third-order and composed of three substeps and four substeps, respectively, are SSP:\begin{array}{|ccc|ccc|c|c|} \hline &\alpha & & & \beta & &\gamma & c_{\textup {os}} \\ \hline \hline 1 & & & 1 & & &0 & \\ \frac{3}{4}&\frac{1}{4}& & 0 ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03772854804992676, 0.05118994414806366, -0.04746593162417412, -0.007207647897303104, -0.013781907968223095, 0.036934904754161835, 0.01741435006260872, -0.0160712618380785, -0.002024170011281967, 0.0035752071999013424, -0.004704621620476246, 0.053418248891830444, -0.0040101842023432255, ...
f782a37941463de9c464ce1e8d0c2a8c95b60fff
subsection
38
84
The key result
We henceforth denotec_\textup {os}:=\inf _{\lbrace \alpha _{ik}\ne 0,\ 1\le k+1\le i\le s\rbrace }\alpha _{ik}\beta _{ik}^{-1}.The following theorem is the main result of this section. [Shu-Osher] Assume that the Runge Kutta method with the Butcher tableau (REF ) is SSP. Let B\subset A be convex. Let u^n\in B and assum...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04331319034099579, 0.026677750051021576, -0.04981474578380585, -0.005116542335599661, -0.008249041624367237, 0.039161961525678635, 0.01181268785148859, 0.008592433296144009, 0.03336245194077492, 0.016269154846668243, -0.006513003259897232, 0.032080452889204025, -0.024876849725842476, 0....
f3d6ab6df8096669a0eb75ada1863070c30c91f2
subsection
39
84
The key result
The assumption (REF ) then consists of stating that + L(t,\cdot ) maps any ball B centered at 0 into B for any s \in [0,_{\max }] and any t\in [0,T]. In particular taking any v\in E and defining B to be the ball of radius \Vert v\Vert _B centered at 0, the assumption (REF ) amounts to saying that \Vert v+ L(t,v)\Vert _...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03630524501204491, 0.021295009180903435, -0.044054433703422546, 0.024834008887410164, 0.0003260607772972435, 0.022347556427121162, 0.0012832684442400932, 0.027884868904948235, 0.05555617809295654, -0.012241579592227936, 0.004698325879871845, 0.059522297233343124, -0.0023892053868621588, ...
be2d8d068c136cc8bfd926ac156194218ecc4175
subsection
40
84
High-order method
The algorithm that we are going to develop in § relies on the construction of the low-order invariant domain preserving solution _i described in §REF -§REF and a high-order solution _i that possibly wanders outside the invariant domain. We are then going to limit the high-order solution by pushing it back into the inva...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04371355473995209, 0.012958383187651634, 0.0010388456284999847, 0.014667851850390434, -0.007929187268018723, 0.04035567119717598, -0.007704056333750486, 0.021154675632715225, 0.007307215593755245, 0.02538255788385868, -0.05506931245326996, 0.006383797153830528, -0.023016775026917458, 0....
f4a47003fc876a11b8a534860dda7c7c85b3d35e
subsection
41
84
Achieving high-order consistency
In this section we describe in broad terms how high-order consistency can be achieved.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.017993951216340065, 0.0031153648160398006, -0.04349682852625847, 0.02995939366519451, -0.014666825532913208, 0.006024692207574844, -0.01555202342569828, 0.026632267981767654, 0.013178776018321514, 0.03772777318954468, -0.008737520314753056, 0.01180519163608551, 0.004429808352142572, 0.0...
a0499be99bb2a7a6cf134ec6239cd98d127f83bc
subsection
42
84
Discretization-independent setting
Independently of the space discretization that is used, we henceforth assume that the high-order update _i is computed as follows:\frac{m_i}{}(_i- _i^{n}) + \sum _{j\in (i)} _{ij}=m_i (_i^n),where the high-order flux _{ij} is assumed to be skew-symmetric; _{ij}=-_{ji} for all i\in , j\in (i) (under appropriate boundary...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06909213215112686, -0.026199467480182648, -0.0243378859013319, -0.013382022269070148, -0.01035313867032528, 0.027557505294680595, -0.0496523380279541, 0.0067596761509776115, 0.052276864647865295, 0.02029428817331791, -0.03451554849743843, -0.02001962810754776, -0.037475768476724625, 0.0...
6ea0328c89cceee3c906a9597d35919164e8a17f
subsection
43
84
High-order algebraic fluxes: Finite Volumes
In the context of finite volume schemes, high-order algebraic fluxes _{ij} are obtained as integrals of high-order numerical fluxes over the interfaces between volumes, _{ij}:= \int _{\Gamma _{ij}} \widehat{}_{ij} s where \widehat{}_{ij} is some numerical flux. For instance, a widely popular choice of algebraic flux co...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.01987609826028347, 0.027350610122084618, -0.02863195538520813, -0.011799050495028496, -0.019677795469760895, 0.012607517652213573, 0.005518172401934862, 0.0027114171534776688, 0.009747372940182686, 0.04820297285914421, -0.03374208137392998, 0.02004389278590679, -0.033467505127191544, 0....
3681da61a0be14d3b86b0003ecc3d59b152fe4a8
subsection
44
84
High-order algebraic flux: Continuous Finite Elements
We now turn our attention to continuous finite elements. In this case high-order consistency can be achieved by using a degenerate graph viscosity d_{ij} such that d_{ij}\ll d_{ij} in smooth regions while d_{ij}\approx d_{ij} near shocks. Of course d_{ij} must also satisfy the conservation constraintsd_{ij}= d_{ji}\ge ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0394514836370945, -0.019191788509488106, -0.040031202137470245, -0.02022918313741684, -0.02280741184949875, 0.017834022641181946, -0.0011089055333286524, -0.005526411347091198, 0.00857375655323267, 0.05824661999940872, -0.03813948482275009, -0.02132759988307953, -0.0077728270553052425, ...
fe02dfe4ac46f0ac51f0f02ddfae32a2621f3421
subsection
45
84
High-order algebraic flux: Continuous Finite Elements
\end{aligned}Then (REF ) holds with the following definition for the high-order algebraic flux:_{ij}:=&{} \frac{m_{ij}}{}(_j-_j^n - _i+ _i^n ) \\ & + ((_j^n) + (_i^n) ) _{ij} - d_{ij}(^n_j - ^n_i).In the context of finite difference methods, a scheme with the above structure is said to be linearly implicit as the numer...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04034770652651787, -0.01394773181527853, -0.00253126909956336, -0.019334547221660614, -0.0061154854483902454, 0.007122652139514685, 0.030291300266981125, 0.02026541344821453, 0.01744229532778263, 0.02591165155172348, -0.0169539712369442, -0.02171512320637703, -0.03573915734887123, 0.053...
04abd6c247ed95e580b94b98fa9d45dcd0c0530b
subsection
46
84
High-order algebraic flux: Discontinuous Finite Elements
Just like for continuous finite elements, high-order consistency is space is obtained for discontinuous finite elements by replacing the low-order graph viscosity d_{ij} by a high-order graph viscosity d_{ij} satisfying the symmetry and positivity properties stated in (REF ). The corresponding flux in (REF ) is_{ij}:= ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03775535151362419, -0.018297763541340828, -0.04794960469007492, -0.02649284154176712, -0.01236129179596901, -0.013536377809941769, -0.000721075339242816, -0.002596252830699086, 0.007164970971643925, 0.05173429474234581, -0.010667337104678154, -0.01913711056113243, 0.006730036344379187, ...
81baeb1a4f4bb2ad6c502e3b63c1947751f3133b
subsection
47
84
Smoothness-based graph viscosity
The objective of this section is to present a method where the high-order graph viscosity in (REF ), (REF ), (REF ), or (REF ) is obtained by estimating the smoothness of some functional (an entropy) of the current solution.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0368136465549469, -0.039866190403699875, -0.012828304432332516, -0.007928975857794285, -0.01633109524846077, 0.0032624034211039543, 0.00905841588973999, -0.0018057689303532243, 0.02945702336728573, 0.0146293044090271, 0.002142502460628748, 0.007574118208140135, 0.03293691948056221, 0.00...
cb4a15eba62e2b65a05085df11de15f5977704c8
subsection
48
84
Principles of the method
Let _h^n = \sum _{i\in } _i^n \varphi _i be the current approximation and let g:\rightarrow be some functional (examples will be given below). We define the smoothness indicator associated to g as follows:\alpha _i^n:= \frac{\left|\sum _{j\in (i)} \beta _{ij}(g(_j^n)-g(_i^n))\right|}{\max (\sum _{j\in (i)} |\beta _{ij}...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03881245106458664, -0.005580052733421326, -0.04061271250247955, -0.030497675761580467, 0.01766698807477951, -0.04500657692551613, 0.019360456615686417, 0.007273520343005657, 0.02563086338341236, 0.030131520703434944, -0.026653047651052475, 0.021603155881166458, -0.010214203968644142, 0....
4119af2cde191573beaa96987ddc3ce444481b87
subsection
49
84
Principles of the method
For discontinuous elements, one could take \beta _{ij} = \int _{K}\varphi _j\varphi _ix - \int _{\partial K} \frac{1}{2} \varphi _j _K \varphi _i x, where K is the unique cell such that i\in (K) and _K is the unit normal vector on \partial K pointing outward K. For finite volumes, one should get \alpha _i^n=0 if a line...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06037917360663414, 0.012141464278101921, -0.04102608934044838, -0.02078777365386486, -0.011622533202171326, -0.04117871820926666, 0.05708243325352669, 0.015110055916011333, -0.005685349460691214, 0.049603719264268875, -0.044292304664850235, 0.0024897251278162003, -0.015980029478669167, ...
edde94e44e277ab6234ca98bbf5e028a044ecf50
subsection
50
84
Stability
We now establish some invariant domain preserving properties associated with the smoothness-based graph viscosity (REF ) when the coefficients \beta _{ij} are positive. We further specialize the setting by assuming that g:\rightarrow is a projection onto one of the scalar components of . Without loss of generality we s...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04833467677235603, -0.00348626053892076, -0.021665163338184357, -0.004191903863102198, 0.037929292768239975, -0.0016048622783273458, 0.02764596976339817, 0.0453137569129467, 0.021405791863799095, 0.03329111635684967, -0.029507342725992203, 0.027539169415831566, -0.024594537913799286, -0...
07c5cb9aff6ca6329cb198e376cff229ef8a254a
subsection
51
84
Stability
Let _i^{n+1} be the high-order update given by (REF ) using either the high-order cG flux (REF ) or the high-order dG flux (REF ) with any graph viscosity \lbrace d_{ij}\rbrace _{j\in (i)\backslash \lbrace i\rbrace } defined by d_{ij}:= d_{ij}\max (\psi _i^n,\psi _j^n) with \psi _i^n,\psi _j^n\in [0,1]. Then,_i^{n+1} &...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05943424999713898, -0.016774073243141174, -0.033395517617464066, -0.03415866941213608, -0.00494522275403142, 0.01376725547015667, -0.020070888102054596, 0.004239307250827551, 0.007616253569722176, 0.028862396255135536, -0.01732354238629341, -0.0033158939331769943, -0.03205236792564392, ...
1e9c13b7d40e15838840748212bb55320fdb4e76
subsection
52
84
Stability
Then setting _i^{*,n} := \frac{1}{\gamma _i^n}\sum _{j\in (i){\setminus }\lbrace i\rbrace }\frac{2d_{ij}}{m_i}\overline{}_{ij}^{n}, we have _i= (1-\gamma _i^n)_i^n + \gamma _i^n _i^{*,n}, and this in turn implies that_i& = (1-\gamma _i^n)_i^n + \gamma _i^n _i^{*,n} + \frac{}{m_i}\sum _{j\in (i){\setminus }\lbrace i\rbr...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ 0.001520685269497335, 0.02606998011469841, -0.05006534233689308, -0.03740409016609192, -0.018778318539261818, 0.028114086017012596, -0.033804021775722504, 0.04091263189911842, 0.0017294815042987466, 0.042651645839214325, -0.020852932706475258, 0.028953084722161293, -0.0528111569583416, 0.0...
f0bcadd8228364d7c977ef20a3ec2c7ed03849e7
subsection
53
84
Stability
With these definitions, the above inequality is rewritten as follows:_i&\le _i^{,n} + (_i^{,n} -_i^{,n}) \bigg ((1-\theta _i^n)(1-\gamma _i^n) - \theta _i^n\frac{}{m_i}\sum _{j\in (i^-)} (d_{ij}- d_{ij})\bigg ).(iii) Using that d_{ij}\ge d_{ij}\psi _i^n and \psi _i^n\ge 0, we infer that -d_{ij}\le -d_{ij}\psi _i^n, whi...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.017429325729608536, 0.031378891319036484, -0.05414995551109314, -0.03864365443587303, 0.0002184389013564214, 0.012278364971280098, -0.001654984662309289, 0.032050423324108124, 0.004914398305118084, 0.03977305069565773, -0.03632381558418274, 0.00412076897919178, -0.029898466542363167, 0....
b5d57224e70ef9e483e503c1f960df6347f8b1d4
subsection
54
84
Stability
Let i\in and n\ge 0. Then, under the local CFL condition \gamma _i^n \le \frac{1}{1+ k_\psi c_\sharp }, where c_\sharp = \varpi ^\sharp \max _{i\in }\textup {card}((i)) (this number is uniformly bounded with respect to the mesh sequence), the scheme is locally invariant domain preserving for the scalar component : _i\i...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.018935484811663628, 0.020674925297498703, -0.054166775196790695, -0.008247082121670246, 0.025038782507181168, -0.02052234299480915, 0.022154973819851875, 0.03518551588058472, 0.029158510267734528, 0.02720545418560505, 0.009521146304905415, 0.05337334796786308, -0.048765357583761215, 0.0...
8731ed794a306022e49b7326c17fe8f844f50435
subsection
55
84
Stability
This inequality in turn implies that1-\alpha _i^n &= 1- \frac{\left|\sum _{j\in (i^+)} \beta _{ij}|_j^n-_i^n| - \!\sum _{j\in (i^-)} \beta _{ij}|_j^n-_i^n|\right|}{\sum _{j\in (i)} \beta _{ij}|_j^n-_i^n|} \\ &\le \frac{\sum _{j\in (i)} \beta _{ij}|_j^n-_i^n| + \sum _{j\in (i^+)} \beta _{ij}|_j^n-_i^n| - \sum _{j\in (i^...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.008209466934204102, 0.027848100289702415, -0.0830102264881134, -0.020477890968322754, 0.01583908312022686, -0.020477890968322754, 0.010727240703999996, 0.034577421844005585, 0.0014505807776004076, 0.030167503282427788, -0.047089993953704834, 0.024384254589676857, -0.021851221099495888, ...
2ce85e276ac1fc3a20af54c91151f773b9abdc87
subsection
56
84
Stability
Similarly, provided again that \gamma _i^{n} \le \frac{1}{1+ k_\psi c_\sharp }, we have\theta _i^n(1-\gamma _i^n) - (1-\theta _i^n)(1- \psi (\alpha _i^n))\tfrac{1}{2} \gamma _i^{+,n} &\ge \theta _i^n(1-\gamma _i^n) - k_\psi c_\sharp \theta _i^n(1- \theta _i^n)\gamma _i^{n} \\ &\ge \theta _i^n(1-(1+ k_\psi c_\sharp (1-\...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0361902080476284, 0.004218630958348513, -0.05153900384902954, -0.007994800806045532, 0.021253354847431183, -0.03393213450908661, 0.03228434920310974, 0.02940072864294052, 0.027051111683249474, 0.024625208228826523, -0.011290367692708969, 0.04906732961535454, -0.022458678111433983, 0.003...
0ca8c4e26d1d5b2e275e3e672fe1a77a54dbf6a1
subsection
57
84
Greedy graph viscosity
We continue with a technique entirely based on the observations made in Lemma REF , irrespective of any smoothness considerations. As in §REF , we specialize the setting by assuming that there is one scalar component of , say , for which the source term is zero, S\equiv 0.Let i\in and n\ge 0. Let \theta _{n}^n, \gamma ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05474631488323212, 0.013564513996243477, -0.03921349719166756, -0.034391917288303375, 0.03518534079194069, -0.03524637594819069, 0.011115577071905136, 0.018370836973190308, 0.019438909366726875, 0.027205318212509155, -0.02027810923755169, 0.04543883353471756, -0.007926619611680508, -0.0...
d76d145a16f5d185fbc57ea726e0b4e743b4478c
subsection
58
84
Greedy graph viscosity
The definition of \psi _i^n in (REF ) implies that \psi _i^n \ge 1-2\frac{1-\gamma _i^n}{\gamma _i^{-,n}}\frac{1-\theta _i^n}{\theta _i^n}, which in turn gives \theta _i^n(\psi _i^n-1) \frac{1}{2} \gamma _i^{-,n} + (1-\gamma _i^n)(1-\theta _i^n)\ge 0. This is the condition in Lemma REF that shows that _i^{n+1}\le _i^{,...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.043994154781103134, -0.005232314579188824, -0.04274328052997589, -0.027671772986650467, 0.011753639206290245, -0.0323091596364975, 0.00925951823592186, 0.022439459338784218, 0.00020915432833135128, 0.025215789675712585, -0.03441429138183594, 0.02526155300438404, 0.003916608635336161, 0....
62957b00f9e8615a3e0b645abbfa2cbde1cc496f
subsection
59
84
Commutator-based graph viscosity
The objective of this section is to construct the high-order graph viscosity so that the method is entropy consistent and close to be invariant domain preserving. In other words, we do not want to rely on the (yet to be explained) limiting process to enforce entropy consistency. For instance one naive choice consists o...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07410561293363571, -0.034641627222299576, -0.01136916782706976, 0.004498070105910301, -0.007191570941358805, -0.02351662889122963, -0.0005846728454343975, -0.014612051658332348, 0.018205929547548294, 0.04678908735513687, -0.009881257079541683, 0.017946498468518257, 0.004208118189126253, ...
edbcb99b783a3c5ed309ffe1b99ff0bf73683947
subsection
60
84
Convex Limiting
In this section we develop a general limiting framework to preserve convex invariant sets and (more generally) quasiconcave constraints. This work is aligned with the ideas presented in KhobalattePerthame1994, PerthameYouchun1994, PerthameShu1996 in the context of finite volume methods. We also refer the reader to Zhan...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.03806217759847641, 0.010896709747612476, -0.0133003955706954, -0.00016382263856939971, 0.013224087655544281, -0.03986303135752678, 0.025837715715169907, -0.00014426885172724724, -0.016879217699170113, 0.026173468679189682, -0.02319747768342495, -0.011949753388762474, -0.01188107579946518,...
b72fc428aec033f74ec0bf91e0ff53ea92c80c40
subsection
61
84
Quasiconcavity
We have seen in § that the low-order solution _i satisfies some “convex bounds” and, in principle, we would like the high-order solution to satisfy these “convex bounds” as well. But, before proceeding any further, we need to define clearly what we mean by convex bounds. We also need to give a precise statement about t...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.023339364677667618, -0.006021708715707064, -0.05046183988451958, 0.02379699982702732, 0.008084878325462341, 0.022973256185650826, 0.016596881672739983, -0.01563584804534912, 0.027198748663067818, 0.04643465578556061, -0.011227302253246307, -0.012386643327772617, 0.0021623235661536455, 0...
0354ab68391e224b2951be34e878f7193dcc075d
subsection
62
84
Quasiconcavity
As a result, \Phi is concave since R() \Psi () and -\chi R() are both concave and the sum of two concave functions is concave (this may not be the case for the sum of quasiconcave functions). Notice also that \min _{j\in } \Phi (_j)\ge 0 because R\ge 0 and \min _{j\in } \Psi (_j)-\chi \ge 0. Hence\Phi \Big (\sum _{j\in...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04353782534599304, 0.017405977472662926, -0.0381985679268837, 0.013454927131533623, -0.003912912216037512, -0.03926641866564751, 0.0289235170930624, 0.025414861738681793, -0.001021132804453373, 0.039052847772836685, -0.016887307167053223, 0.004912116099148989, -0.0007570303278043866, -0...
f503b033c42b3e3261fe6daf8ddc8b7e8a949d7c
subsection
63
84
Quasiconcavity
A direct computation shows that the functional \varepsilon :\rightarrow has a negative semi-definite Hessian for every equation of state, thereby proving that \varepsilon is concave, hence quasiconcave.Let us now illustrate the use of Lemma REF with R()=\rho .[Specific internal energy] Let :=\lbrace =(\rho ,,E)^\in ^m...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.011383375152945518, 0.026474740356206894, -0.005016467534005642, 0.024033265188336372, 0.01211581751704216, -0.05334622040390968, 0.03357027471065521, 0.0034943604841828346, -0.008026347495615482, 0.04058951511979103, -0.013611220754683018, 0.000052632374718086794, -0.0072633870877325535,...
c51083a1271285d9181bc45380f5cc1cc4c5982a
subsection
64
84
Quasiconcavity
Let k\in be such that \Psi (_k) := \min _{j\in } \Psi (_j). Then, for any j\in , we have \Psi (_k) \le \Psi (_j), which implies that L\circ \Psi (_k) \le L\circ \Psi (_j). Hence L (\min _{j\in }\Psi (_j)) = L(\Psi (_k)) = \min _{j\in } L(\Psi (_j)). In conclusion L\circ \psi (\sum _{j\in } \lambda _j_j) \ge \min _{j\in...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06287120282649994, 0.04642092436552048, -0.010674371384084225, 0.009590910747647285, -0.010933791287243366, -0.021745502948760986, 0.030275842174887657, -0.018830841407179832, -0.020845161750912666, 0.06616736948490143, -0.033327843993902206, 0.010399690829217434, -0.023225722834467888, ...
dc9f88d4548cd55cfd85726221fe7191141a1014
subsection
65
84
Bounds
In this section we define the bounds that we are going to use to limit the high-order solution. The following result will play a key role in the rest of the paper, since it tells us precisely what are the “convex bounds” that the low-order solution produced by the GMS-GV scheme satisfies.[Natural bounds on the GMS-GV s...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05347776040434837, -0.003304205136373639, -0.032782599329948425, -0.004338200204074383, -0.001334463944658637, 0.012850534170866013, -0.0036933841183781624, -0.00044879576307721436, 0.05482080578804016, 0.0531725212931633, -0.05707957223057747, 0.0017341353232041001, -0.025487402454018593...
eb851564c7daace0dfbab87f3eea09dca1e0869e
subsection
66
84
Bounds
The second statement \Psi (_i) \ge \Psi _i^{\min } is a local bound that can be viewed as a local “generalized minimum principle.” This bound cannot be made uniform; it is local in time and space, since \Psi _i^{\min } depends on i and n.[Relaxation] The reader must be aware that in general the bound \Psi _i^{\min } de...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04394324868917465, 0.006599116139113903, -0.02221575379371643, -0.015273330733180046, 0.017180589959025383, -0.03936582803726196, -0.00296388054266572, 0.008819165639579296, 0.035825952887535095, 0.018965784460306168, -0.06890545785427094, -0.006541898474097252, -0.030806047841906548, 0...
2631efd8215af935d7bd843a2accbbc5756f7781
subsection
67
84
Abstract Framework
In the sections §REF , §REF and §REF we have seen that most high-order methods can be written in the algebraic form\frac{m_i}{}(_i-_i^n) + \sum _{j\in (i)} _{ij}= m_i(_i^n),with _{ij}\in ^m satisfying the skew-symmetry constraint _{ij}= -_{ij} for all j\in (i) (whether we use the consistent mass matrix for the discreti...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0235120989382267, -0.024061374366283417, -0.024580135941505432, -0.027936827391386032, 0.012160377576947212, 0.04809223487973213, -0.023405294865369797, 0.017561599612236023, 0.014578721486032009, 0.03963947668671608, -0.040799058973789215, 0.0021131897810846567, -0.032041147351264954, ...
a585c836e49c1b8d827135341a8b62dc331a068b
subsection
68
84
Convex limiting
Without loss of generality, we consider a family of quasiconcave functionals \lbrace \Psi _i \rbrace _{i \in }, \Psi _i: \rightarrow where \subset ^m is a convex set and \Psi _i(_i)\ge 0 for each i \in . Or goal is to modify the high-order update so that the modified high-order update satisfies the same quasiconcave c...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0851031169295311, -0.010180017910897732, -0.03250890225172043, 0.019490079954266548, 0.015628693625330925, 0.007852503098547459, 0.011088130995631218, 0.014224553480744362, -0.0032547060400247574, 0.04435252398252487, -0.055097248405218124, -0.02417563460767269, -0.036232929676771164, 0...
cafe6d8084e3441386a3a9644dfc43e40ade5629
subsection
69
84
Convex limiting
We rewrite (REF ) as follows:_i^{n+1} = \sum _{j \in (i)\backslash \lbrace i\rbrace } \lambda _j (_i+_{ij}_{ij}^n), \qquad \text{with} \qquad _{ij}^n := \frac{1}{m_i \lambda _j} _{ij}^n,where \lbrace \lambda _j\rbrace _{j \in (i)\backslash \lbrace i\rbrace } is any set of strictly positive convex coefficients (see Rema...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07117822766304016, 0.03912971168756485, -0.010247894562780857, -0.011232241988182068, 0.006905692629516125, 0.04221247509121895, 0.005299451295286417, 0.021213065832853317, 0.015299351885914803, 0.05753472074866295, -0.03668792173266411, 0.012384463101625443, -0.014635489322245121, 0.04...
b5b5723c32f5b8c78a29a2d09d9d047b6298fe37
subsection
70
84
Convex limiting
\end{array}\right.}The following two statements hold true: (i) \Psi _i(_i+ _{ij}^n)\ge 0 for every \in [0,^i_j]; (ii) Setting _{ij} = \min (^i_j, ^j_i), we have \Psi _i(_i+ \ell _{ij}_{ij}^n)\ge 0 and _{ij}=_{ji}. (i) First, if \Psi _i(_i+ _{ij}^n)\ge 0 we observe that \Psi _i(_i+ _{ij}^n)\ge 0 for any \in [0,1] beca...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.0718965083360672, 0.023802749812602997, -0.051267459988594055, 0.021880218759179115, 0.0012120710453018546, 0.0011472238693386316, 0.008613238111138344, 0.007472689729183912, 0.01162672694772482, 0.059659454971551895, -0.01733328402042389, 0.006091825198382139, -0.02575579471886158, 0.0...
7e4dee136bac77c898fa21c32faec0636e3a0614
subsection
71
84
Convex limiting
It might be interesting though to explore this question further; for instance, other choices of convex coefficients could help preserve some symmetries.[Multiple limiting] In general we have to consider families of quasiconcave functionals \lbrace \lbrace \Psi _i\rbrace _{i \in }\rbrace _{l\in }, \Psi _i^l: ^l \rightar...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.06175851449370384, 0.025158051401376724, -0.03347286954522133, 0.005210111849009991, -0.01221286877989769, 0.006217044312506914, 0.04518990218639374, -0.0015809600008651614, 0.03216080740094185, 0.03594443202018738, -0.021816866472363472, 0.0003125208895653486, -0.041833460330963135, -0...
f4366e0d20e8c00300b55fcb1196431316edeac4
subsection
72
84
Convex limiting
Since is a convex invariant set for (REF ), the CFL assumption together with Theorem REF implies that _i\in for all i\in . Then Theorem REF can be applied because \Psi _i^l()\ge 0. This theorem then implies that \Psi ^l(_i^{n+1})\ge \Psi _i^{l,\min } for all ł\in . Moreover, _i^n+2(_i^n) \in and \overline{}_{ij}^n\in ,...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04268398508429527, 0.017528198659420013, -0.022242048755288124, 0.039632949978113174, 0.003607849357649684, 0.019587647169828415, 0.03231046721339226, 0.01595691591501236, 0.021341992542147636, 0.010716762393712997, 0.009595506824553013, 0.028008505702018738, -0.039236314594745636, 0.03...
bdec1d52c5844d819f6304bcdb5d91960cd9f7ae
subsection
73
84
Implementation details
The objective of this section to give further details on the convex limiting technique introduced above in order to help the reader to implement it.
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.019826272502541542, 0.007009404245764017, -0.030449125915765762, 0.013660134747624397, -0.006555338390171528, 0.01001997571438551, 0.021489908918738365, 0.05244270712137222, 0.027778150513768196, 0.023764055222272873, -0.03809574991464615, -0.012950417585670948, -0.02811392955482006, 0....
fb996bc073d035cf4d317ac495bb749576946fd1
subsection
74
84
Pseudocode of the limiting algorithm
Given a set of quasi-convex functionals \lbrace \Psi _i\rbrace _{i \in }, \Psi _i:\rightarrow , such that \Psi _i(_i) \ge 0 with convex set , Algorithm REF enforces the quasi-concave constraints \Psi _i(_i^{n+1}) \ge 0 for each i \in . This pseudocode attempts to reflect as accurately as possible the way convex limitin...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.04219254106283188, 0.012462511658668518, -0.04423657804727554, 0.010403222404420376, 0.003434055019170046, 0.059795647859573364, -0.005472369492053986, 0.028631743043661118, 0.018914951011538506, 0.03874513879418373, -0.04582299292087555, -0.00650582741945982, -0.02164541557431221, 0.01...
821f0d0d0651862ecfd8161fc30a075061223b2c
subsection
75
84
Transforming
As mentioned in the previous subsection, the line-search invoked in line REF and line REF of Algorithm REF could be computationally expensive. However, it happens sometimes that the constraint of interest \Psi _i()\ge 0 can be transformed into \tilde{\Psi }_i()\ge 0 where \tilde{\Psi }_i is a quadratic function, not ne...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.020766161382198334, 0.02244454436004162, -0.037656787782907486, 0.010108436457812786, -0.0006565715884789824, 0.0023134970106184483, 0.031339969485998154, -0.0016459586331620812, 0.03634459897875786, 0.04562147334218025, -0.034757763147354126, -0.010611951351165771, -0.0025976775214076042...
cdb80a9de9bc3481b7c12205ddbbaf287a0107d6
subsection
76
84
Transforming
Let _j^{i} := \min (\ell ^{\min },\ell ^{\max }), then \Psi (+ )\ge 0 for all \in [0, _j^{i}]. Let us first observe that \tilde{\Psi }(+ ) = a\ell ^2 + b\ell + c\ell =:g(\ell ) for all \ell \in [0,\ell ^{\max }]; hence, \Psi (+ )\ge 0 iff g(\ell )\ge 0 for all \ell \in [0,\ell ^{\max }]. If there is no positive root ...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.05598783493041992, 0.03569796681404114, -0.018840592354536057, 0.021220456808805466, 0.011464538984000683, -0.03847447410225868, 0.05937456712126732, 0.03331810235977173, 0.011739138513803482, 0.03392832353711128, -0.012143410742282867, 0.021006880328059196, -0.026941286399960518, 0.001...
52cd646d21d1fea5441e52ed430cf20d15ed5b93
subsection
77
84
Transforming
Note that \tilde{\Psi }_i(+)\ge 0 iff \Psi _i(+)\ge 0 provided \rho (+)\ge 0. Hence before applying Lemma REF , one must compute the limiter ^{\max }, which depends on and , such that \rho (+)\ge 0 for all \in [0,^{\max }]. This technique has been introduced in GuermondQuezadaPopovKeesFarthing2018 in the context of th...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.050254352390766144, 0.011320197023451328, -0.03774416074156761, 0.005812661722302437, 0.024699999019503593, -0.010618405416607857, 0.016171710565686226, 0.017560036852955818, 0.00856642983853817, 0.044090792536735535, -0.014623192138969898, -0.006575478240847588, -0.03322828561067581, 0...
5421ca2fc9cd550bce8f0896f6f4a9b22a562033
subsection
78
84
Transforming
Observe also that that -\Psi ^{\min }R:\rightarrow is concave if R:\rightarrow is convex and \Psi ^{\min }\ge 0. Hence the second assertion is just a consequence of the concavity of \Phi : \rightarrow .[Specific entropy] Let us illustrate the use of Lemma REF with the compressible Euler equations. Assume to simplify th...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.059788644313812256, 0.030092701315879822, -0.00833195447921753, 0.004543662071228027, -0.003782570129260421, -0.022554265335202217, 0.027223823592066765, 0.02429390512406826, 0.013596652075648308, 0.04797741025686264, -0.04062209650874138, 0.007404910400509834, -0.024736445397138596, -0...
9b92e12e5010c8eaaf24411bea262c3d7b182a59
subsection
79
84
Line-search: The Newton-secant solver
Unless the function g(\ell ) := \Psi _i(+ \ell _{ij}^n) has a special structure (say, linear or quadratic), the line-searches invoked at lines REF and REF in Algorithm REF require the use of an iterative procedure. Without claiming originality, we now show how the line-searches can be done by using the Newton-secant al...
{ "cite_spans": [] }
10.1016/j.cma.2018.11.036
1807.02563
Invariant domain preserving discretization-independent schemes and convex limiting for hyperbolic systems
[ "Jean-Luc Guermond", "Bojan Popov", "Ignacio Tomas" ]
[ "math.NA" ]
2,018
en
Mathematics
[ -0.07512563467025757, 0.008307455107569695, -0.0168742798268795, -0.01830844022333622, 0.02717277780175209, 0.008932993747293949, 0.0018451475771144032, 0.016569139435887337, -0.0052026486955583096, 0.04226198419928551, -0.01464675273746252, 0.002546017523854971, -0.008498167619109154, 0.0...