Contents lists available atScienceDirect
Journal of Computational Physics
www.elsevier.com/locate/jcp
Conservative discontinuous Galerkin methods for the nonlinear Serre equations
Jianli Zhao
a, Qian Zhang
a,1,2, Yang Yang
b,3, Yinhua Xia
a,∗,4aSchoolofMathematicalSciences,UniversityofScienceandTechnologyofChina,Hefei,Anhui230026,PRChina bDepartmentofMathematicalSciences,MichiganTechnologicalUniversity,Houghton,MI49931,USA
a r t i c l e i n f o a b s t r a c t
Articlehistory:
Received10January2020
Receivedinrevisedform17July2020 Accepted19July2020
Availableonline24July2020
Keywords:
Serreequations
DiscontinuousGalerkinmethods Conservativeschemes Well-balancedschemes
Inthispaper,wedevelopthreeconservativediscontinuousGalerkin(DG)schemesforthe one-dimensional nonlineardispersiveSerre equations,includingtwo conserved schemes for the equations in conservative form and a Hamiltonian conserved scheme for the equationsinnon-conservativeform.Oneoftheschemesownsthewell-balancedproperty viaconstructing ahighorderapproximation tothe source termforthe Serreequations with anon-flatbottom topography.By virtueofthe Hamiltonianstructure ofthe Serre equations,weintroduceanHamiltonianinvariantandthendevelopaDGschemewhich canpreservethediscreteversionofsuchaninvariant.Numericalexperimentsindifferent casesareperformedtoverifytheaccuracyandcapabilityoftheseDGschemesforsolving theSerreequations.
©2020ElsevierInc.Allrightsreserved.
1. Introduction
WeintroduceandstudythediscontinuousGalerkin(DG) methodsfortheSerreequations
⎧ ⎪
⎨
⎪ ⎩
ht
+ (
hu)
x=
0,
ut
+
uux+
ghx+ β
h−1(
h2γ )
x=
0, γ =
h(
ux2−
uxt−
uuxx),
(1.1)
whereh(x,t)representsthetotalwaterdepth,u(x,t)istheaveragedvelocityofthefluid, gisthegravitationalconstant,β isaparameterand
γ
isthefluidverticalaccelerationonthefreesurface.TheSerreequations(1.1) can beused todescribethe shallowwaterwithlarge amplitudeonthefree surface. In[40], weknowthatEulerequationsoftheincompressiblefluidflowdescribethepropagationofthefullwaterwavesonthefree surface. Sincethe difficulty insolving thisproblemis mainlycaused bythe freesurface, various approximatemodelsfor thiscasearepresented,andthecorresponding analyticaltheoriesandnumericalsimulations arealsodeveloped.Inthese
*
Correspondingauthor.E-mailaddresses:zjl88914@mail.ustc.edu.cn(J. Zhao),gelee@mail.ustc.edu.cn,qian77.zhang@polyu.edu.hk(Q. Zhang),yyang7@mtu.edu(Y. Yang), yhxia@ustc.edu.cn(Y. Xia).
1 Currentaddress:DepartmentofAppliedMathematics,TheHongKongPolytechnicUniversity,HongKong.
2 Researchsupportedbythe2019HongKongScholarProgram.
3 ResearchsupportedbyNSFgrantDMS-1818467.
4 ResearchsupportedbyNationalNumericalWindtunnelgrantNNW2019ZT4-B08andNSFCgrantNo.11871449.
https://doi.org/10.1016/j.jcp.2020.109729
0021-9991/©2020ElsevierInc.Allrightsreserved.
models,thepropagationofthewavesundertheassumptionofshallowwaterwithlargeamplitudeisgovernedby theSerre equations,whileforsmallamplitudeorweaklynonlinearshallowwaterwavesthemodelreducestotheclassicalBoussinesq system[40].TheSerreequationscanberegardedasanasymptoticapproximationofthetwo-dimensionalEulerequations.
Itisafullynonlineardispersivesystemthatmodelstwo-waypropagationoflongwavesonthesurfaceofanidealfluidina channel.Andifweignoreallthedispersiveterms,itwillbecomeahyperbolicsystemofshallowwaterwave.Thedispersive termscombinedbythetemporalandspatialderivativesarethemajordifficultyinsolvingtheSerreequations[16].
TheSerreequationswere firstderivedbyFrancoisSerre[36] in1953andthenrestudiedbySuandGardner[39],Green [22],Seabra-Santosetal.[35],GreenandNaghdiin[23] derivedthesystemwithtwospatialdimensionsforgeneraluneven bottomsin1976.LannesandBonnetonin[24] providedtheformalderivationoftheequations fromEulerequationsandthe justificationforseveralapproximativewaterwavemodelsofEulerequations.Furtherstudiescanbegotfromsomerecent literaturesbyBarthélemy[3],DiasandMilewski[16],CarterandCienfuegos[9],andthereferencestherein.
TherearevariousnumericalmethodsdealingwiththenonlinearSerreequationsincludinghigherorderdispersiveterms.
In [26], a pseudo-spectral methodwas proposed andused to solvethe one-dimensional Serreequations forfree surface waves in a homogeneous layer of an ideal fluid. In [29,7,8,2,18], finite difference and finite volume methods were also applied in simulating all the processes of wave transformation with nonlinear dispersive effect. For the equations with non-flat bottoms (Green-Naghdiequations),the well-balancedcentralDG method was proposed in[25],inwhich theel- liptic partwas solved by the continuous finiteelement method.A localDG (LDG)scheme was presentedin[31] for the one-dimensional Green-Naghdiequationsbasedonrotationalcharacteristicsinthevelocity field.Forthetwo-dimensional problem,DuranandMarchedevelopedaLDG methodbyrewritingtheequationsasahyperbolic-ellipticsystemin[17],and ahybridizedDGmethodwasintroducedin[34] forthesplittingsystem(thehyperbolicpartanddispersivepart).Asforthe shallowwaterequationsandotherhyperbolic conservationlaws,theconservative schemeisdemanded whenthesolution isnotsmoothenough.Inthispaper,wewilldeveloptheLDG methodsfortheSerreequationsbasedontheirconservation lawsandalsowell-balancedpropertyasintheshallowwaterequations.
TheDG methodwhichisaclassoffiniteelementmethodswasfirstintroducedbyReedandHill[33] forsolvingtheneu- trontransportlinearequationin1973.Completelydiscontinuouspiecewisepolynomialspaceisusedforthetestfunctions andthenumericalsolutionsinthismethod.Later,CockburnandShuhavemadeagreatcontributionindevelopingtheDG methodinaseriesofpapers[11,13–15].TheyusedthestablehighorderRunge-Kuttamethodsfortemporaldiscretization and theexact or approximateRiemann solvers asthe interface fluxesin spatial discretization.In addition, they alsode- signedtheLDGmethodforsolvingconvection-diffusionproblemscontainingsecondderivativeterms,whichwasmotivated by BassiandRebay[4] insolving thecompressibleNavier-Stokesequations.Atpresent, theDGmethodiswidely usedin many areas,such as gasdynamics,electromagnetism, magnetohydrodynamics, modeling ofshallow water,etc. Moreover, theDGmethodenjoysmanyattractiveproperties.Firstwecangethighorderaccuracyforthenumericalsolutionbyusing element-based test functions. Second theDG methodcan be appliedon arbitraryunstructuredmesheswiththe suitable choices ofthe interelement boundary fluxes. Finally,parallel computing caneasily be applieddue to localdata commu- nications. In addition to the advantages given above, the LDG method contains auxiliary variables that approximate the gradientsoftheprimarysolutionstotheexpectedorderofaccuracy.Thentheauxiliaryvariablescanbelocallyeliminated.
TheLDGmethodshavebeengeneralizedtononlinearwaveequationsin[49,50,43–47,51],suchasKdVequation,Camassa- Holm equationandnonlinearSchrödinger equation,etc.Werefertothereviewpaper[48] ofLDGmethodsforhigh-order time-dependentpartialdifferentialequationsandthereferencestherein.
Mostrecently, aseriesof schemesso calledstructure-preservingschemesfornonlinear waveequations havebeenat- tracted considerable attention. For some integrableequations, such as KdVtype equations[6,27,52,53], Zakharovsystem [42], Schrödinger-KdV system [41], etc., the authors proposed various conservative numerical schemes to “preserve the structures”. Forthe Serreequations,thealiasing orspurious dissipativeeffects willresultinsignificant errors forsolving large amplitudeor rapidly oscillating waves, such as the dispersive shockwaves. Therefore, Mitsotakis andDutykh [30]
constructedanon-dissipativefiniteelementmethodtoapproximatetherapidlyoscillatingwavesaccurately.Afundamental propertyoftheSerreequationsisitsHamiltonianstructure.Fromthefollowingconservationlawsderivedbytheequations (1.1),
(
hu)
t+ (
hu2+
12gh2
+ β
h2γ )
x=
0,
(1.2)(
hu− β(
h3ux)
x)
t+ (
hu2+
12gh2
−
2β
h3u2x− β
h3uuxx−
3β
h2hxuux)
x=
0,
(1.3)(
u− β
h−1(
h3ux)
x)
t+ (
12u2
+
gh−
12h2u2x
− β
uh−1(
h3ux)
x)
x=
0,
(1.4)(
12hu2
+
12
β
h3u2x+
12gh2
)
t+ ((
1 2u2+
12
β
h2u2x+
gh+ β
hγ )
hu)
x=
0,
(1.5) wehavethecorrespondingconservativequantities E0, E1,E2andHamiltonianHintherelevantdomain I,E0
=
I
hudx
,
E1
=
I
(
hu− β(
h3ux)
x)
dx,
E2
=
I
(
u− β
h−1(
h3ux)
x)
dx,
H
=
I
1
2
(
hu2+ β
h3u2x+
gh2)
dx.
Thenumericalschemescanbeestablishedbasedontheaboveconservedquantities.Fortheform(1.2),themaindifficulty indesigning E0 conservednumericalschemeliesinthemixedspatial andtemporalderivativesin
γ
.Therefore,wedesign andimplementthreekinds ofnumericalschemeswhich areall basedon theDGmethods for theone-dimensional Serre equations.ThefirsttwoareconstructedforE1andE2 intheconservativeform,andthelastonepreservestheHamiltonian invariantHinanon-conservativeform.TheE1conservedschemecanmaintainthestill-waterstationaryforthewaterwave equationsovernon-flatbottomtopography,socalledwell-balancedscheme.FortheHamiltonian(H)conservedDGscheme, we takeadvantage oftheDGmethodforHamilton-Jacobi equations[10] todeal withthe nonlinearterms.Moreover,the stabilityoftheHamiltonianfortheSerreequationswouldbeprovedtheoretically.Thepresentpaperisorganizedasfollows.InSection2,theE1andE2conservedDGschemesaredesignedindetail.For the E1 conservedscheme,we alsodiscussandproveits well-balancedpropertywhenthe bottomisnotflat.Section 3is devotedtoelaborate theHamiltonian(H) conservedDGscheme.Moreover, thestabilitybasedon theHamiltonianisalso giveninthispart.InSection4,numericalexperimentsindifferentcircumstancesareprovidedtoillustratetheaccuracyand capabilityoftheseschemes.ConcludingremarksaregiveninSection5.
2. TheLDGdiscretizationinconservativeform
Inthissection,wedevotetodesignLDGschemesfortheSerreequationsby adoptingtheconservationlaws E1 andE2 in(1.3) and(1.4) respectively,andthengivethecorrespondingalgorithmflowcharts.
Todiscretizethespatial derivatives,we proceedasfollows.Foreach partitionoftheinterval I= [a,b],wesetthecells and cell centers as Ij= [xj−1
2,xj+1
2], xj=12(xj−1 2 +xj+1
2) for j=1,· · ·,N, and denote the mesh size x= max
1≤j≤Nxj withxj=xj+12 −xj−12. The meshis assumedto be quasi-uniform, namelythere existsa positive constant c such that x≤c min
1≤j≤Nxj.
Weseekthenumericalapproximationsbelongingtothefiniteelementspace Vh
= {
v(
x) :
v(
x) ∈
Pk(
Ij)
forx∈
Ij,
j=
1,· · · ,
N},
where Pk(Ij) denotespolynomialsin Ij ofdegree upto k. Ittranspiresthat the functionsbelonging to Vh could bedis- continuousacross thecell interfaces.Forv∈Vh,thevaluesofv atxj+1
2,fromtheleftandrightcells Ij andIj+1,canbe regardedas v−
j+12
and v+ j+12
,respectively.The notations[[v]]=v+−v− and{{v}}= 12(v++v−)areusedto representthe jumpandthemeanofthefunctionatthecellinterfaces.
2.1. E1conservedDGscheme
Inordertotreatthemixedspatialandtemporalderivativesin(1.1),weadopttheconservationlaw(1.3) anddevelopthe E1 conservedDGschemeforthefollowingsystem
ht
+ (
hu)
x=
0,
(2.1a)zt
+ (
uz+
12gh2
−
2β
h3q2)
x=
0,
(2.1b)hu
− β(
h3q)
x=
z,
(2.1c)q
−
ux=
0,
(2.1d)with periodic boundary conditions.To design the LDG method,we accomplish thecomplete schemes for the system as follows.Forsimplicity,westilluseh,u,zandqtodenotethenumericalsolutions.
Forthetime-dependentequations(2.1a) and(2.1b),assumethat uandq areknownandhandz aretobesolved,then wecanformulatetheDGmethodasfollows:Findh,z∈Vh suchthat,foralltestfunctionsθ,
ϕ
∈Vh,Ij
ht
θ
dx−
Ij
(
hu)θ
xdx+ (
f1θ
−)
j+12
− (
f1θ
+)
j−12
=
0,
(2.2)Ij
zt
ϕ
dx−
Ij
(
uz+
12gh2
−
2β
h3q2) ϕ
xdx+ (
f2ϕ
−)
j+12
− (
f2ϕ
+)
j−12
=
0,
(2.3)∀j=1,· · ·,N. The “hat” terms f1 and f2 appear on the cell boundaryterms are the so-called numericalfluxes, which are single valuedfunctionsfrom integrationby partsandshould be designedto ensurethe stability ofthe schemes.For conservationlaws,thenumericalfluxesareusuallyLipschitzcontinuousfunctionsandpreservethemonotonepropertyand consistentproperty,namedthemonotonefluxes[12].Here,wecanchoosethefollowingdissipativenumericalflux
f1= {{
hu}} − α
2
[[
h]],
(2.4) f2= {{
uz+
12gh2
−
2β
h3q2}} − α
2
[[
z]] ,
(2.5)where
α
=max(|u| +gh)isthemaximumtakenintherelevantinterval.Wehaveomittedthehalf-integerindices j+12 asallquantitiesin(2.4) arecomputedatthesamecellinterfaces.
Forthetime-independentequations(2.1c) and(2.1d),assumethathandz areknownandu andqaretobesolved,the LDGschemecanbeformulatedasfollows:Findu,q∈Vh suchthat,foralltestfunctions
ρ
,φ∈Vh,Ij
hu
ρ
dx+
Ij
β
h3qρ
xdx− (
h3qρ
−)
j+12
+ (
h3qρ
+)
j−12
=
Ij
z
ρ
dx,
(2.6a)Ij
q
φ
dx+
Ij
u
φ
xdx− (
uφ
−)
j+12
+ (
uφ
+)
j−12
=
0.
(2.6b)Herewetakethealternatingfluxes[48] suchas
h3q
= (
h3q)
−,
u=
u+.
(2.7)2.1.1. Algorithmflowchart(I)
Thispartshowstheimplementationaboutthefully-discrete E1 conservedscheme.Forsimplicity,we choosetheEuler forwardmethodto illustratethenumericalevolutionfromtimeleveltn totime leveltn+1.Innumericalexperiments,the explicitSSP/TVDRunge-Kuttamethods[37,38] areadopted,whichcanbeviewedastheconvexcombinationsoftheEuler forwardmethod.
Leth,u,zandqdenotethevectorscontainingthedegreesoffreedomforthenumericalsolutionsh,u,zandq,respec- tively.Giventhesolutionhn,un,znandqnatthetimeleveltn,thealgorithmcanbedividedinto twosteps.
StepI:Fromtheequations(2.2) and(2.3) equippedwiththecorrespondingnumericalfluxes(2.4) and(2.5),wehavethe fullydiscreteschemeinthematrixform:
hn+1
−
hntn
=
res1(
hn,
un),
(2.8)zn+1
−
zntn
=
res2(
hn,
un,
zn,
qn),
(2.9)bytheEulerforwardmethodwiththetimesteptn=tn+1−tn.
StepII:After weget hn+1 andzn+1 fromthe above,basedontheschemes (2.6a) and (2.6b) withthefluxes(2.7),we cansolveun+1,qn+1inthefollowingmatrixform
A
(
hn+1)
un+1+
B(
hn+1)
qn+1=
zn+1,
(2.10)qn+1
+
Cun+1=
0,
(2.11)where the matrices A, B and C are the corresponding discrete operators in the scheme. Then we get all the numerical solutionsattimeleveltn+1.
2.1.2. Thewell-balancedscheme
In thissubsection,weturnto presentthe well-balancedpropertyof E1 conserved DGschemefortheSerreequations withanon-flatbottom,alsoknownasGreenandNaghdiequationsin[23].Thereinto,hcanberepresentedasthesumofd and
η
,whereddenotesthebottomtopographyandη
isthefreesurfaceelevation.Inconsiderationoftheone-dimensional shallowwaterflowswithnon-flatbottomtopography,themodelcanbeformulatedasfollows:ht
+ (
hu)
x=
0,
(2.12a)(
hu)
t+ (
hu2+
12gh2
+ β
h2γ +
12h2
)
x= (
gh+
12h
γ +
h)
dx,
(2.12b)with
= −
dxut−
dxuux−
dxxu2.
(2.13)Itcanbeseenthatfortheflatbottom,meaningd(x)isaconstant,theequation(2.12b) isequivalenttotheequation(1.2).
Thewell-balancedmethods[5] areintroducedtoexactlypreserve thestill-waterstationarysolutionatadiscrete level.
FortheSerreequations,thestill-waterstationarywavesolutioncanberepresentedasfollows,
u
=
0 andη =
h−
d=
constant.
(2.14)Inordertodealwiththemixedspatialandtemporalderivativesin(2.12b),weintroduceanewunknownvariable
ω = −β(
h3ux)
x+
h(
1−
hxdx−
12hdxx
+
d2x)
u.
(2.15)Noticethat
ω
=zwiththeflatbottom.Theequation(2.15) canbediscretizedsimilarlyastheLDGscheme(2.6a) and(2.6b) forz.Thusweomitthedetailshere.Asin[25],werewritethetime-dependentpartofthesystem(2.12a)–(2.12b) intoabalancelaw
η
t+ (
hu)
x=
0,
(2.16)ω
t+ ( ω
u+
12gh2
−
2β
h3u2x−
h2uuxdx)
x=
ghdx+
12h2uuxdxx
+
hu2dxdxx.
(2.17) Denoting U=(η
,ω
)T astheunknowns,weformulatethe E1 conservedschemeforthebalanceequationsasfollows:Findη
,ω
∈Vh suchthat,foralltestfunctionsθ1,θ2∈Vh,Ij
Ut
·
dx−
Ij
F
·
xdx+ (
F·
−)
j+12
− (
F·
+)
j−12
=
Ij
S
·
dx, ∀
j=
1, · · · ,
N,
(2.18)whereequals(θ1,θ2)T,thefluxistakenas F
(
U) = (
hu, ω
u+
12gh2
−
2β
h3u2x−
h2uuxdx)
T,
andthesourcetermisgivenby S
= (
0,
ghdx+
12h2uuxdxx
+
hu2dxdxx)
T.
Ingeneral, the numericalscheme(2.18) does not havethewell-balanced property.Based onthe ideain [32], wedis- cretizethesourcetermasfollows:
Ij
Ut
·
dx−
Ij
F
·
xdx+ (
F·
−)
j+12
− (
F·
+)
j−12
=
Ij
S1
·
dx+
Sj,
(2.19)with
S1
= (
0,
12h2uuxdxx
+
hu2dxdxx)
T,
and
Sj= (
0, ˜
sj)
T,
˜
sj=
12g
( −
Ij
d2
(θ
2)
xdx+ {{
d2j+1 2}} (θ
2)
−j+12
− {{
d2j−1 2}} (θ
2)
+j−12
)
+
gη ¯
j(−
Ij
d
(θ
2)
xdx+ {{
dj+1 2}}(θ
2)
−j+12
− {{
dj−1 2}}(θ
2)
+j−12
)
+
Ij
g
( η − ¯ η
j)
dxθ
2dx,
(2.20)wheres˜j isthehighorderapproximationofthesourceterm
Ijghdxθ2dxand
η
¯j=I1jIj
η
dxdenotesthemeanofη
over Ij.F isamonotoneflux,e.g.theLax-Friedrichsflux.In ordertoensuretheunique solvabilityoftheequation (2.15) in thestill-waterstationary,thefollowingassumptions areneeded[25]:
(A1) Thewaterheighthinthestill-waterstationarysolutionconsideredherehasastrictlypositivelowerbound.
(A2) 1−12hdxx>0.
Nowweturntoprovethewell-balancedpropertyofthemodifiedscheme(2.19).
Proposition2.1.(Well-balancedproperty)Undertheassumptions(A1)and(A2),themodifiedE1conservedDGschemedescribedby (2.19) fortheSerreequationswithnon-flatbottomhasthewell-balancedproperty.
Proof. Let Un (n∈N) be the numericalsolution of (2.19) at time level tn.For brevity, herewe consider Eulerforward time discretization,whichcanbeeasilygeneralizedtotheSSP/TVDRunge-Kuttamethods.Thenweapplytheminto(2.19), leadingto
Ij
Un+1
·
dx=
Ij
Un
·
dx+
tn(
Ij
Fn
·
xdx− (
Fn·
−)
j+1 2+(
Fn·
+)
j−12
+
Ij
Sn1
·
dx+ (
Sn)
j).
(2.21)Byinduction,firstweassume
Un
= (
C0,
0)
T and un=
0,
(2.22)whereC0 isaconstant.Fromthisassumptionwewanttodeducethatthenumericalsolutioncomputedfrom(2.21) satisfies
Un+1
= (
C0,
0)
T and un+1=
0.
(2.23)Byvirtueoftheassumptions(2.22),theequation(2.21) becomes
η
n+1= η
n=
C0,
(2.24)Ij
ω
n+1θ
2dx=
tnIj
1
2g
(
hn)
2(θ
2)
xdx− (
12g
{{(
hn)
2}}(θ
2)
−)
j+1 2+ (
12g
{{ (
hn)
2}} (θ
2)
+)
j−1 2+ (
sn)
j.
(2.25)Denotingthesumoftheright-handsideofequation((2.25))by Rj,wecanget Rj
=
tnIj
1
2g
(
hn)
2(θ
2)
xdx− (
12g
{{ (
hn)
2}} (θ
2)
−)
j+1 2+ (
12g
{{ (
hn)
2}} (θ
2)
+)
j−1 2+ (
sn)
j=
tnIj
1
2g
((
hn)
2− (
dn)
2−
2η ¯
njdn)(θ
2)
xdx−
12g
({{(
hn)
2j+12
}} − {{(
dn)
2j+12
}} −
2η ¯
nj{{
dnj+12
}})(θ
2)
−j+12
+
12g
( {{ (
hn)
2j−12
}} − {{ (
dn)
2j−12
}} −
2η ¯
nj{{
dnj−12
}} )(θ
2)
+j−12
+
Ij
g
( η
n− ¯ η
nj)
dnxθ
2dx=
tnIj
1
2g
( η
n)
2(θ
2)
xdx−
1 2g{{( η
n)
2j+12
}}(θ
2)
−j+12
+
1 2g{{( η
n)
2j−12
}}(θ
2)
+j−12
=
0,
(2.26)wherewetakeuseoftherelationshiph=d+
η
andthefactη
n= ¯η
nj=C0 by(2.22).Thentheequations(2.24)–(2.25) lead to Un+1=(C0,0)T.Astheassumptions (A1)and(A2)ensure theuniquesolvabilityofthefiniteelementdiscretization,it impliesun+1=0.Thiscompletesourproof.2.2. E2conservedDGscheme
SimilartothedevelopmentoftheE1 conservedscheme,weadopt(1.1) and(1.4) toconstructtheE2conservedscheme.
TheSerreequationsarerewritten asfollows:
ht
+ (
hu)
x=
0,
(2.27a) zt+ (
uz−
12u2
+
gh−
12h2q2
)
x=
0,
(2.27b)u
−
h−1p=
z,
(2.27c)p
− β(
h3q)
x=
0,
(2.27d)q
−
ux=
0.
(2.27e)Without ambiguity, we still use h,z,u,p,q to denote the numerical approximations. The E2 conserved DG scheme for (2.27a)-(2.27e) canbeformulatedasfollows:Findh,u,q,z,p∈Vh,suchthat,foralltestfunctionsθ,
ϕ
,ρ
,ψ andφ∈Vh,Ij
ht
θ
dx−
Ij
(
hu)θ
xdx+ (
f1θ
−)
j+12
− (
f1θ
+)
j−12
=
0,
(2.28)Ij
ztϕ
dx−
Ij
(
uz−
12u2
+
gh−
12h2q2
) ϕ
xdx+ (
f4ϕ
−)
j+12
− (
f4ϕ
+)
j−12
=
0,
(2.29)Ij
p
ρ
dx+
Ij
β
h3qρ
xdx− (
h3qρ
−)
j+12
+ (
h3qρ
+)
j−12
=
0,
(2.30)Ij
u
ψ
dx−
Ij
h−1p
ψ
dx=
Ij
zψ
dx,
(2.31)Ij
q
φ
dx+
Ij
u
φ
xdx− (
uφ
−)
j+12
+ (
uφ
+)
j−12
=
0,
(2.32)wherethenumericalfluxes arechosenas
f1= {{
hu}} − α
2
[[
h]] ,
(2.33) f4= {{
uz−
12u2
+
gh−
12h2q2
}} − α
2
[[
z]],
(2.34)h2q
= (
h3q)
−,
u=
u+,
(2.35)where
α
=max(|u| +gh)isthemaximumtakenintherelevantinterval.
2.2.1. Algorithmflowchart(II)
Similarto thefullydiscrete E1 conservedscheme,the E2 conservedschemecoupledwiththeEulerforwardtimedis- cretizationcanbedividedintwosteps.Giventhesolutionhn,un,z˜n,pn andqn atthetimeleveltn,thealgorithmcanbe dividedinto twosteps.
StepI:Fromthetimedependentequations(2.28) and(2.29) withthefluxes(2.33) and(2.34) wehavethefullydiscrete schemeinthematrixform:
hn+1
−
hntn
=
res1(
hn,
un),
(2.36)˜
zn+1
− ˜
zntn
=
res4(
hn,
un,
z˜
n,
qn),
(2.37)bytheEulerforwardmethodwiththetimesteptn=tn+1−tn.
StepII:Afterwegethn+1 andz˜n+1fromthelaststep,basedonthetimeindependentschemes(2.30),(2.31) and(2.32) withthenumericalfluxes(2.35),wecansolvethecorrespondinglinearsystemtogetun+1,pn+1,qn+1.
Thusweobtainallthenumericalsolutionsatthenexttimesteptn+1.
3. TheDGdiscretizationinnon-conservativeform
Inthissection,wedesignandproveaHamiltonian(H)conservedDGschemeandgiveitsalgorithmflowchart.
Combining (1.1) and (1.3) and after some manipulation, we can obtain the Serre equations in the following non- conservativeform:
ht
+ (
hu)
x=
0,
(3.1)hut
− β(
h3ux)
xt+
huux− β(
u(
h3ux)
x)
x+
12g
(
h2)
x−
2β(
h3u2x)
x=
0.
(3.2) Introducingsomeauxiliaryvariablesm,p,q,s,w,k,v,r,m
−
hu=
0,
p−
12u2
=
0,
s−
h3q=
0,
q−
ux=
0,
w−
sx=
0,
k− (
hu)
x=
0,
v−
32h2qk
=
0,
r−
12h2qmx
−
h2mqx−
hqmhx=
0,
(3.3)
theSerreequationscanbewrittenasfollows:
ht
+ (
hu)
x=
0,
(3.4)hut
− β
wt+
hpx+
12g
(
h2)
x− β(
v+
r)
x=
0.
(3.5)We willintroduce thestandard L2 projection operatorP ofafunction
ω
intothe finiteelementspace Vh,i.e., forall v∈Pk(Ij),Ij
(P ( ω ) − ω )
vdx=
0, ∀
j=
1,· · · ,
N,
(3.6)anddefinethediscretederivativeoperator Dj
( ω ; ρ ) =
Ij
ωρ
xdx− ( ωρ
−)
j+12
+ ( ωρ
+)
j−1 2,
where the numerical flux
ω
ˆ is to be determined. Then for the equations (3.4), (3.5) and the auxiliary equations (3.3), the H conserved DG scheme can be constructed as follows: Find h,u,w,q,k,r∈Vh such that, for all test functions θ,φ,ζ,τ
,γ
,μ
∈Vh,Ij
ht
θ
dx−
Dj(
hu; θ ) =
0,
(3.7)Ij
hut
φ
dx− β
Ij
wt
φ
dx+
Dj(
hφ ;
p) −
12g Dj
(
h2; φ) + β
Dj(
v+
r; φ) =
0,
(3.8)Ij
q
τ
dx+
Dj(
u; τ ) =
0,
(3.9)Ij
k
γ
dx+
Dj(
hu; γ ) =
0,
(3.10)Ij
w
ζ
dx+
Dj(
s; ζ ) =
0,
(3.11)Ij
r
μ
dx−
12Dj
(
h2qμ ;
m) −
Dj(
h2mμ ;
q) −
Dj(
hmqμ ;
h) =
0,
(3.12)withm=P(hu),p=P(12u2),s=P(h3q)andv=P(32h2qk)andthenumericalfluxes
hu
= {{
hu}} ,
hφ = {{
hφ }} ,
h2=
h+h−,
(3.13)v
+
r=
v++
r+,
s=
s+,
u=
u−,
(3.14)h
2qμ = {{
h2qμ }},
h2mμ = {{
h}}{{
hμ }}{{
m}},
hmqμ = {{
hμ }}{{
m}}{{
q}}.
(3.15)Herethechoicesofthenumericalfluxesintheschemearenotunique,whichcanbeseenfromtheproofofitsconservation ofHamiltonian.
3.1. H-stability
Inthispart,weprovetheHamiltonianconservationoftheabovescheme.Forbrevity,wedefine
(
v,
w) :=
N j=1Ij
v wdx
,
D(
v;
w) :=
N j=1Dj
(
v;
w).
(3.16)Proposition3.1.(H-stability)Thesolutiontothescheme(3.7)–(3.12) withthenumericalfluxes(3.13)–(3.15) satisfiesthefollowing H-stability:
d dt
I
1
2
(
hu2+
gh2+ β
h3q2)
dx=
0.
(3.17)Proof. SumminguptheDGschemes(3.7)–(3.12) over jfrom1 toN,wehave
(
ht, θ ) −
D(
hu; θ ) =
0,
(3.18)(
hut, φ) − β(
wt, φ) +
D(
hφ;
p) −
12g D
(
h2; φ) + β
D(
v+
r; φ) =
0,
(3.19)(
w, ζ ) +
D(
s; ζ ) =
0,
(3.20)(
q, τ ) +
D(
u; τ ) =
0,
(3.21)(
k, γ ) +
D(
hu; γ ) =
0,
(3.22)(
r, μ ) −
D(
12h2q
μ ;
m) −
D(
h2mμ ;
q) −
D(
hmqμ ;
h) =
0.
(3.23) Byintroducinganewauxiliaryvariableν
=P(32h2q2)andtakingthetestfunctionsθ=p+gh−βν
in(3.18) andφ=u in (3.19),itfollowsthat(
ht,
p+
gh− β ν ) −
D(
hu;
p+
gh− β ν ) =
0, (
hut,
u) − β(
wt,
u) +
D(
hu;
p) −
12g D
(
h2;
u) + β
D(
v+
r;
u) =
0.
Ifthefollowingidentitieshold D
(
hu;
gh) +
12g D
(
h2;
u) =
0,
(3.24)D
(
hu; ν ) +
D(
v;
u) =
0,
(3.25)D
(
r;
u) =
0,
(3.26)thenwehave