c c ========================================================= subroutine rpn2eurhok(ixy,maxm,meqn,mwaves,mbc,mx,ql,qr,maux, & auxl,auxr,wave,s,amdq,apdq) c ========================================================= c c # solve Riemann problems for the thermally perfect 1D multi-component c # Euler equations using a Roe-type approximate Riemann solver. c # Scheme is blended with HLL for robustness and uses a multi-dimensional c # entropy correction to prevent the carbuncle phenomenon. c c # On input, ql contains the state vector at the left edge of each cell c # qr contains the state vector at the right edge of each cell c # This data is along a slice in the x-direction if ixy=1 c # or the y-direction if ixy=2. c # On output, wave contains the waves, c # s the speeds, c # amdq the left-going flux difference A^- \Delta q c # apdq the right-going flux difference A^+ \Delta q c c # Note that the i'th Riemann problem has left state qr(i-1,:) c # and right state ql(i,:) c # From the basic clawpack routine step1, rp is called with ql = qr = q. c c # Copyright (C) 2002 Ralf Deiterding c # Brandenburgische Universitaet Cottbus c implicit double precision (a-h,o-z) c include "ck.i" c dimension ql(1-mbc:maxm+mbc, meqn) dimension qr(1-mbc:maxm+mbc, meqn) dimension s(1-mbc:maxm+mbc, mwaves) dimension wave(1-mbc:maxm+mbc, meqn, mwaves) dimension auxl(1-mbc:maxm+mbc, maux) dimension auxr(1-mbc:maxm+mbc, maux) dimension amdq(1-mbc:maxm+mbc, meqn) dimension apdq(1-mbc:maxm+mbc, meqn) c c # local storage c --------------- parameter (maxm2 = 10005) !# assumes at most 10000x10000 grid with mbc=5 parameter (minm2 = -4) !# assumes at most mbc=5 common /comroe/ u(minm2:maxm2), v(minm2:maxm2), u2v2(minm2:maxm2), & enth(minm2:maxm2), a(minm2:maxm2), g1a2(minm2:maxm2), & dpY(minm2:maxm2), Y(LeNsp,minm2:maxm2), pk(LeNsp,minm2:maxm2) logical efix, pfix, hll, roe, hllfix c c define local arrays c dimension delta(LeNsp+3) dimension rkl(LeNsp), rkr(LeNsp) dimension hkl(LeNsp), hkr(LeNsp) dimension fl(minm2:maxm2,LeNsp+4), fr(minm2:maxm2,LeNsp+4) c data efix /.true./ !# use entropy fix data pfix /.true./ !# use Larrouturou's positivity fix for species data hll /.true./ !# use HLL instead of Roe solver, if unphysical values occur data roe /.true./ !# turn off Roe solver when debugging HLL c c # Riemann solver returns fluxes c ------------ common /rpnflx/ mrpnflx mrpnflx = 0 c if (minm2.gt.1-mbc .or. maxm2 .lt. maxm+mbc) then write(6,*) 'need to increase maxm2 in rpA' stop endif c c # set mu to point to the component of the system that corresponds c # to momentum in the direction of this slice, mv to the orthogonal c # momentum: c if (ixy.eq.1) then mu = Nsp+1 mv = Nsp+2 else mu = Nsp+2 mv = Nsp+1 endif mE = Nsp+3 mT = Nsp+4 c c # note that notation for u and v reflects assumption that the c # Riemann problems are in the x-direction with u in the normal c # direciton and v in the orthogonal direcion, but with the above c # definitions of mu and mv the routine also works with ixy=2 c # and returns, for example, f0 as the Godunov flux g0 for the c # Riemann problems u_t + g(u)_y = 0 in the y-direction. c c c # compute the Roe-averaged variables needed in the Roe solver. c # These are stored in the common block comroe since they are c # later used in routine rpt2eurhok to do the transverse wave splitting. c do 20 i=2-mbc,mx+mbc rhol = 0.d0 rhor = 0.d0 do k = 1, Nsp rkl(k) = qr(i-1,k) rkr(k) = ql(i ,k) rhol = rhol + rkl(k) rhor = rhor + rkr(k) enddo if( rhol.le.1.d-10 ) then write(6,*) 'negative total density, left', rhol stop endif if( rhor.le.1.d-10 ) then write(6,*) 'negative total density, right', rhor stop endif c c # compute left/right rho/W and rho*Cp c rhoWl = 0.d0 rhoWr = 0.d0 do k = 1, Nsp rhoWl = rhoWl + rkl(k)/Wk(k) rhoWr = rhoWr + rkr(k)/Wk(k) enddo c c # calculate left/right Temperatures c rhoel = qr(i-1,mE)-0.5d0*(qr(i-1,mu)**2+qr(i-1,mv)**2)/rhol call SolveTrhok(qr(i-1,mT),rhoel,rhoWl,rkl,Nsp,ifail) rhoer = ql(i ,mE)-0.5d0*(ql(i ,mu)**2+ql(i ,mv)**2)/rhor call SolveTrhok(ql(i ,mT),rhoer,rhoWr,rkr,Nsp,ifail) c Tl = qr(i-1,mT) Tr = ql(i ,mT) pl = rhoWl*RU*Tl pr = rhoWr*RU*Tr c c # compute quantities for rho-average c rhsqrtl = dsqrt(rhol) rhsqrtr = dsqrt(rhor) rhsq2 = rhsqrtl + rhsqrtr c c # find rho-averaged specific velocity and enthalpy c u(i) = (qr(i-1,mu)/rhsqrtl + ql(i,mu)/rhsqrtr) / rhsq2 v(i) = (qr(i-1,mv)/rhsqrtl + ql(i,mv)/rhsqrtr) / rhsq2 u2v2(i) = u(i)**2 + v(i)**2 enth(i) = (((qr(i-1,mE)+pl)/rhsqrtl & + (ql(i ,mE)+pr)/rhsqrtr)) / rhsq2 c c # compute rho-averages for T, cp, and W c T = (Tl * rhsqrtl + Tr * rhsqrtr) / rhsq2 W = rhsq2 / (rhoWl/rhsqrtl + rhoWr/rhsqrtr) c c # evaluate left/right entropies and mean cp c call tabintp( Tl, hkl, hms, Nsp ) call tabintp( Tr, hkr, hms, Nsp ) do k = 1, Nsp Y(k,i) = (rkl(k)/rhsqrtl + rkr(k)/rhsqrtr) / rhsq2 enddo Cp = Cpmix( Tl, Tr, hkl, hkr, Y(1,i) ) gamma1 = RU / ( W*Cp - RU ) gamma = gamma1 + 1.d0 c c # find rho-averaged specific enthalpies, c # compute rho-averaged mass fractions and c # compute partial pressure derivatives c tmp = gamma * RU * T / gamma1 ht = 0.d0 do k = 1, Nsp hk = (hkl(k)*rhsqrtl + hkr(k)*rhsqrtr) / rhsq2 pk(k,i) = 0.5d0*u2v2(i) - hk + tmp / Wk(k) enddo c c # compute speed of sound c dpY(i) = 0.d0 do k = 1, Nsp dpY(i) = dpY(i) + pk(k,i) * Y(k,i) enddo a2 = dpY(i) + enth(i)-u2v2(i) g1a2(i) = 1.d0 / a2 a(i) = dsqrt(gamma1*a2) c 20 continue c c do 30 i=2-mbc,mx+mbc c c # find a1 thru a4, the coefficients of the Nsp+3 eigenvectors: c dpdr = 0.d0 drho = 0.d0 do k = 1, Nsp delta(k) = ql(i,k) - qr(i-1,k) drho = drho + delta(k) dpdr = dpdr + pk(k,i) * delta(k) enddo delta(mu) = ql(i,mu) - qr(i-1,mu) delta(mv) = ql(i,mv) - qr(i-1,mv) delta(mE) = ql(i,mE) - qr(i-1,mE) c a2 = g1a2(i)*(dpdr - ( u(i)*delta(mu) + v(i)*delta(mv) ) & + delta(mE) ) a3 = delta(mv) - v(i)*drho a4 = 0.5d0*( a2 - ( u(i)*drho - delta(mu) )/a(i) ) a1 = a2 - a4 c c # Compute the waves. c # Note that the 1+k-waves, for 1 .le. k .le. Nsp travel at c # the same speed and are lumped together in wave(.,.,2). c # The 3-wave is then stored in wave(.,.,3). c do k = 1, Nsp c # 1-wave wave(i,k,1) = a1*Y(k,i) c # 2-wave wave(i,k,2) = delta(k) - Y(k,i)*a2 c # 3-wave wave(i,k,3) = a4*Y(k,i) enddo c # 1-wave wave(i,mu,1) = a1*(u(i) - a(i)) wave(i,mv,1) = a1*v(i) wave(i,mE,1) = a1*(enth(i) - u(i)*a(i)) wave(i,mT,1) = 0.d0 s(i,1) = u(i)-a(i) c c # 2-wave wave(i,mu,2) = (drho - a2)*u(i) wave(i,mv,2) = (drho - a2)*v(i) + a3 wave(i,mE,2) = (drho - a2)*u2v2(i) & - dpdr + dpY(i)*a2 + a3*v(i) wave(i,mT,2) = 0.d0 s(i,2) = u(i) c c # 3-wave wave(i,mu,3) = a4*(u(i) + a(i)) wave(i,mv,3) = a4*v(i) wave(i,mE,3) = a4*(enth(i) + u(i)*a(i)) wave(i,mT,3) = 0.d0 s(i,3) = u(i)+a(i) c 30 continue c c # compute flux differences as c # (+/-) c # A (Ur-Ul) = 0.5*( f(Ur)-f(Ul) +/- |A|(Ur-Ul) ) c -------------------------- c call flx2(ixy,maxm,meqn,mbc,mx,qr,maux,auxr,apdq) call flx2(ixy,maxm,meqn,mbc,mx,ql,maux,auxl,amdq) c do 35 i = 1-mbc, mx+mbc do 35 m=1,meqn fl(i,m) = amdq(i,m) fr(i,m) = apdq(i,m) 35 continue c if (roe) then do 40 i = 2-mbc, mx+mbc do 40 m=1,meqn amdq(i,m) = 0.5d0*(fl(i,m)-fr(i-1,m)) 40 continue c do 50 i = 2-mbc, mx+mbc do 50 m=1,meqn sw = 0.d0 do 60 mw=1,mwaves sl = dabs(s(i,mw)) if (efix.and.dabs(s(i,mw)).lt.auxl(i,ixy)) & sl = s(i,mw)**2/(2.d0*auxl(i,ixy))+ & 0.5d0*auxl(i,ixy) sw = sw + sl*wave(i,m,mw) 60 continue amdq(i,m) = amdq(i,m) - 0.5d0*sw apdq(i,m) = amdq(i,m) + sw 50 continue endif c if (hll) then do 55 i = 2-mbc, mx+mbc c # set this to hllfix = .true. when debugging HLL hllfix = .false. if (.not.roe) hllfix = .true. c rhol = 0.d0 rhoWl = 0.d0 do k = 1, Nsp rkl(k) = qr(i-1,k) + wave(i,k,1) rhol = rhol + rkl(k) rhoWl = rhoWl + rkl(k)/Wk(k) enddo rhoul = qr(i-1,mu) + wave(i,mu,1) ul = rhoul/rhol rhovl = qr(i-1,mv) + wave(i,mv,1) rhoEl = qr(i-1,mE) + wave(i,mE,1) Tl = qr(i-1,mT) rhoel = rhoEl - 0.5d0*(rhoul**2+rhovl**2)/rhol call SolveTrhok( Tl, rhoel, rhoWl, rkl, Nsp, ifail) rhoCpl = avgtabip( Tl, rkl, cpk, Nsp ) gammal = RU / ( rhoCpl/rhoWl - RU ) + 1.d0 pl = rhoWl*RU*Tl al = dsqrt(gammal*pl/rhol) if (rhol.le.0.d0.or.pl.le.0.d0) hllfix = .true. c rhor = 0.d0 rhoWr = 0.d0 do k = 1, Nsp rkr(k) = ql(i ,k) - wave(i,k,3) rhor = rhor + rkr(k) rhoWr = rhoWr + rkr(k)/Wk(k) enddo rhour = ql(i ,mu) - wave(i,mu,3) ur = rhoul/rhol rhovr = ql(i ,mv) - wave(i,mv,3) rhoEr = ql(i ,mE) - wave(i,mE,3) Tr = ql(i ,mT) rhoer = rhoEr - 0.5d0*(rhour**2+rhovr**2)/rhor call SolveTrhok( Tr, rhoer, rhoWr, rkr, Nsp, ifail) rhoCpr = avgtabip( Tr, rkr, cpk, Nsp ) gammar = RU / ( rhoCpr/rhoWr - RU ) + 1.d0 pr = rhoWr*RU*Tr ar = dsqrt(gammar*pr/rhor) if (rhor.le.0.d0.or.pr.le.0.d0) hllfix = .true. c if (hllfix) then c if (roe) write (6,*) 'Switching to HLL in',i c rhol = 0.d0 rhoWl = 0.d0 do k = 1, Nsp rkl(k) = qr(i-1,k) rhol = rhol + qr(i-1,k) rhoWl = rhoWl + qr(i-1,k)/Wk(k) enddo ul = qr(i-1,mu)/rhol Tl = qr(i-1,mT) pl = rhoWl*RU*Tl rhoCpl = avgtabip( Tl, rkl, cpk, Nsp ) gammal = RU / ( rhoCpl/rhoWl - RU ) + 1.d0 al = dsqrt(gammal*pl/rhol) c rhor = 0.d0 rhoWr = 0.d0 do k = 1, Nsp rkr(k) = ql(i ,k) rhor = rhor + ql(i ,k) rhoWr = rhoWr + ql(i ,k)/Wk(k) enddo ur = ql(i ,mu)/rhor Tr = ql(i ,mT) pr = rhoWr*RU*Tr rhoCpr = avgtabip( Tr, rkr, cpk, Nsp ) gammar = RU / ( rhoCpr/rhoWr - RU ) + 1.d0 ar = dsqrt(gammar*pr/rhor) c sl = dmin1(ul-al,ur-ar) sr = dmax1(ul+al,ur+ar) c do m=1,meqn if (sl.ge.0.d0) fg = fr(i-1,m) if (sr.le.0.d0) fg = fl(i,m) if (sl.lt.0.d0.and.sr.gt.0.d0) & fg = (sr*fr(i-1,m) - sl*fl(i,m) + & sl*sr*(ql(i,m)-qr(i-1,m)))/ (sr-sl) amdq(i,m) = fg-fr(i-1,m) apdq(i,m) = -(fg-fl(i ,m)) enddo amdq(i,mT) = 0.d0 s(i,1) = sl s(i,2) = 0.d0 s(i,3) = sr endif 55 continue endif c if (pfix) then do 70 i=2-mbc,mx+mbc amdr = 0.d0 apdr = 0.d0 rhol = 0.d0 rhor = 0.d0 do k = 1, Nsp amdr = amdr + amdq(i,k) apdr = apdr + apdq(i,k) rhol = rhol + qr(i-1,k) rhor = rhor + ql(i ,k) enddo do 70 k=1,Nsp if (qr(i-1,mu)+amdr.gt.0.d0) then Z = qr(i-1,k)/rhol else Z = ql(i ,k)/rhor endif amdq(i,k) = Z*amdr + (Z-qr(i-1,k)/rhol)*qr(i-1,mu) apdq(i,k) = Z*apdr - (Z-ql(i ,k)/rhor)*ql(i ,mu) 70 continue endif c return end c c c ========================================================= double precision function Cpmix( Tl, Tr, hl, hr, Y ) c ========================================================= implicit double precision(a-h,o-z) include "ck.i" c dimension Y(*) dimension hl(*), hr(*) data Tol /1.d-6/ c if( dabs(Tr-Tl).gt.Tol ) then Cp = 0.d0 do k = 1, Nsp Cp = Cp + (hr(k)-hl(k)) * Y(k) enddo Cp = Cp / (Tr-Tl) else T = 0.5d0*(Tr+Tl) Cp = avgtabip( T, Y, cpk, Nsp ) endif Cpmix = Cp c return end