| 1 | <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> |
|---|
| 2 | <html><head> |
|---|
| 3 | |
|---|
| 4 | |
|---|
| 5 | |
|---|
| 6 | <meta http-equiv="CONTENT-TYPE" content="text/html; charset=windows-1252"> |
|---|
| 7 | |
|---|
| 8 | |
|---|
| 9 | |
|---|
| 10 | |
|---|
| 11 | |
|---|
| 12 | |
|---|
| 13 | <title>PALM chapter 3.8</title><meta name="GENERATOR" content="StarOffice 7 (Win32)"> |
|---|
| 14 | |
|---|
| 15 | |
|---|
| 16 | |
|---|
| 17 | <meta name="AUTHOR" content="Siegfried Raasch"> |
|---|
| 18 | |
|---|
| 19 | |
|---|
| 20 | |
|---|
| 21 | <meta name="CREATED" content="20040809;13460943"> |
|---|
| 22 | |
|---|
| 23 | |
|---|
| 24 | |
|---|
| 25 | <meta name="CHANGED" content="20041112;15085727"> |
|---|
| 26 | |
|---|
| 27 | |
|---|
| 28 | |
|---|
| 29 | <meta name="KEYWORDS" content="parallel LES model"> |
|---|
| 30 | |
|---|
| 31 | |
|---|
| 32 | |
|---|
| 33 | <style> |
|---|
| 34 | <!-- |
|---|
| 35 | @page { size: 21cm 29.7cm } |
|---|
| 36 | --> |
|---|
| 37 | </style></head> |
|---|
| 38 | <body style="direction: ltr;" lang="en-US"> |
|---|
| 39 | |
|---|
| 40 | |
|---|
| 41 | <h3 style="line-height: 100%;">Topography flags</h3> |
|---|
| 42 | |
|---|
| 43 | |
|---|
| 44 | <p style="line-height: 100%;">Starting from version 3.4a |
|---|
| 45 | a flag array has been introduced in PALM for representing the topography. The new array<span style="font-family: Courier New,Courier,monospace;"> flags </span>is a 3D-<span style="font-family: Courier New,Courier,monospace;">INTEGER</span> array dimensioned (<span style="font-family: Courier New,Courier,monospace;">nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1</span>) where each value describes the features of the respective grid point with respect to the topography. Strictly speaking, <span style="font-family: Courier New,Courier,monospace;">flags</span> is a <span style="font-family: Courier New,Courier,monospace;">POINTER</span> which by default points to the real array<span style="font-family: Courier New,Courier,monospace;"> wall_flags_1</span>. In case of using the multigrid method as pressure solver, additional arrays<span style="font-family: Courier New,Courier,monospace;"> wall_flags_2</span>,<span style="font-family: Courier New,Courier,monospace;"> wall_flags_3</span>, |
|---|
| 46 | etc. are used containing the respective flag informations for the |
|---|
| 47 | coarser grids. (Attention: The finest grid always has the largest |
|---|
| 48 | number, i.e. if four multigrid levels are used, then the flags for the |
|---|
| 49 | finest grid are stored in array<span style="font-family: Courier New,Courier,monospace;"> wall_flags_4</span>.)</p><p style="line-height: 100%;">Each |
|---|
| 50 | INTEGER value has to be interpreted bitwise. As to the FORTRAN model |
|---|
| 51 | bits are counted from right to left starting with bit 0. So far, the |
|---|
| 52 | first seven bits are used for storing the following informations (a bit value of "<span style="font-style: italic;">1</span>" means that the appropriate flag is set):</p><br><table style="text-align: left; height: 220px; width: 496px;" border="1" cellpadding="2" cellspacing="2"><tbody><tr><td style="width: 85px; text-align: center;"><span style="font-weight: bold;">bit position</span></td><td style="width: 393px;"><span style="font-weight: bold;">meaning</span></td></tr><tr><td style="width: 85px; text-align: center;">0</td><td style="width: 393px;">wall to the bottom</td></tr><tr><td style="width: 85px; text-align: center;">1</td><td style="width: 393px;">wall to the top (this cannot be used in the code so far)</td></tr><tr><td style="width: 85px; text-align: center;">2</td><td style="width: 393px;">wall to the south</td></tr><tr><td style="width: 85px; text-align: center;">3</td><td style="width: 393px;">wall to the north</td></tr><tr><td style="width: 85px; text-align: center;">4</td><td style="width: 393px;">wall to the left</td></tr><tr><td style="width: 85px; text-align: center;">5</td><td style="width: 393px;">wall to the rigth</td></tr><tr><td style="width: 85px; text-align: center;">6</td><td style="width: 393px;">inside topography / building</td></tr></tbody></table><p style="line-height: 100%;"><span style="font-weight: bold;">Example:</span><br>If a gridpoint has two neighbouring walls to the bottom and to the left, then the flag bit value is "<span style="font-style: italic;">...010001</span>" which gives the<span style="font-family: Courier New,Courier,monospace;"> INTEGER </span>value "<span style="font-style: italic;">17</span>" = 1*2<sup>4</sup>+1*2<sup>0</sup>.</p><p style="line-height: 100%;">For a gridpoint in the free flow without any neighbouring wall no bits are set, i.e. the<span style="font-family: Courier New,Courier,monospace;"> INTEGER </span>value of the respective array element of<span style="font-family: Courier New,Courier,monospace;"> flags </span>is "<span style="font-style: italic;">0</span>".</p><p style="line-height: 100%;">The intrinsic FORTRAN functions <span style="font-family: Courier New,Courier,monospace;">IBSET</span>, <span style="font-family: Courier New,Courier,monospace;">IBITS</span>, <span style="font-family: Courier New,Courier,monospace;">IAND</span>, <span style="font-family: Courier New,Courier,monospace;">IOR</span>, etc. can be used to manipulate the bits.</p><p style="line-height: 100%; font-weight: bold;"></p><p style="line-height: 100%; font-weight: bold;">Examples:</p><div style="margin-left: 40px;"><span style="font-family: Courier New,Courier,monospace;">flags(k,j,i) = IBSET( flags(k,j,i), 6 )</span></div><p style="line-height: 100%;">sets the sixth bit of array element <span style="font-family: Courier New,Courier,monospace;">(k,j,i)</span> to "<span style="font-style: italic;">1</span>".</p><div style="margin-left: 40px; font-family: Courier New,Courier,monospace;">IBITS( flags(k,j,i), 4, 1 )</div><p style="line-height: 100%;">gives the<span style="font-family: Courier New,Courier,monospace;"> INTEGER </span>value "<span style="font-style: italic;">1</span>" if the fourth bit of array element <span style="font-family: Courier New,Courier,monospace;">(k,j,i)</span> is "<span style="font-style: italic;">1</span>", otherwise "<span style="font-style: italic;">0</span>".</p><p style="line-height: 100%;">Here is an example from routine <span style="font-family: Courier New,Courier,monospace;">redblack</span> (within <span style="font-family: Courier New,Courier,monospace;">poismg.f90</span>) which shows how <span style="font-family: Courier New,Courier,monospace;">IBITS</span> is used for implicit setting of Neumann boundary conditions:</p><p style="line-height: 100%; margin-left: 40px; font-family: Courier New,Courier,monospace;">p_mg(k,j,i+1) + IBITS( flags(k,j,i), 5, 1 ) * &<br> ( p_mg(k,j,i) - p_mg(k,j,i+1) ) </p><p style="line-height: 100%;">If there is no wall to the right of gridpoint <span style="font-family: Courier New,Courier,monospace;">(k,j,i)</span> the term results to <span style="font-family: Courier New,Courier,monospace;">p_mg(k,j,i+1)</span>, otherwise it gives <span style="font-family: Courier New,Courier,monospace;">p_mg(k,j,i)</span>.</p><br>The flag array <span style="font-family: Courier New,Courier,monospace;">flags</span> is calculated in routine <span style="font-family: Courier New,Courier,monospace;">init_grid</span> using the topography information stored in <span style="font-family: Courier New,Courier,monospace;">nzb_local</span>. |
|---|
| 53 | Here also the flag arrays for the coarser grids needed for the |
|---|
| 54 | multigrid scheme are calculated. This requires informations from a |
|---|
| 55 | distance of more than one grid spacing away from the current gridpoint. |
|---|
| 56 | Therfore, array <span style="font-family: Courier New,Courier,monospace;">nzb_local</span> had to be extended for further ghost point levels (see <span style="font-family: Courier New,Courier,monospace;">init_grid</span>).</body></html> |
|---|
