-- Using accurate, user-performed calculation of friction lengths.

WFRICTN
--wname diam roughness flowscalefact
  WPH2   0.31  0.01  /
-- ic jc kc  Ddist1 Ddist2 penetrDir ijkEnd Diam
   13 17  3  0       51.7  /
   13 16  3  1*     155.2  /
   13 15  3  1*     258.7  /
   13 14  3  1*     360.0  /
   14 14  3  1*     362.2  /
   14 13  3  1*     465.7  /
   14 12  3  1*     569.2  /
   14 11  3  1*     672.7  /
   14 10  3  1*     775.0  /
   15 10  3  1*     776.2  /
   15  9  3  1*     979.7  /
   15  8  3  1*    1180.0  /
   16  8  3  1*    1183.2  /
   16  7  3  1*    1386.7  /
   16  6  3  1*    1585.0  /
   17  6  3  1*    1590.2  /
   17  5  3  1*    1793.7  /
   17  4  3  1*    1995.0  /
   18  4  3  1*    1997.2  /
   18  3  3  1*    2200.7  /
   18  2  3  1*    2404.2  /
/

-- Geometry of well (in layer 3)
--
-- ix 13 14 15 16 17 18 
-- jy                  
--  2                 C
--  3                 x
--  4              x  x
--  5              x
--  6           x  x
--  7           x
--  8        x  x
--  9        x
-- 10     x  B
-- 11     x
-- 12     x
-- 13     x
-- 14  x  x
-- 15  x
-- 16  x
-- 17  A

-- Cell sizes (get from FloViz), DX = 100m in relevant area
-- DY = 200 m for J = 1 - 10, DY = 100 m for J = 11 - 20.

-- Distance beween points A and B is 776.2 m (Dist. in X-dir 200m,
-- in Y-dir 750 m. Using Pythagoras, AB = SQRT(200**2 + 750**2)
-- Similarly, BC = 1627.9 m.

-- Then the segment length for each cell is computed as the cell's
-- contribution to the segment. E.g., cell (13, 17) has a perforation
-- length of 50 m in the Y-direction. As the distance AB along the Y-axis
-- is 750 m, the friction length in the cell is 50 * 776.2 / 750 = 51.7 m,
-- the first entry in the table.

-- Same procedure for rest of well.

-- We have chosen to define the friction lengths for each cell, as this is
-- the safest. We could have used ranges, but may then either "lose" some
-- perforations, or get inconsistent lengths, both of which Eclipse complains.
-- So this method is a little more laborious, but usually ends up correct.

-- Note that for this particular well, the approximate eclipse method
-- would give a total well length of 2350 m, so the difference isn't that
-- great.