Study on the numerical simulation of flow pattern in freshening reservoir using satellite image data
Grid generation
The grid generation used the transfinite interpolation for the different times in Yongam lake and applied to boundary adaptive curvilinear coordinate grid. The outline of the grid generation is shown in Fig. 2 and these steps are follows.
First, the representative sides are generated by linking the representative points decided to locations which have topographical features and the changes of water flows. Second, The grid points are generated onto the representative sides of equal and unequal intervals.. If it have an equal interval, the grid number is divided in proportion to distance between the representation point of both sides. However, the division of unequal intervals is used the linear interpolation using Robert's equation. Third, smoothing is executed using Lagrangian interpolation on the whole area. The smoothing has two kinds of method whether smoothing will perform or not for grid point onto the boundary. Fouth, the Jacobian evaluation of the grid generation put into practice to examine overlapping representative sides on the same axis. Finally, the masking is described as a black mark these not included computational area.
The grids generated by these steps are 194 points in vertical plane and 55 points in horizontal plane for the direction of the mouth of a river in the area before enclosure, 210 points and 56 points separately in the area after enclosure, as shown in Fig. 3.
Fig3. Gird generated in the area of A) before and B) after enclosure
Fig4. Contour line of eater depth in the area
Numerical simulation of inflow flows
The long wave equation utilized as equation for the numerical analysis of flow fields is defined as follows
Where
h : displacement of water surface
q: liner discharge vector,
t : time ,
h: water depth,
g: acceleration of gravity,
n: kinetic viscosity
f
e : coefficient of friction,
Ñh
a,
Ñh,
Ñh2. divergence of horizontal level and gradient, Laplace operator respectively. The numerical analysis ois applied to the method associated time integral method with four-step Runge-Kutta scheme and full implicit method. The representative values are shown in Table 1, where L
0 means the average of water rapidly by using the value of the water depth in Yongam lake. The viscosity 1.0 X 10
3[m
2/s] so that grid number is increased rapidly by value of the water temperature of 20°C, 1.0 X 10
3[m
2/s].
| Length [L0] | 10.0[m] |
|
Velocity [U0] | 9.899[m/s] (=ÖgL0) |
| Time [T0] | 1.0102[s] (=L0/U0) |
Table 1 representative values
