Table 1 The dimension size in Fig.2 (m)

Inlet Stage 1 Stage 6 Stage 8 Stage 10 Outlet

S0 L1, D

L1, R S1

L6, D

L6, R

S6 L8, D

L8, R

S8 L10, D

L10, R

S10

Lout

0.0060 0.0040 0.0070 0.0069 0.0040 0.0050 0.0060 0.0029 0.0035 0.0075 0.0042 0.0038 0.0115 0.0055

In  the dimensions in  Table 1,  we  define   S0    to

the channel is relatively small (only 0.0040 m in the prototype) as compared with the length of the flow

L6, R

as  the  “series  passage”,  where  there  is only a

passage, so the 2-D and single-precision solver is cho-

single way in and out, while    after

L6, R ,  the  flow is

sen. The implicit scheme of the segregated   algorithm

pided into two parallel parts, which is defined as the

is adopted. Then the standard

k - turbulence model,

“parallel passage”, and it is shown that the  narrowest in the “series passage” is 0.0040 m, while it is only 0.0025 m in the “parallel passage”. So it is very diffi- cult to mount the sensors to test the pressure and the velocity in the passage. Therefore, two alternative so- lutions might be adopted, one is the model experiment, which will scale and enlarge the prototype passage, and the other way is to use the CFD method for a nu- merical simulation. In the following sections, the model experiment results will be used to validate the CFD results and provide a series of simulation results to analyze the flow characteristics in the prototype labyrinth passage.

2. Validation of the numerical method

In order to validate the current numerical method, a model experiment is conducted by selecting a part of labyrinth passage just as shown in Fig.3. For the con- venience of mounting the pressure sensors, the passage is scaled to 4 times of the prototype, the inlet and the outlet are also shown in Fig.3.

Fig.3 A model passage used in experiment (scaled 4 times for the mounting pressure sensor conveniently)

There  are  total   7   YOKOGAWA   EJA110A (0 kPa-500 kPa) pressure differential transmitters used in the experiment as shown in Fig.4(a). To avoid the severe vortex area and also to ensure the locations not far away from the pressure wave section, the  pressure

the second-order upwind scheme, and the SIMPLE al-

gorithm are used for the solutions. Figure 4(b) shows the pressure contours of the selected model passage.

Fig.4(a) The pressure sampling points in the model passage

Fig.4(b) Pressure contours of model passage

Fig.4(c) The simulation vs. the experiment results at different points

Figure 4(c) depicts the simulation and experime- nt results for the static pressure at each sampling point, where   p   is the pressure of monitoring stations  sele-