Hello,
I was trying to run my 3D hydrodynamics model for a longer simulation period but it kept crashing after one day. In my steering file, I set
MAXIMUM NUMBER OF ITERATIONS FOR DIFFUSION OF VELOCITIES = 201
MAXIMUM NUMBER OF ITERATIONS FOR PROPAGATION = 202
MAXIMUM NUMBER OF ITERATIONS FOR PPE = 103
MAXIMUM NUMBER OF ITERATIONS FOR PROJECTION = 104
MAXIMUM NUMBER OF ITERATIONS FOR VERTICAL VELOCITY = 105
MAXIMUM NUMBER OF ITERATIONS FOR DIFFUSION OF TRACERS = 206
MAXIMUM NUMBER OF ITERATIONS FOR DIFFUSION OF K-EPSILON = 107
So I can tell which part was going wrong. Then I got the errors from the screen. It said:
PROCESS BOUNDARY DATA AT 55110.0000000000 seconds
PROCESS BOUNDARY DATA AT 55120.0000000000 seconds
PROCESS BOUNDARY DATA AT 55125.0000000000 seconds
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 206 RELATIVE PRECISION: NaN
PROCESS BOUNDARY DATA AT 55130.0000000000 seconds
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 202 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 103 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 206 RELATIVE PRECISION: NaN
PROCESS BOUNDARY DATA AT 55135.0000000000 seconds
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 201 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 202 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 103 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 206 RELATIVE PRECISION: NaN
PROCESS BOUNDARY DATA AT 55140.0000000000 seconds
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 201 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 202 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 103 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 107 RELATIVE PRECISION: NaN
GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 206 RELATIVE PRECISION: NaN
It seems the k-epsilon model crashed first. In order to find where these infinite values occur, I modified the steering file so it can write out results at every time step (every five seconds). But the most strange thing is that all the nodes changed at the same time so I can't tell where the crash happened first. For example,
1. both of k and epsilon crashed at step 2802 and the values of all the nodes suddenly became 1.0E-10.
2. Salinity crashed at step 2802, too. All the values became -1.#IND at the same time.
3. Velocity crashed at step 2803. All the values became -1.#IND at the same time.
In the previous run, it didn't crash even after 2 days. The differences between these two runs are in two subroutines, tfond.f and corstr. In this crashed run, the UETCAR became quite small in the shallow water due to the damping factor.
Besides, I also found large velocities in shallow water. For example, the velocity at node 24190 is 2.34m/s but the water depth is only 0.033m. There are many other nodes like this.
I have no idea what went wrong in my model. Is it caused by the steep slope in the mesh?
Kind regards,
Qilong
