Learning Forum

[Constantin Bach]

I am doing a frequency statistical analysis using the Hydrodynamics Response module from the workbench. I am doing it for zero speed and with forward speed. I am using irregular wave and have set the direction, Hs, Tp and Gamma as design parameters (design points). I have these combinations of sea spectra from 0-360 deg with an increment on 15deg in direction. 

I get the results for all the design points for zero speed. But for forward speed case, I get the below error at some of the design points.

I am using the non-linear roll damping using bilge properties (taken from actual vessel). I notice that the solver stops after 30 iterations for the design points which gives this error. How to resolve this issue? 

I also want to know if there is any way to increase the number of iterations in the frequency statistics analysis from 30 to some higher number. 

P.S: I know there is some command CALFER 0 0 1000 to increase the number of iterations, but not know how to use it or is this the right command. I treid using the CALFER in ANALYSIS.dat file of the design point with the error, but after saving it when I run for same design point in the frequency statistical analysis, the analysis.dat file gets overwritten. 

[Shuangxing Du]

To add in the Aqwa debug option, you could use External Operation before Solving (See Aqwa User's Manual, Section 5.9.1. Specifying External Pre- or Post-Solve Operations). By setting this function as Run a Python Script, and in your Python script, you could add in the following lighlighted items into the input data file in (your project)\dp0\AQW-1\AQW\AQ\Analysis\Frequency.dat

 

[Constantin Bach]

Thank you Sxdu for your response.

I tried editing the Frequency.dat file as you suggested but for me the location was "dp0\AQW-4\AQW\AQ\Analysis\Frequency.dat". After editing the .dat file, I set the external operation before solving as "run a python script" and don't provide path to the .dat file as it can be seen in the image below.


Here is how my Frequency.dat file looks like for dp0.

Unfortunately, after performing the recommended steps, I see the solver still closes after 30th iteration and the number of maximum iteration is not increased. 

[Mike Pettit]

Hello Constantin,

You need to use the pre-solve Python script to modify the Aqwa input file, rather than modifying it manually. Please copy and paste the following code into the pre-solve Script Editor in Aqwa Workbench:

import os

AqwaInputFile = os.environ["AQWA_INPUT_FILE"]
lines = []

with open(AqwaInputFile, "r") as fin:
    lines = fin.readlines()

optsWritten = False
dbugWritten = False

with open(AqwaInputFile, "w") as fout:
    for line in lines:
        if "OPTIONS" in line and not optsWritten:
            fout.write(line)
            fout.write("OPTIONS DBUG\n")
            optsWritten = True

        elif "RESTART" in line and not dbugWritten:
            fout.write(line)
            fout.write("CALFER        0    0 1000\n")
            fout.write("END\n")
            dbugWritten = True

        else:
            fout.write(line)

The script reads the input file generated by Aqwa Workbench, then re-writes it with the additional 'DBUG' and 'CALFER' entries before the solver is launched.

I hope this helps; please let me know if you have any problems or questions.

Mike

[Constantin Bach]

Hi Mike, 

Thank you for your very helpful response. I am able to increase the number of iterations using the python script and it seems to have solved my issue. However, I am facing another issue now and that is shown in the figure below.

I see this error for some design points and when I rerun the specific design points, sometimes it solves and sometimes it doesn't. I have contact my IT team and they are of the opinion that it is not a license issue but some file system is not working properly. Is there any way to fix this issue? Sometimes when I see the above error, I see a message before, given below for your reference. 



But this message is not shown for all the dps that give the "workbench communication error". Awaiting your helpful response. 

[Constantin Bach]

Hi Mike, 
An update to the AQWA-FER failed to converge error. After increasing the number of iterations, some of the design points converge within 100 iterations but there are still some points which don't converge even at 2000 iterations. 

[Constantin Bach]

Hi Mike, Some updates for your reference. 

Some Design Points Not Converged: I am using Nonlinear Roll Damping in my analysis. I am using the Bilge type as “Bilge Vortex Shedding and Bilge Keel Roll Damping”. I have noticed that when I use the actual bilge parameters, I get many design points that don’t converge. But as soon as I move bilge keel further away from the centerline, some design points gets converged. I see the warning as well that damping is less than 0.5% critical damping and adjusted automatically. However, I have a few questions as well regarding defining the bilge keel. 

 

1. What is the difference when we select bilge type as “Bilge Vortex Shedding Only” and “Bilge Vortex Shedding and Bilge Keel Roll Damping”? Does it mean for vortex shedding only, it only accounts for the eddy part of the damping and doesn’t consider the damping part due to pressure difference on the sides of the bilge keel?

 

2. Bilge Radius: I am using the one at the midship section, is this the right approach?

 

3. Depth To Bilge: When I use the depth to bilge till the point where the BK is attached to the hull i.e 5.85 m, I have more points that don’t converge, but when I use the “depth to the bilge” to the bottom corner of the bilge i.e 6.142 m, then the number of unconverged points is less. I assume as the damping increase, it is easily linearized by the solver. What is the right approach, which depth should be used bottom corner of the Bilge keel or the point where the BK is attached to the hull.

 

4. Offset of Bilge from Central Line: What exactly is this distance from the centerline? To the center of the red cylinder that is formed and shown in the image or is this the distance to the point where the BK is attached to the hull?  I notice that this distance play an important role in linearization, the higher the distance, the more points converged.

 

5. Bilge Keel Breadth: It is the bilge keel height that I am using.

 

Below are the actual values from the vessel for the BK which gives less converged points with depth to bilge until the point attached to the hull and offset of Bilge from center line till the point bilge keel attached to the hull (12.201m). Looking at the images, this configuration seems to be realistic as we have on our vessels.

Below are the modified BK parameters which gave more converged points. Depth to bilge set to the bottom point of BK. Offset from center taken as 13.99 m (12.201 m + 1.798 m). Not a realistic scenario but gives only 2 unconverged points rather than 30 points. 

Here is my outline tree, I don’t have any additional loads or non-linearities. Just the irregular wave spectrum is included. 

Looking forward to your helpful response regarding the convergence issue.