Problem creating input HDF5 file

General Announcements (New Downloads, etc)
samanthansm@gmail.com
MOHID User
Posts: 28
Joined: Thu Jan 01, 1970 00:00
Location: CPTEC-INPE-São José dos Campos -SP - Brasil

Problem creating input HDF5 file

Postby samanthansm@gmail.com » Thu Oct 30, 2008 16:32

Dear users
I created a .hdf5 file to use as an input in InterfaceWaterAir module.
This file has the following structure:
- A group called \"Grid\" where I put Datasets of ConnectionX, ConnectionY, Latitude and Longitude
- A group called \"Results\", with two other groups inside: (Wind_X and Wind_Y). Inside these groups, I put Datasets: Wind_X_000001 to Wind_X_000004 and the same to Wind_Y.
- A last group, Time, with the Datasets of each time (4 were used).

[the contents of this hdf5 file can be seen here: http://docs.google.com/View?docid=dcd7qp2z_0dpdt35gp ]


When I run Mohid, some errors appeared:

    #000: H5D.c line 1169 in H5Dopen(): not found
    major(15): Dataset interface
    minor(48): Object not found
    #001: H5G.c line 1783 in H5G_find(): object not found
    major(10): Symbol table layer
    minor(48): Object not found
    #002: H5Gtraverse.c line 397 in H5G_namei(): component not found
    major(10): Symbol table layer
    minor(48): Object not found
    #003: H5Gstab.c line 161 in H5G_stab_find(): not found
    major(10): Symbol table layer
    minor(48): Object not found
    HDF5ReadDataR8_1D - ModuleHDF5 - ERR20


What's wrong with my input hdf5 file? Could you help me?

Thank you.

Samantha Martins

samanthansm@gmail.com
MOHID User
Posts: 28
Joined: Thu Jan 01, 1970 00:00
Location: CPTEC-INPE-São José dos Campos -SP - Brasil

I've done some adjusts

Postby samanthansm@gmail.com » Fri Oct 31, 2008 12:20

Dear users

I am here again because I've worked on my problem since yesterday. I've changed the properties descriptions in InterfaceWaterAir.dat and in Atmosphere.dat.
The contents of these files can be seen in the following links:

InterfaceWaterAir.dat > http://docs.google.com/Doc?id=dcd7qp2z_1hjgs48gg
Atmosphere.dat > http://docs.google.com/Doc?id=dcd7qp2z_3gx8fx7dq

[in my original docs, the properties are aligned]

Now I got another error, a different one:

Code: Select all

forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image              PC        Routine            Line        Source
MohidWater         08238FE4  Unknown               Unknown  Unknown
MohidWater         08238995  Unknown               Unknown  Unknown

Stack trace terminated abnormally.


Do you have any idea what could it mean? :roll:

Thank you

Samantha Martins

samanthansm@gmail.com
MOHID User
Posts: 28
Joined: Thu Jan 01, 1970 00:00
Location: CPTEC-INPE-São José dos Campos -SP - Brasil

Postby samanthansm@gmail.com » Fri Oct 31, 2008 15:16

I am still working on my problem. I decided to take all 'wind stress' away.
Now my InterfaceWaterAir.dat file is empty and the option WIND in hydrodynamic is =0.
I have my Atmosphere.dat file exactly like before:
Atmosphere.dat > http://docs.google.com/Doc?id=dcd7qp2z_3gx8fx7dq

And when I run Mohid, I got the same mistake that I got in my first message:

[code]


HDF5-DIAG: Error detected in HDF5 library version: 1.6.7 thread 0. Back trace follows.
#000: H5D.c line 1169 in H5Dopen(): not found
major(15): Dataset interface
minor(48): Object not found
#001: H5G.c line 1783 in H5G_find(): object not found
major(10): Symbol table layer
minor(48): Object not found
#002: H5Gtraverse.c line 397 in H5G_namei(): component not found
major(10): Symbol table layer
minor(48): Object not found
#003: H5Gstab.c line 161 in H5G_stab_find(): not found
major(10): Symbol table layer
minor(48): Object not found
HDF5ReadDataR8_1D - ModuleHDF5 - ERR20




What is wrong with my file? Could you help me, please?

Thank you!

acanas.maretec@taguspark.pt
MOHID Advanced user
Posts: 103
Joined: Thu Jan 01, 1970 00:00
Location: Instituto Superior Técnico - Technical University of Lisbon

Re: Problem creating input HDF5 file

Postby acanas.maretec@taguspark.pt » Wed Nov 05, 2008 18:16

Hi Samantha,

Sorry for the late answer.
Apparently you lack a variable in the HDF5 file which the program is expecting to find there. The error you get originates because the program cannot find a data set in the file.

I noticed you lack the WaterPoints variable, which defines the land and water cells in your domain. It also appears that connection x/y and latitude/longitude are defined for cells center instead of cells limits, is that so? The program is expecting values for cells limits.

How did you created the HDF5 file?

Regards,

Angela

samanthansm@gmail.com
MOHID User
Posts: 28
Joined: Thu Jan 01, 1970 00:00
Location: CPTEC-INPE-São José dos Campos -SP - Brasil

Postby samanthansm@gmail.com » Thu Nov 06, 2008 11:54

Thanks for helping, Angela.
I finally could make Mohid read my HDF5 file.
The problem was the Atmosphere_1.dat. It must be written like that:

Code: Select all

<beginproperty>
NAME         : wind velocity X
UNITS         : m/s
DESCRIPTION      : wind velocity X
DEFAULTVALUE      : 0.0
FILE_IN_TIME      : HDF
FILENAME      : /home/mag/d-estudo_hdf5/cond_in_golfo.h5
REMAIN_CONSTANT      : 0
OUTPUT_HDF      : 0
I<endproperty>


For each property.

So when I compared the results in 2 experiments (one with and another without Atmospheric forcing) I noticed differences in the velocities.

I also concluded that the option WIND must be = 1 in Hydrodinamic_1.dat and I must have the file InterfaceWaterAir with wind stress (calculated).

I created my Hdf5 file with hdf functions, using a program.

I am thinking about what you wrote:

I noticed you lack the WaterPoints variable, which defines the land and water cells in your domain. It also appears that connection x/y and latitude/longitude are defined for cells center instead of cells limits, is that so? The program is expecting values for cells limits.


And I will do it and compare the results with the others.

I have another question: do you know if I have to have an HDF5 file with the same grid of my simulation?

I mean, lets suppose I had some results of a atmospheric model (WRF, MM5, for example). And its grid differ from the grid of Mohid. Can I create a HDF5 file anyway?

Thanks a lot.

acanas.maretec@taguspark.pt
MOHID Advanced user
Posts: 103
Joined: Thu Jan 01, 1970 00:00
Location: Instituto Superior Técnico - Technical University of Lisbon

Postby acanas.maretec@taguspark.pt » Fri Nov 07, 2008 11:51

Hi,

You must have the same grid in your model and in atmospheric forcing file, because the MOHID Water uses the grid indexes to get the atmospheric forcing for each model cell.

If your atmospheric forcing data has a different grid from your MOHID Water model grid you can interpolate the data to your model grid using the MOHID Support Tool ConvertToHDF5 and action Interpolate Grids. Check the wiki page (http://www.mohid.com/wiki/index.php?tit ... erpolation) for instructions of how to use.

Hope this helps.

Best regards,

Angela


Return to “MOHID General”

Who is online

Users browsing this forum: No registered users and 2 guests