[mesa-users] [MESA]: Question Astrophysical Factor
Gonçalo Andrade
gnmandrade at hotmail.com
Fri Dec 9 06:42:57 EST 2016
Hi
Thank you Rob, I think it's finally working.
I've run some tests and it's running well with fair results.
Thank you for your help.
Gonçalo
________________________________
De: Robert Farmer <rjfarmer at asu.edu>
Enviado: 29 de novembro de 2016 22:34
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
So whats happening is that when mesa sees a rate for the first time it creates a cache file in $MESA_DIR/data/rate_tables/cache then the next time it loads the rate with the same name, it uses the cache file not the original. So lets say you run your test4 then test5 runs, test4 creates a cached rate based on its rate for the pp reaction, then when you run test5 it sees a rate with the same name so it loads the cached version not the altered one in your folder.
The solution to this is, is to add:
rates_cache_dir = './rate_cache'
to your star_job inlist. this creates the cached files locally so different runs with different rates wont interfere with each other. Then you only need to delete that folder when you change the rate in that folder. ie running test4 then test5 you wouldn't need to delete the folder (in either folder), but if you run test4, then change the rate in test4 you need to delete the rate_cache folder.
Rob
On Mon, Nov 28, 2016 at 1:32 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
I forgot the attachment , I'm sorry.
Gonçalo
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 28 de novembro de 2016 17:39
Para: Gonçalo Andrade
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
Please send requests for help via the mesa-users mailing list.
>I send you two of my test projects,
There where no files attached to this email
Rob
On Sun, Nov 27, 2016 at 4:38 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi Robert
I thought the code was doing what I wanted it to do, but I still can't test different nuclear rates and I can't understand why.
I already tried a lot of things, I used a rate_tables directory in the work folder, but also tried changing the files in
$mesa_dir/data/rates_data/rate_tables/
but the output files always came exactly the same. I've even checked the test rates in
$mesa_dir/rates/test
and it was working, but there's no difference in the output. I think this makes no sense since I'm changing (for test purposes) the pp reaction by a lot, it's supposed to have a major impact in the star...
I've already downloaded a new version of the code, re-installed it some times and it still doesn't work.
I send you two of my test projects, I simply picked the 1M_pre_ms_to_wd from the test_suite and put in new rate tables and asked it to stop at 1E10 yrs.
The ziped folder has 2 directories, they are equal except in the rate file.
What I ask you, if possible, is that you run both of them in your computer and then tell me if there is any difference between their history.data output files.
I'm sorry for bothering you again with this problem...
Thanks for your help.
Gonçalo
________________________________
De: Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>>
Enviado: 9 de novembro de 2016 18:02
Para: Robert Farmer
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
I think it is working. When you said, in the last email, that it worked on your computer I thought it could be a problem with the version of MESA I had. I downloaded the most recent one and installed it. It all worked well, I can even put the rates folder in the working directory, it works just as you said it should.
Thanks a lot for all your help, I hope I can continue my work without these major problems from now on.
Thank you again and good luck with your research!
Gonçalo Andrade
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 8 de novembro de 2016 17:40
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net>
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
hi
So things work fine for me (once i put the test_nuclear_rates_tables into the test_nuclear folder as its missing from there)
quick test, do:
rm $MESA_DIR/data/rates_data/cache/*
to remove MESA's cached version of the rate tables, then try your code again, otherwise try re-downloading mesa and installing it again to clean up any changes you might of made.
Rob
On Mon, Nov 7, 2016 at 5:20 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
That's really weird. I've been trying those 3 nets that appear in the inlist. They all work if I don't include my rates' tables and the last one I used was my modified net.
I attached my working folder and my rates_tables folder (test_nuclear_rates_tables) that I had to put in
$MESA_DIR/data/rates_data/
Also, referring to the last email, I can generate the rates_tables based on modified S-factors. My objective is to use them but right I'm just trying to run the code and making sure he uses my tables instead of the pre-defined ones, so I hadn't that much care choosing the values. I'll try to follow your advice and add some points with lower rates at smaller temperatures, but it shouldn't matter that much since it works on your system.
Gonçalo Andrade
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 7 de novembro de 2016 23:42
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net>
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
So i just ran your inlists as is and everything works fine (did you change the nets since you last sent them?), mesa picked up the new rate in the local folder and there where no convergence issues getting out of the pre-MS.
can you do a:
rm -rf LOGS photos
to get rid of any data files (or move them to another if you want to keep them) and then send a zipped copy of your work folder with your inlists and everything else in, so i can see exactly what you have.
Rob
On Mon, Nov 7, 2016 at 4:21 PM, Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>> wrote:
Hi
A convergence problem is not a bug. What its telling you is that MESA has struggled and failed to find a way to satisfy all the stellar structure equations given the constraints that your inlist sets. Updating to a new version may help, and is worth a shot, as changes are made to the code but i don't know of anything particular that will has been fixed that would help
>This might work, but then I cannot specify the mass and the Z right?
You can specify mass but not Z, that is pre-set to solar for the pre created pre-MS models
The next thing to think about is, are your rates sensible? If we look at MESA's rate for rpp_he3, to do this we need to look at the cached data mesa written out:
cd $MEA_DIR/rates/test
./show_rates $MESA_DIR/data/rates_data/cache/rpp_to_he3_1.bin > rpp_to_he3_1.txt
Then inspecting rpp_to_he3_1.txt shows that for T<10**7K the rates are ~<10^-20. So maybe a trial for you would be to add some more low temperature data points, with a very small rate factors such that the rpp_he3 rate is almost 0 for the temperatures on the pre-MS and early evolution until the star gets hotter. This is the E (experiment) part of MESA.
Why the rates folder isn't working properly i'll have a look into.
Rob
On Mon, Nov 7, 2016 at 3:52 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
------
Does having the rates_table option as:
rate_tables_dir = './rates_tables' (put a ./ in front of the folder name)
Work for you? for loading the rates from a local folder called rates_table
------
I've tried that, the code looks for the tables in the non existing directory
$MESA_DIR/data/rates_data/./rates_tables
So I gave up on that and added my tables on a "test_nuclear_rates_tables" folder inside rates_data, using then
rate_tables_dir = 'test_nuclear_rates_tables'
----------
First test, does the model work okay without the new rates, given all the other options you have set?
----------
If I use the same inlist but comment just the line
rate_tables_dir = 'test_nuclear_rates_tables'
It runs without any problem.
-----------
Some options that may or may not work for you:
Don't make your own pre-MS star, (create_pre_main_sequence_model = .false.) MESA will load a pre-built model and then continue the evolution with your inlist options
-----------
This might work, but then I cannot specify the mass and the Z right? I needed to model a real observable star, so it would be useful to have freedom in these parameters.
-----------
set pre_ms_relax_num_steps = 100 this makes MESA take more steps during the pre-MS so the change per-step is smaller and it might be easier to converge the results
-----------
I've tried it at 100, and in several values as high as 25000, still doesn't converge.
-----------
Make a pre-MS model with the rate changes off, save a photo, stop the run then restart from the photo with the rates on (a more complicated and manual way of running the first option)
-----------
I've tried this. I used the same inlist commenting the line that calls my tables and using the comands
Lnuc_div_L_zams_limit = 0.99d0
stop_near_zams = .true.
to stop and save the ZAMS model.
When I restart with the new reactions on the last photo it has the same convergence problem, but with an even smaller dt.
-----------
Change mixing_length_alpha (again you can always change it after the pre-MS has finished by stopping the run, changing the file and restarting from a photo)
-----------
I've tried a wide variety of values, between 0.005 and 500, they all had the same convergence problem.
Having tried all the suggestions, is it possible that this is a bug that has been fixed in more recent versions?
My MESA version is the 8118. Current version is the 8845. Is it worth to try installing it all again with the new version?
Gonçalo Andrade
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 7 de novembro de 2016 18:35
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net>
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
Does having the rates_table option as:
rate_tables_dir = './rates_tables' (put a ./ in front of the folder name)
Work for you? for loading the rates from a local folder called rates_table
Convergence issues are hard to fix normally and even harder to do during the pre-MS.
First test, does the model work okay without the new rates, given all the other options you have set?
Some options that may or may not work for you:
Don't make your own pre-MS star, (create_pre_main_sequence_model = .false.) MESA will load a pre-built model and then continue the evolution with your inlist options
set pre_ms_relax_num_steps = 100 this makes MESA take more steps during the pre-MS so the change per-step is smaller and it might be easier to converge the results
Make a pre-MS model with the rate changes off, save a photo, stop the run then restart from the photo with the rates on (a more complicated and manual way of running the first option)
Change mixing_length_alpha (again you can always change it after the pre-MS has finished by stopping the run, changing the file and restarting from a photo)
Rob
On Sun, Nov 6, 2016 at 4:45 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi Robert
Since our last email I've been trying to solve this problem.
I've fixed some errors but others have appeared.
1) I was misunderstanding how MESA uses the nuclear tables directory. By looking at the error messages I saw that MESA instead of using the folder inside the working directory (where I have my inlists and stuff like that), it ads whatever I put as argument in the
rate_tables_dir = 'rates_tables'
to the folder
$MESA_DIR/data/rates_data/
So I need to place my tables in this directory.
2) After "solving" this problem, the code stopped ignoring my tables and started to output error messages saying "segmentation fault". This was easy to solve, it was only due to errors in the files' names.
3) I now think I've managed to get the code reading the tables, but each time I change a reaction the code doesn't run because it has convergence problems. I've already tried a wide range of values in the rate files.
I've changed some things, so I send you as an attachment my current inlist files, my rate files and a print of the error message.
Thank you,
Gonçalo
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 1 de novembro de 2016 17:35
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net> users
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
So looking over your files
When the run starts you see MESA prints out a list of isotopes and abundances, there is no h2 printed at all, this means there is no h2 in your nuclear networks. You need to remove the comment in you add_test_nuclear file on the add_isos(h2) file.
Also maybe just a copy error, but in the rate file the temperature should be in units of 10^8 not 10^9 as your comment in the file suggests
Rob
On Tue, Nov 1, 2016 at 3:59 AM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
I've fixed that detail on the rates_list.txt, but the result was the same. It doesn't change a bit and I've put some really big numbers on the rates of the pp reaction so I think it should have a major impact on the model.
I now tried changing the first line of rpp_to_he3.txt from low rate to high rate, but I'm still waiting for the results.
Thank you,
Gonçalo
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 31 de outubro de 2016 20:06
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net> users
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
your rates_list.txt has the line
r_pp_to_he3 'rpp_tp_he3.txt'
guessing it should be:
r_pp_to_he3 'rpp_to_he3.txt'
Rob
On Mon, Oct 31, 2016 at 1:00 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
>So you have a modified network, does this network include the reaction you want? you have the list_net_reactions = .true. so check that list to see if your desired reaction is there.
I've tried with the other networks (commented) and It seemed not to work, so I tried modifying it but the result was practically the same.
> What does rate_tables/rates_list.txt look like? What does the rates file you are using look like? What does your network look like?
I think I attached all the relevant files. I created a folder called "nets" for the networks.
Gonçalo
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 31 de outubro de 2016 19:50
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net> users
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
>new_net_name = 'mod_cno_extras_o18_and_ne22.net<http://mod_cno_extras_o18_and_ne22.net>'
So you have a modified network, does this network include the reaction you want? you have the list_net_reactions = .true. so check that list to see if your desired reaction is there.
What does rate_tables/rates_list.txt look like? What does the rates file you are using look like? What does your network look like?
Rob
On Mon, Oct 31, 2016 at 12:28 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi
> Please keep replies on the mesa users list, others may be able to help and may be helped by any response.
Thanks for you advice!
> Did you alter your star_job inlist to point to the rates folder? I would need to see your inlists, folder structure and rates files to see whats gone wrong.
I think I did. I added the line:
rate_tables_dir = './rate_tables'
on my inlist and put the files on that folder.
I send my inlist as an attachment, maybe that helps.
It also has a commented section using the special rate factor command, I tried it because of another member's advice, but I think it doesn't allow me to test the astrophysical as I need.
Thanks for the rest of the information.
Gonçalo Andrade
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 31 de outubro de 2016 18:18
Para: Gonçalo Andrade; mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net> users
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
Please keep replies on the mesa users list, others may be able to help and may be helped by any response.
>I've created the rate_tables directory and put the files there as you said but the code seems to ignore my tables.
Did you alter your star_job inlist to point to the rates folder? I would need to see your inlists, folder structure and rates files to see whats gone wrong.
>1) If I do what you said for an existing reaction, like rpp_to_he3, the code is supposed to ignore the default tables and use mine right? Or it adds a new reaction with my tables?
It should use yours, if its been set up correctly
>2) On the table files (txt), what are the units of the "Rates" column?
Reactions per second
Rob
On Sun, Oct 30, 2016 at 10:59 AM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Hi Rob
Thank you for your help. It took me a long time to produce properly the rate tables from the s-factor so I could only try your method a few weeks ago.
I've created the rate_tables directory and put the files there as you said but the code seems to ignore my tables. I even tried to put really big numbers just to check if it was using the tables and the result was no different from what I had before.
Is there any step missing? I always do "./clean && ./mk && ./rn" when running the code.
Two more questions:
1) If I do what you said for an existing reaction, like rpp_to_he3, the code is supposed to ignore the default tables and use mine right? Or it adds a new reaction with my tables?
2) On the table files (txt), what are the units of the "Rates" column?
Thanks a lot for your help.
Gonçalo Andrade
________________________________
De: Robert Farmer <rjfarmer at asu.edu<mailto:rjfarmer at asu.edu>>
Enviado: 16 de setembro de 2016 00:59
Para: Gonçalo Andrade
Cc: mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net>
Assunto: Re: [mesa-users] [MESA]: Question Astrophysical Factor
Hi
You cant change the s-factor directly, however you can give mesa a file with a customized set of nuclear reaction rates where you have changed the s-factor yourself.
What you need to do (and example inlists and rate files can be found here http://mesastar.org/results/properties-of-carbon-oxygen-white-dwarfs-from-monte-carlo-stellar-models) follows:
Add this to the star_job section of your inlists
rate_tables_dir = './rate_tables'
Then make a folder rate_tables in your mesa work directory. Inside that folder make a file rates_list.txt file with the following line:
r_h1_pg_h2 'r_h1_pg_h2.txt'
This says we want to alter the reaction h1+h1 -> h2 by looking for a file called r_h1_pg_h2.txt where we will specify the rate as a function of temperature
Then make a file called r_h1_pg_h2.txt in (rate_tables), where the first two lines are:
# r_h1_pg_h2 low rate:
# T8 RATE
Then you need to add a line with the number of rate points you have. Then give the temperature in units of 10**8 K and a rate with one rate per line. An example file might look like this though you will want many more rate points than this:
# r_h1_pg_h2 low rate:
# T8 RATE
5
0.01 0.0
0.5 1.0
1.0 2.0
10.5 3.0
100.0 10.0
Then run mesa and it should be using your new nuclear rate.
Rob
On Thu, Sep 15, 2016 at 4:25 PM, Gonçalo Andrade <gnmandrade at hotmail.com<mailto:gnmandrade at hotmail.com>> wrote:
Dear Sir or Madam,
I'm a MSc student and I'm doing my master thesis on asteroseismology of red giant stars. I'm using MESA to do models of these stars.
I found this email on the MESA community website and I decided to try it as a last resource.
The thing I need to do now in my project is to change the Astrophysical Factor (S-Factor) from the pp nuclear reaction to try to find out what changes does it have in the star and if these can be measured using seismology. The problem is I've been been searching online for some weeks and I can't find out how to change this factor. I've tried several things, the closest to what I wanted, I think, was to change the value on the file $MESA_DIR/rates/private/ratelib.f90 , but after that change the code no longer compiles...
I a few words, I just need to change the S-Factor of the pp reaction.
I thank you in advance and hope I'm not being inconvenient.
Yours faithfully,
Gonçalo Andrade
------------------------------------------------------------------------------
_______________________________________________
mesa-users mailing list
mesa-users at lists.sourceforge.net<mailto:mesa-users at lists.sourceforge.net>
https://lists.sourceforge.net/lists/listinfo/mesa-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20161209/52d03239/attachment.html>
More information about the Mesa-users
mailing list