[mesa-users] convergence problems

[Francis Timmes]

> integer, parameter:: Y=0.4 

0.4 is a real number, not an integer.
you want to declare Y to be a double precision real number.


> Any help would be greatly appreciated

taking josiah's advice to write out values of interest would be prudent.

fxt






> On Jul 5, 2016, at 9:51 AM, zoe ansari <zoemc.ansari at gmail.com> wrote:
> 
> 
> ---------- Forwarded message ----------
> From: Josiah Schwab <jwschwab at berkeley.edu>
> Date: Tue, Jul 5, 2016 at 9:15 PM
> Subject: Re: [mesa-users] convergence problems
> To: zoe ansari <zoemc.ansari at gmail.com>
> 
> 
> Please resend this message to mesa-users and not just me personally.
> 
> Josiah
> 
> 
> On Jul 5, 2016 9:08 AM, "zoe ansari" <zoemc.ansari at gmail.com> wrote:
> Hi Josiah
> 
> What I was trying to do is modifying Lane-Emden equation with adding this extra term i.e. (-Y/4)*(rho(r))*(G(r))*(M"(r))*(1/M'(r)) to dp/dm.
> 
> I changed integer Y=6d-6 to integer, parameter:: Y=0.4 ( I wanted Y=0.4 at first place I just changed it to Y=6d-6 to see if it works with a lower Y) any way, I get all values of dp/dm which contains Y, as zero. so I just set Y in the expression( (Y/4) = (0.1)) and  it does kind of solve the problem by now.
> 
> Then I changed the "rr" and determined dp/dm in 3 do-loop as you can see in run_star_extras.f.
> 
> When I run the program without "other_momentum" subroutine it gives me different values for dp/dm, which contains an extra term of dp/dm of newtonian detemination.
> 
> But when I add "other_momentum" subroutine , with new defenition for extra_dpdm, it does not run and gives convergence problem as you can see by my inlistis and run_star_extras.f in attached file.
> 
> 
> 
> Any help would be greatly appreciated
> 
> Zoe
> 
> 
> On Sun, Jul 3, 2016 at 10:57 PM, Josiah Schwab <jwschwab at berkeley.edu> wrote:
> Hi Zoe,
> 
> > I am trying to add an expression to "extra_dpdm" in other_momentum
> > subroutine which has been added to run_star_extras.f of my work
> > directory.
> 
> In general, it is a good idea to briefly describe your physical
> motivation.  For example, that lets us understand whether you need to
> include this term during the pre-main sequence evolution of the star.
> As indicated in the error message, MESA is struggling while constructing
> the PMS model.
> 
> > Here is my inlists http://www.uploadman.com/5482b953b1dbf96e or
> > http://dsa.uploadman.com/5482b953b1dbf96e?download_token=d735d3feb5bc080405aafe5c3f265a8327d5c4ea385d0ad7685f051395bf809d
> 
> I looked at your run_star_extras.f.
> 
> I would venture a guess that if you print the value of s% extra_dPdm
> you'll be surprised :)
> 
> I'll give you two hints to start.  The line
> 
>   integer :: k, Y=6d-6
> 
> and the lines like
> 
>   rr= (s% r(k+1) - s% r(k))*(1/4)*(pi)*s% r(k)**2
> 
> almost certainly don't do what you think they do.
> 
> Why don't you try printing the values of Y and rr?
> 
> Josiah
> 
> P.S.  There's no need to include the full contents of your working
> directory.  We don't need the 20 MB of executable and restart files.
> The inlists and run_star_extras files are only a few KB.  That means you
> will be able to directly attach them to your messages, making it even
> easier for people to take a look.