search

SunPower / PVMismatch

An explicit Python PV system IV & PV curve trace calculator which can also calculate mismatch.
http://sunpower.github.io/PVMismatch/
BSD 3-Clause "New" or "Revised" License
79 stars 30 forks source link

Fix Temperature Sensitivity of MPP #67

Closed rayhickey closed 6 years ago

rayhickey commented 6 years ago

Fixes I_mp and V_mp sensitivity to temperature caused by selecting Isys and Vsys values at the maximum Psys index. This is accomplished by linear interpolation between the points surrounding MPP. Previously: i_mp_temp_old Now: i_mp_temp_new

rayhickey commented 6 years ago

@mikofski is this noise expected between a pvcell and pvsystem? imp_cell_vs_syst

mikofski commented 6 years ago

@rayhickey can you go through test_setSuns(), substitute the newly calculated values, and make sure that they make sense? This was a tricky issue before, #9, and why we dropped the interpolation to begin with, so IMO make sure it's working well before moving forward.

I'm not sure why you're seeing differences between PVcell and PVsystem. How are you getting Imp for the cell? I wouldn't expect to see any differences in reality, but it might be due to the interpolation. Since both curves have the same number of discrete points, but the range in voltage is greater in the system, then there would be greater voltage difference between points, so possibly a greater opportunity for errors due to the interpolated estimate?

Did you try the parabola option? probably overkill, but may do better than linear interpolation. But maybe take too long, so if linear is okay, and noise is small enough then stick with that? Does noise go away if you increase number of points? Maybe use parabola for system, and linear for cells? Are you implementing something for cells? I didn't look too closely at your changes. @chetan201 what do you think? Try to get @bmeyers to weigh in too.

mikofski commented 6 years ago

also, sorry to nit, and this is just my humble opinion, but IMHO in general you should not work out of master, and you should not submit PR's from master, because your master branch should be a clean, perfect, mirror of the upstream master branch, and the master branch should be ready to deploy. You can read more about it on my old blog in "winning workflow" and in a talk I gave last year on OSS at UCBIDS.

You can fix this easily from your laptop. Just run these commands:

path\to\pvmismatch (master) $ git status  # make sure your repo is clean, commit changes first
path\to\pvmismatch (master) $ git checkout -b fix_mpp_temp_sensitivity
path\to\pvmismatch (fix_mpp_temp_sensitivity) $ git push -u remotes/origin fix_mpp_temp_sensitivity
# this assumes that the remote for your GitHub repo is called origin

Pause here to go online to GitHub and change your PR to come from the new feature branch instead of master.

path\to\pvmismatch (fix_mpp_temp_sensitivity) $ git checkout master
# this next command resets master to the upstream version
path\to\pvmismatch (master) $ git reset --hard remotes/upsteam/master
# assuming your remote to SunPower/PVMismatch is called upstream
path\to\pvmismatch (master) $ git push -f origin master  # use -f to force overwrite

you could drop off the prefix remotes/ I just put it there so you know these are remotes.

:)

mikofski commented 6 years ago

closed in favor of #68