ENH: implementing pvsyst recombination loss current for CdTe and a:Si #504

mikofski · 2018-07-14T00:50:47Z

closes Adding New Single Diode Model For CdTe #163
add constant VOLTAGE_BUILTIN = 0.9 (V) to singlediode_methods.py
add arguments d2mutau=0, voltage_builtin=VOLTAGE_BUILTIN, and
cells_in_series to calculate i_recomb and v_recomb
add check for d2mutau to calc i_recomb, v_recomb, and gradients
add terms to current and gradients

pvlib python pull request guidelines

Thank you for your contribution to pvlib python! You may delete all of these instructions except for the list below.

You may submit a pull request with your code at any stage of completion.

The following items must be addressed before the code can be merged. Please don't hesitate to ask for help if you're unsure of how to accomplish any of the items below:

Closes issue Adding New Single Diode Model For CdTe #163
I am familiar with the contributing guidelines.
Fully tested. Added and/or modified tests to ensure correct behavior for all reasonable inputs. Tests (usually) must pass on the TravisCI and Appveyor testing services.
Updates entries to docs/sphinx/source/api.rst for API changes.
Adds description and name entries in the appropriate docs/sphinx/source/whatsnew file for all changes.
Code quality and style is sufficient. Passes git diff upstream/master -u -- "*.py" | flake8 --diff
New code is fully documented. Includes sphinx/numpydoc compliant docstrings and comments in the code where necessary.
Pull request is nearly complete and ready for detailed review.

Brief description of the problem and proposed solution (if not already fully described in the issue linked to above):

* closes pvlib#163 * add constant VOLTAGE_BUILTIN = 0.9 (V) to singlediode_methods.py * add arguments d2mutau=0, voltage_builtin=VOLTAGE_BUILTIN, and cells_in_series to calculate i_recomb and v_recomb * add check for d2mutau to calc i_recomb, v_recomb, and gradients * add terms to current and gradients

* also set v_recomb to Inf, so that 1/v_recomb is always zero * but precalculate grad_2i_recomb only if d2mutau > 0, otherwise, use zero

mikofski · 2018-07-29T04:16:56Z

@cwhanse the difference between DeSoto/CEC versus PVSyst with recombination losses seem correct now. See below for plots and and the next comment for the PVSyst First Solar FS-495 parameters used.

mikofski · 2018-07-29T04:21:40Z

@cwhanse if you think the ranges for these differences seems okay, then it is working for bishop88.

So, do you and @wholmgren think we should now add the vbi and d2mutau terms to the other singlediode methods like *_from_*() and mpp()?

Also, @cwhanse sorry to raise a dead issue, but Will and I both agree that singlediode.py is fine without _methods append to the end. Do you have an opinion?

cwhanse · 2018-07-29T19:42:22Z

Can you post the other PVsyst parameters for this module, and I'll check against a MATLAB implementation?

My quibble with changing singlediode_methods.py to singlediode.py is that singlediode is a function in pvlib.pvsystem. We'd have both pvlib.pvsystem.singlediode and pvlib.singlediode. Personally I find the re-use of names to be awkward.

mikofski · 2018-07-29T20:45:53Z

For DeSoto I got the parameters from pvsystem.retrieveSAM('CECMod'). I used reference conditions for POA=1000.0 [W/m^2] and TCELL=25.0.

In [8]: fs_495
Out[8]:
BIPV                  N
Date          9/18/2014
T_NOCT             44.6
A_c                0.72
N_s                 216
I_sc_ref           1.55
V_oc_ref           86.5
I_mp_ref            1.4
V_mp_ref           67.9
alpha_sc       0.000924
beta_oc        -0.22741
a_ref            2.9482
I_L_ref           1.563
I_o_ref        2.64e-13
R_s               6.804
R_sh_ref         806.27
Adjust           -10.65
gamma_r          -0.264
Version          NRELv1
PTC                89.7
Technology         CdTe
Name: First_Solar_FS_495, dtype: object

For PVSyst First Solar FS-495 I used this:

PVSYST_FS_495
Out[14]:
{'d2mutau': 1.31,
 'alpha_sc': 0.00039,
 'gamma_ref': 1.48,
 'mu_gamma': 0.001,
 'I_o_ref': 9.62e-10,
 'R_sh_ref': 5000,
 'R_sh_0': 12500,
 'R_sh_exp': 3.1,
 'R_s': 4.6,
 'beta_oc': -0.2116,
 'EgRef': 1.475,
 'cells_in_series': 108,
 'cells_in_parallel': 2,
 'I_sc_ref': 1.55,
 'V_oc_ref': 86.5,
 'I_mp_ref': 1.4,
 'V_mp_ref': 67.85,
 'temp_ref': 25,
 'irrad_ref': 1000,
 'I_L_ref': 1.5743233463848496}

Ok, I'll await further changes to the thin film recombination loss current until I get a thumbs up.

RE: the other matter, I guess let's leave it as singlediode_methods.py then? Unless anyone else can think of a better name? Like maybe cells.py or modules.py or dc_energy.py - and I'm sorry for crossposting here and elsewhere, I'm just too shy to create another issue, maybe this topic would be suitable for google groups?

mikofski · 2018-07-31T20:35:41Z

@wholmgren , @cwhanse after talking with @jdnewmil I realized I should not have used parameters from CEC database - and also I used the wrong method, I should have used the new calcparams_pvsyst, so sorry for wasting your time.

I've revised the test and plots now, and posted them above. Thanks!

cwhanse · 2018-08-01T17:05:26Z

@mikofski I'm having trouble interpreting your figures above. Does DeltaI compare the Desoto model to the Pvsyst model? I would like to verify that the results with the recombination term appear reasonable, and to do that it would help to compare two IV curves computed with the same model (Pvsyst) without and with the term activated.

mikofski · 2018-08-01T17:52:43Z

OK, but AFAIK you can't use PVSyst this way, b/c the parameters would be different ie using the PVSyst database parameters for FS-495 without the recombination loss will not fit the measured IV curve or nameplate output power. What my plots above show is that the PVSyst model agrees with the DeSoto model at the critical points (short circuit, max power, open circuit) when the correct parameters for each are used respectively, but that the shape of the IV curve is different, especially between Vmp and Voc which is where Merten observed differences between model and measurements.

Here are the deltas for PVSyst model with and without recombination loss current:

delta Pmax = 5.797158
delta Isc = 0.022876
delta Voc = 0.544359

and plots comparing PVSyst model with and without recombination loss current:

NOTE: The same PVSyst model parameters were used for both conditions, however, PVSyst model is not intended to be used without the recombination loss current for CdTe or amorphous-silicon modules, therefore the calculate IV curve and max power without recombination loss are wrong.

cwhanse · 2018-08-01T18:52:29Z

thanks @mikofski I'm more interested to qualitatively verify the with/without recombination term calculation than to compare Desoto and Pvsyst for nominally the same module. The results in the Pvsyst with/without figures look correct.

Comparing Desoto to Pvsyst brings in other assumptions in addition to the recombination current, not least of which is the method used to derive the model parameters.

I'm ready to merge this, pending concurrence from @wholmgren

wholmgren · 2018-08-01T18:59:17Z

@mikofski let me know when the PR template boxes are checked and I will review.

mikofski · 2018-08-01T23:51:06Z

Hi @wholmgren, I don't think this is ready yet.

I still need some direction on how many other methods these parameters should be added to, and what type of text we want to put in the documentation for users.

I think we would need to add voltage_builtin (aka vbi), d2mutau, and cells_in_series to each of the following:

singlediode_methods.bishop88_i_from_v()
singlediode_methods.bishop88_v_from_i()
singlediode_methods.bishop88_mpp()
singlediode_methods._lambertw_v_from_i()
singlediode_methods._lambertw_i_from_v()
singlediode_methods._lambertw()
pvsystem.PVSystem.singlediode()
pvsystem.PVSystem.i_from_v()
pvsystem.singlediode()
pvsystem.max_power_point()
pvsystem.v_from_i()
pvsystem.i_from_v()

I think in addition to docstring API changes that describe the new arguments, we need some sternly worded messages that warn users:

Do not use d2mutau with CEC coefficients.

... and ...

Usage of d2mutau with PVSyst coefficients is required for CdTe and a:Si modules.

... and ...

For PVSyst CdTe and a:Si modules, the cells_in_series parameter must only account for a single parallel sub-string if the module has cells in parallel greater than 1.

EG: FS-495 has two parallel sub-strings with 108 cells each, therefore cells_in_series should be 108.

Can someone comment to let me know if these changes are aligned with their thoughts? thx!

wholmgren · 2018-08-02T01:30:22Z

@mikofski ah, I haven't followed closely and was only responding to @cwhanse's comment regarding ready to merge. Maybe there is a mismatch in anticipated scope?

Re singlediode.py vs singlediode_methods.py... We've previously discussed breaking pvsystem.singlediode into a couple of functions. pvsystem.singlediode could disappear entirely if/when that happens. I rarely see module names as long as singlediode_methods or module names with underscores. So, my vote would be change the module name to singlediode.

mikofski · 2018-08-02T02:04:49Z

Ok, so then what should we do with this PR and do I need to open a few new issues to start discussion of these remaining tasks?

cwhanse · 2018-08-02T14:32:59Z

pvlib/singlediode_methods.py

@@ -62,7 +64,9 @@ def estimate_voc(photocurrent, saturation_current, nNsVth):


 def bishop88(diode_voltage, photocurrent, saturation_current,
-             resistance_series, resistance_shunt, nNsVth, gradients=False):
+             resistance_series, resistance_shunt, nNsVth, d2mutau=0,
+             cells_in_series=None, voltage_builtin=VOLTAGE_BUILTIN,


A nit - re-order the arguments to keep the non-recombination current terms first, followed by d2mutau and voltage_builtin - just move cells_in_series up one position.

cwhanse · 2018-08-02T14:34:07Z

pvlib/test/test_singlediode_methods.py

+
+def pvsyst_fs_495():
+    """
+    PVSyst First Solar FS-495 parameters.


Calculate PVsyst parameters for First Solar FS-495 module.

Actually I would prefer to simply hard-code the dict of results, and provide a brief statement for the source of these values - where does the initial fs_495 data come from?

✔️ the initial data comes from the PVSyst v6.7.2 database, I don't think any of the values were altered but it was vetted internally.

I've removed the method, and hard coded the values with a comment re: initial source.

Not having a current PVsyst license, I'm guessing that it does NOT permit uploading that database to pvlib?

Maybe not the whole database, but I'm guessing the relevant values come from pan files generated by FS.

cwhanse · 2018-08-02T14:38:28Z

pvlib/test/test_singlediode_methods.py

+PVSYST_FS_495 = pvsyst_fs_495()  # PVSyst First Solar FS-495 parameters
+
+
+def test_pvsyst_fs_495_recombination_loss():


I think the test is more general than just the fs_495 module - I'd drop that nomenclature from the function name.

✔️ agreed, changed to test_pvsyst_recombination_loss

cwhanse

OK, we can rename the module to singlediode. We might want to refactor the function singlediode but I am in favor of keeping it as a wrapper that accepts the coefficients for the single diode equation and return the IV curve. That functionality is generic to many of the DC models we implement. We could rename the function, I suppose.

I'm in favor of merging this pull request with its current scope (after approvals). We can open separate issues to 1) add voltage_builtin ('Vbi') and d2mutau to the component functions, 2) clarify the cells_in_series (aka Ns) argument. We currently pass the product nNsVth into singlediode and related functions, rather than passing n, Ns, and Vth separately. With the recombination current, the only place Ns is needed separately is with the per-cell 'Vbi' to get a module 'Vbi'. I think a more consistent method would be to pass 'NsVbi' as the argument, rather than 'Vbi'. That keeps all the single diode equation parameters as 'per-module' quantities, rather than a mix of per-cell and per-module.

I'm not sure we need those 'sternly worded' warnings. We could test inside calcparams_pvsyst for cell type and non-zero d2mutau and warn the user when the recombination current term isn't set up right. But to me, the real issue is abeting the view that the parameters in the CEC model are somehow generic for any single diode model. I think a way to ameliorate that risk is tie the SAM parameter reader function to the CEC model, so that it is awkward to invoke the PVsyst model (or other single diode model) on a set of parameters read from the SAM file.

cwhanse · 2018-08-02T14:40:38Z

pvlib/test/test_singlediode_methods.py

+    )
+
+    # test expected change in max power
+    assert np.isclose(max(desoto[2]) - max(pvsyst[2]), 0.01949420697212645)


This test passes if the Desoto model and the PVsyst model are 'close', meaning that the test would fail is the CEC parameter database entry for this module changes, if there is an error in the Desoto functions, or if the PVsyst function has an error. I'd change the test condition to compare the PVsyst output to known values calculated by a separate method.

✔️ I've changed the test to be independent of DeSoto, now it only tests 2 conditions:

reference condition, should agree with its own parameters

another condition, (888[W/m^2], 55[degC]), for which I calculated the expected output using calcparams_pvsyst and bishop88

But I should note that the output at (888[W/m^2], 55[degC]) doesn't agree well with DeSoto at the same conditions:

In [1]: from pvlib import pvsystem In [2]: CECMOD = pvsystem.retrieve_sam('cecmod') In [3]: fs_495 = CECMOD.First_Solar_FS_495 In [4]: il, io, rs, rsh, nnsvt= pvsystem.calcparams_desoto(888, 55, ...: alpha_sc=fs_495.alpha_sc, a_ref=fs_495.a_ref, I_L_ref=fs_495.I_L_ref, ...: I_o_ref=fs_495.I_o_ref, R_sh_ref=fs_495.R_sh_ref, R_s=fs_495.R_s, ...: EgRef=1.475, dEgdT=-0.0003) In [5]: pvs = pvsystem.singlediode(il, io, rs, rsh, nnsvt, method='lambertw') In [6]: pvs Out[6]: OrderedDict([('i_sc', 1.4020527838304147), # PVSyst: 1.3868344548308347 ('v_oc', 75.41704855182024), # PVSyst: 79.29198489135703 ('i_mp', 1.2599626119945482), ('v_mp', 57.70813302772382), ('p_mp', 72.71009002293975), # PVSyst: 76.26211540956041 ('i_x', 1.3606253515830833), ('i_xx', 0.8340096657730403)])

According to @jdnewmil this is possibly due to a setting in PVSyst that adjusts d2mutau to a fixed max power temp coeff instead (or in addition???) to the adjustment on gamma. Even though I can see these check boxes in PVSyst, I'm not sure to what degree we want to copy or guess what PVSyst does here in PVLib (or in SAM either @timorichert)

@cwhanse if we wanted to try to emulate this 2nd order d2mutau behavior, then IMO it should go into calcparams_pvsyst, and so those values: d2mutau (and voltage_builtin if multi junction?) should be outputs from there.

Agree. I don't know what PVsyst might be doing behind the interface when d2mutau is in action, however, so I can't help clarify here.

If it isn't clear what PVsyst is doing, then I suggest mailing Bruno to get clarification. I think they're quite open about the calculations.

cwhanse · 2018-08-02T14:41:25Z

pvlib/test/test_singlediode_methods.py

+    return desoto, pvsyst
+
+
+if __name__ == '__main__':


Although great for development and testing, I believe we don't keep if __name__ == '__main__': blocks in modules.

✔️ agreed, removed

* change order of bishop88 args: cells_in_series to be first before d2mutau and voltage_builtin * hardcode PVSYST_FS495 coefficients * don't compare pvsyst to desoto, instead compare pvsyst to fixed values at two conditions: reference and 888[W/m^2],55[degC] * rename test to be agnostic to module SKU * remove testing script if __name__ == '__main__': section, not for release

…ation loss

mikofski · 2018-08-02T23:49:32Z

okay, @wholmgren @cwhanse and anyone else, I think this is ready for review and merging if there are no more comments. thanks

mikofski · 2018-08-03T00:33:10Z

Evidently Python-2 doesn't like "André Mermoud" because it had utf-8, let me add encoding to the file and see if that fixes it

mikofski · 2018-08-03T07:19:38Z

emphasis mine:

I'm not sure we need those 'sternly worded' warnings. We could test inside calcparams_pvsyst for cell type and non-zero d2mutau and warn the user when the recombination current term isn't set up right. But to me, the real issue is abetting the view that the parameters in the CEC model are somehow generic for any single diode model. I think a way to ameliorate that risk is tie the SAM parameter reader function to the CEC model, so that it is awkward to invoke the PVsyst model (or other single diode model) on a set of parameters read from the SAM file.

I think a possible solution to this situation is for calcparams_pvsyst to return both d2mutau and NsVbi which are then passed to singlediode and its delegates (bishop88) and proxies (PVSystem.singlediode). If this is what you are suggesting then I agree.

NOTE: having calcparams_pvsyst handle NsVbi also takes care of multi-junction cells, which is another caveat that users would have to deal with since the default 0.9-V is only for single junction, and although rare, tandem and triple a:Si cells would have a builtin voltage of 1.8-V and 2.7-V respectively.

And this could also avoid any confusion regarding CIS/CIGS which should not use the thin-film recombination loss at all.

I'm in favor of merging this pull request with its current scope (after approvals). We can open separate issues ...

I agree with everything in the second paragraph.

OK, we can rename the module to singlediode. ...

Thank you. So shall this also be a separate issue out of scope for this PR?

* need to indent bullets in warning * only need cells in series for PVSyst thin-film recombination loss * escape math latex twice inside docstrings without raw prefix, eg: \\tau without extra escape is a tab character, ha ha ha * in rst, italics is a single asterisk, not underscore, thats markdown

mikofski · 2018-08-03T08:04:26Z

failures are do to Siphon, not related to doc changes in 1e9434a

cwhanse · 2018-08-03T14:33:44Z

I think a possible solution to this situation is for calcparams_pvsyst to return both d2mutau and NsVbi which are then passed to singlediode and its delegates (bishop88) and proxies (PVSystem.singlediode. If this is what you are suggesting then I agree.

We agree.

Thank you. So shall this also be a separate issue out of scope for this PR?

Yes, please.

wholmgren

@mikofski this looks good. Only minor changes requested. I appreciate the limited scope!

wholmgren · 2018-08-03T17:28:53Z

pvlib/singlediode_methods.py

@@ -86,6 +95,18 @@ def bishop88(diode_voltage, photocurrent, saturation_current,
    nNsVth : numeric
        product of thermal voltage ``Vth`` [V], diode ideality factor ``n``,
        and number of series cells ``Ns``
+    cells_in_series : int


None or int

wholmgren · 2018-08-03T17:29:47Z

pvlib/singlediode_methods.py

+    .. warning::
+       * Do not use ``d2mutau`` with CEC coefficients.
+       * Usage of ``d2mutau`` with PVSyst coefficients is required for CdTe and
+         a:Si modules.


I wonder about the short hand for module type here. I suppose this is fairly low level function so we should not worry too much.

wholmgren · 2018-08-03T17:30:41Z

pvlib/singlediode_methods.py

+    The PVSyst thin-film recombination losses parameters ``d2mutau`` and
+    ``voltage_builtin`` are only applied to cadmium-telluride (CdTe) and
+    amorphous-silicon (a:Si) PV modules, [2]_, [3]_. The builtin voltage should
+    account for all junctions. *EG*: tandem and triple junction cell would have


EG -> For example

wholmgren · 2018-08-03T17:34:12Z

pvlib/singlediode_methods.py

+    i_recomb, v_recomb = 0, np.inf
+    if d2mutau > 0:
+        v_recomb = voltage_builtin * cells_in_series - diode_voltage
+        i_recomb = photocurrent * d2mutau / v_recomb


I would go with

else: i_recomb, v_recomb = 0, np.inf

instead of defining them first, then potentially redefining them. Seems cleaner, no?

wholmgren · 2018-08-03T17:38:13Z

pvlib/singlediode_methods.py

    """
+    # check if need to calculate recombination loss current
+    i_recomb, v_recomb = 0, np.inf
+    if d2mutau > 0:


d2mutau : numeric implies that it can be an array (see here). This if statement will error if d2mutau is an array. Perhaps the docstring should be changed to scalar?

wholmgren · 2018-08-03T17:41:39Z

pvlib/singlediode_methods.py

+        grad_i_recomb = grad_2i_recomb = 0
+        if d2mutau > 0:
+            grad_i_recomb = i_recomb / v_recomb
+            grad_2i_recomb = 2 * grad_i_recomb / v_recomb


here, too, I prefer an else statement.

wholmgren · 2018-08-03T17:42:18Z

pvlib/singlediode_methods.py

    v = diode_voltage - i * resistance_series
    retval = (i, v, i*v)
    if gradients:
+        # check again if need to calculate recombination loss current gradients
+        grad_i_recomb = grad_2i_recomb = 0


No offense, but I don't care for this assignment pattern. Two lines, please.

wholmgren · 2018-08-03T17:47:44Z

pvlib/test/test_singlediode_methods.py

+#     Vd = I_sc_ref * R_s
+#     Id = I_o_ref * (exp(Vd / nNsVt) - 1)
+#     Ish = Vd / R_sh_ref
+PVSYST_FS_495 = {


dangerous to use a module level dictionary in tests because it can be modified. Suggest turning this into a pytest fixture

@pytest.fixture def PVSYST_FS_495(): return {...}

probably with a lowercase name but no big deal here.

wholmgren · 2018-08-03T17:52:15Z

pvlib/test/test_singlediode_methods.py

+      (5e-4, 0.04)),
+     (POA, TCELL, {'pmp': 76.26, 'isc': 1.387, 'voc': 79.29}, (1e-3, 1e-3))]
+)  # DeSoto @(888[W/m**2], 55[degC]) = {Pmp: 72.71, Isc: 1.402, Voc: 75.42)
+def test_pvsyst_recombination_loss(poa, temp_cell, expected, tol):


if PVSYST_FS_495 becomes a fixture, add it as a parameter so that the value can be accessed in line 161 below.

wholmgren · 2018-08-03T17:57:07Z

pvlib/test/test_singlediode_methods.py

+    'poa, temp_cell, expected, tol',
+    [(PVSYST_FS_495['irrad_ref'], PVSYST_FS_495['temp_ref'],
+      {'pmp': PVSYST_FS_495['I_mp_ref']*PVSYST_FS_495['V_mp_ref'],
+       'isc': PVSYST_FS_495['I_sc_ref'], 'voc': PVSYST_FS_495['V_oc_ref']},


I can't recall if these dict look ups will work if you turn PVSYST_FS_495 into a fixture. If not, I guess it's ok to leave as is for now.

* expand module type abbrev. in warnings * also expand EG to For example, too terse * cells in series can be int _or_ `None` * since d2mutau and vbi are arrays, then ambiguous to test for zero, so use np.where to assign conditionally instead (in two places)

* dangerous to use module-level dictionary inside test, b/c dictionary values could be changed inadvertently, but not a fixture * add fixture as argument in test * change parametrize to call fixture instead of module-level dictionary

* related to pvlib#516, the use of cells_in_series here is problematic * to calculate arrays of different cells or modules with different parameters we use np.where(is_recomb, ..., ...) 3d0deb2 but what should cells_in_series be if d2mutau is zero? doesn't make physical sense for it to be zero, but None doesn't work with numbers * remove cells_in_series arg from bishop88 * remove last warning about setting cells_in_series correctly, move it to Notes, but worded differently in the context of setting NsVbi * replace voltage_builtin with NsVbi, default to np.inf * fix docstring parameter descriptions

mikofski · 2018-08-08T03:38:19Z

Test failure for AppVeyor Python 3.6 due to test_forecast is unrelated. All other jobs pass and Travis all pass. Is there anything more anyone is expecting here? @cwhanse do you approve this PR?

wholmgren · 2018-08-08T03:47:36Z

@mikofski sorry I did not realize this was ready. Thanks for making the changes. Looks great. np.where is a good choice. Only catch is that your function will always return np.arrays for the gradients. I'm guessing that is no big deal.

I restarted the build and the tests all pass.

wholmgren force-pushed the master branch from 7d27df9 to 2498040 Compare July 14, 2018 15:27

ENH: need to set grad_i_recomb if d2mutau not given

2bd8b13

* also set v_recomb to Inf, so that 1/v_recomb is always zero * but precalculate grad_2i_recomb only if d2mutau > 0, otherwise, use zero

ENH: fix need to multiply vbi by Ns, add test for recombination

1650bc6

WIP: ENH: TST: BUG: in test compare desoto vs. pvsyst

9d96585

cwhanse reviewed Aug 2, 2018

View reviewed changes

mikofski added 3 commits August 2, 2018 13:34

DOC: ENH: update bishop88 documentation for pvsyst thin-film recombin…

a62759e

…ation loss

DOC: ENH: update what's new

bfa9c5f

DOC: add utf-8 file encoding

79fde5b

mikofski changed the title ~~WIP: ENH: implementing pvsyst recombination loss current for CdTe and a:Si~~ ENH: implementing pvsyst recombination loss current for CdTe and a:Si Aug 3, 2018

mikofski mentioned this pull request Aug 3, 2018

ENH: rename singlediode_methods.py to singlediode.py and discuss pvsystem.singlediode refactor #515

Closed

wholmgren reviewed Aug 3, 2018

View reviewed changes

This was referenced Aug 3, 2018

ENH: calcparams_pvsyst should output d2mutau and vbi #516

Open

ENH: model pvsyst thin-film recombination for CdTe and a:Si in pvsystem.singlediode() #517

Open

mikofski added 4 commits August 3, 2018 11:40

Merge branch 'master' into pvsyst_cdte_recombination_loss_current

adde5bf

cwhanse approved these changes Aug 8, 2018

View reviewed changes

wholmgren merged commit 2a436c6 into pvlib:master Aug 8, 2018

mikofski deleted the pvsyst_cdte_recombination_loss_current branch August 8, 2018 19:40

		PVSYST_FS_495 = pvsyst_fs_495() # PVSyst First Solar FS-495 parameters


		def test_pvsyst_fs_495_recombination_loss():

ENH: implementing pvsyst recombination loss current for CdTe and a:Si #504

ENH: implementing pvsyst recombination loss current for CdTe and a:Si #504

Conversation

mikofski commented Jul 14, 2018 • edited Loading

pvlib python pull request guidelines

mikofski commented Jul 29, 2018 • edited Loading

mikofski commented Jul 29, 2018

cwhanse commented Jul 29, 2018

mikofski commented Jul 29, 2018 • edited Loading

mikofski commented Jul 31, 2018 • edited Loading

cwhanse commented Aug 1, 2018

mikofski commented Aug 1, 2018 • edited Loading

cwhanse commented Aug 1, 2018

wholmgren commented Aug 1, 2018

mikofski commented Aug 1, 2018 • edited Loading

wholmgren commented Aug 2, 2018

mikofski commented Aug 2, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cwhanse Aug 2, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cwhanse left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mikofski commented Aug 2, 2018

mikofski commented Aug 3, 2018 • edited Loading

mikofski commented Aug 3, 2018 • edited Loading

mikofski commented Aug 3, 2018

cwhanse commented Aug 3, 2018

wholmgren left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mikofski commented Aug 8, 2018 • edited Loading

wholmgren commented Aug 8, 2018 • edited Loading

mikofski commented Jul 14, 2018 •

edited

Loading

mikofski commented Jul 29, 2018 •

edited

Loading

mikofski commented Jul 29, 2018 •

edited

Loading

mikofski commented Jul 31, 2018 •

edited

Loading

mikofski commented Aug 1, 2018 •

edited

Loading

mikofski commented Aug 1, 2018 •

edited

Loading

cwhanse Aug 2, 2018 •

edited

Loading

mikofski commented Aug 3, 2018 •

edited

Loading

mikofski commented Aug 3, 2018 •

edited

Loading

mikofski commented Aug 8, 2018 •

edited

Loading

wholmgren commented Aug 8, 2018 •

edited

Loading