Trying to model AIKO NeoStar 2S+ 455 and Trina Vertex S+ 445

[ROB-ernst]

Hi all,

I am running into the same problem as in Q&A #2015 - I cannot seem to get fit_cec_sam to converge for the two module types I am using.

Parameters

AIKO NeoStar 2S+ A455-MAH54Db

Pmp = 455 
Voc = 41.00 
Isc = 14.22 
Vmp = 34.56
Imp = 13.17
alpha_isc_rel = 0.0005
beta_voc_rel = -0.0022
gamma_pmp_rel = -0.0026

cells_in_series = 108

Datasheet at https://aikosolar.com/wp-content/uploads/2024/10/Neostar-2S_Plus_188-AIKO-A-MAH54Db-440-470W_1757x1134x30mm_DS_EN_2405_V1.5.pdf

Trina Vertex S+ TSM-445 NEG9R.28

Pmp = 445
Voc = 52.6
Isc = 10.71
Vmp = 44.3
Imp = 10.05
alpha_isc_rel = 0.0004
beta_voc_rel = -0.0024
gamma_pmp_rel = -0.0030

cells_in_series = 144

Datasheet at https://static.trinasolar.com/sites/default/files/NEG9R.28_DE.pdf

Test function call

alpha_isc = Isc * alpha_isc_rel        # A / °C
beta_voc  = Voc * beta_voc_rel         # V / °C 
gamma_pmp = gamma_pmp_rel              # 1 / °C   

cec_params_test = pvlib.ivtools.sdm.fit_cec_sam(
                celltype='monoSi', 
                v_mp=Vmp,
                i_mp=Imp,
                v_oc=Voc,
                i_sc=Isc,
                alpha_sc=alpha_isc,
                beta_voc=beta_voc,
                gamma_pmp=gamma_pmp,
                cells_in_series=cells_in_series
            )

The function always fails with

6parsolve simulation error : Could not solve, sanity check failed (-33): abs((P - Pmp) / Pmp) > 0.015

I also tried a grid search with +/- 0.2 V resp. A around the values of Vmp and Imp, no success ...

Any suggestions what to do?

Thank you!

[markcampanelli]

Just looking at your code, is it possible that you need to multiply you max power relative temperature coefficient by the max power to get the absolute temperature coefficient?

[cwhanse]

For the AIKO I think all three temperature coefficients are incorrect. Pvlib matched the units requested by the underlying SAM code: A/C, V/C and for power %/C. I'd expect Idc to be about 0.007/C, Voc to be about 0.08 V/C and power to be 0.26/C.

I do didn't look at the second module.

[ROB-ernst]

Oh boy, thank you --- solved the issue. The issue was that the power coefficient was not in %/K, but I scaled it to a factor. So, instead of gamma_pmp_rel = -0.0026 it should be gamma_pmp_rel = -0.26. Now it works. Too many Christmas sweets, probably ...