TM3 weather data errors - one year works, another doesn't

I have two weather files, which have data in the TM3 format and are saved as .csv

When I run a default SAM setup for:
Photovoltaics > Flat Plate PV > Residential,

the simulation runs fine with the weather data from 1991, but fails for 1992 with the error message shown below. Oddly, both of the weather files work fine for a default PVWatts model. Both weather files are about the same format, so I'm not sure what's going on - maybe SAM is having some issues with the data. Anybody have an idea as to why one fails and the other doesn't? I'm using the 2012.5.11 version of SAM.




Unrelated: I don't know why you can't attach .tm3 files to this forum, since they seem to be commonly used.

---

SAMSIM MESSAGES

---

MSG[0]: failed to compute irradiation on surface (code: -105)
MSG[1]: MISSING OUTPUT system.monthly.e_net
MSG[2]: MISSING OUTPUT system.monthly.dcpower
MSG[3]: MISSING OUTPUT system.monthly.inc_total
MSG[4]: MISSING OUTPUT system.monthly.inc_beam
MSG[5]: MISSING OUTPUT system.annual.incident_beam
MSG[6]: MISSING OUTPUT system.annual.input_radiation
MSG[7]: MISSING OUTPUT system.annual.dc_nominal
MSG[8]: MISSING OUTPUT system.annual.dc_gross
MSG[9]: MISSING OUTPUT system.annual.dc_derated
MSG[10]: MISSING OUTPUT system.annual.ac_gross
MSG[11]: MISSING OUTPUT system.annual.ac_derated
MSG[12]: MISSING OUTPUT system.annual.e_net
MSG[13]: MISSING OUTPUT system.annual.fuel_usage
MSG[14]: MISSING OUTPUT system.annual.water_usage
MSG[15]: MISSING OUTPUT system.performance_factor
MSG[16]: MISSING OUTPUT system.use_lifetime_output
MSG[17]: MISSING OUTPUT system.heat_rate
MSG[18]: MISSING OUTPUT system.use_recapitalization
MSG[19]: MISSING OUTPUT 6par.a
MSG[20]: MISSING OUTPUT 6par.Io
MSG[21]: MISSING OUTPUT 6par.Il
MSG[22]: MISSING OUTPUT 6par.Rs
MSG[23]: MISSING OUTPUT 6par.Rsh
MSG[24]: MISSING OUTPUT 6par.Adj

---

TRNSYS LOG FILE

---

I tried using SAM's TMY3 editor to copy the data from the files giving SAM errors into a functional TMY3 (in this case the TMY3 file for Los Angeles Airport). I was able to run the simulation without errors after copying in all columns except for DHI; if DHI data is entered I get an error.

Does anybody know how important DHI is for the model? I might eliminate it or estimate it based on other responses in this forum.

Don't know if this is on-point or not. I installed a weather station here in Paauilo HI because I didn't think the Hilo airport would be characteristic of our elevation higher up the mountain. I've just (as of July 25) completed a year of data collection. The station includes a standard pyranometer, which gives me hourly measured GHI.

To make my local TMY3 file, I had to generate DNI from GHI using NREL's DISC code, and then had to generate DHI using NREL's GHI/DNI-to-DHI code. The DISC code occasionally calculates a negative number for DNI in the early morning or late afternoon, and this bad DNI data caused SAM (version 2012.5.11) to give the Simulation Error message. I found this out by process of elimination.

I went back to my local data spreadsheet and used an IF statement to find and scrub the negative DNI values, and then used TMY3 Creator (using Hilo as the base file) to make my local TMY3 file. Simulations using this file run fine in SAM.

So I suggest checking for oddball data in DNI and DHI (and possibly other columns) -- I suspect maybe all it takes is one cell with a screwy datum to cause a simulation to fail. You could also try NREL's GHI/DNI-to-DHI code to generate an alternative DHI column, paste that in and see how it works.

Good luck!

Denning :-)

Denning,

We figured out the problem, and it was a similar issue to what you had. The data set we were using had some values in DHI as 0, which were causing SAM to throw an error (usually 2-5 of these entries per data set). We just assumed that the value would be an average of the previous and next hour timestamps for the purpose of simulations and it worked fine.

Thanks for the help.

-Andrew