PV Smoothing

[sjanzou]

lib_dispatch_pv_smoothing_fom shared library and test
No weather file forecast until merged into develop

[brtietz]

SSC code looks good overall, mostly some cleanliness requests regarding the transition from development and debugging to production. Nice work!

[brtietz]

Travis seems to be having trouble with the tests, and I'm also having trouble locally:

From PV_Phoeniz_all_on:

    !! This test has probably CRASHED !!
    Test output:
    
    C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30037\include\vector(1566) : Assertion failed: vector subscript out of range

I dug into this on Generic_w_PV_input_all_on and it looks like the error is in the LCOS calculations in common_financial:

charged_pv only has length 1. batt_annual_charge_from_system appears to be empty.

I'm not sure whether this means the test needs to change to ensure this value is reported or the code in the financial models needs to be more robust. This metric appears to work in the GUI.

[sjanzou]

@mjprilliman, why is the generic system failing here? The tests run fine on my Windows box. Thanks!

[mjprilliman]

Travis seems to be having trouble with the tests, and I'm also having trouble locally:

From PV_Phoeniz_all_on:

    !! This test has probably CRASHED !!
    Test output:
    
    C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30037\include\vector(1566) : Assertion failed: vector subscript out of range

I dug into this on Generic_w_PV_input_all_on and it looks like the error is in the LCOS calculations in common_financial:

charged_pv only has length 1. batt_annual_charge_from_system appears to be empty.

I'm not sure whether this means the test needs to change to ensure this value is reported or the code in the financial models needs to be more robust. This metric appears to work in the GUI.

Must be something to do with analysis_period being set to 1. When that happens the array doesn't seem to have a year 0 like I was assuming so there's the size 1 array @brtietz. My code assumes we skip the first index in a lot of places because that is year 0. Is that the correct array sizing behavior @sjanzou?

Sorry I was right the first time. The array goes to size 1 rather than size 2 (expected) and thus shows up in Single Values rather than Annual Values in the GUI

[mjprilliman]

@brtietz @sjanzou

This is in cmod_battery.cpp. Why make it 1 rather than 2?

[sjanzou]

Agreed that should be 2. Same issue in cmod_fuelcell that prevented single year analysis periods to be run.

[sjanzou]

PVSmoothing tests working fine on Windows and Travis-CI

[sjanzou]

Latest Windows results with all test (usual widows weather_file_inputs test failure - usually different each time)