search

dpinney / omf

The Open Modeling Framework for smart grid cost-benefit analysis.
https://omf.coop
GNU General Public License v2.0
112 stars 60 forks source link

2. Test Weather.py output with a GLM #233

Closed dpinney closed 10 years ago

dpinney commented 10 years ago

Note: this should all be done in a .py script. The script doesn't have to handle every possible input, it just needs to make these steps doable on command.

  1. Use weather.py to create a climate CSV for Washington DC over a year.
  2. Take a test GLM (13 Node IEEE--or can you recommend something more realistic?) and make that weather play into a simulation.
  3. Run the GLM for 1 year.
  4. Plot result time series. Power flowing in to feeder from substation, PF at substation, energy consumed by loads, energy consumed by losses, min/max/avg voltage at each meter, all variables from GLD Climate object. This is basically the same charts as come out of /omf/models/gridlabMulti.py.
  5. Take the same GLM and run it with the TMY weather for the same period, plotting the same results.
  6. Pair the charts and let's see what they look like.

Feel free to chart in Excel.

By the way, what happens when you run a glm with TMY and simulate more than a year? Does the weather loop?

Immediate todo: how long will this take to accomplish?

jcfuller1 commented 10 years ago

By the way, what happens when you run a glm with TMY and simulate more than a year? Does the weather loop?

Yes. It will just loop the data indefinitely.

dpinney commented 10 years ago

ETA 21 July.

dpinney commented 10 years ago

We're 4 days late on this -- you guys need help? Questions?

dpinney commented 10 years ago

Okay, looks like weather.py has some crazy insolation code in it.

Two paths forward look promising:

  1. Just play TMY3 historical data (note: from the dataset, not the typical aggregation) into models.
  2. Find a better source of insolation data.
jcfuller1 commented 10 years ago

New weather.py posted last week. Comments?

dpinney commented 10 years ago

This commit?

I see that you guys prefer Dulles instead of National.

I see that some new weather conditions were described.

How have the test outputs changed?

Are TMY3 historical data or other insolation sources being addressed?

nikhilgupta10 commented 10 years ago

Got Dulles TMY2 data, so was easier to compare against WU data plots. All the weather conditions/irradiance ratios are calculated (used TMY3 data and WU specified weather conditions) and changed. Compared the weighted averages for tmy2, and WU data. Solar Direct beam, diffuse and, global irradiance values changed (more realistic now). Ran it against 3 different sites for about 3 years and verified. (by comparing against TMY2 plots-weighted averages)

dpinney commented 10 years ago

Could I see the new output plots?

nikhilgupta10 commented 10 years ago

weather_climate_data_iad_2010

dpinney commented 10 years ago

Nice, solar global looks less boxy. Test process sounded good.

Are we done with weather.py?

jcfuller1 commented 10 years ago

Almost. We had some others pinging on the script and they ran into a couple bugs at a certain airport location. Also noticed that the header information had been stripped off, which is needed to calculate solar time. We're fixing those now.

dpinney commented 10 years ago

Extra fixes welcome.

Let's ignore TMY3/other sources of insolation data for now. Our solar guy doesn't think they are the best approach for our use cases.

dpinney commented 10 years ago

Sri still working as of today.

nikhilgupta10 commented 10 years ago

The issues are already fixed. Tested and running well at the airports we had issues last week

dpinney commented 10 years ago

Nice. Great improvements.