Every so often I receive questions about making your own custom .epw weather file from onsite weather station data. I haven't done this in a while, so others may have better methods. However, since I was typing this up anyway, I thought I'd post it.
Tips for making your own weather file for simulation from a weather station:
There are multiple ways to make your own weather file from a weather station, for example, you can use the EnergyPlus Weather Tool. If you're doing it by hand, here are some tips based on my experience.
1. Start with an .epw which is as close as possible to your custom file, then revise it with your custom data.
2. Info on what each of the columns in the weather file means can be found in the EnergyPlus user guide.
3. You must convert the timesteps. If your weather station logs every 5 mins and you're using hourly data in your simulation, you must convert the data. You'll need to decide how to deal with wind direction, which isn't so easy to average. You could use the dominant wind direction over the hour, or consider each wind direction and speed as a vector and then mathematically combine them, etc.
4. Make sure you're not missing any data entries. If you have any blanks in the file it won't run. Out-of-range data will also cause errors.
5. If there is data that you need in the file but don't have via your weather station, you have multiple options. Here are some shortcut solutions. (Yes, I acknowledge that these shortcuts cause inaccuracies. I recommend doing a sensitivity analysis with high and low values to see if the data really has a big impact on your simulation. If not, use the shortcut.) A. You could keep the data from the original .epw file. B. You could select some realistic value and use it for every time step. I do not recommend this approach for solar radiation. See below.
6. Unless you have a fancy weather station, you will probably need to convert from global horizontal radiance to indirect and direct normal radiance (and possibly other variables). Multiple people have suggested equations for this based on empirical findings. Each empirical formula has its own inaccuracies. I'd recommend picking one that was created based on research in a climate similar to yours. I've used one from Duffy and Beckman's Solar Engineering of Thermal Processes 1991.
7. Make small changes to the file and run it often, to facility easy debugging. If you have a problem, also try running portions of the year to pinpoint the problem.
8. In case you don't know how to convert between an Excel file and .epw formats, I wrote the directions below a while back.
Best of luck,
Holly
I'm sure there are a lot of ways to convert between .epw and excel. This is one
way that works:
1. Open your EPW as a CSV in Excel.
2. Make the desired edits, such as plugging in your weather station data.
(Even if you're not using EnergyPlus, the Auxiliary EnergyPlus Programs manual
will explain the components of the EPW file.)
3. Save the file as a CSV. I usually save only the header first, deleting the
body. Then only the body, deleting the header. Otherwise, you end up with a
lot of extra commas in the header. This method is quicker than deleting the
commas.
4. Then I open the files in Textpad or Notepad and combine the header and body
and make sure everything looks OK. Then save it. CSV or TXT format is fine.
5. Then I manually change the file name from .txt or .csv to .epw.
6. Then the EPW file works in EnergyPlus. I'm assuming it would work in other
programs as well. As long as all the values are OK I haven't had any problems
with conversions.
As an alternative, you can use the EnergyPlus Weather Tool to compile custom
data into an EPW format.
Hope this helps.
Holly
Holly W. Samuelson, DDes, LEED, AIA
Assistant Professor Harvard Graduate School of Design
