CompareToEnergyPlus

Comparison to EnergyPlus

The unit tests are doing a good job of ensuring specific outputs of the functions. They will continually be improved. In addition, the demos on the wiki are demonstrating the functions over time with explanations. As a final check before these are considered validated, we will run EnergyPlus and compare the outputs.

EnergyPlus Model Setup

A shoebox model has been created and the east and west walls will be used as the demonstrative surfaces. The relevant solar angles will be reported from EnergyPlus, which has been validated extensively, and used as a verification (near-validation) source. The model will include the following characteristics:

• Design day located in Golden, CO
• Site:Location:Latitude: 39.57 degrees NORTH
• Site:Location:Longitude: 104.85 degrees WEST
• Winter Date: 12/21
• Summer Date: 7/21

Case 1: Site-specific angles

Three site-related angles are reported from EnergyPlus

• Site Solar Azimuth Angle
• Site Solar Altitude Angle
• Site Solar Hour Angle

Here is the comparison for the Winter day:

Note that I had to negate the hour angle for the comparison because EnergyPlus uses the convention of positive hour angle in the morning and negative in the afternoon. Overall the results match up very closely.

Here is the comparison for the Summer day:

Note that I left daylight savings time OFF because in EnergyPlus, design days do not have daylight savings time enabled (at least not by default).

Case 2: Solar Incidence Angle

EnergyPlus is capable of reporting the cosine of the solar angle of incidence on a particular surface. For this case, we will add that variable to the east and west facing surfaces of the shoebox model. The resulting comparison is here:

Note that the trend is a good match, but the curves are slightly off. Initially, I tried re-enabling daylight savings time, thinking it was possibly treated different from the other angles, and that shifted the curves such that they were off by a similar amount, but in the opposite direction. The data is basically off by 30 minutes, instead of a full hour, which got me thinking this had to do with how EnergyPlus interpolates within the hour.

I went back and re-ran the library, calling it at the 30-minute mark instead of the top of the hour. Here is the result:

This looks so much better. The only real deviation is that EnergyPlus reports a positive value at the 5th hour shown. This is because EnergyPlus is actually calculating this value at each time step (15 minutes), and then averaging to report the hourly value. The library just samples once, in this case at the 30 minute mark of the hour. So this is showing a difference because the sun must be slightly up at the :45 mark of the hour, so EnergyPlus shows a positive value. This could be remedied by calling the library on a 15-minute interval.

I will not be going back and re-creating the plots to account for this 30-minute and reporting variation. This essentially validates the solar calculation library.