I am modeling a multispeed coil (Coil:Cooling:DX:Multispeed) and a variable speed fan using the Fan:OnOff object (I want it to be able to cycle). According to the documentation, if you wish to model a variable speed fan you need to provide a fan curve that represents the ratio of actual fan power to rated fan power when a change in fan speed occurs. If my understanding of the documentation is correct, for example, if my rated fan power is 1000W and that my fan is running at a PLR of 0.6 and that my curve returns a value of 0.4 my actual fan power should be 400W. I tried with several different curves but I couldn’t get my fan power to be calculated according to my interpretation.
To get a better understanding of what EnergyPlus is doing, I took a quick peek at the source code. I noticed that on line 1926 through 1930 there is an if/else statement that sets the fan run time fraction equals to the part load ratio if “OnOffFanPartLoadFraction” is equal to 1. In the case where my fan’s “OnOffFanPartLoadFRaction” is 1 but is running at a PLR < 1, then the fan runtime fraction used in the fan power calculation is going to be equal to my PLR and used as is in the fan power calculation. I ran hourly reports and for all hours where my fan PLR was < 1 but the fan RTF = 1, I divided the fan power calculated by EnergyPlus by the hourly PLR and got the fan power I was expecting to get, which for the previous example (rated fan power 1000W, at a PLR of 0.6 output of the curve 0.4) EnergyPlus is returning a fan power = 1000 * 0.6 * 0.4, whereas if my understanding of the documentation is correct it should return 1000 * 0.4.
To get what I wanted, I created a new curve based on the regression of the output of my original curve divided by the PLR as a function of the PLR. However, I am still curious to know what is going on, if I am just not interpreting the documentation properly or if there is something going on with the code or if something else in my model could be causing this behavior.behavior.
__
EDIT:
@rraustad: Thank you for providing an explanation of the code. Perhaps my question was not well written because I’m not sure you understood what I am trying to do. In short, I would like the fan power to be calculated using a fan power modifier that is function of the fan part load ratio. In DOE-2, this can be done using the FAN-EIR-FPLR keyword which is defined as follows:
FAN-EIR-FPLR is a “curve that gives the ratio of fan electric energy to full-load fan electric energy, as a function of part-load ratio […]”
In my EnergyPlus model I have to use a Fan:OnOff object since I want my fan to be able to cycle. According to the documentation the Fan:OnOff object can be used with variable speed systems. In order to model a variable speed fan I need to provide a “Fan Power Ratio Function of Speed Ratio Curve”. This curve is defined in the documentation as follows:
Fan Power Ratio Function of Speed Ratio Curve “represents the ratio of actual fan power to rated fan power when a change in fan speed occurs.”
To me, this definition is identical to the FAN-EIR-FPLR since according to the affinity laws, fan speed and airflow are proportional.
I would like to model the following fan power modifier curve in EnergyPlus: $\frac{P_{fan}}{P_{rated}} = 1.7647 (PLR) – 0.7647$, with a minimum output of 0.4. I used this curve (exponential version of it) in EnergyPlus and noticed that the calculated fan power was always lower than expected. Here’s an example:
The rated fan power in this example is ~1.309 kW, the PLR on the 5/4 at 3pm is 0.66, the output of the curve should be 0.40 and so the fan power calculated by EnergyPlus should be 1.309 * 0.4 = ~0.53 kW however, as the example shows, the fan power calculated by energy plus is 0.35 kW.
I tried several different curves but the fan power calculated by EnergyPlus was never what I was expecting it to be. I decided to use a constant curve to try to see if I could find a relationship between the fan PLR and the calculated fan power. I noticed that the output of the curve was somehow multiplied by the fan PLR in the fan power calculation. Here’s an example:
The rated fan power is ~1.309 kW, the PLR on the 5/4 at 3pm is 0.68, the output of the curve is 1 so the actual fan power should be the same as the rated fan power. However, it is equal to the rated fan power multiplied by the PLR: 1.309 * 0.68 = 0.89 kW.
Upon noticing that, I created a new curve based on a regression of the output of the original curve divided by the PLR as a function of the PLR. Doing so helped me obtain the expected fan power in Energy Plus. Here’s an example:
- Original curve: $\frac{P_{fan}}{P_{fan_rated}} = 1.7647 (PLR) – 0.7647$
- New curve: $\frac{P_{fan}}{P_{fan_rated}} = 1.1587 (PLR) – 0.1405$
On the 5/4 at 3pm, the fan PLR is 0.67 which gives a rated fan power modifier of 0.4 according to the original curve which results in a fan power of 1.309 * 0.42 =~ 0.55 kW.
Additionally, if we look back at the first example and take the expected fan power (~0.53kW) and multiply it by the PLR (0.66) we get what EnergyPlus calculated, ~ 0.35kW.
I created this post in hopes of learning why I wasn’t getting the results that I was initially expecting. I am not sure how the affinity laws would play a role here. In DOE-2, the user can use any FAN-EIR-FPLR curve no matter how crazy or wrong it is, the power will be calculated according to the output of the curve.
Even if my “work-around” seems to give me the answer that I am expecting, I am still curious to know why I had to go this route or if I am missing something somewhere along the way.
