Discrepancy in values in running both Conventional and VRF systems in 2 different locations.

edit

Thank you, Richard, for sharing that information. I have had the opportunity to read some of your papers for FSEC.

If the above is true, I would assume that the "Design Size Gross Rated Total Capacity" should be different for all locations? In my case I found one location having the same "Design Size Gross Rated Total Capacity" for both VRF and Conventional systems, but when the location changed there was a drastic difference between VRF and Conventional system?

Another difference is the performance curves used for each coil model. The CapFT curve is used during sizing to adjust coil size after the size is selected. For example, if the CapFT result at the cooling coil inlet WB and outdoor temperature used for sizing = 1.1, then the coil size is reduced 10%. I assume you are not using the same coil performance curves and therefore the coil size for a specific location could be different.

That does make sense, But in my case, I have used the same performance curves (FSEC) and still found the total capacity value to be different.

What other system are you using. I may need the inputs files to decipher this further.

I misunderstood your statement about the performance curves. Yes, you are right, the model uses different performance curves for VRF and Conventional systems. However, I am not yet sure why one of the locations the VRF and conventional models gave the same values for "Design Size Gross Rated Total Capacity" whereas the other locations shown a drastic difference in the values. I could share the CBECC-Com project files for the locations. Is there an email to which I can forward it to? Thanks!

I just need the 4 idf files. 2 for one location, 2 for the other. email is in profile. I might not get to this until next week.

Did you not see these errors, in the err file? Something is wrong with your performance curves.

** Warning ** GetDXCoils: Coil:Cooling:DX:VariableRefrigerantFlow="COILCOOLING 11" curve values
**   ~~~   ** ... Cooling Capacity Ratio Modifier Function of Temperature Curve Name = COILCLGVRFCLGQRATIO_FTWBTDBSI output is not equal to 1.0 (+ or - 10%) at rated conditions.
**   ~~~   ** ... Curve output at rated conditions = 15.000

Well I just checked and that performance curve doesn't give 15 at rated conditions. Something else is going on.

Since I was reading the values of CBECC-Com, I did not run into the error file. All of the curves within the model are the default coefficients itself as well.

From a review of the 2 files I see differences between the amount of outdoor air the DX coils see and are therefore sized with. The air loops have mechanical ventilation controllers and use OA specs of 0.000762 m3/m2 of OA per floor area (results in about 20% OA). The VRF system will autosize the OA flow rate to be the maximum of the cooling or heating flow rates (results in 100% OA).So without actually comparing the results I suspect the VRF coils are larger than the conventional coils.

The perf curve coeffs look good but min curve output = 15 and max curve output = 23.89 and min WB (x) = 20. These need fixing.

Curve:Biquadratic,
  CoilClgVRFClgQratio_fTwbTdbSI,  !- Name
  0.058588,  !- Coefficient1
  0.05874,  !- Coefficient2 x
  -0.00021,  !- Coefficient3 x**2
  0.010937,  !- Coefficient4 y
  -0.000122,  !- Coefficient5 y**2
  -0.000525,  !- Coefficient6 x*y
20,  !- Min Value of x
  43.33,  !- Max Value of x
  0,      !- Min Value of y
  100,   !- Max Value of y
15,     !- Min Curve Output
23.89;  !- Max Curve Output

Make sure you check all the curves since I only highlighted problems with 1 curve.

Thanks Richard! I will look into them. I believe this resolves my issue. Thanks a lot!

Responding to this comment: The perf curve coeffs look good but min curve output = 15 and max curve output = 23.89 and min WB (x) = 20. These need fixing.

Having looked into that, it seems those values are present within the idf files generated by CBECC-Com. So should the CBECC-Com not update them automatically?

Responding to comment:
From a review of the 2 files I see differences between the amount of outdoor air the DX coils see and are therefore sized with. The air loops have mechanical ventilation controllers and use OA specs of 0.000762 m3/m2 of OA per floor area (results in about 20% OA)....

I have manually changed the OA specs for VRF to m3/m2, did not cause any change in the cooling/ heating/ fan loads either. Is that of concern?

Funny coincidence, I found this very same issue in the CBECC-Com curves two weeks ago. I've created a ticket for the development team to address this in our next release. Not sure how this error did not get noticed sooner. after I fixed it locally, I did not see a dramatic change in energy use. @bmjohn, after fixing the min/max curve outputs, did you see a big change in energy use?

1) The change in OA quantity for one file should have changed the coil size. If this change did not make coil sizes converge then there is still something causing coil size differences. 2) the CAPFT performance curve is used in sizing and the coil size and CAPFT curve work together during simulation. One expected change is coil size, another is a possible change in energy use depending on how the simulation rides the curve.

Also note that zone equipment with a local OA mixer that does not have economizer controls will usually size coils differently than air loop equipment with OA systems that use economizers since the OA quantity used during sizing may be different. Setting OA quantities to 0 in both files highlights differences caused by OA.

@DavidReddy, We manually deleted the max and min curve outputs for one of the curves and there was a slight change. I would not say it is drastic. I have not tried deleting the max and min curve outputs for the remaining curves to see if the change in energy usage would would be cumulative.

@rraustad, Responding:The perf curve coeffs look good...These need fixing. Is it a mandate to input the curve limits, as this auto-populated by CBECC-Com? On removing it, avoids warnings related to them. Would you call that a wrong approach when running idf files?