Getting some very unusual results

[phaniarvind]

There are three issues I'm facing which I can't find a solution for on my own. Attached is a zip folder that contains a standard simulation script of one building located in Denver. I built the HPXML file from BEopt as suggested in your documentation.

1. Temperature does not track the setpoint:
I am running the simulation in timesteps of 10 minutes which means that use_ideal_capacity is set to True. But the indoor temperature is tracking a setpoint temperature that is slightly above the given setpoint. The figure below shows the heating/cooling setpoint and the indoor temperature and there is a clear offset. Is there a reason for this?

2. Cooling in winter is non-zero
The HVAC Cooling Electric Power for the building in winter during the first ten days of the year is non-zero. It remains close to a steady value of 50W. Is there a reason for this? Some base value used to make sure calculations don't fail? Also will this make a difference in the bigger picture when it comes to say annual total electric energy consumption.

3. Cooling energy comparison with OS-HPXML simulation is way off
This one is pretty self-explanatory. Looking at the graph there is a huge difference between what OS-HPXML predicts and what OCHRE predicts for the same HPXML file. Is there a calculation error somewhere on my end or some other issue?

OCHRE_BaselineIAT_Test.zip

Thank you so much for your patience and help!

[jmaguire1]

Number 2 is a crankcase heater https://en.wikipedia.org/wiki/Crankcase_heater. Any time the outdoor temp is below 55 F, the heater will be on at a small amount of power to prevent damage to the AC, and it gets reported as cooling energy use. See

OCHRE/ochre/Equipment/HVAC.py

Line 944 in 4946b56

self.electric_kw += self.crankcase_kw * self.space_fraction

[jmaguire1]

For number 3: That's certainly an unexpectedly large difference. Some things I'd start with for debugging:

• Make sure the setpoints are the same. If for some reason the cooling setpoint is different, that'd easily do it.
• This is SO different I'd just confirm it's the same weather file.
• Try to isolate if it's a difference in the loads vs. the AC performance. Given how large a difference it is, that's almost certainly the case. Looking at output variables for the loads, instead of energy consumption, will let you be sure.
• One big known difference is with how ducts are modeled. You could try making a version of this without ducts, to see how much of the difference is just down to how the duct models work.

[jmaguire1]

For 1: It'd be helpful if you can also plot the HVAC energy consumption along with temperature, and if you could say what the HVAC equipment/efficiency is here. Something does indeed look suspicious here, but I want to see how the HVAC is modulating in this case to know for sure what's going on.

[phaniarvind]

Thanks for solving 2! That makes sense.

I suspect #1 and #3 both are related to each other. I have ensured that the setpoints are the same and so is the weather file as well as system capacities and efficiency. In fact these were my first ideas as well. Comparing loads is a great idea but I am not sure if it's feasible to do so correctly. Here is the plot with indoor temperature and the HVAC electric energy consumption.

Looking at this surely there is a load difference in between OCHRE and OSHPXML simulations For instance the indoor temp drops on its own accord significantly as per the OCHRE simulation in the cooling case. Now to see why I want to calculate the load differences which spawns the following questions:

(a) In OCHRE, to calculate the total heat gained by the space I calculated heat gain due to temperature difference between the surface and zone temperature using the corresponding film coefficient for each surface. This includes the convective and radiative components both right? The value obtained here is what OSHPXML defined as "Component Loads". I assume the LWR part is included already in this in the film coefficient.

(b) OSHPXL report "Component Loads" but I assume what they really mean is heat gained due to each source (per the doc). I plotted the net sensible + latent heat gain with cooling load delivered from OSHPXML and it looks like this:

Assuming the comparison is legitimate, I am not sure why OCHRE is predicting a negative load in summer. I assume this is why the temp is dropping below setpoint on its own without the AC.

(c) Finally, to check the value of heat gain in OCHRE, I plotted the individual sum of all the gains I calculated using method in (a) and plotted it against the sum of Net Sensible and Latent Heat Gains. The two are not close, so it seems my interpretation of these quantities is wrong. Could you help clarify?