HPWH COP Adjustment in confined space

[yzhou601]

Pull Request Description

Fixes #2009

Checklist

Not all may apply:

• Schematron validator (EPvalidator.sch) has been updated
• Sample files have been added/updated (openstudio tasks.rb update_hpxmls)
• Tests have been added/updated (e.g., HPXMLtoOpenStudio/tests/test*.rb and/or workflow/tests/test*.rb)
• Documentation has been updated
• Changelog has been updated
• openstudio tasks.rb update_measures has been run
• No unexpected changes to simulation results of sample files

[yzhou601]

@shorowit @jmaguire1
This PR is ready for review, I looked at the results and discussed with Jeff about the decreased heating and increased cooling after reducing COP here. The reason is that lower COP means less heat being extracted from air, while more heat being added by consuming more electricity, so it reduces the amount of cooling being added to the zone, which impacts the HVAC heating/cooling loads. We also verified that the base-dhw-tank-heat-pump-ef.xml shows the same trend on the heating/cooling impact (this test file has a different tank volume so we manually changed it to be the same as base-dhw-tank-heat-pump.xml, and compared them).

[jmaguire1]

See individual comments: a few name changes and some additional handling of edge cases suggested.

[jmaguire1]

@shorowit: Reviewing this reminded me of a long wish listed improvement for HPWH interaction factor, because if it's in a confined space and this penalty applies we know there's a lot of localized cooling and the interaction factor should be < 1. Last time I talked to Ben Larson about what HPWHSim does he thought is a reasonable confined guess if you have 0 information about the installation location, but HPWHSim still uses 1.0 everywhere.

Maybe we should just start an issue if this is too much for this PR or a follow up PR, but I double checked our implementation of it in the past and it's really simple: we just multiply these two lines in our HPWH internal gains EMS by the interaction factor: https://github.com/NREL/OpenStudio-HPXML/blob/5141bcd2f2a2c607b72dde834e2aaf79cc617d54/HPXMLtoOpenStudio/resources/waterheater.rb#L1290

Here's what the PNNL lab home testing had:

If I were to propose something, we could default to 0.75 (or 1, to preserve our current behavior) and use 0.5 if you're in a confined space based on what PNNL saw in their lab home. It depends on a lot of things we still don't have information on, but we do at least expect if you're in a confined space and this correction applies it's not going to be 1.0 (or the cooling wouldn't be localized enough that we need this adjustment).

[yzhou601]

@shorowit: Reviewing this reminded me of a long wish listed improvement for HPWH interaction factor, because if it's in a confined space and this penalty applies we know there's a lot of localized cooling and the interaction factor should be < 1. Last time I talked to Ben Larson about what HPWHSim does he thought is a reasonable confined guess if you have 0 information about the installation location, but HPWHSim still uses 1.0 everywhere.

Maybe we should just start an issue if this is too much for this PR or a follow up PR, but I double checked our implementation of it in the past and it's really simple: we just multiply these two lines in our HPWH internal gains EMS by the interaction factor:

https://github.com/NREL/OpenStudio-HPXML/blob/5141bcd2f2a2c607b72dde834e2aaf79cc617d54/HPXMLtoOpenStudio/resources/waterheater.rb#L1290

Here's what the PNNL lab home testing had:

If I were to propose something, we could default to 0.75 (or 1, to preserve our current behavior) and use 0.5 if you're in a confined space based on what PNNL saw in their lab home. It depends on a lot of things we still don't have information on, but we do at least expect if you're in a confined space and this correction applies it's not going to be 1.0 (or the cooling wouldn't be localized enough that we need this adjustment).

I've addressed other comments but this, I felt that we probably need more discussions around this (e.g. what volume to consider what factor to use, definition of confined space?). The equation to calculate relative volume uses 1500cbft as base volume but the standard specifies 1000cbft enclosed space as one of the mitigation, which volume is considered as the definition of confined space? Should we consider implementing it in this PR or with a follow-up issue?

[shorowit]

@jmaguire1 Feel free to make an issue out of that, but I don't want to hold up this work for it. (I am doubtful that it would ever make it into ANSI 301, there was a lot of objections to it. The interaction factor can be wide range of values and is incredibly sensitive to distance between water heater and thermostat, which is an input we don't ask for. If the distance is really small -- e.g., HPWH co-located with HVAC system -- the interaction factor can even be more than 1! I struggle to see how users would have any idea what value to use.)

[jmaguire1]

I'll make a follow up issue to track interaction factor, ResStock folks have asked for it before so I don't want to lose track of how we might implement it (but it'd have to be something I think they pay for if this audience won't use it). What value to use is always a hard question where the ResStock team might want to take rough guess. But lets leave that as future and get this in for now.

[jmaguire1]

Thanks for making these changes @yzhou601! Latest documentation looked good to me, and we'll do a follow up issue for HPWH interaction factor.

[shorowit]

Great job, @yzhou601. Nice implementation, excellent tests and documentation. I just pushed a few minor changes to improve the docs. Other than that, I just had a couple other suggestions below.

[shorowit]

I did some runs without (first set of columns) and with (second set of columns) the additional adjustment to the heat extraction:

RESNET_HPWH.xlsx

After discussing w/ @jmaguire1, we are going to revert the additional adjustment. He will provide additional context.

[jmaguire1]

What we did here is adjusting based on the kWh consumed, rather than the heat added to the HPWH. So really each of the points where there's element usage in this chart should be using the power consumed by the HP multiplied by the average COP. If we call that a COP of 2-2.5:

That would give us a multiplier of something like 0.875 at 250 ft^3, so it's not a huge adjustment. The interaction factor (see #2090) is probably a bigger impact than this, and we're not trying to estimate that due to lack of data and it not being a high priority for the RESNET working group on this. So I think we can revert this rather than trying to propose it in the standard, or later have to explain our slightly different results. I want to keep track of this in case we ever want to try to do something with this if/when we do look at interaction factor, but that's not high priority for anyone right now so lets just revert this and keep this in mind for the future if we do a deeper dive on space interactions for HPWHs.

[shorowit]

Looks great to me. Thanks, @yzhou601! Will merge assuming the CI is still happy.