I'm not drawing conclusions from this based on incomplete data, but thought some might find it interesting. http://jmpco.squarespace.com/home/2013/2/22/ashrae-living-laboratory-showcases-geothermal-efficiency-at.html?goback=%2Egde_1203517_member_217400468 (Edit: changed title from "mini split" to variable air source split system)

Is a 38 ton multisplit Variable Refrigerant Flow fan coil unit on a commercial building now being considered a "mini split", or did I go to the wrong place?

I was familiar with this project since my Daiken training. It is their stated intention to demonstrate that splits are competitive with geo on this project.
Folks routinely try to assign "mini-split values" to multi head systems around here. Interesting that's all you took from that report. I will see if I can change the thread title.

FWIW, the link you referenced said the climatemaster geo system outperformed the VRF system most noticeably during the cooling season. From climatemaster's writeup it seems that's backwards:

"Based on data collection from 2010 through 2013,
the ClimateMaster geothermal heat pump system
performed notably more efficiently than the VRF system
during the heating season, while cooling performance
was ultimately comparable. Data specifically exemplified
more than a 50 percent energy savings from the
geothermal heat pump system when comparing the
average heating/cooling performance of both systems
since their installation. " http://jmpco.squarespace.com/storage/LC999_-_ASHRAE_HQ.pdf

I'm not really a student of ATL weather but my first impression would be that it leaned toward cooling dominated. Which would mean most money savings would be cooling
I would be curious to see the zone loads. I read the Climatemaster report as well and found lots of % savings claims but no hard numbers in support.

Again it was Daiken's stated intention to use this project to show that air source was competitive with water source on op cost side. Of course they have introduced a whole new line of water source products since then. Dunno if they are just jumping on the band wagon as the water source market has exploded or if they are back-paddling on a previous assertion.
I think most of us agree that it is installation cost of GSHP's that makes us constantly re-visit the air-source question. We generally agree that ground source costs less to operate in most cases, but op cost savings does not always justify extra initial expense.

Wouldn't it be nice if we were told how much each system cost to install? Kinda hard to measure cost v benefit with out that detail.

It is a shame we don't have an air source heat pump forum somewhere with somebody with "Doc"-like diligence, running monitors on their equipment to substantiate efficiency claims.

another report someone pointed us to is http://acemechanical.com/vrfvsgeo.pdf

If I look at the data, and take the modeling for lets say Chicago, and it read about 8200 KWH for heating and cooling, but also also spend almost as much, 7200 KWH annually, for the fan and the circulation pumps, then I know I built a very inefficient system. I don't know why they so proudly present that kind of data. I'd be ashamed making a geo system that inefficient. Weird modeling.
However, the building appears to be strikingly cooling dominated and not well balanced. The superior performance in the heating season steams partly from the high EWT during the heating season, and the loopfield will continue to heat up more and more with the years to come. We will see how it performs in 20 years from now, although they put a nice amount of spacing between the boreholes.

I'm not suprised that a busy office building in Atlanta is heavily cooling dominated.

Absent from both published links seems to be info on design load. Data about kwh / SF may seem enlightening but absent heating and cooling load information kWh / SF is info in a vacuum. If the load per square foot is higher, then kwh / SF would be expected to be higher.

It is a shame we don't have an air source heat pump forum somewhere with somebody with "Doc"-like diligence, running monitors on their equipment to substantiate efficiency claims.

It's hard to imagine that the lack of "having a forum" here at GBT would stop someone who was interested in that.  "General Forum - Residential" would be suitable to post those results.

Posted By engineer on 03 Mar 2013 10:53 PM
I'm not suprised that a busy office building in Atlanta is heavily cooling dominated.

Absent from both published links seems to be info on design load. Data about kwh / SF may seem enlightening but absent heating and cooling load information kWh / SF is info in a vacuum. If the load per square foot is higher, then kwh / SF would be expected to be higher.

I guess my point was that the modeling assumes pumping power and fan power to be almost as as high as the compressor power usage. That is how you make a system inefficient, irrelevant of the loads. But you are right, without the loads you only know half the story about the inefficiency.

Loads for GSHP system
– Heating: 308.2 kBtu/hr
– Cooling: 288.6 kBtu/hr
• All zones on floor 2 and a corridor zone on floor 1
• Loads for VRF system
– Heating: 326.9 kBtu/hr
– Cooling: 377.6 kBtu/hr
• All zones on floor 1 (minus corridor zone) and the learning
center
This was buried in the second link I offered

I actually saw that later

I wonder how the results pan out adjusted for load

Posted By engineer on 04 Mar 2013 03:32 PM
I actually saw that later

I wonder how the results pan out adjusted for load

Yeah I wonder about lots things. There is even one rumor that ventilation load is carried by Air Source zone.....I sent an email to ASHRAE to ask. Not a very well supported report.

Ahh...what else is rotten in Denmark?

Dunno. As I said heard the rumor from someone who claimed it was third hand. I don't expect a reply to my queery.
What I do find compelling is the fact the Daiken who set out to demonstrate parity has no such report.

Posted By joe.ami on 06 Mar 2013 08:19 AM
Dunno. As I said heard the rumor from someone who claimed it was third hand. I don't expect a reply to my queery.
What I do find compelling is the fact the Daiken who set out to demonstrate parity has no such report.

But in the Climate Master report they state there on the bottom of page 2:

"Based on data collection from 2010 through 2013, was ultimately comparable."

And the borehole field is still young... 

Parity in cooling mode seems already apparent, but maybe not in heating mode. We won't know for sure where it really lives until it has more than 3 years of temperature stabilization behind it.   For sure the winter & summer air should continue to be pretty much the same in decade's time, climate change trends notwithstanding. Time will tell how that local water/dirt/clay/granite is going to fare but for the time being it looks like "advantage GSHP".

The relevance of large scale full-custom systems to "system in a can" mini-splits is tenuous at best, despite some similar features, with many ways for the designer to fail in the former, and where 99% of the design is "already in there" for the latter.  It could be that there were design errors and compromises taken by the engineers on this system that would be different in another system/configuration, and they may have blown something fundamental. I'm sure this is a subject of intense discussion within the design group at Daikin if it's really that far off the mark in heating mode.

The GSHP contractors took direction directly from the engineering staff at ClimateMaster

"We installed the loop field according to design Phillips, field engineer at Geo-Energy Solutions. "

Which was clearly the right thing to do, but again, in no way comparable to pre-packaged residential-scale mini-splits.

It would be of little surprise if with all factors taken into account the GSHP system proves to have the performance edge in this application and it would surprise me if the VRF system from Daikin actually blew the ClimateMaster away on raw performance.  (That's assuming it actually was engineered by Daikin, whose name does not appear in any of the web links but does appear on a wall plaque in the video  if you keep your eyes peeled- the first cameo comes ~55 seconds into it.)   Neither system is likely to be cheap here, but the allegation that it required drilling through at least some "...good old Georgia granite..." makes me think the ClimateMaster system may have been substantially more expensive up front, but maybe not. We're talking a order of magnitude larger size and complexity than any typical residential application and it's hard to tell how it all scales.

Unlike a 1800' cape style house with sub-25KBTU/hr heating & cooling loads, it's well worth spending some serious engineering effort to get the last bit of efficiency out of the mechanical systems.  Also unlike single family residential buildings, simply reducing the loads with building envelope upgrades to something tiny is not usually an option in commercial buildings of that scale, but there's no indication that any of those factors were being cost-optimized (LEED certification notwithstanding.)

What I do find compelling is the fact the Daiken who set out to demonstrate parity has no such report.

I'm trying to find out about that and I am unable to confirm anything or anyone who claims that was Daikin's aim. Did you have a source you can share with us? I've been in communication with the Daikin engineers over the last couple years and I would be very surprised if they intended those two systems to be "at parity" and subsequently, fell short by that margin.

But note, it in fact DOES meet parity on the cooling side of the equation (either by design or dumb luck. :-) ) But that would indicate it's doing downright horrible on the heating side of the equation- way worse than single-head mini-splits would in an Atlanta climate.

Atlanta's climate is still heating dominated at a 65F residential type balance point, but higher-performance (they're going for LEED, after all) commercial space will often balance somewhere between 50-55F. The mean annual outdoor temp in Atlanta is about 65F, so for commercial buildings (even non-LEED code-min buildings) will have more annual cooling hours than heating hours, even with pretty-good heat rejecting windows and a high-albedo roof.

http://weatherspark.com/#!dashboard;a=USA/GA/Atlanta

The more detailed discussion of the systems here start on page 38 of this document.  Both the heating and cooling loads for the Daikin VRV system are somewhat higher than for the geo system but the heating loads are within 10-15% of each other. (The cooling load on the VRV system is 30% greater than on the geo, but the June-August cooling power use was at most 20% higher.) But a quick look at the January power use number for the VRV relative to the geo system looks insane.  No matter WHAT you assume for a reasonable COP for the geo system, if the loads are comparable it means the heating performance of the VRV system is half that, maybe less.  This would never be the case for a single-head mini-split in an Atlanta climate. Looking at the January 2010 weather data from the weatherspark site, it was colder than the average January, with a mean temp of about 38-39F rather than the 45-46F typ, but nothing too radica. A single head minisplit would be running north of 3.0 for a COP in that 38-39F kind of weather, and close to 4 at the typical ~45F if reasonably sized for the load.  If one assumes that the GSHP system's COP was running in the neighborhood of  4.0 (probably not more than 4.5, or less than 3.5) would appear that the gazillion-0head VRV system is running maybe ~1.8.

That is a COP that a Mitsubishi -FExxNA would achieve even at -10F outdoor temps at max compressor & blower speeds(!), according to 3rd party bench testing (see Figure 10 p14, p22 in .pdf pagination.)  Comparable Fujitsu units have a similar (somewhat better) profile, down to it's -4F min-rated operating point (see Figure 5.)

The reasons why the VRV is performing this poorly in heating mode might be identifiable in the system design, or there may be implementation/control errors, but there's something definitely wrong here. Ducted 2-stage air source heat pumps can usually hit that performance mark in heating mode at those temperatures.  The fact that they have no fewer than 5 (!) compressor units being cycled in/out rather than a single continuously-variable speed compressor on a mini-split surely complicates the control picture to the Nth degree(!), but I'm a bit surprised that it's heating mode efficiency fails so miserably in comparison to it's cooling mode performance.

"I'm trying to find out about that and I am unable to confirm anything or anyone who claims that was Daikin's aim. Did you have a source you can share with us? I've been in communication with the Daikin engineers over the last couple years and I would be very surprised if they intended those two systems to be "at parity" and subsequently, fell short by that margin."

Well you know at least 1 person who claims it was Daiken's aim (me). My source was my Daiken mini-split trainer in ~08. I would not expect them to restate the position if they didn't like the results.

There are some fuzzy numbers and lop sided presentation in the CM presentation, and in fairness I shared another rumor I heard (which mitigates CM claims). Whether you want to believe an employee of Daiken stated the goal in my presence or not means little to me. I'm simply sharing what I heard.

Yeah, that's understandable; you can't do anything about that. You can be on a strategic mission with handpicked team members and one of your own guys, within arms reach, can get off message. And, this was back in -08, right? Time can blur the way we remember things....