Garn WHS3200 Performance Experience - Part 1 & 2 merged

I had the opportunity in March 2010 to measure performance of a Garn WHS3200 under conditions which tested its full capacity. The Garn was installed in the summer of 2009 to be available for the 2009-10 heating season. The Garn replaced existing LP boilers, which were left in place to provide backup heat. The facility being heated is 52,000 square feet and is located in north-central Minnesota. The facility also installed a Wood Gun E500. I could not make similar performance measurements on the Wood Gun due to unusually warm weather in March and after which did not permit continuous operation of the Wood Gun without extensive idle cycles. Had weather been colder, I also could have made measurement which tested the full capacity of the Wood Gun. Those measurements will have to wait for next heating season.

The facility heat emitters mostly are fan-coil units. These units are designed to operate with hot water supply in the 160-180F range. During “warmer” winter periods, supply as low as 140F can reasonably meet system demand.

Garn does not provide a BTUH output rating. Garn advertises the WHS3200 to have a Burn Rate of 950,000 BTU/H. Literature provided by Garn states that “... approximately 97% of the wood is converted to energy and 87% of that energy is stored for use.” This might be interpreted to mean that the Garn should be able to provide approximately 800,000 BTUH output, although experience reported by other Garn users in their posts suggests that the advertised “burn rate” is a good measure of expected output and in fact may be conservative.

SUMMARY. Based on my measurements, under full output demand conditions, the Garn was able to provide about 375,000 BTUH continuous average output, or somewhat less than 40% of the “burn rate.” In follow-up discussions with Garn, Garn has advised that the most efficient discharge temperature of the WHS 3200 is about 130F, and that efficiency would decline substantially at discharge temperatures above 140F.

DATA. Data was collected from 8:44 pm on March 19 to 10:50 am on March 20. 2010. Dallas DS18B20 sensors were fixed to the piping under insulation on the inlet and outlets of both Sides A and B of the plate heat exchanger interfacing the Garn to the system. The heat exchanger was sized based on recommendations from Garn. Temperatures were logged at the rate of 46 readings per minute, and then reduced to one reading every five minutes. The results are shown on the two charts.

GARN SIDE OF THE HEAT EXCHANGER. The upper [blue] line is the temperature of the hot water from the Garn entering the heat exchanger. The lower [green] line is the temperature of the water leaving the heat exchanger and returning to the Garn. The upward spikes in the lower [green] line indicate periods when the system was not demanding heat. The areas of wide spread between the upper and lower lines indicate times when heat is being sent to the system.

The chart evidences when wood fuel was added to the Garn. This is shown by the large upward slopes in the upper [blue] line (rising temperature).

The chart shows by way of the lower [green] line spikes that at outside temperatures between 28F and about 21F, and Garn hot water supply temperatures between about 153F and 165F, the Garn met the heat demand from the system. No lower [green] line spikes appear after outside temperatures dropped below 21F, and after Garn hot water supply temperature dropped below 150 to 155F. This means that the system continued to demand heat which the Garn was unable to satisfy.

When the operator arrived at the facility around 6:00 am on March 20, he noted that system hot water supply temperatures were low, and he added wood to the Garn in an attempt to raise the hot water temperature to meet system heat demand. Following a brief and small rise in Garn hot water supply temperature, the Garn hot water supply temperature continued to drop. The system was pulling more heat from the Garn than the Garn was supplying from the burning wood.

To remedy the insufficient hot water heat supply, the operator then fired the Wood Gun at 9:00 a.m. System hot water temperatures immediately started to rise, rising to the point that by about 10:00 a.m. hot water supplied by the Wood Gun to the system was hotter than the hot water being supplied by the Garn, as evidenced by the lower [green] line rising above the upper [blue] line. As system hot water temperatures had risen and outside temperatures also were rising, the Wood Gun was allowed to burn out its load of wood, and the Garn again took over supplying heat to the system.

Except for about a one hour period when the Garn supplied hot water just under 140F, the facility operated the Garn above the efficient discharge temperature advised by Garn.

The period from 4:00 to 9:00 am provides information on Garn BTU/H input to the heat exchanger. At a calculated flow of 75 gpm on Side A of the HX, BTU/H = 500 x 75 x ▵T (temperature change). ▵T is close to 10F or a little better average during this period. Therefore, output by the Garn into the HX is about 375,000 BTU/H. This must be regarded as somewhat approximate, as the flow rate is derived from calculated pump head and the pump curves, and the ▵T is an observation from the graph.

SYSTEM SIDE OF THE HEAT EXCHANGER. The upper [red] line is the temperature of the hot water from the heat exchanger going to the system. The lower [pink] line is the temperature of the cooler return system water entering the heat exchanger. The spikes and slopes closely mimic the Garn side chart and correspond to similar events.

The period from 4:00 to 9:00 am also provides information on Garn BTUH output to the system via the heat exchanger. At a calculated flow of 75 gpm on Side B of the HX, and using the same formula, with an average ▵T of about 9F, BTUH output through the HX to the system is about 337,500, or about 90% of input, which would be about the expected number based on heat exchanger efficiency. For this system, however, this needs to be adjusted downward to account for use of 30% glycol on the system side. The adjustment factor for glycol at 30% concentration is about 8%. This reduces effective output to about 310,500 BTUH. The reduction in effective output by reason of use of a heat exchanger and glycol is not attributable to Garn performance. The output calculation also must be regarded as somewhat approximate, as the flow rate is derived from calculated pump head and the pump curves, and the ▵T is an observation from the graph.

WOOD GUN E500. By way of anecdotal operator reports, the Wood Gun E500, which is rated at 500,000 BTUH, was able to meet full system heat demand at outside temperatures below +20F and down to some point below 0F. The operator learned by experience that at about +20F the Garn no longer could meet system heat demand. Therefore, the Wood Gun was fired, and by experience the operator learned that the Wood Gun had repeated idle cycles while the Garn still was being fired. Therefore, the operator would allow the Garn to burn out, and the operator observed that the Wood Gun, operating alone, could meet the entire system heat demand down to some point below 0F. When the Wood Gun alone no longer could meet system heat demand, the operator again would fire the Garn and operate both boilers together to meet system demand.

All good questions, some of which I cannot answer. The owner of the building made the decision to install the Garn and the Wood Gun with the goal of substantially reducing LP usage. As you suggest, the heat loss obviously is greater than the Garn alone can meet, depending on outside temperature. It is interesting that the Garn alone could meet the heat demand within the range of outside temperature indicated (above 21F). But as outside temperature dropped below 21F, the Garn no longer could meet the demand. In other words, as temperature dropped below 21F, even with fueling the Garn (at about the 395 and 550 minute marks), 100%+ of Garn BTUH output was being used, and the Garn supply temperature to the HX was falling.

It also is interesting that the Wood Gun was able to meet 100% of demand at outside temperatures below that at which the Garn could meet demand, even though it is rated at only 500,000 BTUH.

I think it is at the point where the Garn no longer can keep up with system demand that calculation of actual Garn usable output can be measured, as 100% of available output then is going to the system. Thus, the BTUH calculation that I indicated appears to be the approximate maximum available continuous output

As to the HX, Garn provided the specs to be used with this system, and the heating contractor slightly over-sized the HX. Garn specified a GEA 10 x 20 x 80 plate, and the contractor installed a GEA 10 x 20 x 90 plate HX. It is a single HX and is installed directly between the Garn output and the Garn return with about as short of piping as possible, with both the piping and HX fully insulated. GEA spec's for the 10 x 20 by 80 plate are: Side A entering temp of 190F, leaving temp of 157.5F, Side B entering temp of 147.6 and leaving temp of 180F, both sides at 60 gpm, pressure drop both sides of 2.5 psi. As my post indicates, actual calculated flow rates are 75 gpm both sides.

The Garn and the Wood Gun both feed a primary loop, and the two structures making up the facility draw from the primary loop as each structure demands. The Garn and Wood Gun are side-by-side, the primary loop is at the rear of both boilers, and all are in very close proximity. The heated buildings are varying distances away.

As to the calculated flow rates, I am quite comfortable with the accuracy of the calculations. The heating contractor provided the initial information, and I then verified that by measuring all the piping and fittings, determining equivalent pipe length, used freecalc.com to calculate head at the stated flow rate, used the GEA software to calculate pressure drop at the stated flow rate, and then used the pump curves at the calculated head to verify the gpm flow. It would be helpful to have an actual flow measurement, but at this time I have no way to do that. I doubt that flows are much different than I calculated, and therefore that BTUH output is not likely to be much different than I calculated.

Wood all was hardwood, mostly oak, estimated MC 25-30%. The same wood that was used in the Garn also was used in the Wood Gun. The Garn was loaded, according to the building owner, in accordance with the Garn manual and information provided by Garn.

I did not monitor or log Garn tank temperature, as I only have a 4 ch data logger, and all channels were used on both sides of the HX. Also, I don't know where a suitable location on the Garn tank might be to fix a Dallas sensor. I did visually observe the analog thermometer on the front of the Garn from time to time, and I did visually observe an analog thermometer provided by the heating contractor on the HWS line to the heat exchanger. Both typically read between 0 and +5F over the output temperature picked up by the Dallas sensor on the HWS to the HX. This doesn't surprise me, as I am used to finding slight variation in reading between different thermometers.

I hope I addressed all of your questions, and if not, ask again. Thanks for the interest.

...and [would] require a lot more btu’s than the numbers quoted.

Click to expand...

In the abstract, a person might be inclined to agree with you. But with the data, not just from the sensors but also verified by the analog thermometers as to the output temperature, I can't find a rational basis to conclude that boiler output is much different than calculated. Even if it could be assumed that flow was greater than the calculated 75 gpm's, the pump curve chart on the Garn side of the HX shows maximum possible flow of about 115 gpm at 0 feet of head, which would equate to 575,000 BTUH. Obviously, pump head is not 0 feet.

I am very open to being corrected, as I am trying to be as objective as possible and base conclusions only on hard data. With that in mind, here is my calculation of pump head and flow on the Garn (supply/return) side of the HX. Please feel free to verify that I have not made any material errors.
Pipe measured at 29' of 2.5" pipe
Nine 90 degree L's at 7' equivalent pipe length per L = 63' total equivalent pipe length.
Total equivalent pipe length = 92'
92' of 2.5" pipe at 75 gpm's = 3.5' of head.
10 x 20 x 90 plate HX pressure drop at 75 gpm's = 3.4 psi = 5.8' of head
CORRECTION: should be 7.9' of head for the HX
Total head at 75 gpm's = 11.4'
UPS 40-160/2 pump curve at 11.4' head = 75-80 gpm at Medium speed.

If you can verify the flow gpm's, then the only other source of error would be the temperature. Input temperature to the HX is secondarily verified by the readings from the analog thermometers. I have no secondary verification of the temperature on the return side of the HX, and therefore no secondary verification of the temperature drop across the HX. I fixed the DS18B20 sensors to the input and output HX pipes in the same manner: cut back the insulation on the pipes, cable tied the sensors firmly to the pipes, and then replaced the insulation over the sensors. Also, I have used these same sensors and data logger on another system that had analog thermometers to verify that the sensors were "accurate," and in that case the sensors were attached to the pipes in the same manner, and the sensors and data logger returned temperatures within 1-2 degrees of the analog temperature readings.

Where else do you find possible errors?

Keep in mind that the data only determines actual usable BTUH provided by the boiler to the system as determined by gpm flow and measured supply/return temperatures. Also keep in mind that Garn advised that efficiency would decline substantially at discharge temperatures above 140F, and that at almost all times the Garn discharge temperature was above 140F (briefly between 135-140F, and as high as about 166F). What exactly Garn means by this advice, I do not know.

There is a second part to this post, which I should have included as a reply, but instead did a separate post. If you haven't seen, here the the link to ......

Moderator note - I used my magic mod buttons, to merge the two threads. The way the software merged them (which I assume is based on comparing time / date stamps) the second part mentioned is now the second post in this thread, so you've already read it...

Gooserider (as Mod)

what output are the original Boilers, when was the building built

Click to expand...

I can find out this info. One thing I know is that considerably more btuh capacity with LP boilers was provided that was actually used, and the LP boilers were staged on a rotation basis.

...building would have to be cutting edge efficient (passive solar?)

Click to expand...

I think I am safe in stating not cutting edge efficiency, although there is some passive solar. Another factor is that the buildings are fairly heavy occupied, on a somewhat variable basis, and there likely is a good quantity of human-provided btu's which may have been present on the test day.

Two things come to mind, one knowing the actual gpm for exact btu calc., and two it almost sounds like the garn firebox was not filled frequently enough to sustain a high burn. I would press garn for a more substantial explanation and advice.

Click to expand...

Actual gpm may be elusive, but again, I would be surprised if it is much greater than calculated. I would have desired the opportunity to manage the Garn for a couple of cold winter days to really determine the capacity, based on my prior experience, not with the Garn, but with my Tarm. I hope to have this opportunity, and also with the Wood Gun, this coming winter. I reserve for the future any additional comments regarding communications with Garn.

What happens when the wg burns down or are you feeding it every 30 minutes. I think that boiler would get a lot of thermal shock with that much draw. or is that what the garn is for, a buffer?

Click to expand...

As my post mentioned, operating experience resulted in not firing the WG when outside temperatures were above about 20F, as the Garn met the demand and the WG would idle extensively. Between about 0-20F only the WG would be fired, and below 0 both would be fired. The WG is provided with return water protection at 160F, so no thermal shock situation.

tappings on the inlet and outlet ... connect some gauges to see what the pressure differential and therefore head/gpm

Click to expand...

I don't remember but will check for tappings. Gauges would provide pressure differential and therefore head, and then the pump curves would provide gpm info. More work may be done on the system this summer, and this may be pursued, if possible, as part of that work. I still doubt that head would fall much below the calculated head to increase flow materially and raise the BTUH calculation materially, and I agree that better and more precise data would be helpful.

I have a hunch there are flow rate and HX sizing issues here because things aren’t adding up. It’s really impossible to say without a valid heat loss of the structure plus an accurate evaluation of the output capability of the connected emitters at a given fluid temp. The info you have posted really doesn’t shed a lot of light on the problem because without knowing flow rates and what’s actually going on in the system load wise the numbers are just numbers.

Click to expand...

I have a hard time getting my head around your hunch. To me it seems that we have a near ideal situation to determine real world performance. Isn't the temperature drop across Side A of the HX and Side A gpm an accurate representation of the load BTU's being drawn off the Garn? If the Garn also is being fired according to Garn recommendations, then don't we also have the measure of actual capacity to deliver BTU's to the system when being operated as Garn recommends? In effect, based on outside temperature and the size of this structure, don't we also have a variable load that, at "mild" temperatures has a demand within the capacity of the Garn, and at "cold" temperatures a load that exceeds that capacity, thus providing a variable heat sink to determine the maximum deliverable BTU's which the Garn is able to provide when being operated as Garn recommends?

For me the most questionable variables are how the Garn actually was fired and was it fired as Garn recommends. The owner advises that it followed Garn recommendation, and at this point I assume this is true. This needs to be verified, and that will have to wait until the coming winter and verification that Garn recommendations were followed.

As to the existence of an "engineered" project, what I know is that the system with LP boilers was pressurized, Garn and its advocates did a good job at selling the 3200 to meet the needs of the system and use of a heat exchanger to interface with the pressurized system, although Garn did want to sell two 3200's, not just one. Garn did spec the HX for the system. At the same time, WG and its advocates did a good job at selling the E500 and direct pressurized connection with the system. Also, the owner is an educational institution, and another contributing factor was the opportunity of being able to operate both types of boilers, side-by-side, and provide information on what was learned.

I am concerned, however, that after installation and operating experience, Garn advised that the most efficient operating temperature of the Garn is at 130F, while the fan coil units definitely need water well above 130F to provide needed output. To the best of my knowledge this information was not provided by Garn before the sale, and if it had been, the purchase decision outcome might have been different.

The owner did not expect that either unit alone would be sufficient to heat the system. The owner well-understood that the system would need the output from both wood boilers plus supplement from the LP boilers in high demand situations. The goal was 75% reduction in consumption of LP. What needed to be learned was how to operate both boilers to best meet the system heat demands. I don't think that learning is done, and operating experience to-date is outlined in my initial posts.

I have no data on the amount of ventilation air moving through the fan coils along with the infiltration factor of the building(s).

Looking at the charts I have to wonder what exactly kicked on and stayed on at roughly 2:00. The HX return drops about 20* in a straight line and then pretty much stays there. That would indicate a pretty substantial load coming on line and remaining there. Any snow melt connected to this system? It would appear from the temp drop that there is a wide variation in system demand from time to time. Correct?

Click to expand...

I talked to the operator shortly after recording this data. The operator also kept a log of when he loaded the Garn, and that log matches the times when the chart shows rising Garn HWS temperatures. Therefore, at about 9:00 pm, midnight, 3:00 am, 6:00 am and 8:00 am the Garn was loaded (blue line). [The chart shows start time of 8:44 am, it should be "pm."] The operator would fire the Garn based on system supply hot water temperatures based on analog gauges on the HWS lines to each of the two buildings. The instruction, as I understand it, was a "rule of thumb" to keep HWS above 140F or higher if occupants reported that space heating was too cool.

What I read from the chart as to what happened at about 2:00 am is that the midnight wood load had substantially burned out, outside temperatures began to drop rapidly, both buildings were calling for heat, and the Garn stored heat was rapidly played out. The 3:00 am firing appeared to provide the needed heat, but the chart shows this to be true only during the initial "high burn" stage of the firing, lasting about one hour, and then temperatures again dropped. Another effort was made at 6:00 am to provide needed heat by again adding wood to the Garn, but this had little impact, as building heat demand was equal to or exceeding output and building heat demand was not being satisfied.

[continued in following post]

[continued]

The spikes on the Garn HWR (green) and spikes on HX HWR (pink) are quite instructive. Recall that there are two building, each calling for heat separately. The low points of the ♠very steep and high spikes I interpret to mean that both buildings are calling for heat at the same time. The lesser spikes are one building or the other calling for heat. Where the spikes approach the supply line chart (blue and red), demand is fully satisfied and there is no call for heat. From about 2:00 am and after, the spikes never again approach the supply; therefore, building heat demand is no longer being fully met after 2:00 am.

There is no snow-melt system.

I’d also really like to see primary loop supply and return temp along with an accurate flow rate in it. Those numbers would give a fairly close approximation of actual load at a given point in time.

Click to expand...

I agree. Unfortunately, my involvement came at a time when we experienced a fast spring warm-up, very unusual, and I had no more opportunity to accomplish any significant data collection. Getting accurate flow rates in the primary supply and return faces the same issues we have in getting accurate flow rates on the Garn side of the system.

Chris@FHS, GARN Rep -- thanks for the info, which I will forward on to the owner. Do you have a page reference in the Garn literature for

“burn rate numbers based on 24” oak with 20% moisture content and a reloading once an hour”

Click to expand...

? I obviously missed that, and I wish that the local rep would have brought that info to the attention of the owner. I assume the literature you are referring to was provided to the owner.

Chris, thanks for the info. I'm really pleased that someone from Garn contributed to this discussion. I expect I will be doing a lot more data collection under different scenarios during the 2010-2011 heating season, and I hope to have the opportunity to post that data as well. Also thanks to others that contributed to this thread. I can't say enough about the positive value that Hearth.com provides to the wood burning community.

Looking back on your chart, the water temperature was never as high as 200, more in the region of the needed temperature, so very little storage.

Click to expand...

I see the explanation as to why water temperature never reached 200 being that the heat load exceeds the output capacity of the Garn. Water temperature can rise only if the Garn was producing output greater than demand. This was known when the decision to purchase the Garn was made, as it was not expected that the Garn alone could meet system demand. The goal with the Garn and the Wood Gun was a 75% reduction in LP consumption.

heaterman said:

"In effect, based on outside temperature and the size of this structure, don’t we also have a variable load that, at “mild” temperatures has a demand within the capacity of the Garn, and at “cold” temperatures a load that exceeds that capacity"

In a word, no. I haven't seen any information on any of the posts that talk about a heat loss. Proceeding with a project of this size without that information documented would be less than prudent. This too makes me wonder about the involvement of an engineer or at least a competent mechanical contractor.

Click to expand...

Heaterman, you're losing me here. The chart clearly shows that at "mild" outside temperature (roughly above 21F), 100% of demand was being met. Therefore, heat loss = BTUH input to the system during these periods; otherwise, the system still would be calling for heat. A heat loss calc would be useful, but just like determining BTUH by doing a pump head calculation and looking at a pump curve, it is only a calculation. It seems to me that measurement of BTUH input to the system at an equilibrium point would establish actual heat loss.

As an outside observer, I likely also would raise questions about engineer or mechanical contractor involvement. Being more inside than outside, however, I am quite aware of the constraints involved in this project. My objective is not to place fault or praise anywhere; rather, it is to objectively analyze the system as installed to achieve the best performance reasonably possible. I hope I am doing that.

We're on the same page. I always appreciate your insight.