Heating Help

quantifying $$$ savings...

Inherently, I truly believe that running at low pressure, proper venting, clean dry steam etc all has a monetary savings attached .. but has anyone ever actually quantified the savings?

or does anyone has anecdotal evidence such as  "before I installed my Gorton #2 antler, I was getting 2 oil deliveries a month, not I only get 1?"

are there any formulas used to project cost savings if this and this is changed?

or does it all simply boil down to timing .. "you mains currently vent in 10mins .. if you do this they will vent in 5 minutes and therefore save you 50%"?

i understand too that not everything performed to "tune" a steam system is simply for a $$$ gain .. often it comes done simply to pure creature comfort...but that's not the question at hand...i'm just trying to get a handle on how to measure expected $$$ saved prior to fine tuning a system.

thanks.

Savings

In Dec 08 I had a new gas boiler installed to replace an ancient Weil-Mclain. At that time I also installed all new vents. My system would run about 35 minutes everytime it ran. After reading about rad covers I decided to open up the lids and see how things worked. That knocked about 10 minutes off of every cycle and the rooms felt more comfortable. So if you have covers they will cause your system to run longer.

Mark

oil savings

I've been tracking the oil usage for my apartment building (approx 22k sq feet, 1930's era construction, 1.4mm btu, oil fired, single pipe steam).

The blue line is a couple years of data with a circa-1970's Heat Timer, with no input for system temperatures.

Last year that heat timer died and was replaced with an Heat Timer MPC Platinum. I also discovered the pressuretrol was set at 4psi cutout and turned it down to approx 2psi. Data for this period is represented by the pink line.

This year (green line) I got serious and have made the following improvements (although the 51 degree reading was pre-venting/insulation).
-Installed 6 pcs Gorton #1 (to replace a single main vent...a hoffman 40!!)...have another 6 pcs to go at various locations.
-Insulated bare basement mains with 1" thick fiberglass (except a portion of asbestos insulated pipes in the boiler room itself)
-Lowered pressuretrol further - although it's still cutting out in the 1.9-2.2psi range.
-Experimented with cycle length on the heat timer. In early fall, we were runnng 120 minute cycles. Presently running 80 minute cycles (default cycle is 60 mins).

Unfortunately, I can't isolate the effect of any of the actions, but collectively they seem to be having a beneficial effect.

 

thanks for sharing ..

is each data point a different year? month?
when is a data point recorded?
it seems as though you only have 2 data points for the green line.
how are you measuring gal/day?

dos anyone else have data like this to share?

data points

Each data point is an oil delivery. To determine gal/day I divide the total delivery volume by the # of days since the previous delivery. This requires an assumption that the tank is fully filled on each delivery (although we do have a petrometer, which allows a gross assessment of the tank level).
I get the average temperature for the period from Weather Underground weather history page.
There are only two green data points because there have only been two deliveries since the start of the 09/10 heating season (we have a 4k gallon tank). I started recording these data points seperately, since I've made a bunch of changes to the system.

excellent...

thanks for the follow-up ..

fyi, although it's been posted in The Wall before .. you may want to check out .. www.degreedays.net .. they have a nice tool to download HDD in several different ways.

as well as an explanation of how to analyze consumption data .. although you seem to have a pretty good handle on this, perhaps it would be useful to someone else. http://www.degreedays.net/regression-analysis ..

although I must advise .. www.degreedays.net data does not seem to match the NOAA data from the same station (KLGA)..it's over by about 1.5%

tackling my analysis project....

I have just ordered 3 USB-Temperature Dataloggers from http://www.lascarelectronics.com/ .. I got 1pc Thermocouple and 2pcs "Ambient" ..

I'm going to mount the thermocouple onto my flue .. in an effort to obtain "on/off" timing of the burner (I already have an hour meter, but it doesn't log) ..
I'm going to place 1 ambient at my t-stat and another ambient outside. The ambients will help me to understand if the true goal of heating is being recognized...keeping up a "comfortable" indoor temperature based on outdoor temp. The logging functions of the ambients will also allow me to see how quickly the temp falls and rises (inside and out).  I will add to this temperature data oil-tank dipstick measurements to track consumption..

(Thinking about it now, I should get another thermocouple and mount it onto my "furthest" rad for better "system" timing.. EDIT: ordered another Thermocouple) .. if I wanted to go crazy I could add more thermocouples ... inside main insulation (1" vs 0.5") .. at return start .. at return end .. at end of mains .. but I'm not that $$$ right now..

(fyi, I know that wind has much to do with heat-loss, but i'm not sure I can properly (consistent and accurate) log that data right now..i will have to settle for time, temp, TFC (Total Fuel Consumption) data and see where that leads me.)

It is my assumption, having read what literature there is on the units, is that they will all be time-synced so data analysis should be pretty easy.

right now it looks like I am burning (based on the most recent 28day sample) 1.91gph or 0.35g/hdd (t-stat was set @ 71Fahr, we have now lowered it to 68/69Fahr)...i really don't have any "pre-changes" data since I just moved in over the summer and started all my steam tinkering right at beginning of heating season.

i will report my progress as appropriate.

On-off time

Why not just sample control loop or burner pump current?

mainly ....

mainly b/c the voltage/current dataloggers are much more expensive than the thermocouple dataloggers. and although the hour-meter provides cumulative data regarding run-time based on current/voltage, I have no indication of on-time vs off-time except for somewhat long periods like a day or a week...e.g. 24hrs in a day and the hour meter showed 6hrs on. 

Big Saving!

If I'm reading your chart correctly it looks like you are saving about 12 gallons a day. In a 30 day month that's 360 gallons!  Figuring $2.75 /gal for fuel oil that's a saving of $990.month and over a 7 month heating season that's $6930!    No too bad!    I'd check your graph figures against your fuel bill and correlating the two should give you a pretty accurate picture.  - Rod

another graph

Here's another graph you guys might find interesting.

System temp is measured by the Heattimer thermocouple (located on the dry return of one of the longer mains). I'm not exactly sure why it never reaches full "steam" temp...in fact, i've measured the pipe location with an IR thermometer and gets up to ~210 degrees. Regardless, it's just a proxy for steam reaching the radiators...so the actual reading is somewhat irrelevant. 

The pink line is pre-insulation/pre-venting. The green line is with 1" insulation and new main vents installed. The dashed red line is the "heat established" temp for the Heat Timer (ie, the temp at which the heating cycle timer is started).

So basically, establishing heat used to take 18-19 minutes, but now takes more like 10-11 minutes. This has been especially helpful in saving fuel in the shoulder season, when the "on" cycle might only be 3 or 4 minutes.  For example:

Old: 18 minute startup + 3 minute cycle = 21 minutes
New: 11 minute startup + 3 minute cycle = 14 minutes
Estimated savings, 33%

On a long, cold weather cycle, the startup time represents a smaller proportion of total run time, but it's still a savings nonetheless.

great chart ..

this chart really illustrates the effect of changes to the system .. i am expecting my temp-dataloggers tomorrow or thurs. I'm already thinking of ways I can use them .. including timing my rads with & without the short-circuit pipe in place (http://www.heatinghelp.com/forum-thread/129044/radiator-short-circuiting)  .. it's nice that the heat-timer equipment allows this data of data capture .. hopefully my (less expensive) equipment allows similar results. (funny .. my aquastat is set at 170Fahr so i guess i'm always ready to go!)

data capture = me + pencil

There are upgrades to the heat timer that allow for longer term data capture, but in my case I just stand by the control unit and jot down the system temps once a minute during a cycle.

Now that I have a low pressure gauge I also get to run back and forth between the boiler and heat timer once a minute to get a pressure reading for my data set. It's like boiler room circuit training! :-) 

Just to be clear, the system temp isn't the boiler water temp - it's the temp (albeit the wrong temp!) of the dry return at the end of the main....which can only be increased by generating steam. Our aquastat is set at 150, so boiler water always at least 150, even as main temps retreat toward ambient.  

version with pressure

This is the "improved" (post-vent, post insulation) cycle with corresponding pressures (on the right hand scale in psi). It was recorded on a cycle which happened to end right after the pressuretrol cut out.

Our system is way oversized (by roughly 40%) for our EDR...which should be obvious once you see how quickly the pressure rises.

The burner (Carlin 801) has a low-high-low feature (with a corresponding automatic air damper adjustment). Currently there is no trigger to return to low fire on pressure rise, so we run on high fire constantly (except for a low fire period of about 30 seconds at startup). I have a vaporstat on order to address that. I'll post new data when its available. The low fire setting uses 1/2 the fuel of the high fire, so I'm optimistic.

got the dataloggers

I got the dataloggers today .. I did a preliminary 3hr side-by-side calibration test .. the 2 thermocouples on the boiler header .. the 2ambients in another part of the basement...they were all in time-sync and they captured a reading every 1min.

I think they did a pretty good job .. have a look at attached chart. (it captured a boiler fire event but I didn't leave it on much after that)

I have now deployed them in their designated places ..
1- Thermocouple (TC) in the Flue
1- Thermocouple at rad vent on furthest rad
1- Ambient at t-stat
1- Ambient outdoors where it will be exposed to sun (if it's sunny)
I plan on logging undisturbed for at least 24hrs 4PM-4PM

I did speak with the customer service rep to find out that the outer sheathing/insulation (glass braided) on the thermocouple probes is rated for 400degC (752Fahr) .. the probe itself that is shipped with the logger is rated to 1350degC...

I will post more as I have more. I'm going to also look into the

interesting...

....but why wouldn't the headers and the far rad reach a temp closer to steam temp (212-215)?
If you don't have one already, I recommend adding an IR thermometer to your tool arsenal. I got a Ryobi model at home depot for ~$30. I can tell that you'll find many uses for it!
[url=http://www.homedepot.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10051&productId=100674438&langId=-1&catalogId=10053&ci_sku=100674438&ci_src=14110944&cm_mmc=shopping-_-googlebase-_-D25X-_-100674438&locStoreNum=6175&marketID=373]http://www.homedepot.com/webapp/wcs/stores/servlet/ProductDisplay?storeId=10051&productId=100674438&langId=-1&catalogId=10053&ci_sku=100674438&ci_src=14110944&cm_mmc=shopping-_-googlebase-_-D25X-_-100674438&locStoreNum=6175&marketID=373

please note ..

in this initial "test" it was to see how far off the "pairs" of thermometers were .. the ambients as a pair and the TCs as a pair .. i just wanted a few hour sample to match them up .. the TCs were both side by side on a big 2in non-insulated union nut on the header .. my guess is that there is natural cooling as well as the metal providing "insulation" (1/2-3/4" of iron with exposed surface area to allow for cooling)  from the live steam temp ..

this was the first time I ever measured anything .. but notice in your own posts, the pipes don't reach 215Fahr.

i'm waiting on at least 1more heating cycle, then I'm going to pull the units and plot the data .. i should have something up by 11PM EST.

that's true...

...i top out around 184F. But if i measure adjacent to the thermocouple using the IR gun thermometer, it's actually 212-215.  I've never been able to figure out exactly why.  I thought about adding some thermal grease (the kind used to attach heatsinks to computer CPUs), but I haven't gotten around to it.

one thing maybe ...

the cust serv rep today said that the "tip" of the probe is where it is most accurate .. i have them taped down so that the side of the probe is making contact. but since i'm really using them as "event timers" and i don't really need the accuracy of temp, i'm not too concerned.

i'm in the middle of a cycle now .. i'll pull them in about an hour. 

first dataset

Here's the first dataset

also in PDF for enlarging below

Nice graphs

How are you superimposing the plots of the data from the 4 loggers on to a single graph? I downloaded the easy-log software, but it seems to only plot one logger at a time. I guess you must be importing the data into another program to generate the graphs.

Lots of useful info available at one time on that graph. Interesting to see the relationship between the time of firing and when steam actually gets to the radiators. Also the long time constant of the drop in radiator temp after steaming ends.

Stack temp seems pretty high. I saw  you noted no baffles.  Were you planning on installing baffles increase heat transfer and reduce the stack temp?

thanks ..

please keep in mind that I feel my system is pretty well balanced .. slightly favoring the 3rd floor .. the rad I'm connected to is the "farthest rad" .. and is on the 3rd floor at last takeoff from main .. however, i'm not sure if it's on my 55ft main or my 65ft main.

i reinserted my baffles last night.. i should have new data today, i'm very curious to see the difference myself..

i renamed the Lascar text files as .csv and brought them into Excel. since I have them all "time delayed" for 4PM log-start, the times are all in sync, so you only need to copy/paste the data column of each and they will match up. (you need the time column from 1 of them) what I'm working on now is finding a way to automatically identify the milestone points.. right now this is a manual process of finding the significant temp changes in the data list. this will become cumbersome if I have more then just a couple of cycles.

i used a little trick to mark out the "milestone" points. if you email me directly, I can send you my excel sheet...it's for the "latest and greatest" excel version and will be delivered as a .xlsx.

looks good

I'm trying to think of a way you could automatically detect those inflection points using excel. Actually, detecting them is easy if you just take the first or second derivative....the hard part is getting excel to reliably measure the time btwn inflections in multicycle data, since you can't you use a MAX function. Maybe the answer will come with more coffee. 

argh .. calculus

I did find a way to somewhat reliably determine inflection points .. but you are correct, not too easy to find duration between cycles programatically... i'm interested in hearing about how to take 1st derivative (i never excelled in calculus)

that makes two of us..

..who aren't so fond of calculus.
Have a look here: [url=http://www.cem.msu.edu/~cem333/POTTITR.html]http://www.cem.msu.edu/~cem333/POTTITR.html

that is certainly

that is certainly not most attractive way to do it....i found another site which has a more attractive "single formula" method .. but i haven't tried yet. http://people.hofstra.edu/Stefan_waner/RealWorld/utilsindex.html (Derivgrapher download) i do have to give more thought to how to pull the timing out and perform those calcs before i work on finding the derivatives. 

2nd dataset...overnight

Some very interesting things showed up in this data:
1) When I moved the outdoor sensor from upon the brickwork beneath window, to a nearby space not on the structure, I recorded proper outdoor temps .. either the brick was holding heat captured through the day or my house is leaking such that it is 10dFahr warmer in that location.

2) When I had my fluebox open to install the baffles the draft was really noticed by the probe.

3) Most intersting and perhaps pointing to a problem .. the rad DID NOT HEAT during 2nd cycle .. however, it may be that the rad didn't heat all across enough for the TC to sense the heat at the inlet side.

4) Baffles decreased flue temp about 50dF.
5) The first run was longer since I forgot to switch boiler on after baffle install so the house cooled further
6) My time to rads went down BY 4MINS
7) cycle time dropped 1min (but was skewed by that longer 1st run) .. actual runs were 45, 33, 35mins (before baffles 37, 41, 38mins)

Below is the most current chart (and I have attached PDF)

Real world example of lowering steam pressure and right size of boiler

I have a 90 year old customer. She calls today worried her duplex next door is all frozen up. I ask why and she says it is because one of the oil tanks only took 2 gallons of oil. The oil guy said it should have taken more oil. Turns out the tanks were filled on December 24, 2009. The building is being kept at 50 during the renovations if no one is working there, The building is an upstairs downstairs duplex for background. 3,000 sq ft total for both units. The lower unit took 125 gallons in its tank but it is not known if it was filled all the way on the 24th. The second floor units tank was filled on the 24th.

3rd data .. next one NO MAIN VENTS

Here is the 3rd dataset .. nothing much has changed since last one .. I have added a few more stats in the margin .. we are getting a real cold whip through right now .. so I am cycling more often ..

NEXT: I have capped off my proper capacity MAIN VENTS .. and the cycle is currently running .. I'm very interested to see the effect this has... EDIT: cycle just finished .. 35min - ruh roh.. I will let another cycle or 2 w/o main vents run.

TOMORROW: I'm stripping the (fiberglass) insulation off my mains! ... stay tuned.

No Main Vents .. no difference?

OK .. so I eliminated my vents from the system (through using a plug .. and unfortunately I don't see a difference. The avg temp was colder, so I ran more frequent but shorter cycles. Elimination of the vents had no effect at all on the arrival of steam to the furthest rad.

This really rocks my understanding of the advantages of main vents. The only things I can think is
1) that there remained 2 Gorton #C's on the old original vent holes .. but I eliminated 1pc g#2 + 2pcs g#1 on Main1 and 1pc each g#2 + g#1 on Main2.
2) The other option is that my main vents are all stuck closed, but the #1s are less than a week old and I have blow-tested them.
3) The last option is that given the volume of steam and the balance of my rads, there is enough extra venting capacity on my rads to handle the troughput. (I have included below my venting capacity table if anyone is curious)

Timing

I have a feeling that there are two different factors here determining the total time from onset of firing to steam reaching the furthest radiator.

First is the venting which determines how fast the air can be purged from the piping and radiators. Removing the vents should tend to increase the time required.

The second factor is how long it takes for the water in the boiler to reach the boiling point after firing begins. This time would be affected by how long the boiler was previously in standby, and how low the water temperature dropped. Since the outdoor temperature decreased  since your last data was acquired,  the the boiler is now running more frequent cycles and therefore the water in the boiler is retaining a higher temperature when the next cycle starts.The net result is that it now takes a shorter time after firing to begin generating steam. This effect may be partially cancelling the longer time necessary for steam to reach the radiators with less venting.

I think that to really accurately judge the effect of the venting, you would have to remove and install the vents during a period when the outdoor conditions were the same, to eliminate the steaming time shortening factor caused by the more frequent cycles.

good point ...

good point .. and exactly why i hope the outdoor temp is relatively the same tonight. the vents are reinstalled as of this morning. i'm hoping to have a repeat night.

i also realize that overnight testing of conditions is really the only way to do proper testing .. too much sun factor during day .. and people and electronics and cooking and doors opening etc ..

Boy, you guys take this seriously :)

What fun. Thanks for sharing your results.

Since you know know that indoor temperature doesn't vary very much, perhaps you could move your indoor sensor to the boiler riser. This would let you determine how long it took to make steam, then how long it took to get to your radiator inlet.

ambient .. need thermocouple

good suggestion .. i wish i had more thermocouples .. but the room temp one is ambient. 

main testing

You need to eliminate the variables from your testing.  You are not taking into account the temp of the boiler and the temp of the mains.  In order for your results to be valid these have to be the same every time you test.  Don't forget the steam is also bringing your mains up to steam temp.  The colder your mains are the longer they'll it will take to get steam to the vents.  I give you an example of what I've observed with my system.  When my system is completely cold boiler and mains at ambient temp of my basement.  It takes 17.5 min to get steam to my last rad.  During the recent cold snap when boiler was running every few hours I got steam at the last rad in 9.5 min.  Everything happens quicker because the system is hotter.  Here is what I would do.  The first test I would run is open pipe.  Let the boiler sit off for a set amount of time say 1 hour.  Time how long it takes steam to arrive at the open pipe.  Thats your baseline can't vent quicker than the open pipe.  Hook up your normal venting arrangement.  let the boiler sit off for 1 hour again.  Time how long it takes steam to arrive at vents.  Next plug all your main venting.  Let boiler sit again for one hour. Time how long it takes steam to arrive at the end of the main. 

Mark

thanks Mark...

valid suggestions and I appreciate the input. those are all three easy and good tests to establish controls and baselines. what I would also like to do, in order to provide more samples, would be to add a thermocouple to my actual boiler water and keep track of that as well as a TC on my header as someone had suggested earlier to keep track of that and another in the central part of my main simply to understand pipe heating and cooling measurement in a central (non-extreme) location.

as I said earlier as well, we are expecting pretty much the same weather tonight (one of the reasons why I added windchill to my charts) and I have the vents in place .. so there is a good possibility that I will obtain some interesting new "with vent" data tonight.

the jury is still out ..

correlating basically the same time period, with the same
number of cycles, on an overnight of approx the same outdoor temp and
changing nothing else... ON AVERAGE, RUNNING WITH VENTS SHAVED 3 MINUTES the end of furthest rad measurement ...but my runs were 3mins longer

trying to make sense of it:

   if, as someone suggested, the furthest rad would be favored by the
lack of vents .. then the steam took a bit longer to get there .. and
then backed up to the next rad down the line, which is the LR Couch rad
which may have had an effect on the t-stat (shorter run) .. similar on
the other main branch .. it would have favored And being the furthest
and then Foy next far then DR which may affect t-stat (shorter run) ..
I must admit that last night I physically felt warmer with the vents on.

  with the vents, all rads would have in theory favored equally (and
got to furthest faster), and thus had less proximity effect on the
t-stat leading to longer runs.

  in order to test this, I would have to run the same test on a middle
rad on 2nd flr...to see if a mid rad timing changes, (or even gets
heated all across)...tonight I will run WITH VENTS TESTING MID RAD MstB.

cumulative data

below is cumulative dataset thus far..i have posted my excel sheet (.xlsx) in its entirety here: http://www.ypgmedia.com/heatinghelp/cum.xlsx .. I welcome any updates, additions, suggestions, etc .. as long as you openly share them .. I will do my best to keep a running cumulative dataset at that location.

BTW, arches, I have figured out a single "scanning" formula for finding the inflections .. and have made it pretty easy to pull them out using "Filtering" .. look at the column called "Milestone"

excel file?

I don't see any .xls or .xlsx files in the zip file at the link you provided. I can't actually open .xlsx files, but they can be converted at zamzar.com

it's there ..

the .xlsx file is there .. what's not there is a .zip file .. so I'm not sure what's up on your end. I just clicked the link above and the excel file opened right up. 

also at docs.google.com

I have uploaded the cumulative data to:
[url=http://spreadsheets.google.com/ccc?key=0AnPQdPnGKv4wdFZmTlRDMmV5a053bDV3eXZOVWZ1MXc&hl=en]http://spreadsheets.google.com/ccc?key=0AnPQdPnGKv4wdFZmTlRDMmV5a053bDV3eXZOVWZ1MXc&hl=en .. but it doesn't do any fancy charting  .. the summary data is on "Sheet1" .. but at least it's a place to download the data if you'd like. i will try to keep both the .xlsx and the docs.google.com data in sync.

EDIT: .. Google docs only allows files upto 1MB .. at this point the file has surpassed this size .. I won't be able to keep an updated copy @ Google Docs .. I will leave up the one above .. but realize that the data will not be refreshed or appended to.

zamzar worked

thanks arches .. i tested the conversion at zamzar and it worked without error .. thanks for that .. sometimes i use little idiosyncratic excel functions that don't convert .. but this was fine. 

Interesting

the way I read this when you put the vents back on your furnace ran longer - using more fuel!  But I assume your house was a more uniform temperature (since steam got to that top radiator faster)

now we're getting interesting ...

(OH SHOOT...I MEANT THIS TO BE IN THE $$ THREAD)



Now that's interesting ... the cycle right after my pressure cycle ..
my rad got hot in 1/2 the normal time .. then I guess the vent closed
and started to cool .. then the vent must have opened again an heated
again .. see the cycle labeled 10:22 at top...and the cycle after that
the rad almost didn't even heat all across.

another intersting
thing is that my rad max temp was only 158dF .. which could be a
function of the way I mounted the TC to this rad as opposed to the
other rad..but any temp over 100 is really valid for determining steam
presence.

without my pressure cycle (which was caused by a
jump in t-stat +3dF) .. all other cycles were about averaged 0:35m ..
with the pressure cycle, the avg was 0:43m

my avg time to rad
was 0:25min where before it had been 0:22min on the far-rad .. but
again, it may be due to the way the TC is mounted and perhaps the rad
has to get a little hotter to register on the TC.

and lastly, my flue temp dropped a few dFahr

to enclosure or not?

after a very spirited discussion abour radiation vs convection over here yesterday .. i decided that a test regarding rad enclosures would be interesting ..

you should know that for the last 9wks, i have been running without the tops on my rad enclosures .. they still have sides and a grilled face ... so last night .. i put the tops all back on .. at least on the 1st floor where the stat would be effected ..

it seems that it caused for only slightly longer runs .. like an extra 3mins per cycle .. where I had not cycled on pressure in several days (during normal runs), when i put the rad tops back on on I cycled 1 blip on 3 of the 4 following cycles .. in addition, the first cycle following the top installs, my rads had to heat up a fair amount more (20dFahr) before satisfying the stat .. but once that first cycle had passed, the rads heated only back to "normal" levels .. in general, adding the covers added about 2mins to each cycle.. or an extra $5 of fuel every 3days. (about 9cycles a day at 1.9GPH @ $2.50).. or $200 for 120day heating season.

You have to control for degree days

Outside temperature could easily be the cause of different run times. I would add up a weeks worth of run time and divide by that week's heating degree days.

thanks for pointing that out....

thanks for pointing that out .. it was actually about 7dF(windchill) WARMER during the longer run times WITH the enclosures. but you're right, my sample size very small, i should run for a week with and week without...but don't forget that HDD is simply a daily average under 65dF.

windchill WITHOUT enclosures:
23:00 - 25d
00:00 - 25d cycle 0:41 -- added water, shut off for 2mins while adding (actual burn 0:39)
01:00 - 24d
02:00 - 24d cycle 0:35
03:00 - 23d
04:00 - 22d cycle 0:34
05:00 - 22d
06:00 - 21d cycle 0:36
07:00 - 21d

windchill WITH enclosures
23:00 - 32d cycle 0:40 (+1min) (+7dF)

00:00 - 30d (+5df)

01:00 - 30d cycle 0:40 (+5min) (+6dF)

02:00 - 30d (+6dF)

03:00 - 30d cycle 0:39 (+5min) (+7dF)

04:00 - 29d (+7dF)

05:00 - 29d cycle 0:37 (+1min) (+7dF)

06:00 - 28d (+7dF)

07:00 - 28d (+7dF)

HDD

I don't understand why wind chill would be an appropriate temperature for comparison. Your house doesn't feel wind chill. Yes, on windy days there will be greater convective heat loss, but the the outside of your house can't get any colder than the actual temperature.

windchill

well as you said, and has been said on the wall before, wind has alot to do with heatloss and so I am using a standard of windchill to factor in the effect of the wind. i think it's a fair standard which brings both wind and temp into the equation .. i also plot the outdoor temperature on two different lines .. one if the NOAA 1-hr data and the other is my outdoor sensor -- located in the sun 1-min data .. i have placed it in the sun specifically b/c my house is in the sun. 

It's still a mistake

Wind has a lot to do with convective heat loss, but what drives heat transfer through your walls is conduction, and that is a function of Tinside - Toutside, Toutside being the actual temperature, not the windchill temperature. You aren't even correctiing for outdoor temperature or windchill temperature so I'm not sure what your point is.
I respect what you're trying to do but it's simply irresponsible to write things like "enclosures cost $200 a season" and so forth when you can't justify it. You're measuring a lot of stuff but there's no science behind your analysis.

fair enough...

i agree that i'm probably not doing the data justice .. and am probably making assumptions and conclusions recklessly. i will refrain from extrapolating conclusions and simply post data of different things as I think to measure them. or use words like "enclosures *may* waste more money" and provide for everyone else the master dataset as I build it.

i will still try my best o make sense of what i see but my apologies if I was illogical in my analysis or made assumptions which I didn't fully understand. and I certainly welcome feedback about ways I can improve my methods so that it is more generally useful or informative overall. thanks David.

but ..

is it not intriguing to find that on a warmer temps day (not the windchil, but actual) than the previous that the system ran longer with the enclosures? there must be some correlation, even if i am unable to understand how to quantify it. 

Enclosures

The simple explanation is that your enclosures reduce the total heat flux. It takes longer to get the amount of heat needed out. Runtimes might increase, but they also might occur less frequently if the enclosures increase system mass.

To quantify the effect you'll have to collect data over a longer period, refrain from making other material changes during that time, and control for heating degree days.