Welcome! Here are the website rules, as well as some tips for using this forum.
Need to contact us? Visit https://heatinghelp.com/contact-us/.
Click here to Find a Contractor in your area.
Degree Days and Fuel Consumption
Mike T., Swampeast MO
Member Posts: 6,928
Thanks for the encouragement Brad. My main goal is a more accurate means of comparing fuel consumption under different reset curves, different indoor temps and even different boilers (I can use either a Vitodens or conventional cast iron). Simply using base 65 degree days makes this rather difficult.
0
Comments
-
Am trying to develop a more structure-specific basis for comparing fuel consumption over time.
I have complete hourly government temperature data collected about 10 miles from my location.
Base 65 degree days seem both somewhat crude and too high for my home. For normal room temps of say 65° - 69° a base of 60 seems much more reasonable.
I rarely use any daily heating setback, instead preferring to adjust TRVs as low as possible for reasonable comfort in various spaces and just letting them be.
So, would it be reasonable to:
1) Use 60 for my degree hour standard based on 67° average indoor temp?
2) Further adjust the degree hour standard (60) based on different average indoor temps? Say 55 for an average indoor temp of 62°?
0 -
Degree Day Pioneer
Mike, I think you are on to something. For some time I have felt that Base 65 Degree Days are too high. They harken from days when homes were less insulated and had fewer internal appliances to generate heat.
(OK we *did* have pilot lights on our gas stoves and probably lost a few brain cells, another topic.)
But computers, TV's in every room, fancier halogen lighting, and tighter insulation standards, better glass and envelopes to hold it all in; we can well tolerate lower outdoor temperatures than before.
Commercial calculations use a Base-50 database by comparison. Beginning to blur is it not?
If you are using the "standard DD formula" :
(BTUH x 24 x DD)/(TD x Eff. x Fuel Units) x Cd factor
then the temperature difference you use for your heat loss cancels out/ist taken into account.
From that I think you are well placed to use a lower base for DD calculations.
How low? I think that your experimentation will tell us faster than my guessing...
Wonder if Brookhaven National Laboratories have anything in the works on this?
Brad0 -
Here is what I would do...
... plug the data into your handy-dandy excel spreadsheet and compare the regression fit at different base points. Adjust for insolation by month if you like....
My Excel/NOAA file is over 20MB but our old apartment fit the 64°F model pretty well (R2 ~ 0.94). I do not have enough data points for the new place (yet), so that calculation will have to wait.
(Hello WEL...)0 -
Constantin-
Just WHAT do you do in your spare time?
0 -
too many variables
I can remember people telling me how it used to be done, with charts and slide rules, etc. I think the degree day method was invented in the 1930's? This method has served its purpose. There has gotta to be a better way because it is essentially guessing based on past usage.
Many people have second homes that aren't lived in all the time. There are a lot of tighter homes but you still have to account for the customers with the older drafty homes.
As to using different base temperatures because 65 is too high, that seems true for a lot of cases but it is also too low for some of the elderly. I have noticed (I deliver heating oil) that many of the runouts in the spring and fall are to the homes elderly people live in.
For a building open during business hours, the base temperature might be 55. Who knows when they actually fire up the heating system for the season. When I deliver oil to one of these in the fall, I assume they are not going to take much oil and most of the time I am correct.
Some computer programs using degree day to calculate the next delivery account for this by using winter K's and fall/spring K's. The winter K's are lower most of the time.
The 65 degree method works best in my opinion for elderly year round residents with older homes. I notice that is the account that gets the most consistent deliveries year round.
You can set the customers who aren't living in their second homes for 55 degrees but what do you do when they decide to live there and crank the heat to 70 for three weeks without telling you?
Then you have secondary sources of heat like fireplaces, etc.
No matter what you do, the degree day method cannot figure out oil consumption for pool heater usage.
I wish someone would come out with a cheaper oil liquid flow meter that has the capability to contact an oil company like a low level alarm does. It is true that this would work only with single oil lines but a lot of people do not like two pipe oil lines anyway. Low level alarms are ok but it is hard to install one in an underground tank. I guess oil burner run clocks would work with single units.
0 -
> 2) Further adjust the
> degree hour standard (60) based on different
> average indoor temps? Say 55 for an average
> indoor temp of 62°?
Yes, of course it changes with indoor temperature. If you hold the house at 70 and the degree-days are measured from 65, then if you hold the house at 60, the ddays should be measured from 55. Roughly.
But ddays are not intended for use in individual houses. The method works very well for the gas company to predict demand, since their demand is an average of thousands of houses, each with a different setpoint and different insulation and different numbers of prople etc. For any single home there are lots of unquantified (in effect, random) variables that are averaged out when looking at a whole town.
> My main goal is a more accurate means of comparing
> fuel consumption under different reset curves,
> different indoor temps and even different boilers (I
If you can't average over many houses to smooth out the random factors, you have to average over time instead. So for any given reset curve and indoor temperature, you have to collect data for months and only then change the curve or indoor temp. Few people have the patience!0 -
The Ubiquitous BTU Meter
for any system with a fixed curve (fixed for a season) compared to actual heating-only fuel usage (no domestic water now....) and the concurrent degree-days, would seem to be the simplest indicator.
(BTU meter by definition to be cumulative flow and concurrent delta-T data)
Sure, you can shoot holes in any method. This one is as imperfect as many, but at least takes into account the BTU's delivered to the space. System efficiency is also absorbed and in the end you get a true AFUE.
I would take it one more and measure BTU's delivered as above and add to that the actual Delta-T indoor to outdoor. You lose solar gain aspects and internal gains but those are wide variables anyway. Probably constant in many cases for the internal gains.
I could see that this method would not be affected by the degree days, whatever the basis. They are just a mean value of the day's temperatures as a point of comparison.
At the end of one year you would have data for... one year. Degree days are generally 30-year averages. But it is a place to start.
My next investment will be in an Onset Computer FlexSmart logger bank with appropriate sensors to do the above, now that I think about it. Just what I needed! 0 -
Regression fit, negotiation fit
There we go, break everything with math. Curve fitting is the right thing to do. I have a question though.
Constantin, you found the 64F model to fit your personal data. Was the 64F model good throughout the season or was the 64F model mostly good at high season while for the shoulder season a lower model would have been a better fit?
The degree day numbers provide data that has become standard over years of use. That's very convenient and consistent. There is one thing that has always bugged me (and others) about the degree day:
It adds degrees like a stack of pennies, giving all the same value and weight. Yet we're talking about a delta T.
It's the same argument union leaders get all worked up about when it comes time to agree on a pay raise. Should it be 10 cents or 10 percent all across the board? 10 percent does not stack up like 10 pennies; this always leads to conflicting results.
Given a day were you burn X amount of fuel to recover 10 degree day, do you then expect to burn double the X amount of fuel for a 20 degree day? No, you're likely to see an exponential increase in fuel usage. I don't know how to patch the two, moving the base level seems good, but in relation to what?
Chapchap shows a good relation to the customer's age, that's both interesting and amusing. You're on to something again, Mike, what's you're age?0 -
> Given a day were you burn
> X amount of fuel to recover 10 degree day, do you
> then expect to burn double the X amount of fuel
> for a 20 degree day? No, you're likely to see an
> exponential increase in fuel usage. I don't know
> how to patch the two, moving the base level seems
> good, but in relation to what?
What you are saying is that the fuel consumption depends on the temperature difference nonlinearly (an exponential function is nonlinear, but not all nonlinear functions are exponential).
It is well known that the rate of heat flow by conduction depends linearly on the temperature difference driving the heat flow: Wikipedia article
This applies to conduction only, not to radiation, convection, air leakage, etc. Conduction (through exterior walls, windows etc) is the dominant mode of heat loss in properly constructed homes. Radiation is small in comparison for the temperature differences normally encountered (<100F). So any deviation from linearity is a matter of whether air infiltration plays a significant role in a specific house.
The use of a number such as 65, rather than the actual higher indoor temperature, is an ad hoc way to account for heat produced indoors by electrical devices, humans, sunny days, etc.0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_____: Feb 6, Dec 14
OUTDOOR TEMP____: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
CALC DESIGN LOAD: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We did adjust the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other measures also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_____: Feb 6, Dec 14
OUTDOOR TEMP____: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
CALC DESIGN LOAD: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We did adjust the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other measures also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)______: Feb 6, Dec 14
OUTDOOR TEMP____: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
CALC DESIGN LOAD: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We did adjust the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other measures also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
CALC DESIGN LOAD: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We did adjust the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other measures also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
EST. DESIGN LOAD: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We did adjust the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other measures also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 18.1%, 22.5%
EST. DESIGN LOAD_: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? We do seem to be wasting a lot of fuel in mild weather.
(Note: We adjusted the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other methods also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Christian, we are mere homeowners, and we appreciate and defer to your thoughtful posts and expertise.
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 3 min, 4.5 min
BOILER RUN %__: 18.1%, 22.5%
EST. DESIGN LOAD_: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? Our boiler runtimes (about 3 minutes at 37F, 4.5 minutes) really don't change much (Ergomax buffer tank is on the way).
(Note: We adjusted the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other methods also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Christian, we are mere homeowners, and we appreciate and defer to your thoughtful posts and expertise.
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 3 min, 4.5 min
BOILER RUN %____: 18.1%, 22.5%
EST. DESIGN LOAD_: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements? Our boiler runtimes (about 3 minutes at 37F, 4.5 minutes) really don't change much (Ergomax buffer tank is on the way).
(Note: We adjusted the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other methods also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 3 min, 4.5 min
BOILER RUN %____: 18.1%, 22.5%
EST. DESIGN LOAD_: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements?
Christian, we are mere homeowners, and we appreciate and defer to your thoughtful posts and expertise. Perhaps a buffer would significantly lengthen runtimes and improve efficiency.
(Note: We adjusted the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other methods also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
One Counterexample
Our 99% design temp is 10F, so 37F is halfway between 65 and 10. Here are results from two boiler clockings (both done after prolonged cold temps, before significant solar gain, etc.):
DATE (2006)_______: Feb 6, Dec 14
OUTDOOR TEMP___: 37F, 8.6F
BOILER RUNTIME__: 3 min, 4.5 min
BOILER RUN %____: 18.1%, 22.5%
EST. DESIGN LOAD_: 66MBH, 42MBH
If part-load efficiency improves as outdoor temp falls (at least for severely over-sized, bang-bang boilers), could that explain our measurements?
Christian, we are mere homeowners, and we appreciate and defer to your thoughtful posts and expertise. Just wanted to throw this out for your consideration.
(Note: We adjusted the boiler output with the best part-load efficiency equation we could find, but that didn't change the numbers much. Other methods also point toward a Heat Load in the 40 to 50 MBH range.)
gf0 -
Clarification...
On Feb 6, the runtimes were 30 SECONDS with cycle lengths of about 3 minutes. (After this, we added a triple aquastat to widen the boiler differential.)
On Dec 14, the runtimes were about 4 MINUTES with cycle lengths of nearly 18 minutes.
Sizing is no longer an issue for us, but we're still curious about the magnitude of the part-load efficiency effect.
gf0 -
Clarification...
On Feb 6, the runtimes were 30 SECONDS with cycle lengths of about 3 minutes. (After this, we added a triple aquastat to widen the boiler differential.)
On Dec 14, the runtimes were about 4 MINUTES with cycle lengths of nearly 18 minutes. (So, the aquastat helped quite a bit.)
Sizing is no longer an issue for us, but we're still curious about the part-load efficiency effect.
gf0 -
The other part of heat loss by conduction
through walls is the time factor. Heat loss is all about speed (rate of heat flow) over an hour. Different materials have different specific heats (ability to absorb and hold heat) which is different than their conductance (k) ability to move the heat through them.
The academic ideal situation depicted in books describing heat loss (ASHRAE sample wall sections for example) is where a structure has soaked all of the heat it can at a given time, things are in equilibrium, steady-state emitting into space from warmer to colder.
Throw in setback of temperature, solar gain on the exterior and the dynamics become, well, quite lively.
To give an example, extreme for purposes of explaining it:
Warm-up in an uninsulated a masonry house (wall U factor not a whole lot better than good glass, maybe 0.35 or so) after a deep setback will take some time as you can imagine.
But the wall is cooler to start, hence the apparent delta-T inside to outside sees an interim temperature of much less.
Thus the outer few inches of the wall will lose heat more slowly than the inner few inches of wall, compared to what your typical "instantaneous" heat loss coefficient would indicate. Add solar gain on that wall and heat flow may well be inward, even on a cold day.
Once the mass is warmed, the pulsing of heating slows down appreciably. After a heating cycle and upon setback resumption, the walls radiate inward as the interior cools. You can see the push and pull of the lowly garden variety BTU here, can't you?
As you said, infiltration is the bigger variable.0 -
I'm 42.
Have the basics of my program written and debugged.
Using hourly averages to compute load instead of 24-hour high minus 24-hour low divided by two (e.g. standard degree day calculation) [seems] more reasonable.
Value is always (in samples I've run so far) lower than standard degree days. It [seems] to better reflect the general temperature of a 24-hour period.
The adjusted degree day basis made BIG changes in therms/degree day consumption. Again, it increased in all circumstances.
BUT, therms per degree day increased faster for the conventional boiler than for the Vitodens. AND the difference between Vitodens at original curve (2004-05) and adjusted curve (2005-06) seems reasonable.
Will crunch many more numbers tomorrow morning.0 -
Sorry, didn't realize you were figuring for one house?
I have a theory that if you are figuring for one house, you can pick the basis temperature by whatever makes the K Factor close to constant. I've never used the kind of data you are using but if you are heating the same envelope with the same heating plant, any other variations beside hot water usage (if applicable) should be changes in temperature settings in the house and NOT the K-Factor. If you add on to the house, add insulation, and/or change the heating system, this would change the K-Factor in my mind. What you need to find is the correct K-Factor.
I suppose you can use old data if you didn't leave the house for some reason (vacation) for days and turned down the thermostat for example. That data would have to be thrown out if you are trying to find your correct K-Factor. I do not think any method for trying to figure out a basis temperature is bullet proof but I can't think of a simpler way. (in my mind anyway) Did I explain this well enough?
0 -
It can be done...
... but who is willing to pay for it? Start with some like a WEL from OurCoolHouse and then go from there...
For example, sensing oil consumption indirectly via current sensing on the burner motor is pretty viable on the oil side of the business since oil burners do not modulate up and down. Thus, noting how much time the burner has run gives you a good idea as to how much fuel it has consumed.
This can be done safely with a non-contact sensor that is optically isolated, etc. For those without an internet connection, perhaps a iButton link is in order, i.e. where you walk up to the home, hold up a keyfob and download the data.
Either way, this is where the industry is going. Our power, water, and gas utilities are already using RF, so it's only a question of time before some enterprising soul does the same for the oil industry.0 -
Good Points...
... One thing I am struggling with is that NOAA gives me very good data, except insolation. the R-squared regression fit so far is still close to 90%, now that I have more data points. I did not find there to be a huge difference at the low vs. the high end of the season, the residuals were pretty much OK.
One other factor to consider is how the gas company estimates vs. reads on a number of months. I tried incorporating wind and precipitation but did not find them to be significant factors.
Once I have the WEL up and running, I may consider getting a 1-wire weather station to give me excellent data to compare my heat loss against.0 -
Fuel Use vs. Degree Days
With gas fueled heating systems it is fairly easy to correlate gas use (as per gas meter) with Heating Degree Days. Attached is a correlation I did a couple of years ago.
This is the daily gas meter readings for our building vs. Heating Degree Days published by the local Utility who gets them from regional weather station.
Excellent linear correlation over a wide range of outdoor temperatures. Some scatter in daily readings due to local weather different than regional, sunshine, DHW and other minor gas uses etc.
Confirms linear relationship between fuel use and Heating Degree Days.
Doug0 -
Welt, if you look at the high end of the curve, you'll see most of the points there are above the straight line. This suggests a small quadratic or exponential component.
But it could be because you have a condensing boiler. A CI boiler will be slightly more efficient on a cold day, but a condensing boiler will be less efficient.0 -
Going on strike for a more radiant deal
Here's how in many ways we shouldn't even care all that much about heat transfer rates (that is conduction, convection, radiation all in and out of our home and all dependant on temperature differences between the home and the environment), all we want to do is keep the inside warm as measured by the thermometer.
Thermometers sing like the sirens, we should not fall asleep while staring at them and the songs they chant are only a fabricated image of an otherwise invisible heat flow. Listen to the Loreley and you'll find yourself stranded on the rock at St Goar along the romantic Rhine.
Someone's been listening a little too much to them sirens! You'll be thinking, he even fell on the rock head first... Read on.
Whether you are a whale or a goldfish, what matters most to you? To know if your aquarium is filled to the 55, the 65 or the 75 mark? Or to know if your aquarium is filled with two million or just two gallons of water? Of course this aquarium leaks at the bottom, but it is also connected to a faucet and as long as everything is more or less in balance, the whale is still mainly concerned about finding the two million gallons, the rest of the story is all fluff.
We buy heat by the gallon, as measured in BTU. We move it around in BTU per hour, BTU/h. We keep track of the temperature level in the aquarium in degrees. Do we have any idea of how much heat a home contains, nope, we're usually clueless about this, yet all our analysis revolves around this conceptual amount. Gallons in an aquarium are easy to measure, but how can we have a happy whale?
Well, let's find ourselves an aquarium, that'll be our home. Then let's drill a neat hole in its bottom so that we have a nice leak, that's our computed R-value heat leakage factor. Then let's turn the faucet on and have our tank filled, that's our boiler at work. The level goes up. Oh look look look the level is zooming past the 35 mark, it's flying past the 45, it's reaching past the 55, yes yes yes, it's nearly approaching the 65, go... go... go... it's creeping to the 75, only an hour or two and it'll get to 85, oh man this is getting boring, when will it ever reach 95? how much water have we wasted down the drain so far?
OK, well, whales are mammals? aren't they? They can breathe in normal air just like any of us. So, turn off the water, I'm hungry now.
Anyway, you see how the heat usage grew in an exponential (actually logarithmic) manner relative to the heat level measured in degrees. This happened in spite of the fact that the input and output to our aquarium may very well have been linearly connected.
You integrate a throughput over a period of time and, bamo, you get a quantity which you may or may not be able to afford. Things like differential gaps (the one between the input and output lines) grow in sometimes very strange ways even if they are linear cousins. Make them distant family with a twisted relationship and who knows how things will evolve.
Same problem can happen at the casino.
You're playing roulette. The game is so designed that for every dollar you plunk down you will get back say 80% (I don't know the actual number). The relation is set in stone, fixed that way by the casino management and is strictly linear. We measure our excitement level, we determine we want things to get hotter. What should we do?
Hey! move away from the dorky sixty-five cents table and mosey up to the big seventy-five thousand dollar table. That'll be exciting, oh yes. Same risk factor, bigger stakes.
How does this move affect your cash input into the game? it became an exponential fury that will be hard to explain when you get back home...
Let's get back to the hot stuff.
Conduction is linear. The rate is proportional to the delta T. So too is convection, being proportional to the temperature difference between the wall skin surface temperature and the environment temperature. Radiation is a bit different. It is proportional not to T1minus T2 but to, T1 to the power of 4, minus, T2 to the power of 4. That makes for some super powerful non-linear relationship. T1 and T2 being respectively the emitting source and the receiving surface temperatures.
This is the big reason steam filled (unlike water filled) radiators have the confidence of a champion. Tsteam**4 is masterfully more gigantic than a mere Ttepid**4, it squashes it like a mere bug. Thus we find the exceptionally efficient heat transmission methods of steam heat. Of course, no one ever talks about this, because no one ever cares to be amazed. It's all dopey special effects on TV anymore.
Radiation is much more powerful than we think. Mark Eatherton posted a thread about nighttime re-radiation, he measured it's surprising effect and convinced me this was not just peanuts. I made my own eupatheoscope with an indoor/outdoor electronic thermometer, a coffee can and some black cloth. This radiation we talk about appears with surprising effect that usually flies right into the face of the conventional thermometer.
Remember how I told you we could hear the singing of the sirens.
I'm begging for immunity from the Loreley for all my ranting.
All the nice plots and data furnished so far on this thread show a plenty nice enough linear relation between degree day and repeatable fuel usage. Enough to strand my theory it would seem, but I suspect the trick is to find the good base temperature. After all, moving the base line up and down affects the needed X and 2X energy quantities in the same way these are affected by the outdoor temperature. This would be like painting two lines on our aquarium, one maximum and one minimum; things happening when moving the water level past each line parallel each other in a mirror image fashion. For good balance, pick a good base line.
What I think I'm saying is that increasing number of degree day causes non linear curving that is counterbalanced by what moving the base line does. The two non linear behaviors conveniently straighten each other out and leave you with a handy linear prediction graph. The trick is to find the good base line for just your application.
So, who actually uses the 65 base line? and how do you come up with the good one?
This whale needs some air, see you later.
CH0 -
Homeowner Bottom Line
1. Whales that we are, we don't care one bit how much heat is stored in our house. We only want it to be comfortable at our local climate extremes and at a reasonable cost.
2. We care about Temperature only as a state variable--don't care how we got there, but *at steady state*, Temperature seems a simple substitute for MRT--rub the numbers off our
thermostat and we might set it a tad higher or lower, but not substantially so.
3. Government may recommend setbacks, but frankly, being whales, we prefer to maintain a steady temperature. (Hence we've insulated to the gills, air sealed, etc.)
4. At steady state, how close to linear, with respect to delta T, is heat loss across a defined system boundary?
5. Does Mike T.'s "structure-specific basis for comparing fuel consumption" (or any other method of heat loss calculation) assume a variable system temperature?
0 -
Thanks for the suggestions gentlemen! I've tried to incorporate them all to some degree.
Am comparing four heating seasons (last one not yet complete).
First two with traditional cast iron; last two with a Vitodens.
For the first year, the Vitodens operated at the same reset curve as the cast iron for the majority of the season.
For the second year, the Vitodens operated at a much lower curve (slope & reset) that is just barely able to maintain TRV settings.
Other than the boiler change, there were no other changes to the heating system.
There have been some envelope changes, but not within seasons. I've done my best to compensate and have K factors that work reasonably well both for seasons as a whole and month-to-month within those seasons. Used 60 as the basis for degree day calculation in all seasons. Used average indoor temps (I have datalogs with a minimum of one sensor per level).
While I know this isn't perfect, it [should] be reasonable with "net system seasonal efficiency" being the last variable.
Net system seasonal efficiencies:
Traditional cast iron - 45%
Vitodens at same curve as cast iron - 88%
Vitodens at optimized curve - 97.5%
Supply temp of the Vitodens has only been above 140° for a few minutes in those two years. Return temp has never been above 104° in those two years.
For this heating season, 80% of the HVAC-calc value (based on completed envelope @ 70°F) seems accurate. Four years ago, 110% was accurate. In the middle two, 85% was accurate. Still some thermal improvements yet to occur--about 3/5 of the ground floor does not have drywall and none of the ground floor windows are fully weatherized as neither the casings, weight pocket insulation or weatherstripping are in place. Nearly all should be completed by next heating season. Will be interesting to see if usage predictions will be reasonably accurate.
0 -
Fuel Use vs. Degree Days 2
Attached is the same chart with a 3rd order polynomial trend line. Improves the correlation coefficient from 93.7% to 94.4%. A better fit of trend line at both ends of the curve.
Note the Base Temperature for the Heating Degree Day calculation is 14.5 degrees C which is 58 degrees F. Our local gas utility uses that Base Temperature for estimating gas consumption based on weather. This Heating Degree Day information is published daily.
The point is that the corrlation is essentially linear over the full range of outdoor temperatures, exactly as predicted from the conduction and convection heat loss equations. Radiation is another story and accounts for most of the scatter in the data. I am using a radiation compensated outdoor temperature sensor to reset the heating system - works great.
Doug0 -
The point is that the corrlation is essentially linear over the full range of outdoor temperatures, exactly as predicted
Just what I [seem] to find as well. The trick [seems] to be finding the proper origin for that correlation and it doesn't [seem] to be what is predicted by Manual J and standard degree day calculation...0 -
Guess I need to
redo my graph. I should also add 2004 to 2005. The first year was Dec to April for the old steam plant.0 -
More Data
I'll throw this year's fuel use data for my house into the mix. I plotted the relationship between outside temp and fuel used per hour, then degree days vs gallons used. I used a base of 67 for my degree-days, which gives the best fit.
I use an hour meter to track burner run time and periodically double check it with tank level measurements. It's usually within 5 gallons over a tankfull, which is pretty accurate. The temperature data I used in these plots is taken from a local weather station and averaged each day at 8am. This causes a problem when I take a reading say at 11pm and the temperature changes a lot between the two days. I now have a weather station and am recording outdoor and indoor temp every 10 minutes, and so far this has reduced the scatter quite a bit.
The house is an old Victorian, I'm only heating the downstairs (about 1500ft2). Minimal insulation, but new windows. The steam boiler is a Weil-Mclain 72 5-section, originally fired at 2.75GPH! It has been downfired to 1.5gph (1.25 nozzle at 150psi). Last efficiency reading was 78% (520F stack temp, 8% CO2). The steam mains are insulated and vented pretty well, the system is quiet trouble-free.
I'm about to have insulation blown in the outside walls, so it will be interesting to see if I can see the difference in the data.0 -
Great Data!
Congrats on the close correlation and the great graphs. I'm on my way to similar joy via the WEL... No before data though!
You might want to investigate if your weather station can accept inputs like closed contact switches. There are some current monitoring switches out there that allow the contact to close whenever the burner is firing, for example. That would allow you to gather all the data you need in one place...0
This discussion has been closed.
Categories
- All Categories
- 86.3K THE MAIN WALL
- 3.1K A-C, Heat Pumps & Refrigeration
- 53 Biomass
- 422 Carbon Monoxide Awareness
- 90 Chimneys & Flues
- 2K Domestic Hot Water
- 5.4K Gas Heating
- 100 Geothermal
- 156 Indoor-Air Quality
- 3.4K Oil Heating
- 64 Pipe Deterioration
- 917 Plumbing
- 6.1K Radiant Heating
- 381 Solar
- 14.9K Strictly Steam
- 3.3K Thermostats and Controls
- 54 Water Quality
- 41 Industry Classes
- 47 Job Opportunities
- 17 Recall Announcements