I cannot find any conversion table that translates the BTU/hr heat loss in a Manual J report to the amount of fuel required to heat the building to the design temperature. In this case I'm using propane and I have calculated the 92,000 BTU/gal using the actual BTU/Hr X Degree Days and end up requiring over 30K gallons of propane for a heating system. This cannot be correct. Any guidance would be greatly appreciated.
Tags:
Hi Dan,
That would be close to correct if every day was at the design temperature. I don't use Manual J, so tell me what you have for the design temperature heat loss per hour and then the total heat loss per year at your local HDD.
Bud
Hi Bud, My total heat loss per hour is 51,000 BTU/Hr with a design temperature difference of 48 degrees. I am using 3900 degree days to figure the heating requirements.
Thanks for any guidance you may provide.
Thanks, Dan
If 51,000 is your actual heat loss at the design temperature then you need a system capacity of 51,000 BTUs running continuously to maintain the set temperature. A 60,000 BTU boiler running at 95% efficiency would just barely keep up. Before I go into a bit of oversizing, we will need to verify that your calculations are correct.
Now, for total fuel use per year we need the total heat loss per year and that is where the Heating Degree Days comes in. I'm looking for a number like 40 or 50 Million BTUs per year. Using 50,000,000 BTUs and 95% of the 92,000 BTUs per gallon for propane you would need (50,000,000 / 87,000) 575 gallons of fuel.
To be sure your numbers are in the ballpark, how many sq ft, is it well insulated and air sealed, and what is the efficiency of the furnace or boiler?
Bud
Dan,
As a ballpark figure, take the design heat loss and divide by design temp difference, this gives you btu’s per °F per hour. Multiply this number by 24 hours (per day) and then multiply by heating degree days to get yearly BTU/s.
Next divide that by (91,000 btu’s/gallon * furnace efficiency)
So 51,000/48 degrees = 1062 btu/°*hr*24 =25,500 btu/°*day * 3900Degree days = 99.5 X10^6 BTU’s per year.
Note that this does not account for solar gain, occupant behavior etc.
99.5X10^6 BTU/s (91000*.93) = 1175 gallons of propane
Generally speaking manual J overestimates heat loss so your actual usage will be likely be less.
The issues is you size for PEAK, not AVERAGE conditions. This will vary by location, some areas have more constant weather than others. I have NEVER seen a heating system that was too small to do the job, I have seen MANY that are 3X the required size.
The problem with oversizing forced air systems is a new 90%+ furnace requires 50% more airflow for the same BTU when compared to the old pilot light furnaces. Old furnaces were good with 100CFM per 10,000BTU, 90%+ furnaces need 150CFM per 10,000BTU. 80% Furnaces 130CFM per 10,000BTU, but they are mostly used in the south where AC typically sets the minimum airflow requirement.
NEVER install a furnace larger than the ductwork can handle, you won't get the rated capacity due to high limit trips. It also will kill a heat exchanger fairly quick. Gone are the days of virtually indestructible cast iron heat exchangers, the new ones are much easier to crack.
Hi eric,
Before we can reverse engineer that peak heat loss number from the manual J (MJ) report we need to know what they included when calculating it. MJ is noted for over-estimating the appliance size, so has the 51,000 btus/hr number already been inflated (I suspect so) or is the HVAC tech supposed to select a system 25% larger (I suppose not). If the latter we would be blaming the techs for over sizing and not MJ.
In addition, the 51,000 number certainly has been adjusted for internal gains, so before we work that number backwards we would need to add back those internal gains. Between the plus 15% for internal gains and some potentially unknown over sizing adjustment I would have little confidence in working backwards as you outlined.
@Dan: Your question is a bit unclear as it sounds like you are asking for the fuel consumption to reach the set point while experiencing the outside design temperature. In most cases we only look at the total fuel consumption per year. Let us know what you are looking for.
Bud
Bud,
The question was how to translate- from design heat loss calculated by using ACCA manual J protocol to a yearly fuel usage. I outlined a crude way to do it. Is it perfect? Nope. There are a lot of caveats to it. Generally speaking when you do it and then also compare it to actual fuel usage, actual fuel usage is lower.
You are asking how can you trust the person who did the MJ8 calculations did a good job and used correct assumptions? I CAN’T ANSWER THAT. It is certainly easy to do a bad job of it.
Cheers,
Eric
Hi again,
Just to be clear, I wasn't questioning the person who did the MJ work, but I was questioning MJ itself as I do not know if or where those factors mentioned are included.
There's another distinction between loss at design temp and total heat load. If they oversize the appliance by overstating the size requirement at design temp that has only a small effect on the total fuel consumption. Your rough number of 1175 gallons might be overstated by 20%.
Do you know (I don't) what/how MJ calculates their outputs?
I'm well aware of all of the equations involved, just relating to the often stated over sizing that results from MJ and trying to get his numbers closer to the real world.
Bud
I use Wrightsoft to design HVAC systems, which is base on ACCA Manual 'J' heat load data.
My experience over the last 40 years has not been that over-sizing results from people depending on any Manual 'J' calculation rather, people use their thumbs and add on to that figure.
The parameters are extensive so an experienced designer should be employed where the loads really matter. No positive results will come from over-sizing any HVAC appliance.
Hi Morgan, I have admitted my lack of knowledge concerning Manual J and even worse, I approach heat loss from an energy auditor perspective, somewhat different from the HVAC approach. Example, we have no need to break down heat loss to room by room numbers.
But I have also read (no source at hand) that when REMRate, Treat, and Energy Gauge were compared to MJ that the results illustrated the oversizing that is often seen with MJ. This comparison didn't seem to involve "rule of thumb" issues, although all heat loss calculations are vulnerable to this.
It has been some time, but I will see if I can locate that article and double check my memory. I seen to need to do that more and more often these days. If anyone else recalls that comparison please post.
Bud
Bud,
I think these 2 articles will help illustrate the difficulties of modeling energy usage.
http://www.nrel.gov/docs/fy11osti/51603.pdf
http://www.sustainablybuilt.com/SB_Modeling_Software_Presentation.pdf
The first speaks to the common practice of manipulating inputs to add a “Safety Factor” to the calculations.
The second shows real world data from different software approaches to modeling heat loss and gain.
In my own very limited experience, I place the most trust in calculating the information from actual fuel and electrical usage if that information is available. Of course this is highly dependant on thermostat settings, and occupant behavior, and the efficiency of existing equipment and ductwork so there is quite a bit of variability in that also. I use REM Design for determining Heat loss, yearly consumption, DHW, and utility bill disaggregation etc. What worries me about it is that it is a bit of a black box. I have convinced myself it is OK by comparing it to a spreadsheet I used to use for calculating heat loss , solar gain and annual fuel usage. Which I admit is a dumb way to decide it is ok.
Generally I find that with REM Design If I measure infiltration and duct leakage, mechanical ventilation, and carefully look at insulation levels, and can figure out what they are using for windows, I can get reasonable results. I arbitrarily define reasonable as within 20% of values calculated from actual fuel usage. In all cases so far actual usage is lower than calculated usage by ~20% I think that there are honestly too many independent variables to do better than that.
I also think that for our purposes, this is probably good enough- on the heating and cooling end of things, the important part is to quit over sizing things by 300%, when you get down to over sizing by 10-30% which manual J tends to do when applied correctly, there is little performance penalty.
In terms of modeling what is cost effective in terms of energy efficiency measures, I model based on calculated values and then adjust based on actual usage so it is more relevant to the actual occupant.
Where I worry most about my numbers is when looking at the effectiveness of heat pumps. There it seems to me that I should be using binned hourly temp info to determine how often any resistance heating elements will be used. I do this separately and crudely in an excel spreadsheet.
Cheers,
Eric
"Actual fuel and electrical usage" is anecdotal at best. Bin is the thing and ASHRAE/Manual 'J' have the average weather data relevant to the building site.
The operate can cheat the numbers to add what ever he likes, but all the programs I have use in the past warn against it.
Burdick says the first step in HVAC design is the load calculation. In residential application this the step most often never taken. Thus the tradition of over-sizing HVAC appliances persists.
This is the "how-to" for Manual 'J' and if followed will result very accurate calculations in both new and old building HVAC work. In my own case I have on several occasions built to design and had customers confirm that the thermostats dropped below IDT when ODT was exceeded.
Zero over-sizing using Manual 'J' 8 as designed.
The only thing to improve a careful Manual 'J' is a blower door to more accurately predict the air changes.
Home Energy Pros was founded by the developers of Home Energy Saver Pro (sponsored by the U.S. Department of Energy,) and brought to you in partnership with Home Energy magazine. Home Energy Pros is sponsored by the Better Buildings Residential Network.
© 2015 Created by Lawrence Berkeley National Laboratory.