BEopt results
Tried using BEopt for my house and the results were a shock, somewhat. I live in Zone 4 in Maryland and was concentrating on the things I know will be occurring in the near future, attic insulation and wall retrofit insulation.
According to BEopt, the biggest insulation bang for the buck was the wall insulation. Putting in R13 of cellulose with Grade 3 installation produced a drop of about 500 therms per year. (I assumed Grade 3 installation was appropriate, since 1) it will be a retrofit and 2) I have a severe lack of faith in contractors.)
When strictly comparing attic insulation, it projected my Natural Gas therms/year would drop from 1358 with no insulation to 1104 with R30, to 1092 with R49 and to 1090 with R60.
I am aware of diminishing returns, but I assumed: 1) I would have received a far greater benefit from the R30 in the attic than R13 in the walls, and 2) that the jump in attic insulation from well R30 to R60 would have seen more than about $20 per year return. Is it possible I was just not using the program correctly, or are these results what I should have always expected?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Jeff,
It's hard to determine from the information you've given whether you made errors with your program inputs. Certainly, if something smells fishy, it makes sense to go back and double-check the inputs.
That said, it isn't particularly surprising that going from R-30 attic insulation to R-60 doesn't save much in your annual energy bill. (Remember, natural gas is really, really cheap.)
I agree that going from an uninsulated attic to an R-30 attic should save a considerable amount of energy. Whether the 254 therm savings shown by BeOpt is accurate is hard to determine from the information you've given.
Jeff, those results seem within the range of possible; diminishing returns are significant--doubling the insulation only cuts the heat flow in half--but something might be off. Even uninsulated walls have some R-value, due to air films on both surfaces and the R-value of the materials in the assembly.
Do you have the house modeled with an "unfinished attic" on the Geometry page and "unfinished attic" selected on the Options page? If different assemblies are selected on those two pages, the output will be wrong.
What is your average ratio of windows to wall? When the ratio is low (less than 5-10%) the amount of wall insulation has a bigger impact than when there is a lot of glass.
Is your house tall and narrow, or low and wide? A tall, narrow house will see more of an impact from wall insulation.
There is usually both more surface area in walls than attic floors, and empty wall cavities are essentially flues with a stack effect driving air infiltration, which is substantially mitigated by retrofit insulation. It's really no surprise that insulating the walls offers a bigger performance gain than increasing the attic insulation to above R30.
I noticed a huge error in that I had not even put in the window area, which had some effect upon the wall calculation. Who knows if I would have spotted it but for everyone having mentioned that topic, so thanks. It definitely took some practice to know things like not to use "none", rather than the correct option of "uninsulated, vented". Running the optimization again, so I'll hopefully have new stats in a couple hours. :) I think they used the NORAD computer in determining they had an acceptable run time. But, it is free so only a tongue-in-cheek complaint.
I re-ran it with the window area being 50 sqft on each side, the therms for R13 Cellulose dropped from 500 therms to 420 therms. The drop from uninsulated attic to R30 is still about 250 therms as expected. One thing; however, I didn't consider was that for existing I assumed 20 ACH dropping to 8 ACH, so that undoubtedly influenced therm estimates for all of the calculations.
Honestly, I have no idea what my current ACH is because the one time I had a blower door test the guy said he "couldn't get a reading". I believe I wisely assume both that the house leaks like a sieve and that he may not have known how to use the equipment.
I've read at least one contractor's blog who bravely conceded that even a 50% drop in ACH is a great result, so my ACH input likely isn't based upon reality either on the existing or intended ACH of an 85 year old cape. Nonetheless, 8 ACH is what I'll at least ASPIRE to get to, meaning while air sealing I'll think in my head, "wouldn't that be nice".
Jeff, you're right about air changes, it does have a significant impact. I use 7 ACH50 when I don't know existing conditions, but it depends a lot on local practices. I usually model the energy impact of getting to 3, 2, 1, and 0.6 ACH50. Sometimes I'll model 0.4 and 0.25 ACH50, but once we're in that territory it indicates that investing in a PHPP model might be worthwhile.
I happen to be modeling a house in BeOpt this morning, about 1500 ft², on a full foundation. As far as BeOpt is concerned it looks a lot like a cape-style house, though in reality it's modern, with opposing mono-pitch roofs and carefully balanced glazing. It will be a Pretty Good House with well-insulated walls, but I ran a few other options for fun. It will be heated electrically, but I'll show the MMBtu numbers for all energy use for the building to show the effect of different wall insulation on a somewhat-typical house in a cold climate (zone 6, ~7500 HDD):
2x4 walls, no insulation: 51.9 MMBtu
2x4 walls, R-13: 39.9 MMBtu
2x6 walls, R-19: 38.0 MMBtu
2x8 walls, R-26: 36.9 MMBtu
Double stud walls, R-33: 36.0 MMBtu
2x6 walls, R-19, plus R-12 exterior insulation: 35.9 MMBtu
Double stud walls, R-42: 35.9 MMBtu
Double stud walls, R-60: 35.3 MMBtu
If you graph those out, there is a nice smooth curve that is steep from zero to R-20, and flattens off above R-30 or so. I doubt BeOpt takes into account convection in empty stud bays, so the uninsulated walls may perform worse than predicted.
As for air changes, I used 3 ACH50 as a baseline; reducing that to 2 ACH50 saved 5% in energy costs. Getting to 1 ACH50 saved 8.6% compared to the baseline. Getting to 0.6 ACH60 saved 9.6% compared to the baseline. So airtightness does have a significant effect on energy use, with diminishing returns like everything else.
Jeff
When BEopt predicts a net $20.00 savings the equation has some things in it you may not realize.
BEopt is considering the higher cost per square foot of insulation over its useful lifetime and assumes you will finance the added cost at your mortgage interest rate you enter into BEopt.
Also the preprogrammed insulation and install costs per square foot in BEopt are sometimes very different than the costs you will get from your local bids that you can enter into BEopt.
Walta
I used the online CoolCalc to calculate my Manual J numbers, and I compared how those numbers changed with different levels of insulation and air tightness.Then I manually estimated how much the extra insulation would cost versus how much money it would save. Frankly, it looked pretty doubtful that the energy savings from exceeding the current mandatory insulation levels would pay for the extra costs involved.
Walter, it depends on how you use BeOpt. I don't use the Optimization mode, because I don't trust their cost numbers or know my builders' costs well enough. I just run it in Design mode, which gives straightforward output, which I then put into my own spreadsheet to analyze. While my approach does not fully utilize BeOpt's potential, and maybe someday I'll have enough of my own data to trust its math, it's still a powerful (and free) tool for comparing options.