The problem with energy modeling net-zero homes

What's the biggest problem? Accuracy. A big part of home energy use depends on the occupant. And those darned occupants don't behave as regularly as we'd like them to. We have quite a bit of information what "average" occupants do and how they behave on an average day. The problem is this: how many people do you know who are home every day at the same times, do the same amount of cooking every day, and use the same appliances? Compare that to the number of people you know who don't cook anything at home for a week straight and then invite 6 friends over and cook up a storm on the weekend.

At best, energy models usually let us schedule internal gains and occupant activity on an hourly basis on "weekday" and "weekend" schedules. So, you have 4 people in the house, they go to school or work, they cook, they watch TV, etc. You have a typical weather year, which we may or may not be actually having. Most of the estimates that I've seen are that a typical home's energy use can vary by about 20-30% based on occupant behavior. When I model homes that are in the "typical" or somewhat less energy than typical range, and when I get follow-up data for comparison, I'm usually pretty close. Often closer than 20-30%. But really, I consider within 20% to be pretty good considering the degree to which the input data is an estimate.

"Net-zero" homes pose a couple of distinct problems. First, the occupants are almost never ordinary. Let's face it - most of us don't live in net-zero homes because we either just plan can't afford it yet, or haven't made it a big enough priority in our lives to figure out how to afford it. Unless we're quite wealthy or have a very unusual site, it always makes sense to spend money and effort on extreme conservation before adding more energy generation to get to net-zero. So, the internal heat loads for net-zero houses aren't going to be well predicted based on "average" occupants. The second major problem is that the lack of precision with which we know how occupants behave becomes much more important when you get into very low net energy situations. To put some numbers on it, let's say that a given homeowner's behavior uses 10% more energy than the average person. If that person lives in a house that is just built to code (HERS rating of 100), then it really functions at 110% (HERS rating of 110) and we're not too far off. But if you take that same person and put them in a house that's close to net zero (say it has a HERS of 5), then it really functions more like a HERS of 15 and we're off by a factor of 3. What gets even more interesting is that unless this person is Mr. Boring, he probably has weeks where he's a 20 and other weeks when he's out of town and is a 0.

This can start to matter a lot when designers try to adopt a strategy that counts on these internal gains as part of a function of the whole system. For example, if you're counting on the internal gains to replace part of all of the heating system in the house, you need them to be there. One of my earliest exposures to green residential building was just such a project. What happens when the family that moves in happen to be smaller than average, cooks less than average, some of them go on business trips, and you get a long stretch of cold weather? You sit around and brainstorm ways to make more internal gains: oven cleaning, anyone?

2x4 or 2x6?

My intention is to use this blog to discuss questions I am asked all the time. So, the 2x4 vs. 2x6 debate seems like a good opening post.

There are 2 major issues: how much wood is used and the R-value of the wall assembly.

In terms of wood used, if both have the same stud spacing, the 2x6 obviously uses more wood. But if you can go up to 24 o.c. for the 2x6, it's about the same volume of wood used. (About 1/3 less due to spacing, but 1/3 more due to thickness). Some drywall installers complain, but I've seen it work just fine.

In terms of R-value, 2x4 will get you between R-13 and R-15 for most cavity insulations, and 2x6 can give you R-19 to 23. From an insulation standpoint, you're obviously better off with the higher R-value. But it turns out that an R-13 wall with R-5 exterior sheathing is almost exactly equivalent (in terms of overall assembly R-value) to R-20 in a 2x6 wall. So, you could get almost the same performance. In the past, R-5 exterior sheathing has been tricky for a number of reasons: structurally you have use OSB at the corners or use metal bracing, and siding guys aren't crazy about having to locate the studs to nail into. But there is a new product out from DOW that combines structural sheathing with an R-5 all in one product - eliminating all these problems. Of course nobody's saying you can't also use R-5 sheathing on a 2x6 wall...