Updated by Adam Stenftenagel 7.13.17

Most residential modeling tools offer some form of “true up” feature to adjust the energy model’s results to match the historical utility bills, so that the resulting savings numbers are more accurate. It is a common reality that traditional energy modeling tools over predict energy savings by as much as 300%, especially on older homes. You first create a model of the home and then look at how far you off you are from the bills and either go back into the model and adjust individual settings or you adjust the overall energy consumption of the base home to get it to match the bills. This “true up” process regularly misses the nuances of existing homes — which are incredibly complex from a heat and moisture flow standpoint.

At Snugg Pro, we take a totally different approach. Instead of asking what R-value the insulation in a specific location is, we ask much more simple questions with a level of uncertainty. For example, we ask, “choose a range of insulation thickness that best matches the observed thickness”, or “how well insulated are the walls?” We use these simple questions so that we can give our modeling engine some flexibility in choosing the best or most accurate “effective” R-value of the existing material. But wait, you say, why is the modeling engine choosing the R-value. How does it know better than me?

Before I tell you how Snugg Pro works, let’s first rewind back to your BPI / RESNET training and talk about heat flow. (If you’re all up to speed on this, you can skip to the next section called Snugg Pro Auto-Calibration). You all remember that there are three methods for heat transfer in a building: convection, radiation, and conduction.

Convection is how heat moves into and out of a building through air movement. This one is pretty simple to measure through a blower door test. Yes, there are still some assumptions about how hard the wind blows predominantly and how much stack effect contributes to the mix, etc, but having that highly precise blower door test takes a lot of the guesswork out of the convection side of things.

Radiation is where heat (typically from the sun) transfers through electromagnetic waves. It’s not simple to measure by any means, but at least its effects are constrained almost entirely to windows. So, by counting up the percent of each wall orientation that is glazing and calling out the window type and the overhang above it, we can get a pretty decent guess of the radiant gains for the whole house.

Now, conduction is a whole different animal. Conduction is heat transfer through contact and is calculated in BTU/hr by the formula: U A ∆T or

U-value (inverse of R-value) x surface Area x Difference in Temperature between inside and outside.

Ok, so how do we figure out these things? Measuring surface area seems easy enough, and according to RESNET HERS rating rules, you have to measure every surface type down to the nearest 6 inches. They do that because that’s the one thing in a RESNET model that you can very accurately measure, so they take it to the extreme. What else is easy to obtain? Well, the concept of Heating Degree Days (HDD) does a pretty great job of estimating the outside air temperature through the course of a year, so we’re good there.

What about the inside surface temperature of that wall or ceiling? Now this is where it gets interesting. When you shoot a wall or ceiling or floor with your thermal camera, what do you see on the temperature readout? Chaos right? How do you get all of that information into an energy model? The short answer is, you don’t. Most modeling engines are going to take a guess at the interior temperature as being a derivative of the thermostat setpoints, but how in the world can the inside temperature of the master bedroom on the far side of the house come even close to matching the thermostat temp in the hallway? This is one part of energy modeling that’s extremely fuzzy. We just don’t know for sure.

Now, how about that R-value? Can’t we just rely on those fancy tests from Oak Ridge National Labs that rate the insulation? The test results are what they use to label the package, right? Well, this one isn’t that simple either. Oak Ridge performs tests of insulation in perfect conditions that are consistent among all insulation types that they test. They don’t test it with wind blowing through or with convection loops or with large temperature differentials. So how do we really know if that insulation is performing at its rated capacity in a home?

You could spend a day in the house testing every single surface area with a fancy R-value tester from Testo (the test, by the way, only works well if there’s a large difference in indoor vs. outdoor temperature). I personally used this R-value tester a while back on a home where I knew the original construction (typical for 1993 in Indiana): a standard R-19 fiberglass batt in a 2x6 cavity with the plastic on the inside of the studs, OSB sheathing, no tyvek and then wood siding. I convinced the owner to use spray foam insulation when putting on an addition many years later. The Oak Ridge rated R-value of the spray foam was R-19 for the cavity as well. So what did my fancy R-value tester tell me on this cloudy 20ºF day with a slight 5 mph wind? The spray foam wall came out at R-17, pretty close to its rated value. The fiberglass wall? Wait for it… are you ready? R-5.

Sooo… in the end when building a standard physics based energy model, we’re really guessing in the wind on a lot of important measurements. So how in the world do we create an accurate energy model without spending days at a house? Let’s go back to those simple questions in our modeling engine.

This is where Snugg Pro really shines through. If you are able to provide our software 12 months of historical utility bills, we can figure out exactly how much energy that home consumes given the outside temperature. We then utilize that information along with the answers you gave to our simple questions to best match where the heat loss/gains are coming from. Even if you don't have detailed bills or have delivery fuels like propane or fuel oil where it's hard to tally up the total annual consumption, you can use the simple bills in Snugg Pro to help true up the model as well.

We’ll run up to 1000 permutations of the model, checking each possible energy value (R-values, equipment efficiencies, thermostat setpoints, etc) in the corresponding range that you specified and test it against the bills. This way, we pick the best fit for every energy value and the resulting whole home energy model has base usage numbers that match the weather normalized historical utility bills. Presto! A super accurate representation of the home as it performs today without having to take detailed measurements of surface temperatures and R-values.

Finally, we take the improvement numbers that we recommend or that you override in the refine screen and we calculate the difference for each recommendation. This gives a great picture of how much money the homeowner might save if they did the work you are recommending.