NatHERS specifications can help boost the energy efficiency of buildings, but some simple possibilities are left untapped. Some pundits have said “all energy models are wrong, but some are useful.”
The article Why Don’t Green Buildings Live Up to Hype on Energy Efficiency?, from a Yale publication, brings that quip to mind. The article details reasons for failing to reach performance targets, and overly optimistic energy models are one reason.
But let’s take a step back from our computers and remember what we’re trying to do: save money, avoid climate change, and make healthier, more durable buildings. The Department of Environment and Energy can initiate some simple changes to NatHERS software that would deliver better building outcomes at very low cost.
Let’s just focus on energy loss through uncontrolled air leakage. CSIRO’s AccuRate software is the benchmark for NatHERS. It calculates air movement based on outside temperature and wind for every room in the house, for every hour of the year. That is an enormous calculation, and for years they have refined their model for better accuracy, precision, and speed. But is this level of detail in a computer model actually useful for changing the way buildings are built?
In AccuRate, you can input the number of vented downlights and unsealed fans and even specify “small,” “medium,” and “large” gaps around windows. In Rating mode, you can only assume “medium” gaps but if the building isn’t built yet, how would you know how big the gap will be?
The process of entering all these leaks is time consuming, tedious and ultimately costly for the energy assessor. So wouldn’t you always assume a better level of sealing if it’s easier, faster and going to help your client achieve their NatHERS rating? And wouldn’t you want to assume a builder would do a good job of sealing?
In real houses there are always going to be leaks that don’t fit into the available energy model categories. Unfortunately no one will ever compare such specifically detailed assumptions to what is actually built. This level of detail is quickly lost in the murky swamp of compliance requirements making it truly impossible to ever verify that what is assumed in the model is actually built.
At a recent NatHERS stakeholder workshop in Melbourne there was a strong theme emerging that NatHERS should better align the design intent with as-built outcomes. In order to improve, we need to be able to both calculate it (virtual world) and measure/verify what was calculated (real world).
Decades ago in the US, they admitted that infiltration is an incredibly difficult thing to model, even for one house, and that getting carried away with it is not helpful. Introducing a simpler method, M.H. Sherman and D.T. Grimsrud of the Lawrence Berkeley Laboratory cited “a model that sacrifices accuracy for versatility and simplicity. Rather than predicting accurately the weather induced infiltration of a particular structure, the model is designed to calculate the infiltration of a general structure.”
In the 1980s, they created an “n-factor” map to convert a number from a simple blower door air leakage test to an estimate of annual average air exchange under ambient conditions, based on the size of the building, its location, and its exposure to wind. While an estimate may not be correct for any particular house, it may be more useful. Australia could make a similar map for itself, or even integrate the Sherman and Grimsrud methodology into NatHERS software allowing for a direct qE50 input into AccuRate.
A simple blower door test target that goes into an energy model also fits better into the building code. A majority of the United States now have a code provision for blower door testing and the number of States grows each year. Having a target number to point to in the energy model would align with the Australian Building Code Board quantification project.
Any qualified technician would tell you that a blower door test is a gross estimate and basic measure of build quality. While it may be possible, even admirable, to create an energy model that accurately predicts the airflow through a house in detail, is it useful? Does it have any relevance to or influence on construction practice? Take a look at the graphic below and I’ll outline some steps that could be taken to improve the relevance of energy models to real-world practice.
- Regulators: Define code-required building sealing with a hard number for a leakage target. Defining a number in the software and a number in the building code allows you to compare what is promised to what is required and then to what is built. Now certifiers can better help enforce building code by looking at a single number on the test report, not narrative, subjectively-evaluated language or a long checklist of leakage points. Right now, the poorly defined requirements in the NCC are incredibly vague and difficult to verify.
- CSIRO: Take their existing energy model and have it report an estimated air exchange rate at 50 Pascals (qE50), a standard metric for comparison. Having this number displayed on the NatHERS Universal Certificate would allow a builder, homeowner or certifier to compare what is actually built with what was assumed in the energy model. They can then answer the question, “Was I promised something better?” The Improving Australian Housing Envelope Integrity report by the AIRAH Special Technical Group on Building Physics calls attention to the fact that much of the time, energy models assume a much greater level of building sealing than is actually achieved. Energy modelers understandably assume that a building will be built with code-required sealing, but unfortunately this is not a safe assumption. CSIRO reported that the air tightness of average Australian houses is often poor, much worse than assumed by standard models. Having the software report what is currently being assumed would be easy and useful because we can learn from it.
- CSIRO: Next, they should create an input to the software so that it can accept a projected or actual blower door test number straight into it. This helps close the loop and set expectations for build quality. It would also set the industry on a path towards innovation. Performance-based metrics provide the flexibility to use a range of new and innovative construction methods or materials to achieve an equivalent outcome.
This software capability would also be a necessary input for assessments of existing houses. Often the most cost-effective thing to do in an extremely leaky house is patch the leaks, and the model can show the cost-benefit to prove it.
In the future, regulators could also allow builders to true-up their energy model predictions with real-world results. If a builder promises a house with basic sealing (qE50 = 9) but does even better than that (qE50 = 7), the model could reflect that the better performance from sealing means they have a little more flexibility with other energy measures. Right now, you get no credit for doing better than code requires and no penalty for doing worse.
The first task above is currently under discussion within the ABCB under recommendation 14 of the National Energy Efficiency Building Project. The second two tasks are for CSIRO, and I know they can do it, because they have some pretty hefty brainpower already working on the question of air leakage. But they need direction from above and pressure from outside to prepare the NatHERS software for 2019 Building Code.
Clear direction from regulators and policymakers can allow CSIRO to confidently implement the simple changes that would make the air sealing code requirement much more enforceable. With energy models, the science should be used to guide policy, then the policy can guide practice.