When you decide to utilise the deemed-to-satisfy route for residential developments, you’ll end up producing a NatHERS (Nationwide House Energy Rating Scheme) model.

Given that it is a building envelope assessment, as part of the process you are required to enter information about the dwelling (be it a single-occupancy house or apartment block) including the geometry, envelope materials and thermal performance.

The energy modeller is given a fair bit of flexibility when it comes to the opaque constructions. However, adding windows presents some limitations. Instead of allowing the energy modeller to simply enter the performance metrics of the windows (as you would in all other energy modelling software on the market) you must pick a product from a database.

There are two databases that you can call on – the defaults and the customs. The defaults consist of 88 window options, developed from the 75th percentile of products available on the Australian market. The database is rather limited, but useful for early stages of design.


The ‘custom’ window library (don’t ask me why it’s called ‘custom’ as there’s nothing ‘custom’ about it from what I can tell; these are off-the-shelf products) is a list of certain windows that are actually on the WERS database ( So far, so good - these systems have been tested based on the required methodology. The list within NatHERS software indicates their total system U-value and total SHGC.

Using either of these databases certainly presents problems when you’re using bespoke, custom (actual custom) systems, as is so often the case for the big multi-residential towers that are popping up left, right and centre throughout Australia. The issues are predominantly due to the severe lack of transparency of the WERS database, as it’s ultimately unclear what’s actually been tested to create the certified product.

So NatHERS energy modellers are presented with a U-value and a SHGC for each window system. You would think that products with similar U-values/SHGCs would produce similar NatHERS results.

Well, I’m afraid you’d be wrong!

Below is a graph of NatHERS star rating results for an apartment in Melbourne that was part of a recent development. The U-value is at 2.8W/m2K (convention across the world for system U-values is to one decimal place, WERS provides to two decimal places) and the SHGC is 0.28 +/- 10 per cent (a range of SHGC should always be provided; I’ve taken 10 per cent in this case).


The results clearly show the vast variation between the different products, even determining whether the apartment passes or fails. A variation of 0.8 stars between a U-value of 2.78W/m2K SHGC 0.29 and a U-value of 2.75W/m2K 0.30? How is this possible when these values are in essence the same? Clearly, there is some trickery going on.

My frustration has encouraged investigations which have lead me to believe that the missing factor is the air infiltration assumption of the WERS database (i.e. how leaky the windows are). Now, all window systems have to achieve an air leakage rate which meet Australian Standards, and this requirement varies depending on whether the development is low-rise or high-rise (high-rise requirements are more onerous/less leaky).

Apart from the fact that we don’t air pressure test buildings in Australia (a highly contentious issue and currently one of much debate), this leaves us energy modellers with truly custom window systems in their design in a bit of a pickle. Do you rort the system and take advantage of these occasional magical glass products which result in a pass mark?

I have a lot of respect for the CSIRO, and clearly a huge amount of work and effort has been put into the NatHERS tool, but I think the process is unnecessarily complex for what it is meant to be doing, particularly if it’s not clear what’s actually going on, leaving energy modellers in the dark yet again. The fact that WERS data and certificates are not freely available makes it a rather useless database as you have no idea what the framing details of window products are that make up that certified product, and whether it’s appropriate for your project.

NatHERS modelling as a black-box calculation is a tiresome process as it is (not to mention the Cert IV shambles), and until these issues get sorted out, with more transparency around the whole process, the tool will continue to lose support from its energy modellers. In the meantime, it’s always worth remembering that the NatHERS pathway is a deemed-to-satisfy solution and instead, alternative verification pathways can typically be utilised in its place.

  • Jessica, I very much agree with your view that more transparency would be enormously beneficial. I have similar comments on the infiltration modelling in the NatHERS software. Are you in contact with anyone at CSIRO about possible revisions over the coming years? Have you heard of the AIRAH group getting together to look at building envelope issues? If not, I'll send you some info.

  • Jessica, thanks for the great article, I hope someone from CSIRO/NatHERS admin reads it!

    I'd just add there's also the underlying fact that these U-values (to my knowledge) are derived from WINDOW testing of a standard window size, so are already inaccurate for any window that isn't actually of those standard test dimensions! I also think that I remember a webinar from CSIRO's Chenath team discussed that due to different solar absorbtances and window/frame ratios, that even products with the same total U-values will produce different results (this doesn't help when your specific product isn't on the list!) – I think the take home was that the window frame type should at least match the window you are modelling. Is the outlier +0.8 stars in your test results a totally different window frame type to the others?