Baz Iyer
May 23, 2024
HEM is a building energy modelling engine developed with the backing of the UK Government, and set to be introduced in 2025 as a major upgrade to SAP. The change in name from SAP to HEM reflects that it has been built ground-up - with fundamental differences to SAP. These changes may make it a more accurate and precise tool for modelling buildings than SAP or PHPP models - and enables eventual convergence with dynamic energy models.
In the coming months, we'll explore HEM and these differences in depth. We’ll also share insights from software we’re building to interface with HEM. HEM is still in development, so we expect we'll be learning alongside others in the industry - so invite your collaboration and views too.
Good design separates buildings that experience extreme operational costs or poor indoor environments, and those that are comfortable at low cost to occupants. HEM was released for consultation alongside the Future Homes Standard, but its applications extend beyond new build - it also has important applications to the huge stock of homes we must retrofit.
The majority of retrofits must occur in the “able-to-pay” market, and these homeowners will only make significant improvements out of choice. Property owners must see a clear benefit from improvements to invest in retrofits, or justify the use of financing. To maximise benefit, retrofits must carefully balance energy efficiency and electrification, airtightness and ventilation, operating cost impacts with budget limitations.
Energy modelling - and scenario evaluation using simulation - is the key to meeting the needs of property owners, within their constraints.
Energy model choice matters. High quality methodologies will make building baselines easier to calibrate with sensors (instead of variances being driven by imprecise modelling). Better modelling methodologies are more useful tools to compare improvement options, and can make it easier to detect discrepancies driven by operational issues to ensure higher quality outcomes.
Building energy models use “time steps” to simulate energy use, gains and losses. Dynamic energy models simulate the physics of a building with half hourly time steps, to approximate the interactions between a building’s environmental energy gain and loss, and its occupation patterns across each day of the year.
SAP and PHPP use monthly time steps, as they are approximate models intended for field use. This low granularity, along with other assumptions forced by the limitations of these models, makes these models less accurate. This is because if they need to model an event that changes in nature across the day, they must do so with arbitrary assumptions - creating imprecision.
In contrast, HEM has a granular half-hourly modelling engine, the same as dynamic models. This is ~1500x the granularity of SAP or PHPP! While HEM is still a static model and uses static assumptions - we’ll cover the impact of this in a future newsletter - granular time steps avoid opaque averaging and provide the opportunity for precise modelling of intraday variables such as occupation patterns, heating controls and tariffs.
While HEM provides a strong foundation, the driver of its success we are most excited by is the move to open source. This is what will be key to fixing gaps identified by the HEM team in their validation studies, and in consultation responses.
We’ve met several technology teams that over the past years have built their own static models from scratch, improving on SAP. These models provide better answers than SAP, but the resulting innovations fall outside the boundary of regulated work. This reveals a weakness in both SAP and PHPP - these are tightly controlled models that change slowly, and therefore contain outdated and inaccurate assumptions. Further, not all their mechanics are visible or queryable by users.
HEM makes this grey market effort unnecessary. It is explicitly open source, and the development team have mentioned they will receive contributions following the consultation. While the exact format of these contributions is not yet clear - the assumptions used by HEM can be clearly understood, and modified outside the model. While HEM can be improved, this is acknowledged, clear and visible to potential collaborators. This may help HEM quickly develop improvements that close the remaining gap to measured performance.
In this issue, we've touched upon two aspects of HEM that excite us. In the weeks and months ahead, we'll delve deeper into the responses to the HEM consultation, validation studies (and our own analysis), and the tangible benefits of HEM's improved outputs. We’ll also describe how our software builds on HEM, and how you can try it. Through this process we hope to not only educate you about this important standards change, but give you the knowledge to better appreciate the science of our homes.
We want to hear from you! If there's a specific topic you'd like us to cover or if you want to share feedback, please reach out to us.