Using hard data to improve sustainability
To measure is to know, especially in the energy transition. But, as Laure Itard argues in her inaugural address, we will only truly benefit if we process the measurement data using automated systems. The smart use of big data can prevent energy systems from continuing to underperform.
Data from smart meters, home automation, intelligent domestic appliances and building management systems can combine to produce huge quantities of data. And all of this enables us to make significant improvements to energy efficiency in buildings. “Big data can give us feedback on what we have designed”, argues Itard. “Without measurement data, we are completely unaware of the real performance of systems.”
This, in a nutshell, is what the new chair in Building Energy Epidemiology is all about. In her new position, Itard will be collecting and processing relevant data about energy consumption, conducting research into suitable algorithms and developing knowledge about interactions between building, system and behaviour. This will not only make data-driven monitoring of the performance of systems and buildings possible, but also enable previous policy to be scrutinised. Are the models it was based on accurate?
The big data revolution seen in the last decade opens up new unprecedented possibilities. This is especially the case in the Netherlands, which is further advanced than any other European country when it comes to collecting data about energy consumption for each residential address. Although this may seem strange from a privacy perspective, it offers huge opportunities. Especially if you combine it with targeted measurement campaigns, as Itard did in her E-Common research project. In it, hundreds of homes in Zuid-Holland were fitted with occupancy sensors and meters for air humidity, temperature and CO2. The researchers also assessed the residents’ comfort levels as part of the project. How exactly do they feel? Linked to data from smart meters, the research into methods and algorithms ultimately aims to deliver automated bespoke advice for every individual home in the future.
By adopting this approach, the chair is borrowing methods from the field of epidemiology – the healthcare specialism that attempts to stem the spread of diseases by analysing data from large groups of people. For example, the chair in Building Energy Epidemiology uses the annual energy consumption data for all Dutch households published by Statistics Netherlands (CBS). Another major source of information is the energy labels database that covers hundreds of thousands of addresses. Itard: “Having access to this data as a researcher is invaluable, because it enables you to link together energy consumption, insulation levels and type of household at address level.”
This makes it possible to precisely scrutinise assumptions about the impact of large-scale efforts to improve the sustainability of housing stocks. On several occasions, this has already led to the conclusion that the models are wrong. How can that be? First of all, it is because people’s behaviour was assumed to be different. For example, the models assume an average household temperature of 18°C, whereas it is actually lower in older properties and probably higher in modern, well-insulated houses. Itard: “It seems to be a question of people simply getting used to a certain level of comfort.” Although new houses are still around 30% more energy-efficient, that is still twice as low as what was assumed. This discrepancy also has repercussions for the time it takes for sustainability investments to pay for themselves.
Another reason why the models can often be inaccurate is because the characteristics of the building structure are unknown. For example, it is not clear who has or has not made efforts to insulate their home. Without that knowledge, energy consumption is difficult to predict. “Based on data, sensors and machine learning, we are exploring how we can automatically identify the status of such things as insulation, air infiltration and other key data about a home.”
If this proves successful, the next step starts to become possible: improving the operation of what are increasingly complex energy systems. Calculations suggest that it may be possible to improve performance by between 20 and 30%. “Our data can provide useful feedback for the designers”, says Itard. “Based on the data, they can build better systems in buildings and monitor the results in real time. That would make real progress possible.”
On Friday 7 February, professor Building Energy Epidemiology Laure Itard gives her inaugural address ‘Energy analytics for sustainable buildings’. More information can be found here.