Like all research projects mine has an extensive list of issues I would have liked to of explored further (given more time). For me the two main ones were:
- testing my prototype in situ over a heating season to assess the energy consumption of homes
- during this period test whether any behavioural interventions implemented resulted in energy savings
One of the main drawbacks of my research is the fact that my prototype was never implemented in a real world scenario. This was due to three reasons; my technical programming ability, the availability of suitably similar houses that would be prepared to trial a system and time to measure the energy consumption, if I ever got over the first two hurdles.
In an ideal world I would have developed the system, written a protocol to get the heating control to work with a boiler effectively. Then found 20 identical homes with willing occupants to test the system and give me feedback which I could monitor their energy consumption for 1-2 heating seasons and then written it all up. Simples!
Of course that isn’t what happened. I managed to generate modelled results and compare them to real-world observations which have some indicative savings of 14.5-15% on hear energy consumption annually. But as any modeller will tell you what you predict doesn’t match the actual performance. People interact with their buildings and products in unexpected ways, for me this is the most interesting bit but more on that aspect in a future post.
Secondly features, which may encourage changes in user behaviour, were not incorporated in the prototype. Such behavioural interventions might have included; feedback on participants heat consumption, comparison with other participants’ consumption and ‘nudges’ when a particularly high temperature or long duration of heating is observed.
Any interventions implemented would aim to provide feedback to engage users and to steer their behaviour towards greater energy efficiency. For example, in the summary screen if a particularly high temperature was set or the heating continued late into the night the user could be warned. This could provide an indication to users that they will consume a large amount of energy prior to doing so.
The implementation of any such intervention would have had to be done after the baseline energy savings, relating to the improved controls had been established. Yet it is thought that the combination of the behavioural steers, feedback at the point of control and an easily programmable system may result in greater savings than improvements in usability alone.
Introducing these features across a series of updates could also allow comparison of the energy impact of different features or combinations of features. One consideration with the introduction of any such features would be not to overwhelm the user. Therefore, the clarity, relevance and volume of information provided on the interface would be of critical importance.
Still the main barrier is to get people to engage with their controls in the first place. Sustaining this engagement is even more difficult, especially without using incentives (see van Dam, Bakker and van Hal, 2011). I was questioned about increasing engagement with controls at my viva, which while a very valid question is a very difficult one to answer. If anyone knows how then please let me know.
While it is clear improving the usability of our building controls could enable more successful interactions with them, it is doubtful this alone will change our practices and behaviours. Changing our behaviour could result in greater energy savings than improving usability alone, however I don’t think you can enable these changes without more usable control systems. So if anyone is looking for a research project, you might find one here.