I read an interesting article yesterday by Yao et al (in Energy and Buildings) about how to create domestic load profiles. This paper is properly the most direct precursor to what I’m trying to do with my modelling. The main behavioural response to PV that I will be examining is a temporal shift and so if I’m going to simulate that change I need to be able to simulate consumption on a comparable time scale. Anyway the paper was interesting but left me wondering:
- How exactly does the model decide what appliances get used at what time during the day? The most useful part and as always happens, it’s not explained.
- Yao et al use per capita appliance loads to simulate specific households. I wonder if these are applicable in an agent-based framework. My hesitation is that they are average figures and I’m building up from the bottom.
So all this is very promising but it raised an interesting point. If I’m going to build a model to simulate a behavioural change, I first need to verify that it can simulate energy use in a base condition. This means generating a profile and then comparing it to Electricity Association data. However these data include hot water and domestic heating loads, which I’d initially tried to avoid to save on coding. But it looks like I’m going to have to model DHW, space heating, and lights and appliances to verify that the total profile is right. Only at this stage can I then reasonably play around with the lights and appliances stuff thinking that I’m on the right track.
