R. Choudhary, D. Foures, Y. Heo, Y. Kiguchi, G. Vettigli. A data driven model for TOU customers behaviourl adaptation. Work in progress.... [PDF]
Abstract: Time-of-use (TOU) tariffs, which have multiple rates during the day, have been designed to reduce the peak demand, and are recognised as novel demand side management technique for the residential sector. TOU tariffs will become available alongside the mass rollout of smart meter installation.
This paper proposes a novel methodology to quantify the shifted consumption after the adaptation of a TOU tariff in residential households. The methodology essentially relies on individual historical half hourly electricity consumption data as collected by smart meters, and therefore harvests the potential of this newly available dataset.
This model has two fundamental pillars. Load flexibility is derived from load probability density at a given time. User adaptation is predicted by maximising a utility function taking into account both the user's financial savings and efforts associated with behavioral change.
The following charts summarise some of the results of our study about behavioural adaptation to TOU tariffs. Each chart shows the average behaviour of a user and the adapted behaviour predicted by our model. The users are drawn from four different groups:
A. Already adapted and low adaptation potential.
B. Already adapted but large adaptation potential.
C. Not adapted and low adaptation potential.
D. Not adapted but large adaptation potential.
You can select a different TOU tariff from the following list and see how the users adapt to it:
In the following two charts we have:
(Leftmost) The projection of the users in the space represented by their flexibility and adaptation with respect to the tariff where the on-peak time is between 8am and 10pm. Points in red represent the users in the invidual charts above.
(Rightmost) The usage fingerprint of the average user. Black is associated with a low probability density while yellow represents a high probability density.