Smart homes, towards a more conscious energy consumption together'

Lead Research Organisation: Imperial College London
Department Name: Imperial College Business School

Abstract

During the PhD I will undertake cutting edge research in my chosen fields of energy and behavioural economics.

This research project will look at energy consumption and specifically on Feedback interventions in hom eenergy usage (water/electricity/heating), Wi-Fi plugs (focussed on east London), Pinnacle power experiments and operationalizing a mediation model for feedback intervention.

Wi-Fi plugs enable users to control their appliances remotely by a) reading how much they are consuming in real time, and b) switching them on and off on demand. Since this piece of equipment is relatively cheap and easily implemented in households, deploying a fleet of Wi-Fi plugs allows researchers to monitor a portfolio of appliances remotely and implement micro ancillary services. In other words, should residents regulate appliances via their Wi-Fi plugs, excessive energy demand can be shifted from peak time to off-peak hours (when electricity typically costs less) by switching off some of the plugs for limited periods of time, thus incentivizing residents to be more flexible.
The potential for large-scale systems of Wi-Fi plugs is tremendous. For example, should plugs be deployed on all domestic fridges in London, which consume between 150W to 400W on average then a large coal-fired power station (2.4GW ) could be displaced during peak hours. Still, the question remains how the population can be incentivized to use Wi-Fi plugs with their appliances that remain fairly flexible during peak hours and allow their remote control.
Empirical papers have often investigated the direct effect of feedback intervention on energy consumption (Delmas et al., 2013) and control for demographics that are expected to explain variance between individual. Still, Shippee (1980) suggested that the effect of feedback intervention is mediated by the 'individual responsibility', i.e. our perception of information and our individual analysis of the consequences of our behaviour, while Faiers et al. (2013) presented evidence that demographics often fail to explain systematic differences between case studies. In common practice, the moderation relationship has been implemented due to the large availability of demographics while the theoretical mediation model should be more appropriate (cf. figure 1). This can be explained in 2 ways: before the big data phenomenon, there was no real-time data collection and processing, and neither was there any mean to maintain a close relationship to consumers in order to capture variations in their individual responsibility (other than pre/post evaluation surveys).

Smart systems that deploy both smart meters and smart applications can remedy this situation, by providing researchers with an online platform that can convey real time data and feedback to consumers, and track which information they access. For example, the number of 'likes' and the time spent on pages related to the environmental impact of energy consumption can be used as a proxy for consumers' environmental awareness. One step further, a smart application (cf. Figure 3) would provide researchers with fast response tool to deploy gamification processes in order to re-evaluate the extent to which online ranking can trigger changes in behaviours. In other words, the evolution of consumers' individual responsibility could be tracked over time using simple proxies and would enable the implementation of a more robust, and theoretically more appropriate, mediation framework for feedback intervention.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509486/1 01/10/2016 30/09/2021
1803446 Studentship EP/N509486/1 01/10/2016 31/03/2020 Quentin Coutellier
 
Description 1st project: We report first results from a large-scale randomized controlled trial of various forms of energy consumption feedback facilitated by smart meters and smart phone apps. Nearly 40,000 customers of a large energy retailer in the UK were exposed to either basic feedback apps---i.e. simply giving consumers access to accurate monthly energy consumption information---or more advanced feedback involving peer group comparisons as well as disaggregation of total electricity consumption. In our intent-to-treat analysis, we find that more advanced feedback can lead to an average consumption reduction of nearly 4\%. Focusing solely on individuals who comply with treatment by signing in suggests the reduction is up to 12\%.

2nd project: Furthermore, while it is assumed that the mechanism by which descriptive norms work to change behaviour is by motivating individuals to align their behaviour (consumption) with that of their peers, this has never been truly tested and confirmed. An alternative explanation for this effect is simply that the descriptive norm acts as a reference point to highlight the feedback-standard gap. If this is the case, this would mean that people would equally benefit from any form of feedback that helps in benchmarking their personal consumption. The current series of studies therefore seeks to determine whether descriptive norms change behaviour through social motivation or simply through acting as a reference point for consumption, and if the latter is true, whether there is a difference between the different forms of comparison feedback interventions.

To test this first aspect, a framed field experiment will be conducted in a student resident building where water, energy and heating consumption are measured at an individual room level. Participants will be randomly assigned to either a social or prescriptive norm incongruous or congruous group. The incongruous group will receive genuine descriptive norms regarding their consumption coupled with a prescriptive norm that is counter to it. For example, an individual receiving feedback that their consumption is similar to the average consumption of their peers will also receive prescriptive feedback that will suggest that their consumption is more than the objectively "optimal" consumption. Alternatively, an individual receiving feedback that their consumption is more than the average of their peers will receive prescriptive feedback that suggests that their consumption is objectively ideal. Those in the congruous group will therefore receive prescriptive feedback that is congruous with the descriptive feedback.

3rd project: Demand response strategies that rely on individual behavior change have consistently demonstrated that flexibility in energy demand exists, but that there are constraints to behavioral demand shifting such as limited attention. This research aims to introduce residential energy demand flexibility using WiFi-enabled smart plugs, largely bypassing the need for home dwellers' attention and effort altogether. In these field trials comprising 150 subjects, we study the adoption and user interaction of devices for automated dynamic energy demand management. The trials allow us to explore whether and how devices and incentive systems may be designed and deployed to improve load balancing in the centralized energy grid. Such a system would reduce the need for inefficient and often carbon-intensive back-up power generators during periods of peak demand, and increase the possible share of energy supplied by intermittent renewable energy sources.
Exploitation Route Outcomes of 1st project:These effects are concentrated among customers served by a particular installation firm, which displays higher capabilities along a number of metrics, suggesting an important role for firms installing smart technologies targeting energy conservation and flexibility. In the UK, smart meters are by default installed with in-home displays (IHDs) that provide real-time feedback on energy use. Some of the customers in our sample did not receive an IHD and we explore if this endogenous intervention had any impact on the effect described above. Cautioning against drawing causal conclusions, we do \emph{not} find any evidence that energy reduction is contingent on IHDs.

Outcomes of 2nd project: Descriptive social norms are presented by utility companies to consumers around the world, with the goal of reducing their consumption. This has been predominantly due to the positive results of this intervention found in several high profile large scale randomised control trials in recent years. Despite the success of these studies, a recent meta-analysis comparing different forms of comparison feedback interventions suggests that social norm comparison may not be the most effective feedback when compared to alternative feedback interventions such as historical consumption comparison or goal comparisons.

Outcomes of 3rd project: The proposed project provides valuable insights into the decision-making process regarding energy consumption and IoT applications in this context. The results bridge the more quantitative research tradition of rational-choice modelling (as employed, for instance, in economics) with the more qualitative research tradition of behavioural decision-making (as employed, for instance, in psychology). The outcomes are thus likely to be valuable for researchers in both camps.
Sectors Energy