"Smart meters, data mining and machine learning: identifying energy use profiles and providing feedback that empowers energy demand reduction"

点击翻译按钮获取中文摘要

"My doctoral research would involve interviews with key stakeholders in the UK smart meter rollout at a critical point in the initiative - the 2020 government deadline - while also using novel methods of data mining and machine learning to generate meaningful energy consumption profiles from large, emerging datasets. The insight from these two strands of research will then be applied to a trial energy feedback intervention. This research will be a valuable contribution to energy social science by both capturing this moment in the UK's ambitious smart meter rollout and informing future public engagement for energy demand reduction through smart grids. Research Questions: This doctoral research would be undertaken through thesis by publication, with each publication exploring one of these three interrelated questions:1. How do key stakeholders - energy suppliers, energy consumers, and various intermediaries - perceive the smart grid in terms of data access, value and intelligibility?2. Can data mining and machine learning translate large datasets of smart meter data intoenergy consumption profiles that identify inefficiency or other patterns like time shiftingand flexibility?3. What feedback and interventions based on half-hourly energy consumption profiles lead to energy demand reduction?Theoretical Framework / Methodology: A starting point for the theoretical framework and methodology of this doctoral research is the socio-technical transitions perspective developed by energy researchers (Geels et al 2018, McKenna et al 2018, Eyre et al 2018). I also have been influenced by a critique of the academic rigour of energy social science (Sovacool et al 2018). As I come from a communications and education background, my research will also be informed by diffusion theory, the elaboration likelihood model, and the theory of planned behaviour (Rogers 2010, Petty and Cacioppo 1986, Ajzen 2005). While behavioural economics provides a theoretical framework that greatly influences public policy in the UK, Hampton and Adams (2018) argue for the incorporation of practice theory into the energy policy discourse. Practice theory has also been used to explore different conceptualisations of energy users in smart grids - energy consumers vs. energy citizens (Goulden et al 2014). Fell and Shipworth (2017) propose a theory-agnostic approach to get past this divide, using a Dimension-Set Framework which characterises demand-side responses by electricity relevant dimensions and a phase state of possible options. By using the tools of data science to identify energy consumption profiles and patterns of potential efficiencies, I'm applying practice theory in the realm of big data."