Socio-Economic Data Analytics and Applications in the Smart Grids

  • Qiuyang Ma

Student thesis: Doctoral ThesisPhD


With the vigorous promotion of Advanced Metering Infrastructure and Half-Hourly Settlement (HHS) reform, the two-way communications between the residential customers and suppliers are built. The market signals are transmitted to the end-users from their accurate energy bill calculated in HHS process. The policymakers expect customers to mitigate the uncertainty in the energy market by modifying their usage behaviour following the market signals, and meanwhile reducing their energy bills. However, the policies also introduce uncertainty in customers’ energy bills. Therefore, the impact of policies on customers from different socio-economic status needs to be assessed. Moreover, to timely launch appropriate interventions to assist the vulnerable customers, the socio-economic data needs to be analysed to obtain a more in-depth understanding of customers' usage behaviour. This thesis fills the research gaps by investigating the effect of socio-economic data from two aspects: 1) investigating the impact of interacted socio-economic data; 2) considering the effect of the collaboration of socio-economic data with other data sources, such as the smart metering data, the Time-Of-Use (TOU) tariff data and so on.
The investigation of the effect of interacted socio-economic data is triggered by the HHS reformation to the energy retail market. The HHS process provides more accurate energy bills to individual customers. Meanwhile, it also introduces uncertainty to customers' future energy bill. Hence, by analysing the effect of interacted socio-economic data on the variation of residential customers’ energy bills, the impact of the HHS reform on customers with different socio-economic status can be assessed. A novel high-dimensional interaction-aware search method has been proposed, which is named the KLAM method. The KLAM method can detect the high-dimensional interacting significant factors, meanwhile minimising the information loss. The interacted significant socio-economic factors could describe the socio-economic characteristics of the new vulnerable customers under the HHS process. Additionally, a novel distribution network pricing method is proposed which removes the cross-subsidies in network cost among customers. The impact of network cost variation on customers in different socio-economic status can be investigated.
Applying the socio-economic data with other data sources can explore the better performance of different demand-side appliances.
1)Socio-economic information can remedy the problem caused by the availability issues of other data sources. For example, the availability limitation of smart metering data for the new switch-in customers is a problem for the customer classification. Therefore, a cost-reflective classification framework has been proposed by collaborating socio-economic data with smart metering data. Three scenarios are established in the novel classification framework to estimate the energy cost level for the customers who 1) only have the smart metering data; 2) only have the socio-economic data;; 3) have both two datasets. The accuracy of energy cost prediction for those three scenarios is 74.88% and 53.31% and 75.00% respectively.
2)Furthermore, a responsiveness pre-evaluating framework has been proposed. This framework aims to identify the significant socio-economic criteria and load characteristics for customers’ responsiveness to different TOU tariffs.
Date of Award24 Jun 2020
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorRan Li (Supervisor) & Furong Li (Supervisor)


  • Socio-economic data
  • Retail energy market
  • Energy settlement

Cite this