Clustering

City:

Kunshan, China

Organization:

School of Management, Hefei University of Technology, Hefei

Project Start Date:

16 November, 2014

Project End Date:

16 December, 2014

Reference:

Yang, T., Ren, M., & Zhou, K., (2014), “Identifying household electricity consumption patterns: A case study of Kunshan, China”, Renewable and Sustainable Energy Reviews 91 (2018) p861–868, https://doi.org/10.1016/j.rser.2018.04.037

Problem:

An important form of the Internet of Things application is smart grid as it is safe, stable operation while remaining a cost effective and environmental friendly option to measure electricity consumption. Nonetheless, existing studies focused on daily load profiles instead of monthly electricity consumption patterns which are more appropriate for medium and long term power system prediction and identify abnormal electricity consumers.

Technical Solution:

Hierarchical clustering method was apply to analyze the above monthly electricity use profiles. The process includes the following steps:

1)    Selection of a clustering method adapted to electricity use large scale and real time data.

2)    Selection of parameters

3)    Data clustering according the similarities in electricity consumption.

4)    Validation and evaluation

5)    Patterns extraction and representation

Datasets Used:

Electricity consumption data of 300 residential users collected by smart meters in Kunshan City from November 16, 2014 to December 16, 2014.

Outcome:

Data Preprocessing and abnormal user identification:

·       Withdraw of 8 users profiles from the pool with high and continuous abnormal electricity consumption.

 

Clustering Results:

·       Remaining 292 users can be divide in 4 groups of electricity consumption according to its increases and volatilities.

·       Positive correlation between the increase of consumption for each group and colder temperatures in Kunshan starting December 2014.

Issues that arose:

Identification of common characteristics across the 4 electricity consumers but this study didn’t include the potential incidence of the 8 abnormal consumers in terms of planning and operation of power systems.

Status:

Terminated

Entered by:

25 September, Elodie Girves, elodie.girves@mail.utoronto.ca