On July 7, Bettergy published an article of scientific research in the magazine Energies, developed in collaboration with the University of Malaga. “Energies” is an open access journal of the “MDPI” (Multidisciplinary Digital Publishing Institute), with great international impact.
In an ideal world, companies should be able to see each client as an individual.
Realistically, this is a very difficult task, because consumers are becoming more sophisticated every day and the competition is increasingly globalized.
Therefore, it is necessary to select groups of clients that share similar characteristics (segmentation) to carry out a commercial activity, offering them exactly what they expect.
To simplify the selection of customers for electric utilities, they rely on the advantage of the extensive existing deployment of smart meters, which allow retrieving consumption information (or load curve) remotely, without having to visit the installation of all customers.
In this study, the characterization of the consumption of the clients is proposed starting from defining the signature concept of the electricity demand, which is a synthesized representation of the information of the daily load curves.
The objective of this study is to provide an alternative for customer segmentation that is capable of managing multiple types of clients.
The approach serves to characterize and group electricity customers according to their demand habits. In addition, the study presents an innovative alternative for the grouping of multiple types of electricity clients based on advanced techniques of Evolutionary Computing.
To evaluate the study proposal, the real anonymized data of 64 buildings located in the Andalusian community were used.
The set includes buildings of different types of consumption and categories of use, representative of the majority of industrial divisions and types of buildings in cities.
In this way, the techniques developed can also be used to segment and select clients in other countries or regions with satisfactory results.
Link to the complete study: http://www.mdpi.com/1996-1073/11/7/1788/htm