AMI Data Analytics; An Investigation of the SelfOrganizing Maps Capabilities in Customers Characterization and Big Data Management

AMI Data Analytics; An Investigation of the SelfOrganizing Maps Capabilities in Customers Characterization and Big Data Management

This paper is aimed at investigating self-organizing map (SOM) capabilities in customers characterization in their electricity consumption behavior. Characterization is based on the recorded data by Advanced Metering Infrastructure (AMI) in smart grid. This investigation regards two aspects of SOM: First, capabilities of SOM in pattern recognition applications and second, capabilities of SOM in big data management. Both of these capabilities are instrumental in the current restructured electricity market. From one aspect, requirements of the market for load profiling by which decision making in energy management programs and other policies is more reliable. From another aspect, the increase in information exchanging in the grid in the presence of AMI which complicates the analysis of data. Applying this algorithm in both two aforementioned aspects has shown persuasive results. A real dataset related to Irish electricity consumption is used to evaluate performance of the proposed procedures.

This paper is aimed at investigating self-organizing map (SOM) capabilities in customers characterization in their electricity consumption behavior. Characterization is based on the recorded data by Advanced Metering Infrastructure (AMI) in smart grid. This investigation regards two aspects of SOM: First, capabilities of SOM in pattern recognition applications and second, capabilities of SOM in big data management. Both of these capabilities are instrumental in the current restructured electricity market. From one aspect, requirements of the market for load profiling by which decision making in energy management programs and other policies is more reliable. From another aspect, the increase in information exchanging in the grid in the presence of AMI which complicates the analysis of data. Applying this algorithm in both two aforementioned aspects has shown persuasive results. A real dataset related to Irish electricity consumption is used to evaluate performance of the proposed procedures.