Smart Profile: Smart Meter Data Oriented Customer Profiling

Abstract

Advanced metering infrastructure (AMI) enabled the collection of consumer status in real-time. That is the reason, this information is very critical and huge to take an important decision regarding many aspects of the smart grid. The data collected from smart meter helps to analyse consumer profiling, energy demand-response, optimization of energy generation and many more. This paper demonstrates the complete process of Energy Data Analytics starting from cleaning of real time data till clustering of customer cased on energy usage over a period of six months.  Clustering is one of the techniques which grouped the data elements based on some common characteristics. This paper exhibits the implementation of the clustering technique to classify the raw data into more meaningful information using Machine Learning (ML) models. Determining the number of clusters for any dataset is one of the crucial tasks. In this paper, techniques for cluster optimization, types of clustering and its implementation are discussed in detail. The dataset for the implementation is real-time data of industrial sector from Himachal Pradesh, Solan (H.P.). Furthermore, the implementation of all the techniques is performed on the energy data set using python software.