Smart Meter Customer Data Synchronization Technique by KMean Clustering

The purpose of this paper is to review the methods of classification of stochastic variables in this case the customer power consumption recorded from smart meters. There are data available in the power distribution system from different sources which are often not synchronized in time. Customer smart meter data is existing and is valuable record which provides information of the parts of the power distribution network not visible to the distribution network operator. The purpose of the work is to find initialization process suitable for customer load demand within an active power distribution system and utilize the model to synchronize the customer smart meter load data with the datasets utilized in the power distribution operation such as data available in SCADA. KMean clustering techniques have been tested using actual customer smart meter data collected from ADRRESS Concept website in Australia. The customer datasets are collected for 7 typical winter days and 7 typical summer days. The KMean clustering technique has been widely used to build models from probabilistic natured variables. It is relatively effective and efficient for large datasets as well as datasets on a smaller time scale.