An approach towards consumer power usage pattern using classification technique
Today's life is embedded with technology that consumes an enormous amount of electricity. To maintain the uninterrupted electricity supply, keeping track of the consumer demand pattern is a dire necessity, particularly in the case of the smart grid system. Using a load usage pattern as well as the timing of peak electricity usage, an approach is proposed in this paper to estimate energy consumption patterns for residential and industrial consumers. Although in past, tariff structures were mainly applied on the kind of activity among consumers, the kind of activity and electrical behavior of the customer has a very poor relationship. Applying clustering techniques to classify customers according to load curves is more efficient. This paper proposes a two-fold classification algorithm followed by a supervised learning technique to classify electric customers. At First, results are obtained from the classification techniques and compared. A decision matrix is projected to reach the goal of predicting power consumption behavior in the second phase.