Load profiling and its applications in power market

In a power retail market, the knowledge of hourly load shape of the customer is often required in order to determine the tariff schedule. At the present time, requiring a meter as a prerequisite for power customers to choose a power supplier is not yet achievable. For customer screening, this paper proposes a technique that uses customers´ monthly energy usage data to perform load profile assignment. Two steps are involved in the proposed method, they are the fuzzy c-means (FCM) method used for load profiling and a decision tree based pattern recognition technique for customer class assignments. Test data are from actual measurements and customer information system (CIS) of Taiwan Power Company (TPC). Promising results on assigning existing 500 customers to their proper load profiles are illustrated.