India faces endemic electrical energy and peaking shortages. These shortages have had a very detrimental effect on the overall economic growth of the country. As total Power losses equals transmission power losses plus distribution power losses. The reasons cited for such high losses are; lack of adequate T & D capacity, too many transformation stages, improper load distribution and extensive rural electrification etc. The sources of transmission power losses may be directly driven by network investment or by network operation. Distribution power losses arise from several areas including theft, un-billed accounts, and estimated customer accounts, errors due to the approximation of consumption by un-metered supplies and metering errors Electricity theft can be in the form of fraud (meter tampering), stealing (illegal connections), billing irregularities, and unpaid bills. Estimates of the extent of electricity theft in a sample of 102 countries are undertaken. The evidence shows that theft is increasing in most regions of the world. Electricity consumer dishonesty is a problem faced by all power utilities. Finding efficient measurements for detecting fraudulent electricity consumption has been an active research area in recent years. Data mining has become increasingly common in both the public and private sectors.
This paper presents a new approach towards Distribution Power Loss analysis for electric utilities using a novel intelligence-based technique like Extreme Learning Machine (ELM), an online sequential learning algorithm for single hidden layer feed forward networks (SLFNs) with additive or radial basis function (RBF) hidden nodes in a unified framework. OS-ELM & Support Vector Machine(SVM). The main motivation of this study is to assist Gujarat Urja Vikas Nigam LTD (GUVNL), GUJARAT, INDIA to reduce its Distribution Power Loss due to electricity theft. This approach provides a method of data mining and involves feature extraction from historical – ustomer consumption data. This model preselects suspected customers to be inspected onsite for fraud based on abnormal consumption behavior. The proposed approach uses customer load profile information to expose abnormal behavior that is known to be highly correlated with Distribution Power Loss activities. The approach uses customer load profile information to expose abnormal behavior that is known to be highly correlated with Power Loss activities. Simulation results prove the proposed method is more effective compared to the current actions taken by GUVNL in order to reduce Power Loss activities.