Classification and modelling of loads in power systems using SVM and optimization approach

This paper proposes a Support Vector Machine (SVM) based methodology to classify the loads into various classes based on the load responses, following a disturbance and other information available from the feeder. The classification is performed after a large disturbance, and the model parameters are estimated by using a variable projection based efficient optimization algorithm. For small disturbances or load changes, which happen frequently in the power systems, the proposed load modelling algorithm can be implemented to update the model parameters almost in real-time. The proposed methodologies are successfully applied on Northern Regional Power Grid (NRPG)-246 bus Indian System.