Available transfer capability estimation using Support Vector Machine

In the deregulated power system, effective management of transactions, to meet the active power requirements of the consumers can be done through pool and bilateral/multilateral markets. One of the main functions of ISO in a day ahead market is assessment of Available Transfer Capability (ATC) and updating this value in Open Access Same time Information Systems (OASIS). The market participants use the value of ATC for their transaction planning. Hence it is essential to estimate ATC accurately with lesser computation time. This paper proposes a Support Vector Machine based method for ATC estimation. To the best of our knowledge this is the first attempt to use SVM for ATC estimation. Two methods are proposed in this paper. The first method uses the real power demands and the source bus injections as inputs to SVM model. The second method uses few indices as inputs to estimate ATC. The proposed methods are tested on IEEE 24 bus Reliability Test System (RTS) and IEEE 118 bus system. The results are compared with Repeated Power Flow (RPF) results.

► Available transfer capability of more than one conditions is estimated simultaneously.
► Transactions are used as an index for outage conditions.
► Three inputs are used to estimate available transfer capability.
► Number of inputs is independent of system size.
► Single AI model is used for both normal and outage conditions.