On-Line Estimation of ATC Using Neural Network With Selective Features

In any power system with deregulated environment, the capability of transferring power -ATC of the system needs to be updated every hour or every day. This will help market participants to plan their transmission transactions. The existing methods of calculating ATC need large computational time, as well as they are not effective for online techniques. Hence, it is significant to evaluate ATC values accurately at a lesser time. Here, neural network approach is used for on-line ATC estimation for both bilateral and multilateral transactions under normal and contingency condition. The main feature of this method is the reduction of size for extending the capability. Overall design is enhanced based on the mutual information feature selection. IEEE bus system is tested and the obtained output is distinguished with Repeated Power Flow (RPF) results. The output reveals that the proposed strategy is best suited for online calculation.