Nodal congestion price estimation in spot power market using artificial neural network

In the world wide increasing trend of restructured power system, open access in transmission system and competition in generation and distribution have introduced a frequently occurring problem of congestion. To establish a proficient price-based congestion management procedure, the nodal pricing strategy is found to be appropriate. From congestion management point of view, the optimal nodal prices are comprised of two basic components. First component is locational marginal price, that is marginal cost of generation to supply load and transmission losses both. Second component is nodal congestion price (NCP), that is the charges to maintain network security. Levenberg-Marquardt algorithm based neural network (LMANN) for estimating NCPs in spot power market by dividing the whole power system into various congestion zones is presented. Euclidian distance based clustering technique has been applied for feature selection before employing LMANN. The purpose of using artificial neural network (ANN) based approach for NCP estimation in spot power market is to exploit the tolerance for any missing or partially corrupted data to achieve tractability, robustness and very fast solution. The proposed ANN method also handles the congestion price volatility by taking continuously varying load and constrained transmission into account. The information provided by the proposed method regarding the formation of different congestion zones and the severity of congestion within a zone instructs both the market participants as well as independent system operator in making effective decisions. The proposed method has been examined for an RTS 24-bus system and is found to be quite promising.