Using two multivariate methods for line congestion study in transmission systems under uncertainty

Two multivariate methods (correlation analysis and principal components analysis) are used to forecast which lines may be simultaneously congested. These statistical methods are applied to a database which takes into account transmission planning uncertainties such as localization of new independent power producers and new eligible customers (i.e., customers who can choose their energy supplier), and level of international exchanges. To face the very large number of possible configurations, a design of experiment is used to create the data base. A complete active/reactive-power flow program is used to simulate the power system. The results show the relevance of the proposed methods as applied to a 21-bus power system test case. Knowing which lines will be simultaneously congested may help the system operator to take decision in short-term operation (congestion management) as well as in long-term planning (grid reinforcement).