Security constrained dispatch solutions using the gradient projection method of optimization

In order to use a security constrained dispatch (SCD) program to aid a power system operator in an online environment, the solution time is of major concern. In this work, an existing SCD algorithm, which uses the gradient projection method to solve a sequence of linearly constrained minimization problems is implemented. The gradient projection method is iterated with the Newton Raphson load flow technique to solve the nonlinear equations representing the constraints. The solution time is reduced by replacing the Newton Raphson load flow technique with a decoupled load flow, and then with a fast decoupled load flow. The new solution times are compared to the original one. In the existing SCD algorithm, the dependent variable constraints are on bus voltage magnitudes and on generated reactive powers. It has been shown that real power generated has little effect on these two variables. Therefore, the dependent variable constraints are reformulated to be on the bus voltage phase angles instead. The solution times for the three different load flow techniques are again compared. For both formulations of the security constrained dispatch problem, the solution time is reduced by up to 70%.<>