Intelligent non-linear dynamic equivalent approach applied to Algerian power system

Aggregate or reduced power systems are often needed by operators and engineers to minimise the interconnection complexity and time simulation, without neglecting the effect of the interconnected area on internal system transient or dynamic stability study. In this work, a non-linear power system dynamic reduction is carried out by combining the diversity of solutions obtained with genetic algorithms and the accuracy of Newtonian methods when one is close to the solution. The network is divided into two parts; the first part called the study system, and the second is the external system, which is to be replaced by an equivalent. Using the proposed equivalent estimation procedure, the network model is kept non-linear in both the study and the external systems; the obtained equivalent model is a non-linear model with online possibility application. The proposed approach is tested with IEEE 10-machine-39Bus, IEEE 16-machine-68Bus and the Algerian power system with 90-machine-673Bus. The results prove that the designed equivalent represents faithfully the behaviour of the original system. The estimated equivalent system is ten times faster than real-time dynamic simulation for the Algerian network.