Neural Network Based Composite Load Models for Power System Stability Analysis

Ali Keyhani, Wenzhe Lu, and Gerald T. Heydt
Ohio State Univ., Columbus, OH, USA ;

This paper appeares in : CIMSA 2005 ¨C IEEE International Conference on Computational Intelligence for Measurement Systems and Applications
Place: Giardini Naxos, Italy
Date: 20-22 July 2005
Pages: 32-37

---

Abstract:
Load modeling is an essential element in power system stability analysis. With the continuing increase of nonlinear and composite loads in power system, the modeling techniques used in the past may no longer be adequate. This paper proposes a methodology for the development of neural network based composite load model which can be applied to power system transient stability analysis. A two-layer neural network has been implemented to estimate the load power (P and Q) from terminal voltage and system frequency. The model has been validated using simulation test bed. The effect of measurement noise on the proposed methodology is also studied.