Parameter Error Identification for Type-4 Wind Turbine Models using Neural Networks

Increasing penetration of inverter-based resources (IBRs) makes the power system more vulnerable to transients and grid events. Therefore, a correct dynamic modeling and simulation of the system is more important than ever. However, the reliability of these simulations is affected by the erroneous parameters and models of the system components. Hence, regular identification and calibration of these defective parameters should be carried out. The offline staged calibration and test for these purposes now can be replaced with online tools. In this paper, a sequential neural network is employed for the identification of the erroneous parameters of Type-4 wind turbine dynamic models. Testing and validation of the proposed method are performed with the generated synthetic data and the identification results reach up to 89% accuracy rate.