Development of a Novel Positive Sequence Contactor Model Using Deep Neural Networks

In this work, a novel positive sequence (PS) contactor model for the protection of the ‘motorc’ model Lesieutre et al. 2008 has been developed using a detailed novel methodology, using linear regression and deep neural networks (DNNs). This model has been developed to identify and incorporate the critical tripping and reconnection characteristics of a newly developed 24 V electromagnetic transient (EMT) contactor Nekkalapu et al. 2022 from an EMT simulator (PSCAD) into a positive sequence simulator (GE PSLF). The impact of the accurate EMT contactor model on the motor stalling phenomenon, to accurately estimate fault induced delayed voltage recovery (FIDVR) type phenomenon for single-line to ground (SLG) faults, is also accurately captured using the proposed methodology.