A study on the prediction of capacitor-switching transient times

This paper presents a unique approach in the analysis of capacitor-switching transient disturbances, in an automated switching scenario. Unlike previous research attempts, which analyze the generated disturbance waveform, this study focuses on modeling the switching action of the network device responsible for the switching operation. Successful models were developed, allowing for the prediction of future switching operations. The determination of the minimum sample size needed for an accurate model, along with the associated prediction error criteria are presented. The results are based on an automated capacitor switching voltage support experiment, constructed at the Center for Electric Power Engineering, Drexel University, using the General Electric Type FKC-2 motorized oil switch. For the extracted transient time data sequences, the models can predict future switching operations within 2.1 msec