| This work presents the application of a system identification technique to extract the parameters of the modified Cassie-Mayr arc model, used to describe arc behavior in the post-current zero period. The proposed technique is based on Gauss-Newton implementation of nonlinear least squares, where device model parameters are estimated from input and output data obtained from a test. The advantage of applying the proposed technique is that it eliminates the need of assumptions made by other methods to obtain arc model parameters, and reduces the need of measured data filtering.; From results, it has been concluded that the Gauss-Newton implementation of nonlinear least squares is a technique with many advantages when it comes to identify the unknown parameters of the model. The main contribution of the study is the application of system identification to arc models in circuit breakers simulations improving the quality of the reproduction of current zero measurements, making predictions more reliable. |
