Abstract—We propose the SCCM (Step Classification and Characterization Method), which can detect the presence, classify the type, and estimate the location and height of a step change in the magnitude or the phase of an AC signal. We report on its classification performance for various locations of the step in the observation window, for a wide range of step depths, and in estimating the location and height of the step, separately for magnitude or phase steps. For AC signals with SNR_dB ≥ 30 dB, @ 4.8 kHz sampling rate, it can estimate the location of a step with an accuracy of 3 samples (≈ 0.63 ms), with confidence level above 95.45%. For the height of a step, the mean relative error is below 4%. We analyze a real-world signal with a voltage sag, to illustrate potential uses of the method to applications such as inception instant estimation of a voltage sag.

Matlab Codes and Scripts:

Reference Implementation of the SCCM:
 
Reference Implementation of the SCCM  and demonstration

Figures:

Script in Matlab to generate Figure 1 of the paper

Script in Matlab to generate Figure 2 of the paper

Script in Matlab to generate Figure 3 of the paper

Script in Matlab to generate Figure 4 of  the paper (SCCM analysis of a real-word voltage sag)

Performance Evaluation of the SCCM:

Scritpt used to run the SCCM step type classification performance (Section VIII.A)  and related data resported in Table II;

Script used to run the SCCM step charactezization performance for correctly classified Magnitude Steps (Case 1) - reproduce the results of Table III (Case 1) and Table V.
 
Script used to run the SCCM step charactezization performance for correctly classified Phase Steps (Case 2) - reproduce the results of Table III (Case 2) and Table VI.