In this work an algorithm is presented that allows to detect, classify, discriminate and locate the faults that can be presented in an energy transport line, based on high frequency information (travelling waves). For this, the Wavelet transform and two types or neural network architectures are used as analysis tools. For the validation of the proposed methodology, a four-bar 400 kV reduced power system was used, which was modeled in the ATP / EMTP electromagnetic transient program, to obtain the signals in the time domain. The results obtained show the validity of the methodology proposed, for the scenarios considered.