This work proposes a method for estimating fault severity in the presence of noise using the measured currents for a 7-phase electrical machine. The method is based on analytical models in stationary reference frames and analysis of the DC and fundamental components in the four fictitious machines. The simulation results show that the estimation errors in the tertiary and homopolar machines are less noise-sensitive than in the principal machine, regardless of the fault severity level. On the other hand, if the fault significantly affects the magnitude of the currents (fault > 10%), the estimation in the secondary machine is less sensitive to noise than in the main machine. If the fault results in a change in the mean value, the estimates using the main machine characteristics are less sensitive to noise. If the fault affects the phase shift, the estimates are less sensitive in the secondary machine than in the main one even if the noise is significant in the measurements.