A contactless ammeter based on GMR magnetometers

This work presents a contactless ammeter prototype based on GMR magnetometers, able to measure direct and alternating electrical currents. Two techniques nonlinear curve fitting and artificial neural networks, were used to solve the inverse problem, which is, given the output voltages of the GMR sensors, to estimate the electric current flowing through the conductor - nonlinear curve fitting and artificial neural networks. The performance of both techniques was compared based on the standard uncertainty of the data calculated by each one. The results obtained with the neural networks are significantly better.