In this paper, fault detection system for the AC side of the microgrid is proposed. The fault detection system is based on the bayesian regularization neural network. The neural network is trained with the electromagnetic field signals instead of using current signal. In conventional way, current transformer is used for sensing the current signal. In the proposed method, magnetic sensor is used for sensing the magnetic field generated due to the flow of current in the conductor. Various fault conditions are taken into account for the analysis. The magnetic field under normal current condition is different than the magnetic field under fault conditions. Hence, this is the application of the pattern recognition artificial intelligence technique. The dataset for training and testing are generated with the help of simulation model of the microgrid. The microgrid is consisting of solar energy source, full bridge dc-ac converter, LC filters and load. The LC filter is used for getting the smooth ac output as there are harmonics present in the signal prior to it. The line to line and line to ground faults are created at the AC side of microgrid. A mathematical model is created for transforming current signal to magnetic field signal. These magnetic field signals are used to train and test the neural network. The testing results shows that the accuracy of the proposed system is 98.5% which is comparable with the existing methodologies. © 2021 IEEE.