Abstarct: This thesis proceeds on introducing novel and intelligent algorithms for fault detection in Diode and Thyristor 12 pulse rectifiers, series, and parallel. Challenges and difficulties around fault detection by considering the type of load and connection has been investigated, also it has been tried to modify and improve previous studies in a way of decreasing the number of sensors and complexity of the algorithm and decreasing the cost of devices. A fault detection algorithm has been proposed for all Q-pulse diode rectifiers. Capability of output Voltage signal depends on the presence of shunt capacitor in its output has been explored. In addition to intelligent algorithms such as neural networks like RBF and MLP network, fault detection in diode rectifiers due to signal processing and unique diode rectifier characteristic has been proposed. Empirical test shows the high performance of this heuristic algorithm. Every year billion dollars are spending by energy loss of data saving which intelligent algorithm highly needs costly software and data. In order to fault detection in the thyristor rectifier again, another heuristic algorithm for high power applications has been proposed. Deponed on the importance and sensitivity of the system under study, different algorithms are provided for fault detection systems. Also new strategy by imposing optimal point of the system in reagard to voltage ripple and current has implemented to reconfig the faulty thyristor rectifier. MATLAB/SIMUlixnk has been used to simulate the system and algorithm.