Published Paper Details:

An industrial load is mostly depending on semiconductor devices; control of that loads is important cause industrial loads sensitive against power system disturbance. Hence, in the distribution system, power quality has most important. Voltage and power quality have more problems due to an increase in the use of nonlinear loads such as microcontrollers, power electronic devices, etc. Thus, it needs to identify that problem and the requirement of compensation. Thus, the solution is a dynamic voltage restorer which identifies voltage sag & compensate that sag fast & accurate. In this paper, the simulation of dynamic voltage restorer (DVR) was made by using an artificial neural network. DVR injects the voltage in the system when voltage sag or swell occurred due to fault. The ANN is trained online by data generated by the uncompensated model. These data should be used as a target for ANN. There are many types of training, here we feed-forward backpropagation, Levenberg- Marquardt training algorithm used. ANN is used as a controller for controlling the DVR (Dynamic Voltage Restorer) through an IGBT gate inverter. For mitigating voltage sag & swell inject voltage by DVR with ANN exactly. Also, another method is Integrated Sliding Mode Control using DVR for compensating voltage sag caused due to three-phase faults. The sag or swell detection method is eliminated by using the DVR-SMC model. Hence, this method increases the dynamic response of the DVR & inject the voltage sag with proper amplitude and phase. The comparison between the ANN-based DVR model and ISMC based DVR model has studied about compensation of voltage sag after a fault occurred. From this, obtain best compensation method.