Performance of SAG events using Artificial Neural Network (ANN) / Norfarizani Nordin

This work presents about the performance of voltage sag events by using Artificial Neural Network (ANN). Voltage sag is one of the Power Quality problems which can occur in equipments or even in the transmission line of power system. Voltage sag may occur in many reasons. In Malaysia, the main cause of voltage sag to happen is because of lightning. When lightning strikes, it carries a very high current and may cause the voltage to drop. Hence, there will be an energy lost in the site or building. This energy lost and the severity of voltage sag can be measured by finding the value of energy lost and voltage sag score. The energy lost and sag score value in every sag events can be calculated. However, the validity of the calculated data is quite questionable. The performance of voltage sag is shown by the value of voltage sag score and energy lost values. Thus the purpose of this paper is to make comparison between the monitored and calculated data and to show the performance of sag score and energy lost in sag events using Artificial Neural Network (ANN)-MATLAB. In order to test the validity of sag score and energy lost, two methods were used in this project; first by calculating voltage sag score and its energy lost during sag event from monitored and captured data, and second method is by feeding the calculated data into the artificial neural network. The network is trained using the data on sag duration and three phase voltages Vi, V2 and V3 as the input, while the energy lost and voltage sag score are used to be the targeted output. Results of calculated data will be compared with the results obtained using ANN. It is hoped that the validity of the calculated sag score and energy lost during sag are accurate for monitoring purposes.