ANN-based reactive power controller with real-time web monitoring

This study introduced a real-time reactive power controller based on artificial neural network with web-based monitoring. A feed-forward employing back-propagation was used as training technique. The inputs to the network were the active and reactive power of each load. The targets were to switch on/off the capacitor banks during normal and abnormal conditions. The network was trained using developed MATLAB program and the weights resulted to minimum mean-square-error were fed to the microcontroller unit. The method was then tested in a three-bus radial distribution system model and implemented using Zilog microcontroller. The system actions were monitored using web-based monitoring application. The method was validated in actual operation of the system using the test data and results were satisfactorily obtained.