ANN Based DSPIC Controller for Reactive Power Compensation

Abstract

In India with a power shortage of about 10% of installed capacity, the transmission and distribution losses are relatively high. One of the reasons for the large loss is the excessive reactive loads on the primary distribution feeder. With the right amount of reactive power compensation, as system load varies, line losses can be considerably reduced. The application of Artificial Neural Network (ANN) is an emerging area where a properly trained ANN can be advantageously used for sufficient amount of reactive power compensation under varying load conditions. This paper describes the methodology adopted for training an ANN network for reactive power compensation without human intervention. The ANN is implemented using a DSPIC 30F2010 (Digital Signal Peripheral Interface Controller) and verified on a simulated laboratory network . The results obtained are satisfactory

Particle swarm optimization tuned unified power flow controller for power oscillation reduction

“…As well as the researches to improve the artificial intelligence are also done like direct adaptive network for training it with backpropagation is carried out in [6] and with feed forward neural network [7], [8] are carried out. And in the power system the improvement in transient stability by using the neural network is proposed in [9]. An automatic regulation of the reactive power in a hybrid power system has also been proposed in [10].…”