Title :
Decoupled Reactive power Compensation Using Neural Network Based control scheme
Author_Institution :
Ind. Electron. Group, Central Electron. Eng. Res. Inst., Pilani
Abstract :
It is proposed to decouple three phase load reactive power in fundamental and harmonic power components and then compensate these using a combination of conventional shunt and parallel hybrid active filter. This approach results in use of low power rated PWM converters and offers many other advantages as compared to conventional shunt compensation. A simple neural network assisted scheme for decomposing of three phase current or voltage waveforms under unbalanced and non-sinusoidal conditions has also been proposed. Detailed mathematical model of decomposing algorithm and its real time implementation on DSP has been presented. Experimental results obtained on a laboratory model under steady state and dynamic conditions under different operating conditions have been analyzed and presented to demonstrate the effectiveness of proposed decoupled reactive power compensation scheme
Keywords :
PWM power convertors; active filters; digital signal processing chips; mathematical analysis; neurocontrollers; power harmonic filters; reactive power; DSP; PWM converters; decomposing algorithm; decoupled reactive power compensation; harmonic power components; mathematical model; neural network assisted scheme; parallel hybrid active filter; shunt compensation; three phase current decomposition; three phase load reactive power; voltage waveforms; Active filters; Digital signal processing; Mathematical model; Neural networks; Power harmonic filters; Power system harmonics; Pulse width modulation converters; Reactive power; Reactive power control; Voltage;
Conference_Titel :
Power Electronics Specialists Conference, 2005. PESC '05. IEEE 36th
Conference_Location :
Recife
Print_ISBN :
0-7803-9033-4
DOI :
10.1109/PESC.2005.1581637