Adaptive Neural Network-Based LMS for DSTATCOM
This paper presents distributed static compensator (DSTATCOM) to eliminate harmonics and reactive power of the nonlinear load. The DSTATCOM is realized using voltage source inverter with DC bus capacitor. The DSTATCOM acts as a harmonics current source and inject reactive compensation current. The control schemes for determining the reference compensating currents of the three-phase DSTATCOM based on least mean square (LMS) algorithms are presented. The performance of LMS algorithm with PI and artificial neural network controller (ANN) is studied. The algorithm used for training ANN controller is Levenberg–Marquardt backpropagation (LMBP). The training data for ANN controller is generated offline. The firing pulses for DSTATCOM are obtained with hysteresis current controller. An extensive simulation study is carried out to test the performance of ANN controller and compared with PI controller.
KeywordsPI controller LMS Active power filter Harmonic
- 7.Singh B, Solanki J (2006) A comparative study of control algorithms for DSTATCOM for load compensation. In Proc IEEE ICIT, 15–17 Dec 2006, pp 1492–1497Google Scholar