No abstract

No abstract

Diode and thyristor-based rectifier circuits have been widely used in the industry. Due to non-linear structures of these circuits, they draw non-sinusoidal current from AC network as well as cause a low power factor in the AC side. The DC-link voltage of rectifier is affected by the changes in AC network or by the load variations on the DC side. Pulse-width modulated (PWM) rectifiers can eliminate the mentioned power quality problems if they control properly. This study proposes a controller with an adaptive and robust structure based on proportional + derivative type-2 fuzzy neural network (PD-T2FNN) for DC-link voltage control of PWM rectifier. Dynamic performance of PWM rectifier using the proposed controller is evaluated via dSPACE based experimental setup under different operation conditions: set-point change, step load change in the DC side of the rectifier, set-point change under load and capacitive operation mode. The experimental results are given for traditional PD and proportional + integral and T2FNN controllers to validity performance of the proposed controller. Performances of controllers are evaluated regarding settling time, overshoot, steady-state error and total harmonic distortion. PWM rectifier with PD-T2FNN DClink voltage controller has superior performance for all operating conditions according to performance criteria when compared with other controllers.

A distribution static compensator (D-STATCOM) is a custom power device connected in parallel to a power system to address electric power quality problems caused by reactive power and harmonics. To obtain high performance from a D-STATCOM, the D-STATCOM's dq-axis currents must be controlled in an internal control loop. However, control of the D-STATCOM's currents is difficult because of its nonlinear structure, cross-coupling effect between the dand q-axis, undefined dynamics, and fast changing load. Therefore, the controller to be preferred for a D-STATCOM should have a nonlinear and robust structure. In this study, a neuro-fuzzy controller (NFC), which is a robust and nonlinear controller, is proposed for dq-axis current control of a D-STATCOM. A DSP-based experimental setup is built for real-time control. The basic building block of the experimental setup is a three-level cascaded inverter. This inverter is constructed by using three IPM intelligent modules. A DS1103 controller card is used for real-time control of the D-STATCOM's experimental setup. The control algorithm is prepared in MATLAB/Simulink software and loaded to the DS1103 controller card. The performance of the NFC current-controlled D-STATCOM is tested for different load conditions: no load to full inductive, no load to full capacitive, full inductive to full capacitive, and full capacitive to full inductive. For this aim, the reactive current setpoint is changed as a step. The experimental results are presented to show the efficiency of the proposed controller under different load conditions.

Abstract:The fault detection process is very difficult in transmission lines with a fixed series capacitor because of the nonlinear behavior of protection device and series-parallel resonance. This paper proposes a new method based on S-transform (ST) and support vector machines (SVMs) for fault classification and identification of a faulty section in a transmission line with a fixed series capacitor placed at the middle of the line. In the proposed method, the fault detection process is carried out by using distinctive features of 3-line signals (line voltages and currents) and zero sequence current. The relevant features of these signals are obtained by using the ST. The obtained features are then used as input to multiple SVM classifiers and their outputs are combined for classifying the fault type and identifying the faulty section. Training and testing samples for the proposed method have been generated with different types of short-circuit faults and different combinations of system parameters in the MATLAB environment. The performance of the proposed method is investigated according to the accuracy of fault classification and faulty section identification. To evaluate the validity of this study, the proposed method is also compared to both ST-neural network and previous studies. The proposed method not only provides a good classification performance for all types of faults, but also detects the faulty section at a high accuracy.