Browsing by Author "Mbukani, Mwana Wa Kalaga"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Modelling and control of doubly-fed induction generator systems in wind turbine applications
    (Stellenbosch : Stellenbosch University, 2017-03) Mbukani, Mwana Wa Kalaga; Gule, Nkosanathi; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
    ENGLISH ABSTRACT: The need to reduce the world's carbon foot print has led to a significant rise in wind energy generation. The doubly-fed induction generator (DFIG) is one of the most popular wind turbine generators partly due to its low rated backto- back converter. A constant DC-link voltage in the doubly-fed induction generator system's back to back converter allows for bidirectional power flow of the rotor power. Hence, effective control of the DC-link voltage is necessary. The presence of the switching elements in the back-to-back converter creates harmonics in the systems. LCL and L filters are mostly used to mitigate the harmonics. DFIGs are mainly connected to the grid, however, they can be used in stand-alone mode in isolated rural areas, where there are low loads with no grid connection. DFIGs in the stand-alone mode have to be controlled such that they provide voltage and frequency stability at varying load conditions and changing wind speeds. In this thesis, the power control of the grid-connected DFIG systems in wind turbine applications is presented. Power factor regulation is conducted since it helps in the reduction of the costs linked to the capacitor bank. Maximum power point tracking is also investigated. DC-link voltage control is analysed whereby the grid-side converter is controlled as a voltage-source converter. A comparative analysis of the LCL filter and L filter for switching frequencies below 5 kHz is done and described in this thesis. This is important for systems with low sampling frequencies. Furthermore, in this thesis, the control of a stand-alone DFIG together with simulation result, is presented. Experimental results are also given to demonstrate the effectiveness of the developed rotor-side control algorithm on a DFIG test bed.
  • Thumbnail Image
    Item
    Position/speed sensor-less control of wind energy conversion systems based on Rotor-Tied Doubly-Fed induction generator systems
    (Stellenbosch : Stellenbosch University, 2019-12) Mbukani, Mwana Wa Kalaga; Gule, Nkosinathi; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
    ENGLISH ABSTRACT: The doubly-fed induction generator (DFIG) is amongst the most popular wind turbine generator in South Africa. This is partly due to the fact that its backtoback power converters are partially rated. More precisely, they are rated at 30% of the generator rated power. A new DFIG topology has been proposed recently. That is the rotor-tied doubly-fed induction generator (RDFIG). In this topology, the rotor side is connected to the grid while the stator side is connected to the power converter. It has been shown that this topology holds the advantage of higher effeciency compared to the standard DFIG topology. High accuracy in all the measurements is required for the optimum operations of wind energy conversion systems (WECSs). The measurement of the rotor position/speed is amongst the most important measurements when it comes to implement any control system for the WECS. The conventional method of measuring the rotor position/speed is to use an electronic/mechanical sensor (encoder or resolver). This measurement involves the use of long cables and in a harsh environment, this can lead to faulty operations of the WECS. In this thesis, several slip speed estimators for sensor-less control of RDFIGbased WECSs are developed and implemented. The proposed slip speed estimators are based on the association of different sliding mode observers and the PLL estimator. The association of the PLL estimator improves the estimation performance by reducing the noise created by the sliding control control functions. Also, the proposed PLL estimator helps in avoiding a phase shift of π in super-synchronous operating conditions. In addition, in this thesis, several sliding mode observers were developed in order to improve the estimation performance. The proposed sliding mode observers were satisfactory for all the operating conditions of the RDFIG-based WECSs. The robustness of the proposed slip speed estimators is validated experimentally under various operating conditions. A 5.5-kW custom-designed gridconnected RDFIG test-bench based on a National Instrument (NI) PXIe-8115 controller is used. The proposed slip speed estimators solve the problem linked to the failure of the electromechanical sensors. The overall sensor-less control strategy provides an alternative to the sensor-based control of the RDFIGs. Also, the proposed sensor-less vector control strategy can be used as a back-up in case the electromechanical sensor fails.