Dynamic Phasor Modeling of Type 3 Wind Farm including Multi-mass and LVRT Effects

The proportion of power attributable to wind generation has grown significantly in the last two decades. System impact studies such as load flow studies and short circuit studies, are important for planning before integration of any new wind generation into the existing power grid. Short circuit modelling is central in these planning studies to determine protective relay settings, protection coordination, and equipment ratings. Numerous factors, such as low voltage situations, power electronic switching, control actions, sub-synchronous oscillations, etc., influence the response of wind farms to short circuit conditions, and that makes short circuit modelling of wind farms an interesting, complex, and challenging task. Power electronics-based converters are very common in wind power plants, enabling the plant to operate at a wide range of wind speeds and provide reactive power support without disconnection from the grid during low voltage scenarios. This has led to the growth of Type 3 (with rotor side converter) and Type 4 (with stator side full converter) wind generators, in which power electronics-based converters and controls are an integral part. The power electronics in these generators are proprietary in nature, which makes it difficult to obtain the necessary information from the manufacturer to model them accurately in planning studies for conditions such as those found during faults or low voltage ride through (LVRT) periods. The use of power electronic controllers also has led to phenomena such as sub-synchronous control interactions in series compensated Type 3 wind farms, which are characterized by non-fundamental frequency oscillations. The above factors have led to the need to develop generic models for wind farms that can be used in studies by planners and protection engineers. The current practice for short circuit modelling of wind farms in the power industry is to utilize transient stability programs based on either simplified electromechanical fundamental frequency models or detailed electromagnetic time domain models. The fundamental frequency models are incapable of representing the majority of critical wind generator fault characteristics, such as during power electronic switching conditions and sub-synchronous interactions. The detailed time domain models, though accurate, demand high levels of computation and modelling expertise. A simple yet accurate modelling methodology for wind generators that does not require resorting to fundamental frequency based simplifications or time domain type simulations is the basis for this research work. This research work develops an average value model and a dynamic phasor model of a Type 3 DFIG wind farm. The average value model replaces the switches and associated phenomena by equivalent current and voltage sources. The dynamic phasor model is based on generalized averaging theory, where the system variables are represented as time varying Fourier coefficients known as dynamic phasors. The two types models provide a generic type model and achieve a middle ground between conventional electromechanical models and the cumbersome electromagnetic time domain models. The dynamic phasor model enables the user to consider each harmonic component individually; this selective view of the components of the system response is not achievable in conventional electromagnetic transient simulations. Only the appropriate dynamic phasors are selected for the required fault behaviour to be represented, providing greater computational efficiency than detailed time domain simulations. A detailed electromagnetic transient (EMT) simulation model is also developed in this thesis using a real-time digital simulator (RTDS). The results obtained with the average value model and the dynamic phasor model are validated with an accurate electromagnetic simulation model and some state-of-the-art industrial schemes: a voltage behind transient reactance model, an analytical expression model, and a voltage dependent current source model. The proposed RTDS models include the effect of change of flux during faulted conditions in the wind generator during abnormal system conditions instead of incorrectly assuming it is a constant. This was not investigated in previous studies carried out in the real-time simulations laboratory at the University of Saskatchewan or in various publications reported in the literature. The most commonly used LVRT topologies, such as rotor side crowbar circuit, DC-link protection scheme, and series dynamic braking resistance (SDBR) in rotor and stator circuits, are investigated in the short circuit studies. The RTDS model developed uses a multi-mass (three-mass) model of the mechanical drive train instead of a simple single-mass model to represent torsional dynamics. The single mass model considers the blade inertia, the turbine hub, and the generator as a single lumped mass and so cannot reproduce the torsional behaviour. The root cause of sub-synchronous frequencies in Type 3 wind generators is not well understood by system planners and protection engineers. Some literature reports it is self excitation while others report it is due to sub-synchronous control interactions. One publication in the stability literature reports on a small signal analysis study aimed at finding the root cause of the problem, and a similar type of analysis was performed in this thesis. A linearized model was developed, which includes the generator model, a three mass drive train, rotor side converter, and the grid side converter represented as a constant voltage source. The linear model analysis showed that the sub-synchronous oscillations are due to control interactions between the rotor side controller of the Type 3 wind power plant and the series capacitor in the transmission line. The rotor side controls were tuned to obtain a stable response at higher levels of compensation. A real-time simulation model of a 450 MW Type 3 wind farm consisting of 150 units transmitting power via 345 kV transmission line was developed on the RTDS. The dynamic phasor method is shown to be accurate for representing faults at the point of interconnection of the wind farm to the grid for balanced and unbalanced faults as well as for different sub- synchronous oscillation frequencies