| dc.description.abstract | Wind energy conversion systems, which are in continuous development, are formed by three main elements, generator, converter, and control algorithms. Each element has its own technical challenges but all together they intend to cover the same objective, which is to get the maximum power from the wind with fast and adaptive control techniques to ensure power quality and grid compliance. This work has a special focus on the doubly fed induction generator (DFIG), which stillholds an important market share in the wind energy industry. The DFIG shows advantages over synchronous generators in terms of weight and size, and favorable characteristics such as decoupled active and reactive power control over a wider speed range, and partial scale of the power converter rating when used at rotor terminals. The full power capability of the generator is the aggregate sum of rotor and the stator when operating at super-synchronous speed and with a bidirectional power interface with low loses, high efficiency and variable output voltage amplitude, frequency, phase, and sequence. The indirect matrix converter (IMC) is light, it can handle high power densities and it can operate in harsh condition environments. The IMC, just as the conventional matrix converter (MC), would require a modulator or a complex conventional control strategy. Model predictive control (MPC) is a relatively new control technique for the IMC, specially due to the use of the discrete nature of power converters and its simplicity for implementation and intuitive approach. Model-based predictive rotor current control (MB-PCC) is proposed for a doubly fed induction machine (DFIM) driven by an indirect matrix converter. In this strategy active and reactive power, as well as the synchronization process, are controlled using the control of rotor currents with a dynamic reference calculated from active and reactive stator power set points and the dynamic model of the DFIMin coordination with the grid parameters. The grid synchronization process is carried out only by setting the P-Q power set points to zero and once completed, the control strategy can be applied in all four P-Q operating regions of the DFIM in a variable speed scenario. This strategy was implemented in a simulation using Gecko Circuits®, and in a 5.5 kW test rig. First, the stator active and reactive power references P , Q are decoupled into a stator current reference; second, the rotor electrical angular frequency ωr and angle θr, the grid voltage amplitude, angular frequency ωg and angle θg are integrated in the DIFG model; third, rotor currents are solved for a stator-fixed time-varying αβ reference frame; forth, the rotor current prediction takes into account the twenty-four valid combinations of the IMC using a cost function; fifth and finally, the optimum switching state of the IMC is selected.
The generation system, more specifically the stator, is connected to the grid in forty milliseconds.
This was the time required to guarantee full synchronization conditions applying zero power reference (P = 0,Q = 0). During the synchronization process, a very small voltage ripple in the stator voltage caused an overshoot in the stator current of only 1 Ampere. Right after the synchronization, the power reference can be changed within the nominal power limits as required. However, in simulation, the power reference was changed from zero to 3 kWand 3 kvar, and in the experimental rig the steps were limited to 500Wand 500 var due to setup limitations and in order to emulate a reduced scale grid. With the latter, the feasibility of the proposed control technique was proven both by simulation and implementation in an experimental rig rated to 5.5 kW. | |
