The research is focused on the modelling of a squirrel-cage induction generator in dynamicudgeneration involving ocean-wave energy. The chosen application includes an oscillatingudwater column fitted with a Wells turbine.udThe modelling approach is based on the evaluation of existing generator models. Theseudinclude the equivalent steady-state and dynamic models which are considered from a timedomainud(differential equation) perspective. Since generation is dynamic in nature, modeludstability is an important component of model evaluation.udThe evaluated models provide information regarding the electrical and mechanicaludoperational variables of the generator. Power flow and energy loss between the mechanicaludand electrical subsystems are easily calculated from these variables.udThe wave-energy converter excluding the induction generator is not explicitly considered.udThe generator models are evaluated by considering typical generator inputs which areudrepresentative of the given application. These dynamics are reproduced experimentally and inudsimulations with a comparison of generator response allowing for a conclusion on modeludperformance. Generator inputs include the stator voltage excitation and turbine torque withudthe generator response given by the stator currents and rotor velocity. Electrical andudmechanical power are also considered.udDynamic generation is broken down into two modes of operation: the first mode involvesudgeneration for a constant sea state and the second mode involves generator operation for audchange in sea state. The dynamics for the first mode involve a set generator speed (set voltageudsupply) and a sinusoidal prime-mover torque. Dynamics for the second operating mode areudnot well-defined owing to system variations. Since only the generator model is considered, anudinformative dynamic is tested providing an indication of possible model performance. Theudtested dynamic involves a sinusoidally-varying stator frequency and prime-mover torque.udThe steady-state model considered from a time-domain perspective is found to be unstableudfor all generating slip values and is, therefore, unsuitable for the given generation application.udThe dynamic model shows good agreement between experimental and simulated generatorudresponse for the two operating modes identified. In conclusion, the model is applicable for audconstant sea state with a wave period of up to 0.2 s. Furthermore, it is suspected that theuddynamic model is applicable in the case of a change in sea state. Cases involving magneticudsaturation and parameter variation are left for future development.udThe dynamic-model evaluation assumes a balanced stator-voltage excitation – strangeudelectrical transients including electrical faults are not considered. An important simulation consideration involves the quantification of state-variable initialudconditions. Initial rotor currents are problematic as these are not easily measured or definedudin a practical squirrel-cage rotor construction. The initial rotor currents are approximated by audphasor analysis of the steady-state circuit model at zero time.udThe use of an inverter-based generator excitation for the experimental work poses an analysisudproblem owing to the pulse-width-modulation-based voltage supply (not truly sinusoidal).udThis is solved by considering only the fundamental component of the stator voltage andudcurrent. Second-order low-pass filters are used to facilitate such measurements.
展开▼
