The squirrel cage induction machine as an alternative to the permanent magnet generator for direct drive tidal turbines

Abstract

This thesis looks at the direct drive power take off options for the OpenHydro tidal energy turbine and includes an investigation of the existing permanent magnet generator. The squirrel cage induction machine was chosen as an alternative design after reviewing the literature. The lack of rare earth magnet content of the induction machine was an opportunity to improve the cost competiveness of the OpenHydro tidal turbine to alternative renewable energy sources. The induction machine has been utilised with a gearbox in tidal turbines elsewhere, but in this work a direct drive design has been created. The generator of the OpenHydro turbine is positioned on the outer rim of the turbine blades, the rotor has a 14m outer diameter. Segmentation of the generator simplifies the construction and assembly of the turbine and also improves the fault tolerance of the turbine, depending on the coil connections. The existing permanent magnet design was evaluated with a simplified linear FEA model which underwent both manual and numerical optimisation. The optimisation found that the magnet mass could be reduced with minimum impact on the performance of the generator. To design and model the induction machine, each individual segment was wrapped to form a rotary induction machine, the conventional design process was then used to design the segment. FEA models of three designs were used to investigate the number of segments, with the best design then manually optimised. The study produced a model that was comparable to the permanent magnet generator performance, but with a lower power factor. The turbine features a buoyant shaft less rotor and as a consequence the generator suffers from eccentricity. The effect of eccentricity on the individual segments of both the permanent magnet and induction machine was investigated using linear FEA models. The eccentricity impacted the permanent magnet design in the voltage and EMF whereas the current was affected for the induction machine design. A numerical model, based on publications, was created to evaluate the effect on the fully assembled generator. The model was able to reproduce the presented inductances. For the comparison of the two designs the material costs of the permanent magnet and induction machine were calculated. The available power of a deployment site was calculated using the simulated efficiencies of the two generators, the models were run across the speed range corresponding to the tidal speeds. The cost of this energy sold was then used to compare the two generators over a 15 year lifetime. The capital cost of the induction machine design was 17% lower than that of the permanent magnet, over a 15 year lifetime the net income of the induction machine was 16% lower at a rated speed of 25rpm. The optimised permanent magnet design has 0.1% higher net income and a 9% lower capital cost than the original permanent magnet design. The profitability index is higher than both the original and induction machine.