Wind Turbine Emulation Based on the Accurate Control of Servomotors

The objective of this work is the creation of a novel Wind Turbine Emulation (WTE) technique to support the needs of a new generation of multi-bladed vertical-axis wind turbine (VAWT) designs aimed at the urban environment. The scheme, presented in this thesis, uses Computational Fluid Dynamics (CFD) data, from the analysis of such devices, as the basis of its operation.

CFD methods are used for the analysis of wind turbine performance but CFD models do not incorporate physical system inertial response or provide a physical test bed. WTE must, therefore, continue to play a role in the support of wind turbine design, research and development. Ongoing work, on enhancing wind turbine designs for use in the urban environment, is leading towards the use of complex, drag-based, multi-bladed, vertical-axis devices to deal with the problems inherent in the urban situation. Current WTE systems are found to be incapable of modelling the complex torque output of these devices adequately, since they are based on the use of a steady-state model modulated by an approximating analytical function. The WTE technique developed in this thesis uses CFD profiles mapped to a two-dimensional array to generate torque coefficient values in real time. Initially a standard inertia-compensation scheme, based on an acceleration observer, is used, but testing shows that this method is inadequate due to the requirement for a low-pass filter in the feedback path. To achieve the performance necessary to accurately model the output from a multi-bladed VAWT, a novel inertia-compensation scheme is designed and implemented. The improved technique eliminates the filter induced performance degradation by dynamically manipulating z-plane pole positions based on real-time observation of system stiffness and incorporating an adaptive `lag-lead' pre-filter at the input to the torque compensation control loop. System behaviour is approximated by an s-plane model and pole manipulation is achieved by dynamic modulation of the wind turbine (WT) moment of inertia quantity used in the feedback path.

Tests show that the novel inertia compensation scheme meets the requirements for accurate emulation of VAWT performance. Mean torque and power output tests confirm that, for all profiles used, the output accurately reflects those predicted by the original CFD analyses. Time and frequency domain analyses of generator load current signals confirm that the technique facilitates the analysis of generator output signals on a WTE test bed for the development of fault diagnosis and Condition Based Maintenance (CBM) strategies.