A Lab-oriented Testing Platform for Emulating a Wind Turbine in a DC Microgrid

This paper describes the process of devising a wind turbine emulator using spare equipment selected out of information obtained from a complete characterization of an actual 600W wind turbine based on a permanent-magnet electrical machine. The challenge is to assemble a working wind turbine emulator reflecting the turbine behavior which consists on that the shaft spin is not fixed for a given wind velocity but depends on the target power to be delivered. This task is achieved via a PID controller that can be tuned after proper modelling of the wind turbine emulator. Since the emulator components are different and include even nonlinear components, a strategy of black box modelling is preferred. Artificial Neural Networks are used due to their capabilities as universal approximators, and supervised learning is chosen as the learning strategy due to its positive results in black box modelling. A DC generator is selected for the emulator which eases its interconnection with a DC microgrid testbench available at a laboratory intended for studying the effect of renewable energies in conventional power systems. The excitation voltage and the RPM range of the DC generator allow to create an emulator with a 10:1 power ratio that in fact agrees with the power handling specifications of the microgrid testbench.