6533b832fe1ef96bd129a570

RESEARCH PRODUCT

An approach to the optimum design of the magnetic circuit of a PM linear electrical generator driven by stochastical forces

subject

description

Linear electrical generator has been recently studied for the exploitation of sea waves energy. In order to maximize the energy extraction, the stochastic features of the energy source must be included in the mathematical model of the system. A PM linear electrical generator has been designed and built. The design has been optimized by including in the mathematical model used for the optimization the following aspects: 1. the dynamic and stochastic features of the waves; 2. a dynamic model of the hydraulic conversion system; 3. a parametric circuit model of the magnetic circuit as well as of the electric windings; 4. a circuit model of the power electronics converter used to connect the generator to the utility.The main innovation of the optimisation approach proposed is that the speed of the machine is not either supposed known either supposed fixed, but the stochastic features of the imposed movement are the factors that mostly influences the optimisation process. The optimization has been performed by finding the maximum of a objective function that has been built on the basis of the mathematical model developed. The variables of the function were related to the magnetic circuit of the converter, the parameters to the stochastic features of the waves as well as to the characteristics of the hydraulic system and of the power electronics converter. The results show several maxima that indicate several possible alternative designs. In order to verify the results, a reconfigurable machine has been built and tested in laboratory. The tests have verified the value of the emf, of the magnetic induction as well as of the power production in several configuration of the machine. In a scaled down version of the machine, it has been possible to generate almost 100 watt and to reach an output voltage of 200 V peak to peak. These values were in agreement with the design values within an error of 15%.