Experimental and simulated modal analyses of stator coils of electric generators

The mechanical behavior of a stator has not been frequently simulated using finite element method (FEM). This may be due to unknown material parameters of the materials used in the stator windings or as well the complex geometry and composite structure of a stator that has made creating a reasonably sized finite element model (FE models) very hard if not impossible. The goal of this study was to find an effective way to simulate the dynamics of stator windings accurately using FEM. The unknown material parameters of the materials used in stator windings were defined as a part of this study. The goals of this study were reached by first measuring the natural frequencies and mode shapes of three different pieces of coil of a stator. Then FE models of the three measured pieces of coil were created and the material parameters in the FE models were adjusted so that the calculated natural frequencies and mode shapes corresponded with the measured. The comparison of the measured and simulated mode shapes was done visually and using MAC. Based on the results of this study it can be said that the goals were reached. An accurate and effective way of modelling stator windings was found and the material parameters of the materials used in the windings and the effect of temperature of the coil to its dynamics were defined. The MAC comparisons showed that the correlation between the experimental and simulated modal analyses was very good for the first two pieces of coil and adequate for the third piece of coil. The differences between the calculated and measured natural frequencies were small in general. However, the maximum absolute difference between the natural frequencies of a natural mode shape pair was 19.8 %. The FE models of the pieces of coil were updated to match the results of the experimental modal analyses by modifying only the value of modulus of elasticity of the materials used in the windings. Although the correlation between the experimental and simulated modal analyses was good, better results could be achieved by adding more variable parameters, such as Poisson's ratio and density or by using anisotropic material definition, in the model updating process. It would be recommended to use an existing routine in the model updating process if the number of the variable parameters is increased