Novel Dynamic Loads Identification Method And Application To Hydroelectric Units

Along with the fast development of hydropower construction in China, the capacity and dimension of the hydroelectric unit increase constantly, and the rotational speed increases correspondingly. The vibrational problems of hydroelectric units emerge, which becomes an important problem to be solved. Severe vibrations of hydroturbine units had occurred in some hydropower plants, with the increase of capacity and the water head. So it is urgent to study the characteristics of the dynamic loads by constructing numerical model to identify the dynamic loads.In order to solve the problems of the current dynamic loads identification in time domain, this dissertation invents a new identification method based on modal coordinate transformation, which is called five-points identification method. The general idea is to use the linear combination of the displacements of the measuring points at t+2Δt, t+Δt, t, t-Δt and t-2Δt to determine the loads at t. The identification coefficients can be given by the calculation using the system parameters. If the system parameters are unknown, an experiment or FEA should be conducted to obtain the displacements under the given dynamic loads, then the loads and the corresponding displacements can be substituted to the identification equations system, and the identification coefficients can be obtained by solving the equations system by least squares method. This identification method requires that the number of the measuring points is no less than that of the used modes which is no less than the number of the loads. Since this identification method is not a recursion-chain type, there is not the error accumulation occurring in the traditional time domain identification method.The validation of the five-points identification method is performed through the calculation for the SOF and MOF systems under loads with various complexity and through the experiment on a simply supported beam, and the results show that this method is very effective. Moreover, this dissertation analyses the effects of the different parameters in this method, including the number of the samples, the number and the locations of the measuring points, the directions of the measured displacements, the completeness of the sample loads, the noise and the loads frequencies. The analysis gives some important characteristics of this method and provides knowledge for improving the identification accuracy.For practical application, this dissertation takes an upper bracket as an example. Firstly, the modal analysis is performed to determine the measuring scheme. Then the loads status during steady operation is analyzed, and the corresponding displacements are computed via FEA method. Finally, the five-point identification method is applied to the dynamic loads identification for the upper bracket. The results show the effectiveness of the five-point identification method.The results of this dissertation make great contribution to the development of the dynamic loads identification.