Model updating methodologies for multibody simulation models: application to a full-scale wind turbine model Simone Manzato, Bart Peeters, Alessandro Toso, Herman Van der Auweraer LMS International Interleuvenlaan 68 B-3001 Leuven, Belgium Richard Osgood National Renewable Energy Laboratory National Wind Technology Center – MS 3911 1617 Cole Blvd. Golden, CO 80401, USA ABSTRACT In the industrial environment, the request for accurate models able to predict the behavior of a structure in different operating conditions is continuously increasing. To analyze the dynamic performances of complex mechanisms, multibody models are widely used, in particular if control laws for these systems need to be defined and tested. Wind turbines represent a typical application in which multibody models are used for control laws development. In this paper, an Experimental Modal Analysis campaign on the CART-3 wind turbine is used as reference to update a tailored multibody model. Standard model updating techniques based on mode shapes and natural frequencies are adapted to be used in a professional CAE environment. The results obtained show a big influence of the correlation indices selected to drive the updating phase. 1. INTRODUCTION In modern analysis of structural dynamics, an always increasing effort is devoted to the derivation of accurate and reliable models which can be used to simulate the real behavior of the system and optimize its performances. By using experimental data measured on the actual structure, parameter identification and model updating techniques can be applied to refine, correct or update a numerical model [1,2]. However, when applying these techniques, it must be remembered that both the numerical and experimental models are subjected to different types of errors. While numerical models are limited by the assumptions introduced to represent the behavior of the real structure, experimental results can be considered valid only for the test conditions and the amount of information is limited by the number and nature of measured points. Other errors can be due to insufficient excitation, response bandwidth or assumptions introduced during the processing of the results [3,4]. Historically, structural identification and model updating techniques were developed to improve the accuracy of Finite Element models, which are nowadays an industrial standard to simulate and analyze the static and dynamic behavior of mechanical systems. An overview of the more common methods for model updating is given in [1,2,5]. In most applications, the correlation between experimental and numerical models is performed by comparing the natural frequencies and the mode shapes or the Frequency Response Functions. Natural frequencies are usually compared in terms of error between the numerical and experimental values. Common correlation indices developed to compare mode shapes or Frequency Response Function are the Modal Assurance Criterion (MAC) and Frequency Response Assurance Criterion (FRAC), to compare each numerical vector with all the experimental one, and the CoOrdinate Modal Assurance Criterion (COMAC) and the Response Vector Assurance Criterion (RVAC) to compare all the response values at each degree of freedom (COMAC) or natural frequency (RVAC). Once the correlation between the models is defined, different techniques can be applied to estimate the value of the parameters of the model or to improve them with model updating techniques. By using direct methods, the elements in the stiffness and mass matrices can be updated in a one-step procedure. The main drawback of this technique is that the noise and inaccuracies in the experimental results are exactly reproduced in the numerical model. Moreover, the updated model won’t maintain the structural connectivity and the corrected parameters could be not physically meaningful. To overcome these limitations, iterative methods were developed, which are usually posed as optimization problems and make use of the T. Proulx (ed.), Linking Models and Experiments, Volume 2, Conference Proceedings of the Society for Experimental Mechanics Series 5, 349 DOI 10.1007/978-1-4419-9305-2_24, © The Society for Experimental Mechanics, Inc. 2011
RkJQdWJsaXNoZXIy MTMzNzEzMQ==