6. CONCLUSIONS This paper presents a methodology to update multibody simulation models using experimental modal data. The proposed methodology is based on common updating and parameter identification techniques developed specifically for Finite Element Models, modified to be used with commercial multibody software such as LMS Virtual.Lab Motion. The reference data used for the updating are the modal vectors and natural frequencies identified on the actual structure after performing an Experimental Modal Analysis campaign. To correlate the numerical and experimental model, different correlation indices can be defined: in this case, the Modal Assurance Criterion and the relative error between the natural frequencies are selected. To perform the updating, a single objective optimization problem can be defined, by combining the two correlation indices into a single objective function. Despite being possible to use standard optimization techniques, when combining different indices some weights need to be defined, influencing significantly the results that could be obtained. To solve this problem, multiobjective optimization techniques can be applied, considering separately each correlation indices. The optimal solution will be distributed along a Pareto front, where weights are automatically assigned to each objective function for all the points depending on their location on the front. The proposed methodology is applied to a full scale multibody model of the NREL CART3 wind turbine. Both the single objective and multi-objective optimization problems are applied to try to identify the unknown parameters and update the model. Due to missing information about the model, some geometrical and structural approximation is introduced to build the numerical model. The tower and the blades are modeled as flexible using a lumped-parameter approach, while the nacelle, generator and drive-train are modeled as a single rigid body since the measurement campaign has been performed with the turbine in parked conditions. The two optimization problems are both solved using Genetic Algorithms, to prevent convergence towards local extrema and increase the robustness of the search. The optimal point obtained from the single objective optimization shows a significant reduction in the errors between natural frequencies, while the MAC remains constant. In the multi-objective optimization, the points on the front represent different trade-off solutions, depending on the weights assigned to the objectives. Lower variation of the MAC can be observed, and to obtain some improvement big error between natural frequencies need to be accepted. Hence, a solution which showed both an increase in MAC and a reduction in the natural frequencies index is selected, showing similar results to the one obtained from the single-objective problem. The main advantage of the proposed methodology is its flexibility and the possibility to apply it to a general multibody model for which experimental modal analysis are available. Moreover, depending on the data available, it can be modified to tackle with the specific problem. Great attention should be given to the selection of the correlation indices and the objective functions, since the selection can greatly influence the results that could be obtained. ACKNOLEDGEMENTS The work presented in this paper was performed within the 6th framework of the Marie Curie Host Fellowships Smart Structures (A Computer-Aided Engineering Approach to Smart Structures) project (MRTN-CT-2006-035559). The authors would like to thank the European Commission for the grant received. REFERENCES [1] Friswell M.I. and Mottershead J.E., Finite Element Model Updating in Structural Dynamics, Kluwer Academic Publisher, 1995. [2] Mottershead J.E. and Friswell M.I., Model updating in structural dynamics: A survey, Journal of Sound and Vibrations, 167(2), 347-375, 1993. [3] Jaishi B. and Ren W., Finite element model updating based on eigenvalue and strain energy residuals using multiobjective optimization technique, Mechanical Systems and Signal Processing, 21, 2295-2317, 2007. [4] Christodoulou K. and Papadimitrou C., Structural identification based on optimally weighted modal residual, Mechanical Systems and Signal Processing, 21(1), 4-23, 2007. [5] Heylen W., Lammens S. and Sas P., Modal Analysis Theory and Testing, Katholieke Universiteit Leuven, Department Werktuigkunde, Leuven, 1997. [6] Craig R.R. and Bampton M.C.C., Coupling of substructures for dynamic analysis, AIAA Journal, 6, 1313-1319, 1968. [7] Brüls O., Fisette P. and Verlinden O., Computer-Aided analysis of rigid and flexible multibody system, GraSMech Course 2009-2010 material, Université Catholique de Louvain, 2010. [8] Griffith T.D., Carne T.G. and Paquette J.A., Modal testing validation of blade models, Wind Engineering, 32(2), 91102, 2008. [9] Noesis Solutions, Optimus Rev. 9 Theory Manual, Leuven, Belgium, 2009. [10] Stol K.A., Geometry and structural properties for the controls advanced research turbine (CART) from model tuning, NREL/SR-500-32087, Colorado, USA, 2004. [11] Wright A.D., Fingersh L.J., and Stol K.A., Designing and testing controls to mitigate tower dynamic loads in the controls advanced research turbine, 45th AIAA Aerospace Sciences Meeting and Exhibit,Wind Energy Symposium, Reno, Nevada, USA, 2007. 357
RkJQdWJsaXNoZXIy MTMzNzEzMQ==