Study Of Model Simplification Method Of Complex Enclosure On Radar Structure

Radar structure is always being analyzed by relative vibration experiment duringstudy. The analysis of vibration experiment is difficult to do for the complexity of radarstructure. In order to improve the efficiency of structure design, a simplified model isneeded to replace the radar structure. Model simplification method of complexenclosure on radar structure is the main research content in the proposed paper.According to the research subject, the basic thought of the method is to simplify thecomplex enclosure to a simple shell with the same shape and dimension. Each modalcharacteristic of the complex enclosure and its simplified model is respectivelyequivalent, which is the basis of the simplification. Their average error must be lessthan10percent. Then the optimal material parameters of the simplified model areobtained.The model simplification methods at home and abroad are being researched. A kindof model simplification method of complex enclosure is designed, based on geneticalgorithm and the finite element model updating. The mixed programming technique ofANSYS and MATLAB is used to realize the method. The method has five basic steps.Firstly, material parameters of the simplified model are selected in the range of designvariable, as the initial population of genetic algorithm by MATLAB. Secondly, materialparameters are assigned to the simplified model after being passed into ANSYS. Thirdly,the simplified model is analyzed to gain its modal results by ANSYS. Fourthly, thefitness values of individuals are calculated based on the modal results, which are passedinto MATLAB. Finally, the optimal material parameters are reserved to realize modelsimplification of complex enclosure by a series of iterative operation. Aftersimplification, the average error is about9percent. And6hours and40minutes areneeded for the method.The first model simplification method needs to be improved for its lower efficiencyand accuracy of the simplification. The second model simplification method is designedbased on improving the first one. In this method, the processes of data transfer and finiteelement analysis are replaced by the BP neural network model. Then the optimalmaterial parameters of the simplified model are worked out to realize modelsimplification. After simplification, the average error is about8percent. And25minutes is needed for this method. The efficiency and accuracy of the second model simplification method are betterthan those of the first one by comparing them. Finally, the second model simplificationmethod is developed into a form of client software, which is convenient to simplify themodel of similar structure, by mixed programming technique of VB.NET andMATLAB.