This paper describes the optimization of the beam stiffeners attached to plates in an eigenfrequency problem. The solution space is estimated using the Kriging method. A finite element analysis is carried out to evaluate the objective function at the sample points used for the estimation. The gradient method is used as a local optimizer. The Kriging estimation incurs relatively low cost, and is easy to combine with the gradient method.In this paper, we solve eigenfrequency optimization problems for a fully supported plate to maximize the minimum eigenfrequency and the difference between the 1st-and 2nd-order eigenfrequency. An optimization problem for an L-shaped plate with an eigenfrequency constraint is also solved. Good solutions are obtained for each example, and all the optimizations for these problems can be done at a lower computational cost. The results highlight the effectiveness of the method to solve the eigenfrequency optimization problems for the stiffened plate.
IntroductionThis paper deals with the optimization of the stiffeners attached to plates in an eigenfrequency problem. The stiffened plate is useful in industry because it can easily improve the mechanical properties of a structure at a lower cost. A beam reinforcement for a plate or other general structure can enhance a mechanical property, and it will be available for higher reliability or longer life of a structure at a lower cost [1]. Recently, the more effective design using such as topology optimization of the stiffeners was reported by Gea [2] or Lee [3], while the easier design method to improve the current performance should be discussed for industrial use. For the purpose of the easier design, the evaluation and design process of the structure should be simplified.Considering a use of the stiffened plate, properties for the bending deformation, buckling load and vibration should be basically improved. Such kinds of mechanical behavior of a structure are often analyzed using numerical analysis method, such as finite element method (FEM). Some kinds of FEM calculation incur higher computational costs, and present some numerical problems. As an Eigenfrequency problem is solved using both iteration and convergent methods such as the subspace method, we consider that numerical errors in addition to increasing the computational cost may be introduced. These problems introduce some difficulties into the structural optimization process. Salajegheh attempted to reduce the cost of optimization with frequency constraints [4].To improve the mechanical property for industry, structural optimization using a gradient method can be applied to engineering design because of the robustness of the solution. Information on the sensitivity of design variables to an objective function, which is very important to industrial design, is also evaluated. However, optimization using a gradient method involves the iterated calling of the solver, such that the computational cost of optimization tends to increase. Additionally, the numerical errors and noises that...