Papers by Keyword: Approximate Model

Paper TitlePage

Authors: Bo Yang, Xing Jun Hu, Lei Liao, Xu Yan, Zhi Ming Zhao, Jian Ye
Abstract: Morphing surface mesh employing a parametric model for the aerodynamic shape optimization of a sedan is presented. This study solved the problem of non-linear deformation in modeling surface morphing by volume mesh, and solved the divergence problem due to bad mesh quality, using automatic grid generation method based on CAD model, in calculation. The optimization process consists of three stages: Design of Experiments (DOE), Approximate Model, and optimization algorithm execution. The optimization process was realized by an automatic optimization system consisting of multiple software platforms. The optimal model's aerodynamic drag coefficient is 10.56% decreased though optimization, and the hood is identified as the major factor in this research.
Authors: Hong Ge Zhao
Abstract: This paper proposes a robust adaptive neural network controller (RANNC) for electrode regulator system. An equivalent model in affine-like is derived for electrode regulator system. Then, the nonlinear control law is derived directly based on the affine-like equivalent model identified with neural networks, which avoids complex control development and intensive computation. Pretraining is not required and the weights of the neural networks used in adaptive control are directly updated online based on the input-output measurement. The proposed nonlinear controller is verified by computer simulations.
Authors: Ning Ning Gong, Xi Ping Li, Can Yang
Abstract: Integration of approximate model and clonal selection algorithm (CSA) is considered as an effective way to solve complicated optimization design problems in engineering field. The principles of approximate model and of clonal selection algorithm were presented in this paper. A variotherm temperature injection mold used to produce a large-size liquid crystal display (LCD) TV panel was demonstrated and an approximate model for optimizing the layout of heating channels of the mold was established. The clonal selection algorithm program was coded according to its principle in order to solve the established approximate model. The layout of the heating channels was optimized and the optimal solutions were obtained. Finite element simulation and industrial injection production indicated that the integration of the approximate model and clone selection algorithm used in this paper to optimize the layout of the heating channels for the injection mold was very effective.
Authors: Xue Wen Chen, Dong Won Jung, Ai Xue Sun
Abstract: Technology and die design are very important in the development of forging products due to its great influence on the quality, cost and manufacturing efficiency of the final products as well as the life of the forging die. In the environment of the severe competition, how to improve the quality of forging technology and die design, to reduce the product cost and ultimately to enhance competitiveness of the forging factory are the problems that forging technology and die designer have to solve. In order to improve the quality of forging technology and die design, a design optimization method based on approximate model (response surface model) and FEM technique for hot forging process is proposed in this paper. During design optimization process, finite element analysis is incorporated to calculate the objective function and check the design alternatives. Design of experiment (DOE) method is used to collect sample points and calculate the polynomial coefficients of response surface model, and approximate model is used to calculate the optimum search direction. Finally, a case study is conducted for a gear workpiece hot forging process. The objective function is the degree of uniformity of equivalent-strain, which can be defined as mean square deviation of the equivalent-strain in each element and the average equivalent-strain of all elements, and the design parameters are the initial H0/D0 ratio of billet and the key dimensions of the die. Then the design optimization mathematical model is established. The result shows that the objective function value is dropped from 0.7914 and converges at 0.4843 within 17 iterations, the optimal design parameters are obtained.
Authors: Duo Nian Yu, Lu Yao Zhou, Li Li, Zheng Cai Hu
Abstract: Head injury of pedestrian is the most common and fatal cause of mortality in vehicle-to-pedestrian crash. And the engine hood is most likely to cause harm to pedestrian head. Efforts to improve engine hood design, which minimize the head injury of pedestrian in vehicle-to-pedestrian crash, are becoming more and more important. In this study, an approximate model of hood thickness for three targets: HIC, mass and modality, is established. In order to meet the requirements of lightweight and reducing vibration and noise, approximate models iterate by the NSGA-II genetic optimization algorithm, and select the Pareto optimal solutions for thickness optimization. At last the study re-simulates the collision between pedestrian head and hood to verify the reliability of the obtained optimization results.
Authors: Chuan Qing Wang, Deng Feng Wang, Shuai Zhang
Abstract: The 100% frontal crash and side impact performances of a passenger car are analyzed and compared with tests. The structural optimization of the Closed Body-in-White (BIW) is divided into two stages which are 100% frontal crash safe part optimization and side impact safe part optimization. Use the Optimal Latin hypercube (Opt LHD) design method to generate sample points. Take the Radial Basis Functions (RBF) neural network method to establish optimization approximation model. The non-dominated sorting genetic algorithm (NSGA-II) was used to conduct multi-objective collaborative optimization design. The results show that the total mass of the closed BIW is reduced 9.745kg; the light weight rate was 10.27%. The Crashworthiness performance of the closed BIW does not change obviously.
Authors: Wei Wei
Abstract: A method of structure optimization for hydraulic press is proposed in order to reduce mass while assuring adequate stiffness. Upper beam and lower beam are determined as optimal objects by mass analysis. Key geometric parameters of upper beam and lower beam which have relatively larger impacts on mass and stiffness are extracted as design variables. In order to research relationship between stiffness, mass and design variables, command batch file is built by python language to implement automatic finite element analysis in ABAQUS. Orthogonal experimental design is used to generate samples of design variables. Calculating data are dealed with second order stepwise regression and mathematical model for structure optimization is established by regression equations. The goal of structure optimization is to decrease total mass of hydraulic press while assuring adequate stiffness. Particle swarm optimization is used to solve the mathematical model. The total mass of hydraulic press is decreased by 3.1% and its stiffness is adequate to ensure the forming precision when solving process is finished.
Showing 1 to 7 of 7 Paper Titles