Machine Learning and Metamodel-Based Design Optimization of Nonlinear Multimaterial Structures

2016 ◽  
Author(s):  
Kai Liu ◽  
Duane Detwiler ◽  
Andres Tovar
Keyword(s):  
Machine Learning ◽  
Design Optimization ◽  
Conceptual Design ◽  
Optimization Method ◽  
Machine Learning Techniques ◽  
Efficient Global Optimization ◽  
Element Analysis ◽  
Multi Objective ◽  
Design Generation ◽  
Optimal Material

This study presents an efficient multimaterial design optimization algorithm that is suitable for nonlinear structures. The proposed algorithm consists of three steps: conceptual design generation, design characterization by machine learning, and metamodel-based multi-objective optimization. The conceptual design can be generated from extracting finite element analysis information or by using structure optimization. The conceptual design is then characterized by using machine learning techniques to dramatically reduce the dimension of the design space. Finally, metamodels are derived using Efficient Global Optimization (EGO) followed by multi-objective design optimization to find the optimal material distribution. The proposed methodology is demonstrated using examples from multiple physics and compared with traditional multimaterial topology optimization method.

Multi-Objective Design Optimization of Cooling Turbine Blade Based on Kriging Model

2012 ◽  
Vol 184-185 ◽  
pp. 316-319
Author(s):  
Liang Bo Ao ◽  
Lei Li ◽  
Yuan Sheng Li ◽  
Zhi Xun Wen ◽  
Zhu Feng Yue
Keyword(s):  
Design Optimization ◽  
Turbine Blade ◽  
Optimization Model ◽  
Sampling Technique ◽  
Latin Hypercube Sampling ◽  
Kriging Model ◽  
Optimization Method ◽  
Approximate Model ◽  
Multi Objective ◽  
Pin Fin

The multi-objective design optimization of cooling turbine blade is studied using Kriging model. The optimization model is created, with the diameter of pin fin at the trailing edge of cooling turbine blade and the location, width, height of rib as design variable, the blade body temperature, flow resistance loss and aerodynamic efficiency as optimization object. The sample points are selected using Latin hypercube sampling technique, and the approximate model is created using Kriging method, the set of Pareto-optimal solutions of optimization objects is obtained by the multi-object optimization model using elitist non-dominated sorting genetic algorithm (NSGA-Ⅱ) based on the approximate model. The result shows that the conflict among all optimization objects is solved effectively and the feasibility of the optimization method is improved.

The Multi-Objective Optimization of the Planet Carrier in Wind Turbine Gearbox

2012 ◽  
Vol 184-185 ◽  
pp. 565-569 ◽  
Author(s):  
Peng Xing Yi ◽  
Li Jian Dong ◽  
Yuan Xin Chen
Keyword(s):  
Finite Element ◽  
Design Optimization ◽  
Response Surface ◽  
Minimum Principle ◽  
Element Model ◽  
Multi Objective Optimization ◽  
Wind Turbine Gearbox ◽  
Element Analysis ◽  
Multi Objective ◽  
Planet Carrier

In order to improve the reliability of a planet carrier, a simulation method based on multi-objective design optimization was developed in this paper. The objective of the method was to reduce the stress concentration, the deformation, and the quality of the planet carrier by optimizing the structure dimension. A parametric finite element model, which enables a good understanding of how the parameters affect the reliability of planet carrier, was established and simulated by ANSYS-WORKBENCH. The efficiency of the design optimization was improved by using a polynomials response surface to approximate the results of finite element analysis and a screening algorithm to determine the direction of optimization. Furthermore, the multi-objective optimization was capable of finding the global minimum results in the use of the minimum principle on the response surface. Computer simulation was carried out to verify the validity of the presented optimization method, by which the quality and the stability of the planet carrier were significantly reduced and improved, respectively. The methodology described in this paper can be effectively used to improve the reliability of planet carrier.

DBD Plasma Actuator Multi-Objective Design Optimization at Reynolds Number 63,000: Baseline Case

2013 ◽  
Author(s):  
Taufik Sulaiman ◽  
Satoshi Sekimoto ◽  
Tomoaki Tatsukawa ◽  
Taku Nonomura ◽  
Akira Oyama ◽  
...  
Keyword(s):  
Reynolds Number ◽  
Design Optimization ◽  
Linear Trend ◽  
Lift Coefficient ◽  
Optimization Method ◽  
Input Power ◽  
Plasma Actuator ◽  
Dbd Plasma ◽  
Multi Objective ◽  
Design Variables

The working parameters of the dielectric barrier discharge (DBD) plasma actuator were optimized to gain an understanding of the flow control mechanism. Experiments were conducted at a Reynolds number of 63,000 using a NACA 0015 airfoil which was fixed to the stall angle of 12 degrees. The two objective functions are: 1) power consumption (P) and 2) lift coefficient (Cl). The goal of the optimization is to decrease P while maximizing Cl. The design variables consist of input power parameters. The algorithm was run for 10 generations with a total population of 260 solutions. Although the number of generations and population size was limited due to experimental constraints, the algorithm was able to converge and the approximate Pareto-front was obtained. From the objective function space, we observe a relatively linear trend where Cl increases with P and after a certain threshold, the value of Cl seems to saturate. We discuss the results obtained in the objective space in addition to scatter plot matrix and color maps. This article, with its experiment-based approach, demonstrates the robustness of a Multi-Objective Design Optimization method and its feasibility for wind tunnel experiments.

scholarly journals Multi-objective Optimization of Compliant Parallel Platform for Desired Stiffness and Workspace

2021 ◽  
Author(s):  
Hongwei Xu ◽  
Haibo Zhou ◽  
Zhiqiang Li ◽  
Xia Ju
Keyword(s):  
Numerical Model ◽  
Optimization Problem ◽  
Structural Parameters ◽  
Optimization Method ◽  
Multi Objective Optimization ◽  
Element Analysis ◽  
Multi Objective ◽  
Performance Indexes ◽  
Parallel Platform ◽  
The Relationship

Abstract Stiffness and workspace are crucial performance indexes of a precision mechanism. In this paper, an optimization method is presented, for a compliant parallel platform to achieve desired stiffness and workspace. First, a numerical model is proposed to reveal the relationship between structural parameters, desired stiffness and workspace of the compliant parallel platform. Then, the influence of the various parameters on stiffness and workspace of the platform is analyzed. Based on Gaussian distribution, the multi-objective optimization problem is transformed into a single-objective one, in order to guarantee convergence precision. Furthermore, particle swarm optimization is used to optimize the structural parameters of the platform, which significantly improve its stiffness and workspace. Last, the effectiveness of the proposed numerical model is verified by finite element analysis and experiment.

Development of a Multi-Objective Preliminary Design Optimization Approach for Axial Flow Compressors

2018 ◽  
Author(s):  
Youwei He ◽  
Jinju Sun ◽  
Peng Song ◽  
Xuesong Wang ◽  
Da Xu
Keyword(s):  
Design Optimization ◽  
Axial Flow ◽  
Optimization Methods ◽  
Optimization Method ◽  
Optimization Approach ◽  
Preliminary Design ◽  
Stringent Requirement ◽  
Stall Margin ◽  
Multi Objective ◽  
Axial Flow Compressors

A preliminary design optimization approach of axial flow compressors is developed. Loss correlations associated with airfoil geometry are introduced to relax the stringent requirement for the designer to prescribe the stage efficiency. In face of the preliminary design complexity resulted from the large number of design variables together with their stringent variation ranges and multiple design goals, the multi-objective optimization algorithm is incorporated. With such a developed preliminary design optimization method, the design space can be then explored extensively and the optimum designs of both high level overall efficiency and wide stall margin can be readily achieved. The preliminary design optimization method is validated in two steps. Firstly, an existing 5-stage compressor is redesigned without optimization. The obtained geometries and flow parameters are compared to the existing data and a good consistency is achieved. Then, the redesigned compressor is used as initial design and optimized by the developed multi-objective preliminary design optimization method, and significant performance gains are obtained, which demonstrates the effectiveness of the developed optimization methods.

scholarly journals SADEA-II: A generalized method for efficient global optimization of antenna design

2016 ◽  
Vol 4 (2) ◽  
pp. 86-97 ◽  
Author(s):  
Bo Liu ◽  
Slawomir Koziel ◽  
Nazar Ali
Keyword(s):  
Data Mining ◽  
Global Optimization ◽  
Design Optimization ◽  
Surrogate Model ◽  
Simulation Models ◽  
Optimization Methods ◽  
Optimization Method ◽  
Antenna Design ◽  
Efficient Global Optimization ◽  
Optimization Framework

Abstract Efficiency improvement is of great significance for simulation-driven antenna design optimization methods based on evolutionary algorithms (EAs). The two main efficiency enhancement methods exploit data-driven surrogate models and/or multi-fidelity simulation models to assist EAs. However, optimization methods based on the latter either need ad hoc low-fidelity model setup or have difficulties in handling problems with more than a few design variables, which is a main barrier for industrial applications. To address this issue, a generalized three stage multi-fidelity-simulation-model assisted antenna design optimization framework is proposed in this paper. The main ideas include introduction of a novel data mining stage handling the discrepancy between simulation models of different fidelities, and a surrogate-model-assisted combined global and local search stage for efficient high-fidelity simulation model-based optimization. This framework is then applied to SADEA, which is a state-of-the-art surrogate-model-assisted antenna design optimization method, constructing SADEA-II. Experimental results indicate that SADEA-II successfully handles various discrepancy between simulation models and considerably outperforms SADEA in terms of computational efficiency while ensuring improved design quality. Highlights An EFFICIENT antenna design global optimization method for problems requiring very expensive EM simulations. A new multi-fidelity surrogate-model-based optimization framework to perform RELIABLE efficient global optimization A data mining method to address distortions of EM models of different fidelities (bottleneck of multi-fidelity design).

Sign in / Sign up

Export Citation Format

Share Document