Problem-independent machine learning (PIML)-based topology optimization—A universal approach

Huang, Mo; Du, Zongliang; Liu, Chang; Zheng, Yonggang; Cui, Tianchen; Mei, Ying; Li, Xiao; Zhang, Xiaoyu; Guo, Xu

doi:10.1016/j.eml.2022.101887

Cited by 30 publications

(4 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mengcheng Huang et al (2022) prepared the substrate provides a Problem-Independent Machine Learning (PIML) technique for reducing finite element analysis computation time, that is a significant obstacle to solving the problem [19]. Responding to earlier studies, the proposed mechanism is really issue-agnostic and applicable to any topology optimization problem during the minimal off-line training has been completed.…”

Section: Literature Surveymentioning

confidence: 96%

Topology Optimization using Nonlinear Finite Element Analysis

Kushwaha

2023

cana

View full text Add to dashboard Cite

Topology optimization is a structural-mechanical investigation that structures optimization function values in an optimization iteration. Topology optimization with nonlinear static behavior is difficult due to several design considerations. Due of significant integrate distortion, low-density finite elements provide significant numerical challenges in the prevailing element density focused topology planning taking nonlinear dynamics under factor. Iterative procedures are used in the Bi-directional Evolutionary Structural Optimization (BESO) technique for reducing waste from a structure while concurrently adding efficient material using a finite element-based topology optimization strategy. Integrating the fully-connected Deep Neural Network (DNN) with the norm-level-set techniques yields a powerful method for optimizing structural topology. Hence, BESO-DNN has been designed spatial optimization of material distribution inside a specified area is the focus of topology optimization is a mathematical approach for minimizing a certain cost function while meeting a set of predefined restrictions. Topology optimization of geometrically non-linear systems is advantageous because of the solutions' lack of intermediate-density components and their great processing efficiency of high-resolution barrier representation can be effective. As a result, topology optimization is ensuring effective and providing fresh and efficient designs, understanding that outcomes needs an integration of credible decision-making, domain knowledge, and numerical analytic skills. The outcome of finite element analysis has to understand and anticipate an object's performance under various physical conditions using mathematical models, and testing.

show abstract

Section: Literature Surveymentioning

confidence: 96%

Topology Optimization using Nonlinear Finite Element Analysis

Kushwaha

2023

cana

View full text Add to dashboard Cite

show abstract

“…PINNs even works in noisy and high dimensional contexts and does not suffer the curse of dimensionality. Another advantage of PINNs that it can use automatic differentiation (AD), which is a technique for computing the derivatives of a function without explicitly writing them out [25]. AD is conveniently integrated into many ML libraries, such as TensorFlow and PyTorch.…”

Section: Preliminariesmentioning

confidence: 99%

Application of Physics-Informed Neural Networks for Simulating Mass-Spring-Damper Systems

Sahoo,

Kumar,

Chakraverty

2024

Preprint

View full text Add to dashboard Cite

Dynamical problems are generally governed by a set of linear/non-linear differential equations (DEs). A large amount of prior physical information in the form of DEs plays an important role in simulation of dynamical systems. However, the traditional data-hungry machine learning models fail to express insightful scientific information from the data. Most of the implementations of neural networks are to perform non-linear mapping from input space to target space. However, Physics-informed neural networks (PINNs) can bridge the gap between scientific computing and data-hungry models. This paper exploits a new application of PINNs for approximating the behaviour of mass-spring-damper systems, showcasing how PINNs can effectively blend scientific principles with data-driven modeling. In this regard, we present solutions of two realistic application problems using PINNs. The accuracy of the predicted displacements of objects is established through results from literatures.

show abstract

“…Since the efficiency of ML depends on the number of design variables, ML is especially suitable for the geometrybased TO methods (e.g., MMC and MMB methods) where the number of design variables are much smaller than conventional wise-pixel TO methods, and some successful applications can be found in [140,141]. Huang et al [142] proposed a problem-independent machine learning technique to reduce the computational time associated with finite element analysis in TO problems. Different from the above methods, Chi et al [143] proposed a new universal ML-based TO to accelerate the design process of large-scale problems, which collected the training samples simultaneously as the TO proceeds.…”

Section: Machine Learning Tomentioning

confidence: 99%

Open-Source Codes of Topology Optimization: A Summary for Beginners to Start Their Research

Wang¹,

Li²,

Long³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

Topology optimization (TO), a numerical technique to nd the optimal material layout with a given design domain, has attracted interest from researchers in the eld of structural optimization in recent years. For beginners, opensource codes are undoubtedly the best alternative to learning TO, which can elaborate the implementation of a method in detail and easily engage more people to employ and extend the method. In this paper, we present a summary of various open-source codes and related literature on TO methods, including solid isotropic material with penalization (SIMP), evolutionary method, level set method (LSM), moving morphable components/voids (MMC/MMV) methods, multiscale topology optimization method, etc. Simultaneously, we classify the codes into ve levels, from easy to di cult, depending on their di culty, so that beginners can get started and understand the form of code implementation more quickly.

show abstract

Problem-independent machine learning (PIML)-based topology optimization—A universal approach

Cited by 30 publications

References 38 publications

Topology Optimization using Nonlinear Finite Element Analysis

Topology Optimization using Nonlinear Finite Element Analysis

Application of Physics-Informed Neural Networks for Simulating Mass-Spring-Damper Systems

Open-Source Codes of Topology Optimization: A Summary for Beginners to Start Their Research

Contact Info

Product

Resources

About