{"title":"平轧过程的精确替代模型","authors":"Kheireddine Slimani, Mohamed Zaaf, Tudor Balan","doi":"10.1007/s12289-023-01744-5","DOIUrl":null,"url":null,"abstract":"<div><p>Surrogate models, both polynomial and ANN-based (artificial neural networks), are developed to predict the rolling load in cold rolling of flat metals. An accurate but fast model was developed to serve as high-fidelity model for the training of the machine learning algorithms, allowing for large sampling sizes (up to 1000 samples) with different sampling methods, a number of eight input parameters, and various configurations of surrogate models. The ANN-based models have shown excellent predictive abilities provided that the training sampling is sufficiently large (more than 500 elements). In contrast, polynomial models converge much rapidly to their optimal accuracy (samplings of tens of elements) but their predictive ability is more limited, unless the order of the polynomials is increased. The latin hypercube sampling was more efficient than the random sampling in all cases.</p></div>","PeriodicalId":591,"journal":{"name":"International Journal of Material Forming","volume":null,"pages":null},"PeriodicalIF":2.6000,"publicationDate":"2023-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s12289-023-01744-5.pdf","citationCount":"3","resultStr":"{\"title\":\"Accurate surrogate models for the flat rolling process\",\"authors\":\"Kheireddine Slimani, Mohamed Zaaf, Tudor Balan\",\"doi\":\"10.1007/s12289-023-01744-5\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>Surrogate models, both polynomial and ANN-based (artificial neural networks), are developed to predict the rolling load in cold rolling of flat metals. An accurate but fast model was developed to serve as high-fidelity model for the training of the machine learning algorithms, allowing for large sampling sizes (up to 1000 samples) with different sampling methods, a number of eight input parameters, and various configurations of surrogate models. The ANN-based models have shown excellent predictive abilities provided that the training sampling is sufficiently large (more than 500 elements). In contrast, polynomial models converge much rapidly to their optimal accuracy (samplings of tens of elements) but their predictive ability is more limited, unless the order of the polynomials is increased. The latin hypercube sampling was more efficient than the random sampling in all cases.</p></div>\",\"PeriodicalId\":591,\"journal\":{\"name\":\"International Journal of Material Forming\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.6000,\"publicationDate\":\"2023-03-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://link.springer.com/content/pdf/10.1007/s12289-023-01744-5.pdf\",\"citationCount\":\"3\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Material Forming\",\"FirstCategoryId\":\"88\",\"ListUrlMain\":\"https://link.springer.com/article/10.1007/s12289-023-01744-5\",\"RegionNum\":3,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"ENGINEERING, MANUFACTURING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Material Forming","FirstCategoryId":"88","ListUrlMain":"https://link.springer.com/article/10.1007/s12289-023-01744-5","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, MANUFACTURING","Score":null,"Total":0}
Accurate surrogate models for the flat rolling process
Surrogate models, both polynomial and ANN-based (artificial neural networks), are developed to predict the rolling load in cold rolling of flat metals. An accurate but fast model was developed to serve as high-fidelity model for the training of the machine learning algorithms, allowing for large sampling sizes (up to 1000 samples) with different sampling methods, a number of eight input parameters, and various configurations of surrogate models. The ANN-based models have shown excellent predictive abilities provided that the training sampling is sufficiently large (more than 500 elements). In contrast, polynomial models converge much rapidly to their optimal accuracy (samplings of tens of elements) but their predictive ability is more limited, unless the order of the polynomials is increased. The latin hypercube sampling was more efficient than the random sampling in all cases.
期刊介绍:
The Journal publishes and disseminates original research in the field of material forming. The research should constitute major achievements in the understanding, modeling or simulation of material forming processes. In this respect ‘forming’ implies a deliberate deformation of material.
The journal establishes a platform of communication between engineers and scientists, covering all forming processes, including sheet forming, bulk forming, powder forming, forming in near-melt conditions (injection moulding, thixoforming, film blowing etc.), micro-forming, hydro-forming, thermo-forming, incremental forming etc. Other manufacturing technologies like machining and cutting can be included if the focus of the work is on plastic deformations.
All materials (metals, ceramics, polymers, composites, glass, wood, fibre reinforced materials, materials in food processing, biomaterials, nano-materials, shape memory alloys etc.) and approaches (micro-macro modelling, thermo-mechanical modelling, numerical simulation including new and advanced numerical strategies, experimental analysis, inverse analysis, model identification, optimization, design and control of forming tools and machines, wear and friction, mechanical behavior and formability of materials etc.) are concerned.