{"title":"利用人工神经网络对 MHD 杰弗里-哈梅尔流进行稳定性分析","authors":"","doi":"10.1016/j.ijft.2024.100834","DOIUrl":null,"url":null,"abstract":"<div><p>The current research explores the useful impact of artificial neural networks (ANN) back-propagation along Levenberg-Marquardt Method (ANN-BLMM), for findng the influence of dimensionless numbers on the flow distribution pertaining magnetohydrodynamics (MHD) Jeffery–Hamel fluid amid two plates which are enclined at angles 2α. We have employed a numerical approach, ANN-BLMM to assess various aspects of our data, including testing, training, validation, Mean Square Errors (MSE), performance, and fitting. The methodology being used has been tested and validated through comparison with other results obtain numerically, showing extreme level of accuracy. Moreover, we have confirmed our findings through error histograms and regression tests. We used OHAM for the data set. Furthermore, we have also explored the influence of Reynolds number (R) on both flow and pressure distribution, visually representing our findings through graphical analysis. We have discussed about the nature and variations of velocity profiles within MHD Jefery–Hamel flow (MHDJHF), taking into account various values of Ha and <em>Re</em> in both convergent and divergent channels. It was discovered that a significant stabilizing influence of an increase in the magnetic field intensity was observed for both diverging and converging channel geometries and as the Hartman numbers rise, so does the fluid velocity. The absolute error is reduced to 10<sup>–2</sup> to 10<sup>–6</sup>.</p></div>","PeriodicalId":36341,"journal":{"name":"International Journal of Thermofluids","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666202724002751/pdfft?md5=42a5f893b69289732b5debb1c3bcfd16&pid=1-s2.0-S2666202724002751-main.pdf","citationCount":"0","resultStr":"{\"title\":\"Stability analysis of MHD Jeffery–Hamel flow using artificial neural network\",\"authors\":\"\",\"doi\":\"10.1016/j.ijft.2024.100834\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>The current research explores the useful impact of artificial neural networks (ANN) back-propagation along Levenberg-Marquardt Method (ANN-BLMM), for findng the influence of dimensionless numbers on the flow distribution pertaining magnetohydrodynamics (MHD) Jeffery–Hamel fluid amid two plates which are enclined at angles 2α. We have employed a numerical approach, ANN-BLMM to assess various aspects of our data, including testing, training, validation, Mean Square Errors (MSE), performance, and fitting. The methodology being used has been tested and validated through comparison with other results obtain numerically, showing extreme level of accuracy. Moreover, we have confirmed our findings through error histograms and regression tests. We used OHAM for the data set. Furthermore, we have also explored the influence of Reynolds number (R) on both flow and pressure distribution, visually representing our findings through graphical analysis. We have discussed about the nature and variations of velocity profiles within MHD Jefery–Hamel flow (MHDJHF), taking into account various values of Ha and <em>Re</em> in both convergent and divergent channels. It was discovered that a significant stabilizing influence of an increase in the magnetic field intensity was observed for both diverging and converging channel geometries and as the Hartman numbers rise, so does the fluid velocity. The absolute error is reduced to 10<sup>–2</sup> to 10<sup>–6</sup>.</p></div>\",\"PeriodicalId\":36341,\"journal\":{\"name\":\"International Journal of Thermofluids\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.sciencedirect.com/science/article/pii/S2666202724002751/pdfft?md5=42a5f893b69289732b5debb1c3bcfd16&pid=1-s2.0-S2666202724002751-main.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Journal of Thermofluids\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2666202724002751\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"Chemical Engineering\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Thermofluids","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2666202724002751","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"Chemical Engineering","Score":null,"Total":0}
引用次数: 0
摘要
目前的研究探讨了人工神经网络(ANN)反向传播与 Levenberg-Marquardt 法(ANN-BLMM)的有益影响,以发现无量纲数对磁流体动力学(MHD)杰弗里-哈梅尔流体在两块围成 2α 角的板之间的流动分布的影响。我们采用了数值方法 ANN-BLMM 来评估数据的各个方面,包括测试、训练、验证、均方误差 (MSE)、性能和拟合。通过与其他数值结果的比较,我们对所使用的方法进行了测试和验证,结果显示了极高的准确性。此外,我们还通过误差直方图和回归测试证实了我们的研究结果。我们使用 OHAM 作为数据集。此外,我们还探讨了雷诺数(R)对流量和压力分布的影响,并通过图形分析直观地展示了我们的研究结果。考虑到会聚和发散通道中不同的 Ha 和 Re 值,我们讨论了 MHD Jefery-Hamel 流动(MHDJHF)中速度剖面的性质和变化。研究发现,无论是发散还是收敛通道几何形状,磁场强度的增加都会产生显著的稳定影响,而且随着哈特曼数的增加,流体速度也会增加。绝对误差减小到 10-2 到 10-6。
Stability analysis of MHD Jeffery–Hamel flow using artificial neural network
The current research explores the useful impact of artificial neural networks (ANN) back-propagation along Levenberg-Marquardt Method (ANN-BLMM), for findng the influence of dimensionless numbers on the flow distribution pertaining magnetohydrodynamics (MHD) Jeffery–Hamel fluid amid two plates which are enclined at angles 2α. We have employed a numerical approach, ANN-BLMM to assess various aspects of our data, including testing, training, validation, Mean Square Errors (MSE), performance, and fitting. The methodology being used has been tested and validated through comparison with other results obtain numerically, showing extreme level of accuracy. Moreover, we have confirmed our findings through error histograms and regression tests. We used OHAM for the data set. Furthermore, we have also explored the influence of Reynolds number (R) on both flow and pressure distribution, visually representing our findings through graphical analysis. We have discussed about the nature and variations of velocity profiles within MHD Jefery–Hamel flow (MHDJHF), taking into account various values of Ha and Re in both convergent and divergent channels. It was discovered that a significant stabilizing influence of an increase in the magnetic field intensity was observed for both diverging and converging channel geometries and as the Hartman numbers rise, so does the fluid velocity. The absolute error is reduced to 10–2 to 10–6.