Article(id=1244321216186663728, tenantId=1146029695717560320, journalId=1244284848500682798, issueId=1244321215637209904, articleNumber=null, orderNo=null, doi=10.16156/j.1004-7220.2025.05.019, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1740499200000, receivedDateStr=2025-02-26, revisedDate=1742745600000, revisedDateStr=2025-03-24, acceptedDate=null, acceptedDateStr=null, onlineDate=1774598896308, onlineDateStr=2026-03-27, pubDate=1759248000000, pubDateStr=2025-10-01, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1774598896308, onlineIssueDateStr=2026-03-27, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1774598896308, creator=13701087609, updateTime=1774598896308, updator=13701087609, issue=Issue{id=1244321215637209904, tenantId=1146029695717560320, journalId=1244284848500682798, year='2025', volume='40', issue='5', pageStart='1079', pageEnd='1366', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=1, specialIssue=null, createTime=1774598896178, creator=13701087609, updateTime=1774599509568, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1244323788452639476, tenantId=1146029695717560320, journalId=1244284848500682798, issueId=1244321215637209904, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1244323788452639477, tenantId=1146029695717560320, journalId=1244284848500682798, issueId=1244321215637209904, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=1222, endPage=1229, ext={EN=ArticleExt(id=1244321216471876404, articleId=1244321216186663728, tenantId=1146029695717560320, journalId=1244284848500682798, language=EN, title=Prediction of Blood Flow Field in Artery Stenosis Based on Hard Boundary-Constrained Physics-Informed Neural Network, columnId=1244321216404767539, journalTitle=Journal of Medical Biomechanics, columnName=Original Articles, runingTitle=null, highlight=null, articleAbstract=
Objective

To address the limitations of conventional physics-informed neural network (PINN) in handling hemodynamic boundary constraints, an improved hard boundary-constrained PINN (HBC-PINN) framework was proposed to achieve precise prediction of blood flow fields within stenotic arteries.

Methods

An idealized stenosed vessel geometry model was established and computational fluid dynamic simulation was performed to obtain a validation dataset. Appropriate boundary dependent trial functions were designed according to the hard constraint method to embed the flow boundary conditions into the network output. Thus, an HBC-PINN model with the hard boundary constraint method was constructed to predict the velocity field and pressure field of stenosed blood flow. Meanwhile, an original PINN model with the soft constraint method was also built for comparison. By evaluating the accuracy of the two models on the validation dataset, the capability of the HBC-PINN model to simulate hemodynamics without using any labeled data for training was verified.

Results

The effectiveness of the HBC-PINN method in predicting hemodynamic parameters in stenosed blood flow tasks was validated. The relative L2 errors of the flow velocity and pressure predicted by the HBC-PINN in two different stenosis scenarios were both lower than 0.5%, representing an improvement of over 48.8% in accuracy compared to the original PINN model. Additionally, the prediction accuracy of the transverse velocity also increased by more than 35.4%.

Conclusions

Implementing hard constraints on boundary conditions in the PINN modeling process can effectively improve the prediction accuracy of hemodynamic parameters and the efficiency of model solving.

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Huaxin XIANG, Jianbing SANG, Jingyuan Wang, Mengqiang JI, Chen ZHANG), CN=ArticleExt(id=1244321232477340316, articleId=1244321216186663728, tenantId=1146029695717560320, journalId=1244284848500682798, language=CN, title=基于硬边界约束物理信息神经网络狭窄动脉血管的血液流场预测, columnId=1244321216576734006, journalTitle=医用生物力学, columnName=论著, runingTitle=null, highlight=null, articleAbstract=
目的

针对传统物理信息神经网络(physics-informed neural network,PINN)在处理血流边界条件约束时的局限性,提出一种基于硬边界约束物理信息神经网络(hard boundary-constrained physics-informed neural network,HBC-PINN)的改进方法,实现对狭窄动脉血管内血液流场的精确预测。

方法

首先建立理想化狭窄血管几何模型并进行计算流体动力学模拟以获得验证数据集。根据硬约束方法设计合适的边界相关试函数以将流动边界条件嵌入网络输出中,从而构建采用硬边界约束方法的HBC-PINN模型预测狭窄血流的速度场和压力场。同时还构建了采用软约束方法的原始PINN模型作为对比,通过评估两种模型在验证数据集上的准确性,验证不使用任何标记数据训练下HBC-PINN模型模拟血流动力学的能力。

结果

确定了HBC-PINN方法在狭窄血流动力学参数预测任务中的有效性。两种不同狭窄情况下HBC-PINN预测的流向速度和压力的相对L2误差均低于0.5%,相比原始PINN模型精度提升了48.8%以上,垂向速度的预测精度同样提升了超过35.4%。

结论

在PINN建模过程中实施边界条件硬约束可以有效提高对血流动力学参数的预测精度和模型求解效率。

, correspAuthors=null, authorNote=null, correspAuthorsNote=
桑建兵,教授,E-mail:
, copyrightStatement=null, copyrightOwner=《医用生物力学》编辑部, extLink=null, articleAbsUrl=null, sourceXml=H2JI0EIy9oX3k+dU9uEDFQ==, magXml=smaktRCOixAn2N8n1+jnnQ==, pdfUrl=null, pdf=k/rq/jt34QBRepTKTeBXHg==, pdfFileSize=8780417, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=T6EgUNhw6McX7PAYVSliOQ==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=kn1umN2HSJmV1gqcJkKEvw==, mapNumber=null, authorCompany=null, fund=null, authors=

作者贡献声明:

桑建兵负责文章设计和论文修改;向华鑫负责模型计算、数据整理;所有作者讨论并共同撰写论文。

, authorsList=向华鑫, 桑建兵, 王静远, 纪猛强, 张晨)}, authors=[Author(id=1244321233991484112, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1244321234096341723, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321233991484112, language=EN, stringName=Huaxin XIANG, firstName=Huaxin, middleName=null, lastName=XIANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1244321234213782245, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321233991484112, language=CN, stringName=向华鑫, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=河北工业大学 机械工程学院,天津 300401, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])]), Author(id=1244321234331222763, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=sangjianbing@hebut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1244321234478023416, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321234331222763, language=EN, stringName=Jianbing SANG, firstName=Jianbing, middleName=null, lastName=SANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1244321234599658244, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321234331222763, language=CN, stringName=桑建兵, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=河北工业大学 机械工程学院,天津 300401, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])]), Author(id=1244321234721293072, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1244321234868093721, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321234721293072, language=EN, stringName=Jingyuan Wang, firstName=Jingyuan, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1244321234972951332, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321234721293072, language=CN, stringName=王静远, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=河北工业大学 机械工程学院,天津 300401, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])]), Author(id=1244321235178472240, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1244321235312689979, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321235178472240, language=EN, stringName=Mengqiang JI, firstName=Mengqiang, middleName=null, lastName=JI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1244321235425936194, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321235178472240, language=CN, stringName=纪猛强, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=河北工业大学 机械工程学院,天津 300401, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])]), Author(id=1244321235526599498, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1244321235660817236, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321235526599498, language=EN, stringName=Chen ZHANG, firstName=Chen, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1244321235774063448, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, authorId=1244321235526599498, language=CN, stringName=张晨, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=河北工业大学 机械工程学院,天津 300401, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])])], keywords=[Keyword(id=1244321235925058403, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, orderNo=1, keyword=vascular stenosis), Keyword(id=1244321236029916011, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, orderNo=2, keyword=hemodynamics), Keyword(id=1244321236122190705, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, orderNo=3, keyword=physics-informed neural network), Keyword(id=1244321236252214137, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, orderNo=4, keyword=hard constraint method), Keyword(id=1244321236340294526, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, orderNo=1, keyword=血管狭窄), Keyword(id=1244321236440957830, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, orderNo=2, keyword=血流动力学), Keyword(id=1244321237573419918, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, orderNo=3, keyword=物理信息神经网络), Keyword(id=1244321237703443345, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, orderNo=4, keyword=硬约束方法)], refs=[Reference(id=1244321239943201759, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2021, volume=36, issue=3, pageStart=389, pageEnd=395, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=熊柱翔, 袁丁, 郑庭辉, journalName=医用生物力学, refType=null, unstructuredReference=熊柱翔, 袁丁, 郑庭辉, . 肾动脉狭窄形态对其血流动力学的影响[J]. 医用生物力学, 2021, 36(3): 389-395., articleTitle=肾动脉狭窄形态对其血流动力学的影响, refAbstract=null), Reference(id=1244321240048059363, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2021, volume=36, issue=3, pageStart=389, pageEnd=395, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=XIONG ZX, YUAN D, ZHENG TH, journalName=J Med Biomech, refType=null, unstructuredReference=XIONG ZX, YUAN D, ZHENG TH, et al. Influences of renal artery stenosis morphology on its hemodynamics[J]. J Med Biomech, 2021, 36(3): 389-395., articleTitle=Influences of renal artery stenosis morphology on its hemodynamics, refAbstract=null), Reference(id=1244321240161305574, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2024, volume=39, issue=5, pageStart=853, pageEnd=859, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=石政加, 孙丽芳, 赵明轩, journalName=医用生物力学, refType=null, unstructuredReference=石政加, 孙丽芳, 赵明轩, . 数据驱动下合并狭窄左冠状动脉瘤血流动力学参数反演方法[J]. 医用生物力学, 2024, 39(5): 853-859., articleTitle=数据驱动下合并狭窄左冠状动脉瘤血流动力学参数反演方法, refAbstract=null), Reference(id=1244321240253580262, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2024, volume=39, issue=5, pageStart=853, pageEnd=859, url=null, language=null, rfNumber=[2], rfOrder=3, authorNames=SHI ZJ, SUN LF, ZHAO MX, journalName=J Med Biomech, refType=null, unstructuredReference=SHI ZJ, SUN LF, ZHAO MX, et al. Data-driven inversion of hemodynamic parameters for combined stenotic left coronary artery aneurysms[J]. J Med Biomech, 2024, 39(5): 853-859., articleTitle=Data-driven inversion of hemodynamic parameters for combined stenotic left coronary artery aneurysms, refAbstract=null), Reference(id=1244321240375215082, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=13, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[3], rfOrder=4, authorNames=WANG L, DONG D, TIAN FB, journalName=Front Physiol, refType=null, unstructuredReference=WANG L, DONG D, TIAN FB. Fast prediction of blood flow in stenosed arteries using machine learning and immersed boundary-Lattice Boltzmann method[J]. Front Physiol, 2022(13): 953702., articleTitle=Fast prediction of blood flow in stenosed arteries using machine learning and immersed boundary-Lattice Boltzmann method, refAbstract=null), Reference(id=1244321240484266989, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2022, volume=11, issue=22, pageStart=3749, pageEnd=null, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=ALI NF, ATEF M, journalName=Electronics, refType=null, unstructuredReference=ALI NF, ATEF M. LSTM multi-stage transfer learning for blood pressure estimation using photoplethysmography[J]. Electronics, 2022, 11(22): 3749., articleTitle=LSTM multi-stage transfer learning for blood pressure estimation using photoplethysmography, refAbstract=null), Reference(id=1244321240610096113, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2022, volume=323, issue=6, pageStart=H1194, pageEnd=H1205, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=SUN H, LIU J, FENG Y, journalName=Am J Physiol Heart Circ Physiol, refType=null, unstructuredReference=SUN H, LIU J, FENG Y, et al. Deep learning-based prediction of coronary artery stenosis resistance[J]. Am J Physiol Heart Circ Physiol, 2022, 323(6): H1194-H1205., articleTitle=Deep learning-based prediction of coronary artery stenosis resistance, refAbstract=null), Reference(id=1244321240706565108, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2019, volume=null, issue=378, pageStart=686, pageEnd=707, url=null, language=null, rfNumber=[6], rfOrder=7, authorNames=RAISSI M, PERDIKARIS P, KARNIADAKIS GE, journalName=J Comput Phys, refType=null, unstructuredReference=RAISSI M, PERDIKARIS P, KARNIADAKIS GE. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. J Comput Phys, 2019(378): 686-707., articleTitle=Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, refAbstract=null), Reference(id=1244321240937251835, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=107, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[7], rfOrder=8, authorNames=KHARAZMI E, FAN D, WANG Z, journalName=J Fluids Struct, refType=null, unstructuredReference=KHARAZMI E, FAN D, WANG Z, et al. Inferring vortex induced vibrations of flexible cylinders using physics-informed neural networks[J]. J Fluids Struct, 2021(107): 103367., articleTitle=Inferring vortex induced vibrations of flexible cylinders using physics-informed neural networks, refAbstract=null), Reference(id=1244321241046303744, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2023, volume=71, issue=3, pageStart=543, pageEnd=562, url=null, language=null, rfNumber=[8], rfOrder=9, authorNames=BAI J, RABCZUK T, GUPTA A, journalName=Comput Mech, refType=null, unstructuredReference=BAI J, RABCZUK T, GUPTA A, et al. A physics-informed neural network technique based on a modified loss function for computational 2D and 3D solid mechanics[J]. Comput Mech, 2023, 71(3): 543-562., articleTitle=A physics-informed neural network technique based on a modified loss function for computational 2D and 3D solid mechanics, refAbstract=null), Reference(id=1244321241172131843, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2022, volume=8, issue=2, pageStart=741, pageEnd=744, url=null, language=null, rfNumber=[9], rfOrder=10, authorNames=OLDENBURG J, BOROWSKI F, SCHMITZ KP, journalName=Curr Dir Biomed Eng, refType=null, unstructuredReference=OLDENBURG J, BOROWSKI F, SCHMITZ KP, et al. Computation of flow through TAVI device by means of physics informed neural networks[J]. Curr Dir Biomed Eng, 2022, 8(2): 741-744., articleTitle=Computation of flow through TAVI device by means of physics informed neural networks, refAbstract=null), Reference(id=1244321241331515401, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=491, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[10], rfOrder=11, authorNames=PEREZ S, MADDU S, SBALZARINI IF, journalName=J Comput Phys, refType=null, unstructuredReference=PEREZ S, MADDU S, SBALZARINI IF, et al. Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems[J]. J Comput Phys, 2023(491): 112342., articleTitle=Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems, refAbstract=null), Reference(id=1244321242816299019, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2023, volume=20, issue=7, pageStart=11545, pageEnd=11567, url=null, language=null, rfNumber=[11], rfOrder=12, authorNames=DU M, ZHANG C, XIE S, journalName=Math Biosci Eng, refType=null, unstructuredReference=DU M, ZHANG C, XIE S, et al. Investigation on aortic hemodynamics based on physics-informed neural network[J]. Math Biosci Eng, 2023, 20(7): 11545-11567., articleTitle=Investigation on aortic hemodynamics based on physics-informed neural network, refAbstract=null), Reference(id=1244321242937933838, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=361, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[12], rfOrder=13, authorNames=SUN L, GAO H, PAN S, journalName=Comput Methods Appl Mech Eng, refType=null, unstructuredReference=SUN L, GAO H, PAN S, et al. Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data[J]. Comput Methods Appl Mech Eng, 2020(361): 112732., articleTitle=Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data, refAbstract=null), Reference(id=1244321243067957267, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=1998, volume=9, issue=5, pageStart=987, pageEnd=1000, url=null, language=null, rfNumber=[13], rfOrder=14, authorNames=LAGARIS IE, LIKAS A, FOTIADIS DI, journalName=IEEE Trans Neural Netw, refType=null, unstructuredReference=LAGARIS IE, LIKAS A, FOTIADIS DI. Artificial neural networks for solving ordinary and partial differential equations[J]. IEEE Trans Neural Netw, 1998, 9(5): 987-1000., articleTitle=Artificial neural networks for solving ordinary and partial differential equations, refAbstract=null), Reference(id=1244321243202174996, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=389, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=15, authorNames=SUKUMAR N, SRIVASTAVA A, journalName=Comput Methods Appl Mech Eng, refType=null, unstructuredReference=SUKUMAR N, SRIVASTAVA A. Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks[J]. Comput Methods Appl Mech Eng, 2022(389): 114333., articleTitle=Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks, refAbstract=null), Reference(id=1244321243273478168, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=479, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[15], rfOrder=16, authorNames=HUANG YH, XU Z, QIAN C, journalName=J Comput Phys, refType=null, unstructuredReference=HUANG YH, XU Z, QIAN C, et al. Solving free-surface problems for non-shallow water using boundary and initial conditions-free physics-informed neural network (bif-PINN)[J]. J Comput Phys, 2023(479): 112003., articleTitle=Solving free-surface problems for non-shallow water using boundary and initial conditions-free physics-informed neural network (bif-PINN), refAbstract=null), Reference(id=1244321243386724380, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, doi=null, pmid=null, pmcid=null, year=2021, volume=147, issue=8, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[16], rfOrder=17, authorNames=RAO C, SUN H, LIU Y, journalName=J Eng Mech, refType=null, unstructuredReference=RAO C, SUN H, LIU Y. Physics-informed deep learning for computational elastodynamics without labeled data[J]. J Eng Mech, 2021, 147(8): 04021043., articleTitle=Physics-informed deep learning for computational elastodynamics without labeled data, refAbstract=null)], funds=[Fund(id=1244321239712515035, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, awardId=A2020202015, language=CN, fundingSource=河北省自然科学基金项目(A2020202015), fundOrder=null, country=null), Fund(id=1244321239821566940, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, awardId=12102123, language=CN, fundingSource=国家自然科学基金项目(12102123), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1244321233815323320, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, xref=null, ext=[AuthorCompanyExt(id=1244321233832100539, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China), AuthorCompanyExt(id=1244321233853072063, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, companyId=1244321233815323320, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=河北工业大学 机械工程学院,天津 300401)])], figs=[ArticleFig(id=1244321237858632598, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 1, caption=General structure of flow field prediction model based on PINN, figureFileSmall=0BR4ijCa2wrGM8ttax9kWg==, figureFileBig=T6EgUNhw6McX7PAYVSliOQ==, tableContent=null), ArticleFig(id=1244321237997044635, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图1, caption=基于PINN的流场预测模型一般结构, figureFileSmall=0BR4ijCa2wrGM8ttax9kWg==, figureFileBig=T6EgUNhw6McX7PAYVSliOQ==, tableContent=null), ArticleFig(id=1244321238210954149, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 2, caption=HBC-PINN framework, figureFileSmall=mD75G1vpehBLeC8HvSZyeA==, figureFileBig=3wDDtufExRoiMWAaKMwEfg==, tableContent=null), ArticleFig(id=1244321238311617445, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图2, caption=HBC-PINN框架, figureFileSmall=mD75G1vpehBLeC8HvSZyeA==, figureFileBig=3wDDtufExRoiMWAaKMwEfg==, tableContent=null), ArticleFig(id=1244321238399697835, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 3, caption=Grid independence verification and local schematic diagram for calculation of grid (a) Pressure distribution curves calculated under different grid sizes, (b) Narrow area calculation grid, figureFileSmall=4K6MgsTwj7WBtFJrAr+/bg==, figureFileBig=YL23lGVD/AzBnQV8Pnbw4g==, tableContent=null), ArticleFig(id=1244321238496166830, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图3, caption=网格无关性验证及计算网格局部示意图, figureFileSmall=4K6MgsTwj7WBtFJrAr+/bg==, figureFileBig=YL23lGVD/AzBnQV8Pnbw4g==, tableContent=null), ArticleFig(id=1244321238559081391, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 4, caption=Comparison of predicted and simulated values for different models under narrow parameter A=12 (a) Contours of the blood flow velocity field and pressure field, (b) Contours of the absolute error, figureFileSmall=Z0D4pGVoTcO1PTYodyTuvg==, figureFileBig=iolMO6ERsuCAsNoIzFZiTQ==, tableContent=null), ArticleFig(id=1244321238668133300, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图4, caption=狭窄参数A=12时不同模型预测值与模拟值对比, figureFileSmall=Z0D4pGVoTcO1PTYodyTuvg==, figureFileBig=iolMO6ERsuCAsNoIzFZiTQ==, tableContent=null), ArticleFig(id=1244321238756213688, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 5, caption=Convergence history for the loss functions of two models during training (a) A=6, (b) A=12, figureFileSmall=CpLPWU6AGeg4OOZQxtLEhw==, figureFileBig=QKbxXR5fbAUTWUml49o22A==, tableContent=null), ArticleFig(id=1244321238869459900, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图5, caption=训练期间两种模型损失函数的收敛历史, figureFileSmall=CpLPWU6AGeg4OOZQxtLEhw==, figureFileBig=QKbxXR5fbAUTWUml49o22A==, tableContent=null), ArticleFig(id=1244321238986900416, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Fig. 6, caption=Comparison of HBC-PINN model prediction results under different working conditions (a) Vertical velocity distribution along the centerline of blood vessels, (b) Relative L2 error of flow field solutions under different viscosities, (c) Pressure distribution along the centerline of blood vessels, (d) Relative L2 error of flow field solutions under different inlet velocities, figureFileSmall=3KwduHX1qY6xsG9sUtw4Uw==, figureFileBig=HUpkd3dh+Izwp/wsuOKZ3A==, tableContent=null), ArticleFig(id=1244321239108535239, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=图6, caption=不同工况下HBC-PINN模型预测结果对比, figureFileSmall=3KwduHX1qY6xsG9sUtw4Uw==, figureFileBig=HUpkd3dh+Izwp/wsuOKZ3A==, tableContent=null), ArticleFig(id=1244321239234364360, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Tab. 1, caption=

Comparison of prediction results between PINN model and HBC-PINN model

, figureFileSmall=null, figureFileBig=null, tableContent=
参数HBC-PINNPINN
A612612
εu/%0.220.50.431.71
εv/%2.543.14.634.8
εp/%0.240.331.573.16
训练时间/s492.1504.5665.4671.6
), ArticleFig(id=1244321239360193483, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=表1, caption=

PINN模型与HBC-PINN模型预测结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
参数HBC-PINNPINN
A612612
εu/%0.220.50.431.71
εv/%2.543.14.634.8
εp/%0.240.331.573.16
训练时间/s492.1504.5665.4671.6
), ArticleFig(id=1244321239460856783, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=EN, label=Tab. 2, caption=

Comparison of the relative L2 error of predicted flow field variables under different working conditions

, figureFileSmall=null, figureFileBig=null, tableContent=
参数黏度/(mPa·s-1入口速度/(mm·s-1
3.23.53.84.1150175200225
εu0.59±0.090.50±0.100.76±0.060.65±0.110.52±0.110.50±0.100.42±0.040.53±0.08
εv3.47±0.213.10±0.493.99±0.333.31±0.752.97±0.413.10±0.492.50±0.122.93±0.72
εp0.36±0.050.33±0.020.29±0.010.37±0.060.28±0.050.33±0.020.30±0.010.33±0.06
), ArticleFig(id=1244321239561520084, tenantId=1146029695717560320, journalId=1244284848500682798, articleId=1244321216186663728, language=CN, label=表2, caption=

不同工况下预测的流场变量相对L2误差对比

, figureFileSmall=null, figureFileBig=null, tableContent=
参数黏度/(mPa·s-1入口速度/(mm·s-1
3.23.53.84.1150175200225
εu0.59±0.090.50±0.100.76±0.060.65±0.110.52±0.110.50±0.100.42±0.040.53±0.08
εv3.47±0.213.10±0.493.99±0.333.31±0.752.97±0.413.10±0.492.50±0.122.93±0.72
εp0.36±0.050.33±0.020.29±0.010.37±0.060.28±0.050.33±0.020.30±0.010.33±0.06
)], attaches=null, journal=Journal(id=1244283803934437421, delFlag=0, nameCn=医用生物力学, nameEn=Journal of Medical Biomechanics, nameHistory1=null, nameHistory2=null, issn=1004-7220, eissn=null, cn=31-1624R, coden=null, periodic=1, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=G8Fk4zRvuEegNfdcEgpIYA==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Medical Biomechanics, journalRemark=null, publicationField=null, createdTime=1774589976534, updatedTime=1774590784682, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=J, firstLetterEn=J, subjectCode=Engineering, subjectName=null, subjectCodeEn=Engineering, subjectNameEn=null, picCn=G8Fk4zRvuEegNfdcEgpIYA==, picEn=Eh34RQA3oTtmzcSmNDdO9Q==, jcr=null, cjcr=null, exts=[JournalExt(id=1244287193645167332, language=CN, name=医用生物力学, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1774590784703, updatedTime=1774590784703, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://www.mechanobiology.cn/yyswlx/author/login, submissionEditorUrl=http://www.mechanobiology.cn/yyswlx/editor/login, submissionReviewUrl=http://www.mechanobiology.cn/yyswlx/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1244287193737442021, language=EN, name=Journal of Medical Biomechanics, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1774590784725, updatedTime=1774590784725, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://www.mechanobiology.cn/yyswlx/author/login, submissionEditorUrl=http://www.mechanobiology.cn/yyswlx/editor/login, submissionReviewUrl=http://www.mechanobiology.cn/yyswlx/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1244284848500682798, websiteList=[Website(id=1244289124061000646, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1244284848500682798, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/yyswlx/CN, language=CN, createTime=1774591244950, createBy=18614031015, updateTime=1774591273991, updateBy=18614031015, name=医用生物力学-中文, tplId=1146099689490845704, title=医用生物力学, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1244289419298063322, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=articleTextType, value=kx, createTime=1774591315340, updateTime=1774591315340, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419268703191, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=banner, value=null, createTime=1774591315333, updateTime=1774591315333, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419340006365, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=grayFlag, value=0, createTime=1774591315350, updateTime=1774591315350, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419260314582, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=logo, value=https://castjournals.cast.org.cn/joweb/yyswlx/CN/file/pic?fileId=pmnHPc+Tur67OsVt9JPkuA==, createTime=1774591315331, updateTime=1774591315331, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419440669663, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=minRunFlag, value=0, createTime=1774591315374, updateTime=1774591315374, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419285480409, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/yyswlx/CN/file/pic, createTime=1774591315337, updateTime=1774591315337, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419428086750, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=silenceFlag, value=0, createTime=1774591315371, updateTime=1774591315371, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419277091800, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1774591315335, updateTime=1774591315335, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419310646235, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=themeColor, value=null, createTime=1774591315343, updateTime=1774591315343, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289419327423452, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124061000646, code=themeStyle, value=null, createTime=1774591315347, updateTime=1774591315347, creator=18614031015, updator=18614031015)]), Website(id=1244289124119720904, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1244284848500682798, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/yyswlx/EN, language=EN, createTime=1774591244964, createBy=18614031015, updateTime=1774591288789, updateBy=18614031015, name=医用生物力学-英文, tplId=1146101810881728533, title=Journal of Medical Biomechanics, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1244289447987098584, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=articleTextType, value=kx, createTime=1774591322180, updateTime=1774591322180, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447966127061, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=banner, value=null, createTime=1774591322175, updateTime=1774591322175, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289448003875803, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=grayFlag, value=0, createTime=1774591322184, updateTime=1774591322184, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447957738452, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=logo, value=https://castjournals.cast.org.cn/joweb/yyswlx/EN/file/pic?fileId=pmnHPc+Tur67OsVt9JPkuA==, createTime=1774591322173, updateTime=1774591322173, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289448020653021, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=minRunFlag, value=0, createTime=1774591322188, updateTime=1774591322188, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447978709975, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/yyswlx/EN/file/pic, createTime=1774591322178, updateTime=1774591322178, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289448012264412, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=silenceFlag, value=0, createTime=1774591322186, updateTime=1774591322186, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447974515670, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1774591322177, updateTime=1774591322177, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447991292889, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=themeColor, value=null, createTime=1774591322181, updateTime=1774591322181, creator=18614031015, updator=18614031015), WebsiteProps(id=1244289447999681498, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1244289124119720904, code=themeStyle, value=null, createTime=1774591322183, updateTime=1774591322183, creator=18614031015, updator=18614031015)])], journalTitle=医用生物力学, weixinUrl=null, journalUrl=http://www.mechanobiology.cn/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Medical Biomechanics, journalPhotoCn=G8Fk4zRvuEegNfdcEgpIYA==, journalPhotoEn=Eh34RQA3oTtmzcSmNDdO9Q==, journalFirstLetter=J, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/yyswlx/CN/10.16156/j.1004-7220.2025.05.019, detailUrlEn=https://castjournals.cast.org.cn/joweb/yyswlx/EN/10.16156/j.1004-7220.2025.05.019, pdfUrlCn=https://castjournals.cast.org.cn/joweb/yyswlx/CN/PDF/10.16156/j.1004-7220.2025.05.019, pdfUrlEn=https://castjournals.cast.org.cn/joweb/yyswlx/EN/PDF/10.16156/j.1004-7220.2025.05.019, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于硬边界约束物理信息神经网络狭窄动脉血管的血液流场预测
收藏切换
PDF下载
向华鑫 , 桑建兵 , 王静远 , 纪猛强 , 张晨
医用生物力学 | 论著 2025,40(5): 1222-1229
收起
收藏切换
医用生物力学 | 论著 2025, 40(5): 1222-1229
基于硬边界约束物理信息神经网络狭窄动脉血管的血液流场预测
全屏
向华鑫, 桑建兵 , 王静远, 纪猛强, 张晨
作者信息
  • 河北工业大学 机械工程学院,天津 300401

通讯作者:

桑建兵,教授,E-mail:
Prediction of Blood Flow Field in Artery Stenosis Based on Hard Boundary-Constrained Physics-Informed Neural Network
Huaxin XIANG, Jianbing SANG , Jingyuan Wang, Mengqiang JI, Chen ZHANG
Affiliations
  • School of Mechanical Engineering, Hebei University of Technology, Tianjin 300401, China
出版时间: 2025-10-01 doi: 10.16156/j.1004-7220.2025.05.019
文章导航
收藏切换
目的

针对传统物理信息神经网络(physics-informed neural network,PINN)在处理血流边界条件约束时的局限性,提出一种基于硬边界约束物理信息神经网络(hard boundary-constrained physics-informed neural network,HBC-PINN)的改进方法,实现对狭窄动脉血管内血液流场的精确预测。

方法

首先建立理想化狭窄血管几何模型并进行计算流体动力学模拟以获得验证数据集。根据硬约束方法设计合适的边界相关试函数以将流动边界条件嵌入网络输出中,从而构建采用硬边界约束方法的HBC-PINN模型预测狭窄血流的速度场和压力场。同时还构建了采用软约束方法的原始PINN模型作为对比,通过评估两种模型在验证数据集上的准确性,验证不使用任何标记数据训练下HBC-PINN模型模拟血流动力学的能力。

结果

确定了HBC-PINN方法在狭窄血流动力学参数预测任务中的有效性。两种不同狭窄情况下HBC-PINN预测的流向速度和压力的相对L2误差均低于0.5%,相比原始PINN模型精度提升了48.8%以上,垂向速度的预测精度同样提升了超过35.4%。

结论

在PINN建模过程中实施边界条件硬约束可以有效提高对血流动力学参数的预测精度和模型求解效率。

血管狭窄  /  血流动力学  /  物理信息神经网络  /  硬约束方法
Objective

To address the limitations of conventional physics-informed neural network (PINN) in handling hemodynamic boundary constraints, an improved hard boundary-constrained PINN (HBC-PINN) framework was proposed to achieve precise prediction of blood flow fields within stenotic arteries.

Methods

An idealized stenosed vessel geometry model was established and computational fluid dynamic simulation was performed to obtain a validation dataset. Appropriate boundary dependent trial functions were designed according to the hard constraint method to embed the flow boundary conditions into the network output. Thus, an HBC-PINN model with the hard boundary constraint method was constructed to predict the velocity field and pressure field of stenosed blood flow. Meanwhile, an original PINN model with the soft constraint method was also built for comparison. By evaluating the accuracy of the two models on the validation dataset, the capability of the HBC-PINN model to simulate hemodynamics without using any labeled data for training was verified.

Results

The effectiveness of the HBC-PINN method in predicting hemodynamic parameters in stenosed blood flow tasks was validated. The relative L2 errors of the flow velocity and pressure predicted by the HBC-PINN in two different stenosis scenarios were both lower than 0.5%, representing an improvement of over 48.8% in accuracy compared to the original PINN model. Additionally, the prediction accuracy of the transverse velocity also increased by more than 35.4%.

Conclusions

Implementing hard constraints on boundary conditions in the PINN modeling process can effectively improve the prediction accuracy of hemodynamic parameters and the efficiency of model solving.

vascular stenosis  /  hemodynamics  /  physics-informed neural network  /  hard constraint method
向华鑫, 桑建兵, 王静远, 纪猛强, 张晨. 基于硬边界约束物理信息神经网络狭窄动脉血管的血液流场预测. 医用生物力学, 2025 , 40 (5) : 1222 -1229 . DOI: 10.16156/j.1004-7220.2025.05.019
Huaxin XIANG, Jianbing SANG, Jingyuan Wang, Mengqiang JI, Chen ZHANG. Prediction of Blood Flow Field in Artery Stenosis Based on Hard Boundary-Constrained Physics-Informed Neural Network[J]. Journal of Medical Biomechanics, 2025 , 40 (5) : 1222 -1229 . DOI: 10.16156/j.1004-7220.2025.05.019
狭窄区域的血流动力学参数的精确预测能够为介入治疗时机选择、支架置入方案优化以及术后再狭窄风险评估提供定量化决策依据[1-2]。传统的计算流体动力学(computational fluid dynamics,CFD)、流固耦合等数值模拟方法由于存在预处理繁琐、计算成本高等缺点,限制了其在临床实践中的应用。深度学习算法因其强大的通用表征能力,在心血管生物力学建模研究中得到了广泛的应用。Wang等[3]利用卷积神经网络(convolutional neural network,CNN)提取狭窄特征属性,将狭窄信息和节点坐标作为神经网络输入,从而构建用于快速流量预测的近似模型。Ali等[4]结合长短期记忆(long short-term memory,LSTM)网络和多阶段迁移学习方法实现了对动脉血压的精确预测。Sun等[5]基于多输入反向传播神经网络构建并训练了根据几何参数和血流量快速准确地预测冠状动脉狭窄阻力的代理模型。上述研究均属于纯数据驱动模型,尽管在心血管生物力学建模研究中展现出巨大的潜力,但因其存在对训练数据的强依赖性、缺乏物理机制的理解和解释以及泛化能力不足等问题,在实际应用中受到一定的限制。作为科学机器学习研究和微分方程建模的最新趋势,物理信息神经网络(physics-informed neural networks,PINNs)通过将具有适当初始/边界条件(boundary conditions,BCs)的偏微分方程(partial differential equations,PDEs)嵌入损失函数[6]中,从而确保解决方案的物理一致性,并摆脱了对大量训练数据的依赖。目前,基于PINN的PDEs求解方法已在多个物理领域的仿真工作中取得显著成功[7-8];在心血管领域,也有学者将其应用于血流动力学参数预测[9]或参数反演[10]。然而,当前PINN建模方法在进行血流动力学相关研究时,往往需要结合少量的观测数据来提高模型准确性和加速模型训练过程[11]
PINN建模作为一个优化问题,约束的实施方式对于模型预测精度至关重要。为PINN制定严格的初始/边界条件约束,已被证明在找到正确PDE唯一解方面具有有效性,尤其是在标记数据非常稀缺或不存在的情况下[12]。一种关于边界条件的精确施加技术被用于将边界条件约束以“硬”方式嵌入PINN结构中,从而确保预测结果自然地满足相应的边界条件。Lagari等[13]提出了边界条件硬约束方法的基本思想,即通过设计边界相关试函数(数学变换)以构建满足边界条件的解。在此基础上,Sukumar等[14]进一步介绍了如何使用近似距离函数在PINN中精确地施加Dirichlet边界条件,并通过超限插值推广到Neuman和Robin边界条件。Huang等[15]还介绍了一种无需边界和初始条件的物理信息神经网络(physics-informed neural network,PINN)方法,该方法利用状态转移函数在构造的满足边界条件的解析变换基础上,将初始条件同样以硬方式嵌入到神经网络输出中。本文基于精确边界条件施加技术构建了一个用于狭窄血流预测的硬边界约束PINN模型,通过将传统PINN方法中边界条件“软约束”转为“硬约束”实施,可以有效提高模型预测精度和训练效率。此外,还讨论了模型对血流入口速度、血液黏度等参数的敏感性。
对于真实血管中存在局部狭窄的情况,其狭窄形状通常为不规则。为便于研究且不失一般性,假设血管内的狭窄呈轴对称分布,血管在不同轴向位置的半径可以通过以下函数表示:
式中:x为到血管入口的轴向距离;R0为无狭窄时血管的半径,R0=3 mm;L为血管总长度,为确保血流充分发展,设置L=20R0,即L=60 mm。3个控制参数Aδσ共同定义了狭窄血管的几何形状。参数A(正值)控制血管收缩程度;δσ分别定义血管收缩位置和形状,δ=30 mm,σ=3 mm。因此,本文中只有A被认为是控制狭窄程度的可变参数。
血液在血管中的流动特性,数学上可以通过N-S方程进行描述。为了简化,仅研究稳态下的狭窄血管血液流动,并假设血液是不可压缩的均质牛顿流体。基于以上假设,由连续性方程和动量方程组成的无量纲形式控制方程表示为:
式中:u为速度矢量,对于二维问题u=(uv);p为血流压力;Re为雷诺数,其定义为Re=ρLrefUref/μρ为密度,μ为黏度,LrefUref分别为特征长度和特征速度,在本文中分别定义为入口边界处无狭窄血管半径和最大速度)。模型中涉及的物理参数和边界条件设置如下:血液密度ρ=1 050 kg/m3,黏度μ=3.5 mPa/s;血管壁假定为刚性、无弹性及厚度;在入口处指定一个最大流向速度umax=175 mm/s的抛物线形速度剖面u(0,y)=umax(1-(y/R02),其中y为离对称轴的径向距离;出口条件为自由压力出口。
PINN是一种融合物理先验知识进行训练的神经网络,其基本思想是通过将偏微分方程定解问题转化为具有物理约束的损失函数最优化问题,进而利用神经网络来逼近该物理问题的解。在Raissi等[6]提出的原始PINN中,其主要由一个全连接神经网络(fully connected neural network,FCNN)部分和一个物理信息部分组成,两者彼此共享参数。如图1所示,对于二维稳态流场预测,FCNN充当代理模型,学习从输入(空间坐标)到输出(流场变量uvp)的映射,中间通过L个完全连接的隐藏层网络(仿射变换与非线性激活)对输入和输出的非线性关系进行表征,数学上可以描述为:
式中:为输入信号,输出信号Y为FCNN预测的目标值;h0为输入层,zl分别表示具有nl个神经元的第l个中间隐藏层的预激活和激活单元;分别表示第l层网络的权重矩阵和偏置向量;ϕ(·)为非线性激活函数。
相比传统深度神经网络,PINNs的改进在于通过物理信息部分将包括PDEs、初始或边界条件在内的物理先验知识嵌入损失函数中。因此,求解式(2)中流动控制方程的PINN模型损失函数定义为:
式中:LPDELBC表示不同的损失函数分量,分别对应于控制方程和边界条件的残差;λ1λ2为惩罚系数,代表损失函数中不同分量之间重要性的权衡,(λ1λ2)=(1,10);NfNu分别表示不同损失项的训练样本数量;表示边界上第i个坐标点处神经网络输出的速度和压力;表示流场变量在给定边界条件下的真实值;en表示第n个方程的残差,其定义为:
训练PINN模型以获得PDEs近似解的目标是找到一组最优网络参数,使得损失函数最小化,即
PINNs中关于边界条件的实施主要有两种方式。第1种是软约束方法,其将边界约束作为附加项添加到损失函数中参与训练,通过监督学习的方式让神经网络在边界处的输出逼近边界值。然而这种软约束方式不能保证边界条件精确地传递到计算域中,甚至会对PINNs的精度和训练稳定性产生负面影响[16]。相比之下,硬约束方法通过修改网络输出使得边界条件在网络训练过程中始终得到满足,在边界处不会产生任何误差。同时,硬边界约束在排除边界损失项的同时还意味着惩罚系数的减少。为了提高模型精度,本文提出在PINN建模中采用硬约束方案强制网络输出自动满足流动边界条件,称为硬边界约束物理信息神经网络(hard boundary-constrained PINN,HBC-PINN)。假设1个PDE系统具有单个Dirichlet BC:
并利用一个输出为的神经网络来逼近潜在解ux),则单个Dirichlet BC硬约束实施过程可以定义为:
式中:为HBC-PINN的最终输出;Ψx)和Φ(x)是为构造自动满足边界条件的解而定义的两个边界相关试函数,分别称为边界函数和特解函数。为了确保目标解自动满足边界条件,试函数Ψx)和Φ(x)需要满足以下约束条件:
显然,由于Ψx)的存在,式(10)确保了最终输出在边界ΓD处退化为特解函数Φx),即。PDE的解被分解为通解和特解,其中通解由神经网络近似,而边界处的特解则通过设计的边界相关试函数给出。注意到,边界函数的形式与边界的几何形状有关,而特解函数则与边界上解变量的分布有关。如果边界包含n个子集[Γ1,…,Γn],则边界函数可以定义为
由于神经网络训练过程中损失函数涉及求导计算,此时对进行自动微分将面临一个复合函数求导问题,因此除了约束条件(8)~(10)之外,Ψx)和Φx)还必须保证同PDE一样对xn阶可导的。根据这一特性,本文对网络架构进行了修改,最终构建了一个用于狭窄血流预测的HBC-PINN模型框架(见图2)。针对每一个流场变量()单独使用一个FCNN进行代理建模,这种网络结构使得不同输出之间的网络参数互不影响,从而避免了对不同流场变量施加的硬约束在反向传播过程中产生不必要的求导计算。
为了将入口处的抛物线速度分布、血管壁上的无滑移条件以及压力出口条件作为硬约束实施,本文构造了以下硬约束方案表达式(已无量纲化):
由于式(11)自动满足流动边界条件(Dirichlet BCs),这意味着HBC-PINN的损失函数可以移除边界条件损失项。因此,HBC-PINN的损失函数可简化为:
式中:θ*为用于学习下标表示的流场变量所对应神经网络的可训练参数;残差en将由()替代()进行计算。
为验证硬边界约束方案对狭窄血流预测结果的影响,分别在相同的参数配置基于PINN模型和HBC-PINN模型进行仿真,模型超参数的设置如下:使用3个完全相同子网络来分别学习血流动力学参数(uvp),每个子网络包括4层隐藏层且每层均具有40个神经元,输入为2,输出为1;选择双曲正切函数tanh作为激活函数。
训练超参数设置如下:针对每种狭窄血管模型,在入口、壁面和出口边界上以1∶2∶1的比例均匀采样Nu=600个数据点用于计算PINN模型中的边界条件损失,同时在求解域内使用拉丁超立方抽样(latin hypercube sampling,LHS)方法采样Nf≈5 000个残差点以计算两种模型中的控制方程损失LPDE。网络训练采用全批量学习,初始学习率为0.001,同时使用指数学习率调度器衰减,衰减率为每1 000步0.95。训练初期使用一阶Adam优化器快速获得较理想的网络参数,随后利用二阶L-BFGS算法精细调整网络以获得更精确的解。此外,还引入了早停机制以比较两种模型的训练过程,即当使用Adam优化后的损失不再持续下降时,及时切换到L-BFGS进一步训练。
数值实验中涉及的神经网络模型搭建和训练均在Python 3.9.15中的开源深度学习框架Tensorflow 2.9.1下实现。
将CFD模拟结果作为参考解来评估不同模型的准确性,首先利用ANSYS 18.0 Fluent进行仿真以创建高分辨率验证数据集,为了保证计算结果精确性并减小计算量,选取沿血管中心线的压力分布进行网格无关性验证。在A=6、μ=3.5 mPa/s、umax=175 mm/s工况下,采用3种不同规模网格计算获得的压力分布曲线如图3(a)所示。通过对比,选取了尺寸为35 500的网格进行数值计算[见图3(b)],其中边界层初始高度为20 μm,共10层,同时对狭窄区域进行网格加密。
为了定量比较不同模型的准确性,使用相对L2误差作为评价指标,其定义为:
式中:qpredict为模型预测的流场变量值;qexact为相应的真实值。
利用构建好的两种神经网络模型分别对A=6和A=12两种狭窄血流进行训练求解,并计算得到PINN和HBC-PINN预测结果与Fluent模拟值的相对L2误差指标(见表1)。结果显示,相比PINN,两种不同狭窄情况下HBC-PINN预测的流向速度和压力的相对L2误差均低于0.5%,精度提升了48%以上,两种模型对垂向速度的预测精度都低于流向速度,但HBC-PINN方法依然表现出更好的预测能力,精度提升了35.4%。此外,对比血流速度和压力预测的相对L2误差,可以看出HBC-PINN对压力预测精度的增幅明显大于对速度的增幅,推测其原因是速度变量还需要额外满足质量守恒方程,从而导致硬约束方案对于速度预测的改进不如对压力的明显。
图4展示了血管狭窄参数A=12时两种模型流场解的预测值与模拟值的对比。血流在通过狭窄区域流速明显加快而压力迅速降低,两种模型都能定性地捕获这一血流特性。同时,从两种方法预测结果的绝对误差云图来看,HBC-PINN模型的误差主要分布在血流特性较为复杂的狭窄区域,说明硬边界约束方案对边界条件的精确施加很好改善了模型的预测性能,因为狭窄区域的误差并不会像PINN那样进一步传播到其他区域。
为了说明硬边界约束方式对求解效率的影响,对比PINN和HBC-PINN损失函数随训练轮次的下降情况。结果表明,对于不同狭窄情况,HBC-PINN在收敛速度方面都比PINN更快,求解效率提高了24.8%以上(见图5)。同时注意到,HBC-PINN在训练初期损失值较高,这是由于硬约束施加的边界相关试函数改变了输出的原始分布造成。
在实际血液循环系统中,血液黏度和血流速度通常在一定范围内波动。因此,有必要分析HBC-PINN预测精度对不同黏度和入口速度的敏感性,以评估本方法对于不同工况的适用情况。分别设置了4组血液黏度(3.2、3.5、3.8、4.1 mPa/s)和4组入口速度(150、175、200、225 mm/s),将其应用于A=12的狭窄情况进行求解。表2汇总了不同黏度和入口速度下HBC-PINN预测的流场变量相对L2误差结果,结果取自5次模型独立运行结果的平均值和标准差。
图6展示了沿血管中心线血流的速度或压力分布,同时还显示了HBC-PINN在不同工况下的误差棒折线。随着黏度的增加,狭窄位置以及下游区域垂向速度波动逐渐幅度减小,可以发现HBC-PINN预测的垂向速度分布与CFD模拟结果基本一致。类似地,由于血管狭窄引起的非线性压降随最大流向速度μmax增加而变大,可以观察到预测的压力分布与模拟值几乎重合。在全局误差方面,虽然测试的不同工况下预测的血流动力学参数相对L2误差有所变化,但基本稳定在一个合理区间内并具有较高的精度。
针对当前PINN方法在不使用任何标记数据训练下难以精确建模复杂动力系统这一问题,本文提出一种在PINN训练过程中将边界条件作为硬约束实施的方法,并给出了适用于狭窄血流预测的建模过程。通过数值案例与CFD模拟、原始PINN模型进行对比分析,主要结论如下:①在测试的两种狭窄血流中,相较于原始PINN,两种不同狭窄情况下HBC-PINN预测流向速度、垂向速度和压力的相对L2误差分别降低了48.8%、35.4%和74.4%以上,并且求解效率提高了24.8%。②在不同血液黏度和入口速度下对所提方法进行数值模拟验证,不同工况下预测的血流动力学参数相对L2误差均稳定在1个合理区间内,表明HBC-PINN在精确模拟血流动力学方面的具有良好的鲁棒性和稳定性。
提出的HBC-PINN方法,通过将边界条件约束嵌入到PINN结构中,从而确保边界条件在模型训练过程中始终得到精确满足。相比采用软边界约束方法的传统PINN模型,HBC-PINN在两种狭窄血流预测任务中均表现出更高的精度并且在模型训练方面具有更好的收敛特性,求解效率提高了24.8%以上。这主要由于其允许控制方程损失项单独存在并排除了边界条件损失项的梯度干扰,避免了训练过程中不满足边界条件的非物理解的产生。HBC-PINN的单次模拟时间达到8 min以上(NVIDIA® Tesla® T4 GPU),虽然在模拟时间上高于CFD方法,但由于HBC-PINN构造了满足边界条件的特解,意味着其可以训练神经网络来近似控制方程的通解,而无需针对不同的物理参数重复求解方程,可实现不同工况下的流场快速预测。
本研究也存在一些局限性:①所提出方法的有效性仅在简化血管模型上得到了验证,在进一步的研究中可以通过监督学习的方式预训练一个以计算域的空间坐标作为输入,而输出为该坐标点到血管壁的欧几里得距离的辅助神经网络作为边界函数的光滑近似,从而实现对真实血管模型中血液流场的预测;②所提出的模型适用于血管几何形状和边界条件单独改变的情况,模型在两者同时改变时的预测性能尚需进一步深入研究。
  • 河北省自然科学基金项目(A2020202015)
  • 国家自然科学基金项目(12102123)
参考文献 引证文献
排序方式:
[1]
熊柱翔, 袁丁, 郑庭辉, . 肾动脉狭窄形态对其血流动力学的影响[J]. 医用生物力学, 2021, 36(3): 389-395.
XIONG ZX, YUAN D, ZHENG TH, et al. Influences of renal artery stenosis morphology on its hemodynamics[J]. J Med Biomech, 2021, 36(3): 389-395.
[2]
石政加, 孙丽芳, 赵明轩, . 数据驱动下合并狭窄左冠状动脉瘤血流动力学参数反演方法[J]. 医用生物力学, 2024, 39(5): 853-859.
SHI ZJ, SUN LF, ZHAO MX, et al. Data-driven inversion of hemodynamic parameters for combined stenotic left coronary artery aneurysms[J]. J Med Biomech, 2024, 39(5): 853-859.
[3]
WANG L, DONG D, TIAN FB. Fast prediction of blood flow in stenosed arteries using machine learning and immersed boundary-Lattice Boltzmann method[J]. Front Physiol, 2022(13): 953702.
[4]
ALI NF, ATEF M. LSTM multi-stage transfer learning for blood pressure estimation using photoplethysmography[J]. Electronics, 2022, 11(22): 3749.
[5]
SUN H, LIU J, FENG Y, et al. Deep learning-based prediction of coronary artery stenosis resistance[J]. Am J Physiol Heart Circ Physiol, 2022, 323(6): H1194-H1205.
[6]
RAISSI M, PERDIKARIS P, KARNIADAKIS GE. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. J Comput Phys, 2019(378): 686-707.
[7]
KHARAZMI E, FAN D, WANG Z, et al. Inferring vortex induced vibrations of flexible cylinders using physics-informed neural networks[J]. J Fluids Struct, 2021(107): 103367.
[8]
BAI J, RABCZUK T, GUPTA A, et al. A physics-informed neural network technique based on a modified loss function for computational 2D and 3D solid mechanics[J]. Comput Mech, 2023, 71(3): 543-562.
[9]
OLDENBURG J, BOROWSKI F, SCHMITZ KP, et al. Computation of flow through TAVI device by means of physics informed neural networks[J]. Curr Dir Biomed Eng, 2022, 8(2): 741-744.
[10]
PEREZ S, MADDU S, SBALZARINI IF, et al. Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems[J]. J Comput Phys, 2023(491): 112342.
[11]
DU M, ZHANG C, XIE S, et al. Investigation on aortic hemodynamics based on physics-informed neural network[J]. Math Biosci Eng, 2023, 20(7): 11545-11567.
[12]
SUN L, GAO H, PAN S, et al. Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data[J]. Comput Methods Appl Mech Eng, 2020(361): 112732.
[13]
LAGARIS IE, LIKAS A, FOTIADIS DI. Artificial neural networks for solving ordinary and partial differential equations[J]. IEEE Trans Neural Netw, 1998, 9(5): 987-1000.
[14]
SUKUMAR N, SRIVASTAVA A. Exact imposition of boundary conditions with distance functions in physics-informed deep neural networks[J]. Comput Methods Appl Mech Eng, 2022(389): 114333.
[15]
HUANG YH, XU Z, QIAN C, et al. Solving free-surface problems for non-shallow water using boundary and initial conditions-free physics-informed neural network (bif-PINN)[J]. J Comput Phys, 2023(479): 112003.
[16]
RAO C, SUN H, LIU Y. Physics-informed deep learning for computational elastodynamics without labeled data[J]. J Eng Mech, 2021, 147(8): 04021043.
2025年第40卷第5期
PDF下载
135
66
引用本文
BibTeX
文章信息
doi: 10.16156/j.1004-7220.2025.05.019
  • 接收时间:2025-02-26
  • 首发时间:2026-03-27
  • 出版时间:2025-10-01
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2025-02-26
  • 修回日期:2025-03-24
基金
河北省自然科学基金项目(A2020202015)
国家自然科学基金项目(12102123)
作者信息
    河北工业大学 机械工程学院,天津 300401

通讯作者:

桑建兵,教授,E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/yyswlx/CN/10.16156/j.1004-7220.2025.05.019
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏