Numerical solution of a non-classical two-phase Stefan problem via radial basis function (RBF) collocation methods

详细信息    查看全文

• 作者：Mehdi Dehghan ; ^{mdehghan@aut.ac.ir" class="auth_mail" title="E-mail the corresponding author} ; ^{mdehghan.aut@gmail.com" class="auth_mail" title="E-mail the corresponding author} ; Mahboubeh Najafi ^{mahboubeh.najafi@aut.ac.ir" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Two-phase Stefan problem ; Boundary immobilization method ; Collocation methods ; Chebyshev spectral collocation method ; Legendre spectral collocation method ; The method of radial basis functions (RBFs) ; Kansa method ; RBF&ndash ; PS method ; RBF&ndash ; QR method
• 刊名：Engineering Analysis with Boundary Elements
• 出版年：2016
• 出版时间：November 2016
• 年：2016
• 卷：72
• 期：Complete
• 页码：111-127
• 全文大小：4496 K

文摘

The aim of this paper is to make a comparative study of some high order methods for the numerical solution of a non-classical one-dimensional two-phase Stefan problem. The moving boundary is captured explicitly via boundary immobilization method. The Chebyshev and Legendre spectral collocation methods as high order mesh-based techniques and some radial basis function (RBF) collocation techniques as high order meshless methods are used for spatial discretization. The considered Stefan problem has two stages: one before the extinction time mmlsi0292" class="mathmlsrc">mulatext stixSupport mathImg" data-mathURL="/science?_ob=MathURL&_method=retrieve&_eid=1-s2.0-S0955799716301795&_mathId=si0292.gif&_user=111111111&_pii=S0955799716301795&_rdoc=1&_issn=09557997&md5=f5e99c378fae37145c7af2e8253fdd9b" title="Click to view the MathML source">(0≤t≤t_m)mathContainer hidden">mathCode"><math altimg="si0292.gif" overflow="scroll"><mrow><mo>(mo><mn>0mn><mo>≤mo><mi>tmi><mo>≤mo><mrow><msub subscriptshift="65%"><mrow><mi>tmi>mrow><mrow><mi>mmi>mrow>msub>mrow><mo>)mo>mrow>math> and one after the extinction time mmlsi0293" class="mathmlsrc">mulatext stixSupport mathImg" data-mathURL="/science?_ob=MathURL&_method=retrieve&_eid=1-s2.0-S0955799716301795&_mathId=si0293.gif&_user=111111111&_pii=S0955799716301795&_rdoc=1&_issn=09557997&md5=a2ff046023eff03b9ebbb95b93ac6ce7" title="Click to view the MathML source">(t_m≤t)mathContainer hidden">mathCode"><math altimg="si0293.gif" overflow="scroll"><mrow><mo>(mo><msub subscriptshift="65%"><mrow><mi>tmi>mrow><mrow><mi>mmi>mrow>msub><mo>≤mo><mi>tmi><mo>)mo>mrow>math>. For this particular model there exists a closed form solution for the former stage but there is no analytical solution for the latter one. Numerical results show that RBF-QR method can attain the accuracy of spectral methods when implemented on Chebyshev grid. The high order accuracy for the two stages shows the superiority of the proposed methods in comparison to the previous works.