Knowledge (XXG)

Spectral element method

Source đź“ť

113:, FEM can be used for detecting large flaws in a structure, but as the size of the flaw is reduced there is a need to use a high-frequency wave. In order to simulate the propagation of a high-frequency wave, the FEM mesh required is very fine resulting in increased computational time. On the other hand, SEM provides good accuracy with fewer degrees of freedom. Non-uniformity of nodes helps to make the mass matrix diagonal, which saves time and memory and is also useful for adopting a central difference method (CDM). The disadvantages of SEM include difficulty in modeling complex geometry, compared to the flexibility of FEM. 933:. Since not all interior basis functions need to be present, the p-version finite element method can create a space that contains all polynomials up to a given degree with fewer degrees of freedom. However, some speedup techniques possible in spectral methods due to their tensor-product character are no longer available. The name 820:
Development of the most popular LGL form of the method is normally attributed to Maday and Patera. However, it was developed more than a decade earlier. First, there is the Hybrid-Collocation-Galerkin method (HCGM), which applies collocation at the interior Lobatto points and uses a Galerkin-like
116:
Although the method can be applied with a modal piecewise orthogonal polynomial basis, it is most often implemented with a nodal tensor product Lagrange basis. The method gains its efficiency by placing the nodal points at the Legendre-Gauss-Lobatto (LGL) points and performing the Galerkin method
108:
over non-uniformly spaced nodes. In SEM computational error decreases exponentially as the order of approximating polynomial increases, therefore a fast convergence of solution to the exact solution is realized with fewer degrees of freedom of the structure in comparison with FEM. In
626: 39:
as basis functions. The spectral element method was introduced in a 1984 paper by A. T. Patera. Although Patera is credited with development of the method, his work was a rediscovery of an existing method (see Development History)
362: 914:, differential quadrature method, and G-NI are different names for the same method. These methods employ global rather than piecewise polynomial basis functions. The extension to a piecewise FEM or SEM basis is almost trivial. 821:
integral procedure at element interfaces. The Lobatto-Galerkin method described by Young is identical to SEM, while the HCGM is equivalent to these methods. This earlier work is ignored in the spectral literature.
764: 1079:
Komatitsch, D. and Villote, J.-P.: “The Spectral Element Method: An Efficient Tool to Simulate the Seismic Response of 2D and 3D Geologic Structures,” Bull. Seismological Soc. America, 88, 2, 368-392 (1998)
1117:
Maday, Y. and Patera, A. T., “Spectral Element Methods for the Incompressible Navier-Stokes Equations” In State-of-the-Art Surveys on Computational Mechanics, A.K. Noor, editor, ASME, New York (1989).
100:. The spectral element method chooses instead a high degree piecewise polynomial basis functions, also achieving a very high order of accuracy. Such polynomials are usually orthogonal 1126:
Diaz, J., “A Collocation-Galerkin Method for the Two-point Boundary Value Problem Using Continuous Piecewise Polynomial Spaces,” SIAM J. Num. Anal., 14 (5) 844-858 (1977) ISSN 0036-1429
233: 1210: 870: 830:
G-NI or SEM-NI are the most used spectral methods. The Galerkin formulation of spectral methods or spectral element methods, for G-NI or SEM-NI respectively, is modified and
98: 1091:
Wheeler, M.F.: “A C0-Collocation-Finite Element Method for Two-Point Boundary Value and One Space Dimension Parabolic Problems,” SIAM J. Numer. Anal., 14, 1, 71-90 (1977)
484: 121:
using the same nodes. With this combination, simplifications result such that mass lumping occurs at all nodes and a collocation procedure results at interior points.
787: 812:
The Hybrid-Collocation-Galerkin possesses some superconvergence properties. The LGL form of SEM is equivalent, so it achieves the same superconvergence properties.
412: 185: 890: 807: 668: 648: 452: 432: 385: 158: 490: 1016:
Muradova, Aliki D. (2008). "The spectral method and numerical continuation algorithm for the von Kármán problem with postbuckling behaviour of solutions".
241: 1675: 1301: 1135:
Young, L.C., “A Finite-Element Method for Reservoir Simulation,” Soc. Petr. Engrs. J. 21(1) 115-128, (Feb. 1981), paper SPE 7413 presented Oct. 1978,
1251: 1680: 1227: 1201: 454:
is no larger than the degree of the piecewise polynomial basis. Similar results can be obtained to bound the error in stronger topologies. If
1685: 1525: 1295: 1595: 1452: 1307: 680: 1172: 1156: 1068: 1520: 1503: 1654: 1440: 899:
SEM is a Galerkin based FEM (finite element method) with Lagrange basis (shape) functions and reduced numerical integration by
1690: 1421: 1410: 1387: 831: 17: 1063:
Karniadakis, G. and Sherwin, S.: Spectral/hp Element Methods for Computational Fluid Dynamics, Oxford Univ. Press, (2013),
1393: 56:
series, a chief advantage being that the resulting method is of a very high order. This approach relies on the fact that
1510: 1475: 110: 1515: 1105:
Young, L.C., “Orthogonal Collocation Revisited,” Comp. Methods in Appl. Mech. and Engr. 345 (1) 1033-1076 (Mar. 2019),
124:
The most popular applications of the method are in computational fluid dynamics and modeling seismic wave propagation.
1194: 1632: 1617: 1493: 1259: 1241: 1279: 1602: 1488: 1218: 57: 53: 1264: 1644: 1622: 1607: 1590: 1498: 1483: 981:
Patera, A. T. (1984). "A spectral element method for fluid dynamics - Laminar flow in a channel expansion".
190: 1564: 1335: 1187: 1167:
P. Ĺ olĂ­n, K. Segeth, I. DoleĹľel: Higher-order finite element methods, Chapman & Hall/CRC Press, 2003.
911: 907: 929:
spans a space of high order polynomials by nodeless basis functions, chosen approximately orthogonal for
1612: 1458: 1374: 893: 29: 1179: 840: 1649: 1322: 990: 101: 67: 1416: 1330: 930: 922: 900: 118: 105: 1639: 1580: 1043: 937:
means that accuracy is increased by increasing the order of the approximating polynomials (thus,
457: 1269: 1168: 1152: 1064: 671: 137: 61: 772: 1585: 1575: 1464: 1432: 1033: 1025: 998: 390: 163: 1627: 1570: 1559: 133: 49: 621:{\displaystyle \|u-u_{N}\|_{H^{k}(\Omega )}\leq C_{s,k}N^{k-1-s}\|u\|_{H^{s+1}(\Omega )}} 994: 1405: 1352: 1148: 918: 875: 792: 653: 633: 437: 417: 370: 143: 1669: 1274: 1002: 835: 357:{\displaystyle \|u-u_{N}\|_{H^{1}(\Omega )}\leqq C_{s}N^{-s}\|u\|_{H^{s+1}(\Omega )}} 1047: 1446: 1363: 1340: 1357: 1235: 21: 1029: 36: 1106: 33: 650:, we can also increase the degree of the basis functions. In this case, if 1151:, Finite element analysis, John Wiley & Sons, Inc., New York, 1991. 1543: 1038: 1382: 953: 926: 387:
is related to the discretization of the domain (ie. element length),
1136: 759:{\displaystyle \|u-u_{N}\|_{H^{1}(\Omega )}\leqq C\exp(-\gamma N)} 1537: 1531: 1346: 1183: 921:
space spanned by nodal basis functions associated with
834:
is used instead of integrals in the definition of the
878: 843: 795: 775: 683: 656: 636: 493: 460: 440: 420: 393: 373: 244: 193: 166: 146: 70: 1211:
Numerical methods for partial differential equations
964:
refinements to obtain exponential convergence rates.
1552: 1474: 1431: 1373: 1321: 1288: 1250: 1226: 1217: 884: 864: 801: 781: 758: 662: 642: 620: 478: 446: 426: 406: 379: 356: 227: 179: 152: 92: 1195: 8: 704: 684: 587: 580: 514: 494: 323: 316: 265: 245: 1223: 1202: 1188: 1180: 1037: 892:. Their convergence is a consequence of 877: 842: 794: 774: 712: 707: 697: 682: 655: 635: 595: 590: 562: 546: 522: 517: 507: 492: 459: 439: 419: 398: 392: 372: 331: 326: 307: 297: 273: 268: 258: 243: 204: 192: 171: 165: 145: 75: 69: 941:) rather than decreasing the mesh size, 140:holds here and it can be shown that, if 973: 160:is the solution of the weak equation, 228:{\displaystyle u\in H^{s+1}(\Omega )} 7: 1453:Moving particle semi-implicit method 1364:Weighted essentially non-oscillatory 1101: 1099: 1097: 1087: 1085: 1059: 1057: 917:The spectral element method uses a 1302:Finite-difference frequency-domain 721: 610: 531: 346: 282: 219: 84: 14: 1107:doi.org/10.1016/j.cma.2018.10.019 956:) combines the advantages of the 1676:Numerical differential equations 983:Journal of Computational Physics 865:{\displaystyle a(\cdot ,\cdot )} 187:is the approximate solution and 1655:Method of fundamental solutions 1441:Smoothed-particle hydrodynamics 927:p-version finite element method 1681:Partial differential equations 1296:Alternating direction-implicit 859: 847: 753: 741: 724: 718: 613: 607: 534: 528: 349: 343: 285: 279: 222: 216: 93:{\displaystyle L^{2}(\Omega )} 87: 81: 28:(SEM) is a formulation of the 18:partial differential equations 1: 1308:Finite-difference time-domain 16:In the numerical solution of 1686:Computational fluid dynamics 1347:Advection upstream-splitting 1003:10.1016/0021-9991(84)90128-1 117:integrations with a reduced 111:structural health monitoring 32:(FEM) that uses high-degree 1358:Essentially non-oscillatory 1341:Monotonic upstream-centered 1707: 1618:Infinite difference method 1236:Forward-time central-space 923:Gauss–Lobatto points 1521:Poincaré–Steklov operator 1280:Method of characteristics 1030:10.1007/s10444-007-9050-7 832:Gauss-Lobatto integration 479:{\displaystyle k\leq s+1} 58:trigonometric polynomials 1538:Tearing and interconnect 1532:Balancing by constraints 132:The classic analysis of 119:Gauss-Lobatto quadrature 52:expands the solution in 1645:Computer-assisted proof 1623:Infinite element method 1411:Gradient discretisation 1137:doi.org/10.2118/7413-PA 952:finite element method ( 782:{\displaystyle \gamma } 128:A-priori error estimate 26:spectral element method 1633:Petrov–Galerkin method 1394:Discontinuous Galerkin 912:orthogonal collocation 886: 872:and in the functional 866: 803: 783: 760: 664: 644: 622: 480: 448: 428: 408: 381: 358: 229: 181: 154: 94: 1691:Finite element method 1613:Isogeometric analysis 1459:Material point method 908:pseudospectral method 903:using the same nodes. 887: 867: 804: 784: 761: 665: 645: 623: 481: 449: 429: 409: 407:{\displaystyle C_{s}} 382: 359: 230: 182: 180:{\displaystyle u_{N}} 155: 102:Chebyshev polynomials 95: 30:finite element method 1650:Integrable algorithm 1476:Domain decomposition 1024:(2): 179–206, 2008. 876: 841: 793: 773: 681: 654: 634: 491: 458: 438: 418: 414:is independent from 391: 371: 242: 191: 164: 144: 106:Lagrange polynomials 68: 1494:Schwarz alternating 1417:Loubignac iteration 995:1984JCoPh..54..468P 931:numerical stability 925:. In contrast, the 816:Development History 104:or very high order 1640:Validated numerics 901:Lobatto quadrature 882: 862: 799: 779: 756: 660: 640: 618: 476: 444: 424: 404: 377: 354: 225: 177: 150: 90: 1663: 1662: 1603:Immersed boundary 1596:Method of moments 1511:Neumann–Dirichlet 1504:abstract additive 1489:Fictitious domain 1433:Meshless/Meshfree 1317: 1316: 1219:Finite difference 885:{\displaystyle F} 802:{\displaystyle u} 672:analytic function 663:{\displaystyle u} 643:{\displaystyle N} 447:{\displaystyle s} 427:{\displaystyle N} 380:{\displaystyle N} 153:{\displaystyle u} 62:orthonormal basis 1698: 1608:Analytic element 1591:Boundary element 1484:Schur complement 1465:Particle-in-cell 1400:Spectral element 1224: 1204: 1197: 1190: 1181: 1175: 1165: 1159: 1147:Barna SzabĂł and 1145: 1139: 1133: 1127: 1124: 1118: 1115: 1109: 1103: 1092: 1089: 1080: 1077: 1071: 1061: 1052: 1051: 1041: 1013: 1007: 1006: 978: 891: 889: 888: 883: 871: 869: 868: 863: 808: 806: 805: 800: 789:depends only on 788: 786: 785: 780: 765: 763: 762: 757: 728: 727: 717: 716: 702: 701: 669: 667: 666: 661: 649: 647: 646: 641: 627: 625: 624: 619: 617: 616: 606: 605: 579: 578: 557: 556: 538: 537: 527: 526: 512: 511: 485: 483: 482: 477: 453: 451: 450: 445: 433: 431: 430: 425: 413: 411: 410: 405: 403: 402: 386: 384: 383: 378: 363: 361: 360: 355: 353: 352: 342: 341: 315: 314: 302: 301: 289: 288: 278: 277: 263: 262: 234: 232: 231: 226: 215: 214: 186: 184: 183: 178: 176: 175: 159: 157: 156: 151: 134:Galerkin methods 99: 97: 96: 91: 80: 79: 1706: 1705: 1701: 1700: 1699: 1697: 1696: 1695: 1666: 1665: 1664: 1659: 1628:Galerkin method 1571:Method of lines 1548: 1516:Neumann–Neumann 1470: 1427: 1369: 1336:High-resolution 1313: 1284: 1246: 1213: 1208: 1178: 1166: 1162: 1146: 1142: 1134: 1130: 1125: 1121: 1116: 1112: 1104: 1095: 1090: 1083: 1078: 1074: 1062: 1055: 1018:Adv Comput Math 1015: 1014: 1010: 980: 979: 975: 971: 874: 873: 839: 838: 827: 825:Related methods 818: 791: 790: 771: 770: 708: 703: 693: 679: 678: 652: 651: 632: 631: 630:As we increase 591: 586: 558: 542: 518: 513: 503: 489: 488: 456: 455: 436: 435: 416: 415: 394: 389: 388: 369: 368: 327: 322: 303: 293: 269: 264: 254: 240: 239: 200: 189: 188: 167: 162: 161: 142: 141: 130: 71: 66: 65: 50:spectral method 46: 12: 11: 5: 1704: 1702: 1694: 1693: 1688: 1683: 1678: 1668: 1667: 1661: 1660: 1658: 1657: 1652: 1647: 1642: 1637: 1636: 1635: 1625: 1620: 1615: 1610: 1605: 1600: 1599: 1598: 1588: 1583: 1578: 1573: 1568: 1565:Pseudospectral 1562: 1556: 1554: 1550: 1549: 1547: 1546: 1541: 1535: 1529: 1523: 1518: 1513: 1508: 1507: 1506: 1501: 1491: 1486: 1480: 1478: 1472: 1471: 1469: 1468: 1462: 1456: 1450: 1444: 1437: 1435: 1429: 1428: 1426: 1425: 1419: 1414: 1408: 1403: 1397: 1391: 1385: 1379: 1377: 1375:Finite element 1371: 1370: 1368: 1367: 1361: 1355: 1353:Riemann solver 1350: 1344: 1338: 1333: 1327: 1325: 1319: 1318: 1315: 1314: 1312: 1311: 1305: 1299: 1292: 1290: 1286: 1285: 1283: 1282: 1277: 1272: 1267: 1262: 1260:Lax–Friedrichs 1256: 1254: 1248: 1247: 1245: 1244: 1242:Crank–Nicolson 1239: 1232: 1230: 1221: 1215: 1214: 1209: 1207: 1206: 1199: 1192: 1184: 1177: 1176: 1160: 1140: 1128: 1119: 1110: 1093: 1081: 1072: 1053: 1008: 989:(3): 468–488. 972: 970: 967: 966: 965: 946: 919:tensor product 915: 904: 897: 894:Strang's lemma 881: 861: 858: 855: 852: 849: 846: 826: 823: 817: 814: 798: 778: 767: 766: 755: 752: 749: 746: 743: 740: 737: 734: 731: 726: 723: 720: 715: 711: 706: 700: 696: 692: 689: 686: 659: 639: 615: 612: 609: 604: 601: 598: 594: 589: 585: 582: 577: 574: 571: 568: 565: 561: 555: 552: 549: 545: 541: 536: 533: 530: 525: 521: 516: 510: 506: 502: 499: 496: 475: 472: 469: 466: 463: 443: 423: 401: 397: 376: 365: 364: 351: 348: 345: 340: 337: 334: 330: 325: 321: 318: 313: 310: 306: 300: 296: 292: 287: 284: 281: 276: 272: 267: 261: 257: 253: 250: 247: 224: 221: 218: 213: 210: 207: 203: 199: 196: 174: 170: 149: 129: 126: 89: 86: 83: 78: 74: 45: 42: 13: 10: 9: 6: 4: 3: 2: 1703: 1692: 1689: 1687: 1684: 1682: 1679: 1677: 1674: 1673: 1671: 1656: 1653: 1651: 1648: 1646: 1643: 1641: 1638: 1634: 1631: 1630: 1629: 1626: 1624: 1621: 1619: 1616: 1614: 1611: 1609: 1606: 1604: 1601: 1597: 1594: 1593: 1592: 1589: 1587: 1584: 1582: 1579: 1577: 1574: 1572: 1569: 1566: 1563: 1561: 1558: 1557: 1555: 1551: 1545: 1542: 1539: 1536: 1533: 1530: 1527: 1524: 1522: 1519: 1517: 1514: 1512: 1509: 1505: 1502: 1500: 1497: 1496: 1495: 1492: 1490: 1487: 1485: 1482: 1481: 1479: 1477: 1473: 1466: 1463: 1460: 1457: 1454: 1451: 1448: 1445: 1442: 1439: 1438: 1436: 1434: 1430: 1423: 1420: 1418: 1415: 1412: 1409: 1407: 1404: 1401: 1398: 1395: 1392: 1389: 1386: 1384: 1381: 1380: 1378: 1376: 1372: 1365: 1362: 1359: 1356: 1354: 1351: 1348: 1345: 1342: 1339: 1337: 1334: 1332: 1329: 1328: 1326: 1324: 1323:Finite volume 1320: 1309: 1306: 1303: 1300: 1297: 1294: 1293: 1291: 1287: 1281: 1278: 1276: 1273: 1271: 1268: 1266: 1263: 1261: 1258: 1257: 1255: 1253: 1249: 1243: 1240: 1237: 1234: 1233: 1231: 1229: 1225: 1222: 1220: 1216: 1212: 1205: 1200: 1198: 1193: 1191: 1186: 1185: 1182: 1174: 1173:1-58488-438-X 1170: 1164: 1161: 1158: 1157:0-471-50273-1 1154: 1150: 1144: 1141: 1138: 1132: 1129: 1123: 1120: 1114: 1111: 1108: 1102: 1100: 1098: 1094: 1088: 1086: 1082: 1076: 1073: 1070: 1069:9780199671366 1066: 1060: 1058: 1054: 1049: 1045: 1040: 1035: 1031: 1027: 1023: 1019: 1012: 1009: 1004: 1000: 996: 992: 988: 984: 977: 974: 968: 963: 959: 955: 951: 947: 944: 940: 936: 932: 928: 924: 920: 916: 913: 909: 905: 902: 898: 895: 879: 856: 853: 850: 844: 837: 836:bilinear form 833: 829: 828: 824: 822: 815: 813: 810: 796: 776: 750: 747: 744: 738: 735: 732: 729: 713: 709: 698: 694: 690: 687: 677: 676: 675: 673: 657: 637: 628: 602: 599: 596: 592: 583: 575: 572: 569: 566: 563: 559: 553: 550: 547: 543: 539: 523: 519: 508: 504: 500: 497: 486: 473: 470: 467: 464: 461: 441: 421: 399: 395: 374: 338: 335: 332: 328: 319: 311: 308: 304: 298: 294: 290: 274: 270: 259: 255: 251: 248: 238: 237: 236: 211: 208: 205: 201: 197: 194: 172: 168: 147: 139: 135: 127: 125: 122: 120: 114: 112: 107: 103: 76: 72: 63: 59: 55: 54:trigonometric 51: 43: 41: 38: 35: 31: 27: 23: 20:, a topic in 19: 1447:Peridynamics 1399: 1265:Lax–Wendroff 1163: 1143: 1131: 1122: 1113: 1075: 1021: 1017: 1011: 986: 982: 976: 961: 957: 949: 942: 938: 934: 819: 811: 768: 629: 487: 366: 131: 123: 115: 47: 25: 15: 1581:Collocation 1149:Ivo Babuška 138:CĂ©a's lemma 37:polynomials 22:mathematics 1670:Categories 1270:MacCormack 1252:Hyperbolic 1039:1885/56758 44:Discussion 1586:Level-set 1576:Multigrid 1526:Balancing 1228:Parabolic 935:p-version 857:⋅ 851:⋅ 777:γ 748:γ 745:− 739:⁡ 730:≦ 722:Ω 705:‖ 691:− 685:‖ 611:Ω 588:‖ 581:‖ 573:− 567:− 540:≤ 532:Ω 515:‖ 501:− 495:‖ 465:≤ 347:Ω 324:‖ 317:‖ 309:− 291:≦ 283:Ω 266:‖ 252:− 246:‖ 220:Ω 198:∈ 85:Ω 34:piecewise 1560:Spectral 1499:additive 1422:Smoothed 1388:Extended 1048:46564029 1544:FETI-DP 1424:(S-FEM) 1343:(MUSCL) 1331:Godunov 991:Bibcode 60:are an 1553:Others 1540:(FETI) 1534:(BDDC) 1406:Mortar 1390:(XFEM) 1383:hp-FEM 1366:(WENO) 1349:(AUSM) 1310:(FDTD) 1304:(FDFD) 1289:Others 1275:Upwind 1238:(FTCS) 1171:  1155:  1067:  1046:  954:hp-FEM 769:where 670:is an 434:, and 367:where 24:, the 1567:(DVR) 1528:(BDD) 1467:(PIC) 1461:(MPM) 1455:(MPS) 1443:(SPH) 1413:(GDM) 1402:(SEM) 1360:(ENO) 1298:(ADI) 1044:S2CID 969:Notes 1449:(PD) 1396:(DG) 1169:ISBN 1153:ISBN 1065:ISBN 960:and 948:The 906:The 136:and 64:for 48:The 1034:hdl 1026:doi 999:doi 736:exp 1672:: 1096:^ 1084:^ 1056:^ 1042:. 1032:. 1022:29 1020:. 997:. 987:54 985:. 950:hp 910:, 809:. 674:: 235:: 1203:e 1196:t 1189:v 1050:. 1036:: 1028:: 1005:. 1001:: 993:: 962:p 958:h 945:. 943:h 939:p 896:. 880:F 860:) 854:, 848:( 845:a 797:u 754:) 751:N 742:( 733:C 725:) 719:( 714:1 710:H 699:N 695:u 688:u 658:u 638:N 614:) 608:( 603:1 600:+ 597:s 593:H 584:u 576:s 570:1 564:k 560:N 554:k 551:, 548:s 544:C 535:) 529:( 524:k 520:H 509:N 505:u 498:u 474:1 471:+ 468:s 462:k 442:s 422:N 400:s 396:C 375:N 350:) 344:( 339:1 336:+ 333:s 329:H 320:u 312:s 305:N 299:s 295:C 286:) 280:( 275:1 271:H 260:N 256:u 249:u 223:) 217:( 212:1 209:+ 206:s 202:H 195:u 173:N 169:u 148:u 88:) 82:( 77:2 73:L

Index

partial differential equations
mathematics
finite element method
piecewise
polynomials
spectral method
trigonometric
trigonometric polynomials
orthonormal basis
Chebyshev polynomials
Lagrange polynomials
structural health monitoring
Gauss-Lobatto quadrature
Galerkin methods
CĂ©a's lemma
analytic function
Gauss-Lobatto integration
bilinear form
Strang's lemma
Lobatto quadrature
pseudospectral method
orthogonal collocation
tensor product
Gauss–Lobatto points
p-version finite element method
numerical stability
hp-FEM
Bibcode
1984JCoPh..54..468P
doi

Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.

↑