Professional Documents
Culture Documents
***
* *
* *
* *
European Commission
STEEL RESEARCH
European Commission
Contract No 7210-SA/508
Final report
Directorate-General XII
Science, Research and Development
1996 EUR 15627 EN
LEGAL NOTICE
ISBN 92-827-6457-5
Printed in Luxembourg
Contents.
2. Experimental p a r t 7
2.1. Geometrical characteristics of the HEB200 profiles. 7
2.2. Mechanical characteristics. 7
2.3. Initial deformations, residual stresses. 8
2.4. Three-point bending test. 8
2.5. Portal frame tests. 10
2.6. Results. 11
III
page
4.1.3. Mechanical characteristics. 44
4.1.4. Residual stresses. 45
4.1.5. Discretisation. 46
4.1.6. Initial deformation. 47
4.1.7 Boundary conditions. 47
4.1.8. Loading. 47
4.1.9. Data for numerical simulations. 47
4.1.10. Numerical simulations and comparison with the experimental tests. 49
6. Conclusions. 76
References.
IV
Summarv.
In the third part of this research, a parametrical study on hypothetic steels has been
performed in order to demonstrate that the requirements given in Eurocode 3 are very conservative.
In conclusion, a new method is proposed allowing to check the possibility of a plastic design
of a structure.
Sommaire.
Cette recherche a montré que les prescriptions de l'Eurocode 3 pour le calcul plastique
d'une ossature ne sont pas représentatives du comportement réel de la formation de rotules
plastiques.
Dans la première partie de cette recherche, 5 essais de flexion sur trois points sur le profilé
HEB 200 ont été simulés par le programme d'éléments finis FINELG. Ces simulations ont permis
de montrer l'influence sur la capacité de rotation de cette poutre, de la nuance d'acier, des
déformations initiales globales et locales et, de la prise en compte des congés de raccordement entre
ailes et âme du profilé dans le modèle numérique. Les simulations numériques ont été calibrées à
partir d'essais pratiques effectués par le CRM (Liège).
Dans la seconde partie de cette recherche, 5 portiques simples ont été simulés par le
programme d'éléments finis FINELG afin de déterminer la capacité de rotation requise d'une
structure. Les simulations numériques ont été calibrées à partir d'essais pratiques effectués par le
CRM (Liège).
Dans la troisième partie de cette recherche, une étude paramétrique sur des aciers
hypothétiques a été conduite afin de démontrer que les prescriptions de l'Eurocode 3 sont trop
conservatives.
En conclusion, une nouvelle méthode est proposée pour la vérification de la possibilité d'un
calcul plastique d'une structure..
VI
Zusammenfassung.
VII
List of Figures page
1. Introduction.
Figure 1.1. Elastic perfectly plastic model. 1
Figure 1.2. Example of stress-strain curve for Steel FeE 235 1
Figure 1.3. Plastic hinge and plastic moment concepts. 2
Figure 1.4. Moment rotation curve - Definition of the rotation capacity. 2
Figure 1.5. Gain of plastic design with respect to elastic design. 3
2. Experimental part
Figure 2.1. Static scheme of a three-point bending test. 8
Figure 2.2. Instrumentation of a beam for plastic hinge testing, (ref 2) 9
Figure 2.3. Static scheme of the portal frame. 10
Figure 2.4. Portal frame. 10
3. Numerical simulations of three-point bending tests.
Figure 3.1. Static scheme of the simulated beams. 12
Figure 3.2. Modélisation of the stress-strain curve. 13
Figure 3.3. Residual stresses. 14
Figure 3.4. Discretisation : without simulation of the fillets : Typ DI 1 15
Figure 3.5. Discretisation : Typ DI 2 : with simulation of the fillets. 15
Figure 3.6. Measurement of the initial deformation 18
Figure 3.7. Example of initial deformation simulated with sinus curves 19
(Steel D3): Typ ID 1
Figure 3.8. Example of global initial deformation: Typ ID 2 20
Figure 3.9. Example of local imperfection: Typ Π) 3 21
Figure 3.10. Global and local initial deformation: Typ ID 3 22
Figure 3.11. Boundary conditions: Typ Bl 23
Figure 3.12. Typ B3. 23
photo 3.1. Antimetrical deformation shape of a test performed by CRM. 27
Figure 3.13. Simulation S 1. Moment rotation curve 30
Figure 3.14. Simulation S 1. Plastic deformation at the maximum of the load 30
Figure 3.15. Simulation S 1. Plastic deformation after the maximum of the load 31
Figure 3.16. Simulation S 1. Plastification at the maximum of the load 31
Figure 3.17. Simulation S 1. Plastification after the maximum of the load 32
Figure 3.18. Simulation S 1, S 2, S 3. Comparison. 32
Effect of the modélisation of the stress-strain curve.
Effect of the residual stresses.
Effect of the initial deformation.
Effect of the boundary conditions
Figure 3.19. Simulation S 4. Moment rotation curve 33
Figure 3.20. Simulation S 4. Plastic deformation at the maximum of the load 33
Figure 3.21. Simulation S 4. Plastic deformation after the maximum of the load 34
Figure 3.22. Simulation S 4. Plastification at the maximum of the load 34
Figure 3.23. Simulation S 4. Plastification after the maximum of the load 35
Figure 3.24. Simulation S 5, S 6, S 7. Moment rotation curve 35
Figure 3.25. Simulation S 5. Plastic deformation after the maximum of the load 36
VIII
page
Figure 3.26. Simulation S 5. Plastification after the maximum of the load 36
Figure 3.27. Simulation S 8. Moment rotation curve 37
Figure 3.28. Simulation S 8. Plastic deformation after the maximum of the load 37
Figure 3.29. Simulation S 8. Plastification after the maximum of the load 38
Figure 3.30. Simulation S 9. Moment rotation curve 38
Figure 3.31. Simulation S 9. Plastic deformation after the maximum of the load 39
Figure 3.32. Simulation S 9. Plastification after the maximum of the load 39
Figure 3.33. Simulation S 10. Moment rotation curve 40
Figure 3.34. Simulation S 10. Plastic deformation after the maximum of the load 40
Figure 3.35. Simulation S 10. Plastification after the maximum of the load 41
Figure 3.36. Simulation S i l . Moment rotation curve 41
Figure 3.37. Simulation S i l . Plastic deformation after the maximum of the load 42
Figure 3.38. Simulation S U . Plastification after the maximum of the load 42
IX
page
Figure 4.28. Simulation Ρ 13. Plastification 64
Figure 4.29. Simulation Ρ 16, Ρ 17. Vert, displ. versus vert, load 65
Figure 4.30. Simulation Ρ 16, Ρ 17. Hor. displ. versus vert, load 65
Figure 4.31. Simulation Ρ 16, Ρ 17. Column base rotation versus load 66
Figure 4.32. Simulation Ρ 16, Ρ 17. Beam rotation at mid span versus load 66
5. Numerical simulations on hypothetic steels.
Figure 5.1. Static scheme of the beams. 67
Figure 5.2. Geometrical properties 67
Figure 5.3. Stress-strain curve. 68
Figure 5.4. General sinusoidal initial deformation. 69
Figure 5.5. Simulations AA to CC 71
Figure 5.6. Plastic deformation at M=Mpl (decreasing part) 72
Figure 5.7. Plastic deformation at M=Mpl (decreasing part) 73
Figure 5.8. Plastic deformation at M=Mpl (decreasing part) 74
Figure 5.9. R, function of fy and fu/fy 75
6. Conclusions.
Figure 6.1. Non-dimensional definition of the rotation capacity. 77
Figure 6.2. Rotation capacity of beams (HEB 200) and portal frames. 77
Liste des Figures.
page
1. Introduction.
Figure 1.1. Modèle élastique parfaitement plastique. 1
Figure 1.2. Exemple de la courbe contrainte-déformation pour l'acier FeE 235. 1
Figure 1.3. Concepts des rotules plastiques et des moments plastiques. 2
Figure 1.4. Courbe moment-rotation - Définition de la capacité de rotation. 2
Figure 1.5. Gain du calcul plastique par rapport au calcul élastique. 3
2. Partie experimentale.
Figure 2.1. Schéma statique d'un essai deflexionsur trois points. 8
Figure 2.2. Instrumentation d'une poutre pour l'essai de rotule plastique, (ref 2) 9
Figure 2.3. Schéma statique du portique. 10
Figure 2.4. Portique. 10
3. Simulations numériques des essais de flexion sur trois points.
Figure 3.1. Schéma statique des poutres simulées. 12
Figure 3.2. Modélisation de la courbe contrainte-déformation. 13
Figure 3.3. Contraintes résiduelles. 14
Figure 3.4. Discrétisation : sans simulation des congés : Typ DI 1. 15
Figure 3.5. Discrétisation : Typ DI 2 : avec simulation des congés. 15
Figure 3.6. Mesure des déformées initiales. 18
Figure 3.7. Exemple de déformée initiale simulée avec des courbes 19
sinusoïdales (acier D3): Typ E) 1.
Figure 3.8. Exemple de déformées initiales globales : Typ ID 2. 20
Figure 3.9. Exemple d'imperfections locales: Typ ID 3. 21
Figure 3.10. Déformées initiales globales et locales: Typ ID 3. 22
Figure 3.11. Conditions aux limites: Typ Bl. 23
Figure 3.12. TypB3. 23
Photo 3.1. Allure de la déformée antimétrique d'un essai effectué par le CRM. 27
Figure 3.13. Simulation S 1. Courbe moment-rotation . 30
Figure 3.14. Simulation S 1. Déformée plastique au maximum de la charge. 30
Figure 3.15. Simulation S 1. Déformée plastique après le maximum de la charge. 31
Figure 3.16. Simulation S 1. Plastification au maximum de la charge. 31
Figure 3.17. Simulation S 1. Plastification après le maximum de la charge. 32
Figure 3.18. Simulation S 1, S 2, S 3. Comparaison. 32
Effet de la modélisation de la courbe contrainte-déformation.
Effet des contraintes résiduelles.
Effet des déformées initiales.
Effet des conditions aux appuis.
Figure 3.19. Simulation S 4. Courbe moment-rotation. 33
Figure 3.20. Simulation S 4. Déformée plastique au maximum de la charge. 33
Figure 3.21. Simulation S 4. Déformée plastique après le maximum de la charge. 34
Figure 3.22. Simulation S 4. Plastification au maximum de la charge. 34
Figure 3.23. Simulation S 4. Plastification après le maximum de la charge. 35
Figure 3.24. Simulation S 5, S 6, S 7. Courbe moment-rotation . 35
XI
page
Figure 3.25. Simulation S 5. Déformée plastique après le maximum de la charge. 36
Figure 3.26. Simulation S 5. Plastification après le maximum de la charge. 36
Figure 3.27. Simulation S 8. Courbe moment-rotation . 37
Figure 3.28. Simulation S 8. Déformée plastique après le maximum de la charge. 37
Figure 3.29. Simulation S 8. Plastification après le maximum de la charge. 38
Figure 3.30. Simulation S 9. Courbe moment-rotation. 38
Figure 3.31. Simulation S 9. Déformée plastique après le maximum de la charge. 39
Figure 3.32. Simulation S 9. Plastification après le maximum de la charge. 39
Figure 3.33. Simulation S 10. Courbe moment-rotation . 40
Figure 3.34. Simulation S 10. Déformée plastique après le maximum de la charge. 40
Figure 3.35. Simulation S 10. Plastification après le maximum de la charge. 41
Figure 3.36. Simulation S 11. Courbe moment-rotation. 41
Figure 3.37. Simulation S U . Déformée plastique après le maximum de la charge. 42
Figure 3.38. Simulation S 11. Plastification après le maximum de la charge. 42
4. Simulations numériques des portiques.
Figure 4.1. Schéma statique d'un portique. 43
Figure 4.2. Joints poutre-colonne dans l'essai. 44
Figure 4.3. Modèle courbe contrainte-déformation. 45
Figure 4.4. Discrétisation: numérotation des éléments. 46
Figure 4.5. Discrétisation: vue en perspective. 47
Figure 4.6. Instrumentation des essais (CRM ref (2)). 49
Figure 4.33. Séquence de formation des rotules plastiques. 52
I o - formation simultanée des rotules plastiques en C et D;
2° - rotule plastique en B;
3° - rotule plastique en A.
Figure 4.7. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dépl. vert, en fonction de la charge vert. 54
Figure 4.8. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dépl. hor. en fonction de la charge vert. 54
Figure 4.9. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dépl. hor. en fonction de la charge vert. 55
Figure 4.10. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 1 en fonction
de la charge. 55
Figure 4.11. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 2 en fonction
de la charge. 56
Figure 4.12. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 3 en fonction
de la charge. 56
Figure 4.13. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 4 en fonction
de la charge. 57
Figure 4.14. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 5 en fonction
de la charge. 57
Figure 4.15. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 6 en fonction
de la charge. 58
Figure 4.16. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 7 en fonction
de la charge. 58
Figure 4.17. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 8 en fonction
de la charge. 59
XII
page
Figure 4.18. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 9 en fonction
de la charge. 59
Figure 4.19. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Jauge de contraintes 10 en fonction
de la charge. 60
Figure 4.20. Simulation Ρ 1, Ρ 2, Ρ 5, Ρ 7, Ρ 8. Dépl. vert, en fonction
de la charge vert. 60
Figure 4.21. Simulation Ρ 1, Ρ 2, Ρ 5, Ρ 7, Ρ 8. Dépl. hor. en fonction
de la charge vert. 61
Figure 4.22. Simulation Ρ 9, Ρ 10. Dépl. vert, en fonction de la charge vert. 61
Figure 4.23. Simulation Ρ 9, Ρ 10. Dépl. hor. en fonction de la charge vert. 62
Figure 4.24. Simulation Ρ 11, Ρ 12. Dépl. vert, en fonction de la charge vert. 62
Figure 4.25. Simulation Ρ 11, Ρ 12. Dépl. hor. en fonction de la charge vert. 63
Figure 4.26. Simulation Ρ 13, Ρ 14, Ρ 15. Dépl. vert, en fonction de la charge vert. 63
Figure 4.27. Simulation Ρ 13, Ρ 14, Ρ 15. Dépl. hor. en fonction de la charge vert. 64
Figure 4.28. Simulation Ρ 13. Plastification. 64
Figure 4.29. Simulation Ρ 16, Ρ 17. Dépl. vert, en fonction de la charge vert. 65
Figure 4.30. Simulation Ρ 16, Ρ 17. Dépl. hor. en fonction de la charge vert. 65
Figure 4.31. Simulation Ρ 16, Ρ 17. Rotation de la base de la colonne
en fonction de la charge. 66
Figure 4.32. Simulation Ρ 16, Ρ 17. Rotation de la poutre à mitravée
en fonction de la charge. 66
5. Simulations numériques des aciers hypothétiques.
Figure 5.1. Schéma statique de la poutre. 67
Figure 5.2. Propriétés géométriques. 67
Figure 5.3. C ourbe contraintedéformation. 68
Figure 5.4. Déformées initiales générales sinusoïdales. 69
Figure 5.5. Simulations de AA à CC. 71
Figure 5.6. Déformée plastique à M=Mpl (partie décroissante). 72
Figure 5.7. Déformée plastique à M=Mpl (partie décroissante). 73
Figure 5.8. Déformée plastique à M=Mpl (partie décroissante). 74
Figure 5.9. R, fonction de fy et fu/fy 75
6. Conclusions.
Figure 6.1. Définition nondimensionnelle de la capacité rotation. 77
Figure 6.2. C apacité de rotation des poutres (HEB 200) et des portiques. 77
XIII
Liste der Abbildungen Seite
1. Einleitung.
Bild 1.1. Elastisch-plastisches FließgelenkModell. 1
Bild 1.2. Spannungs-Dehnungslinie für FeE 235 Stahl. 1
Bild 1.3. Konzepte für plastisches Fließgelenk und plastisches Moment. 2
Bild 1.4. Moment-Rotationslinie - Definition des Rotationskapazität. 2
Bild 1.5. Gewinn des plastischen Designs im Vergleich zum elastischen Design. 3
2. Experimental Teil.
Bild 2.1. Statisches Schema der drei-punkt Biegebeanspruchung. 8
Bild 2.2. Instrumentierung der Träger für plastischen Fließgelenk Test (ref 2). 9
Bild 2.3. Statisches Schema von dem Portalrahmen. 10
Bild 2.4. Portalrahmen. 10
3. Numerische Simulierung der drei-punkt Biegebeanspruchung.
Büd 3.1. Statisches Schema der simulierten Träger. 12
Büd 3.2. Modell der Spannungs-Dehnungslinie. 13
Bild 3.3. Eigenspannungen. 14
Büd 3.4. Diskretisierung : ohne Simulieren der Ausrundungen : DI 1 Typ. 15
Bild 3.5. Diskretisierung : mit Simulieren der Ausrundungen : DI 2 Typ. 15
Büd 3.6. Bemessung der Anfangsverformung 18
Büd 3.7. Beispiel der Anfangsverformung simuliert mit sinus Linie 19
(Stahl D3): ID 1 Typ.
Bild 3.8. Beispiel der gesamten Anfangsverformungen : ID 2 Typ. 20
Bild 3.9. Beispiel der lokalen Anfangsverformungen : ID 3 Typ. 21
Bild 3.10. Gesamte und lokale Anfangsverformungen: ID 3 Typ. 22
Bild 3.11. Randbedigungen: Bl Typ. 23
Büd 3.12. B3 Typ. 23
Foto 3.1. Antimetrische Verformungform von eine Versuch des CRM. 27
Büd 3.13. Simulieren S 1. Moment-Rotationslinie. 30
Bild 3.14. Simulieren S 1. Plastische Verformung bei Lastmaximum. 30
Büd 3.15. Simulieren S 1. Plastische Verformung nach Lastmaximum. 31
Büd 3.16. Simulieren S 1. Plastizierung bei Lastmaximum. 31
Bild 3.17. Simulieren S 1. Plastizierung nach Lastmaximum. 32
Bild 3.18. Simulieren S 1, S 2, S 3. Vergleich. 32
Effekt der Modellierung der Spannungs-Dehnungslinie.
Effekt der Eigenspannungen.
Effect der Anfangsverformungen.
Effekt der Randbedigungen.
Bild 3.19. Simulieren S 4. Moment-Rotationslinie. 33
Bild 3.20. Simulieren S 4. Plastische Verformung bei Lastmaximum. 33
Bild 3.21. Simulieren S 4. Plastische Verformung nach Lastmaximum. 34
Bild 3.22. Simulieren S 4. Plastizierung bei Lastmaximum. 34
Bild 3.23. Simulieren S 4. Plastizierung nach Lastmaximum. 35
Bild 3.24. Simulieren S 5, S 6, S 7. Moment-Rotationslinie. 35
XIV
Seite
Bild 3.25. Simulieren S 5. Plastische Verformung nach Lastmaximum. 36
Bild 3.26. Simulieren S 5. Plastizierung nach Lastmaximum. 36
Bild 3.27. Simulieren S 8. MomentRotationslinie. 37
Bild 3.28. Simulieren S 8. Plastische Verformung nach Lastmaximum. 37
Bild 3.29. Simulieren S 8. Plastizierung nach Lastmaximum. 38
Bild 3.30. Simulieren S 9. MomentRotationslinie. 38
Bild 3.31. Simulieren S 9. Plastische Verformung nach Lastmaximum. 39
Bild 3.32. Simulieren S 9. Plastizierung nach Lastmaximum. 39
Bild 3.33. Simulieren S 10. MomentRotationslinie. 40
Bild 3.34. Simulieren S 10. Plastische Verformung nach Lastmaximum. 40
Bild 3.35. Simulieren S 10. Plastizierung nach Lastmaximum. 41
Bild 3.36. Simulieren S 11. MomentRotationslinie. 41
Bild 3.37. Simulieren S i l . Plastische Verformung nach Lastmaximum. 42
Bild 3.38. Simulieren S i l . Plastizierung nach Lastmaximum. 42
4. Numerische Simulierung der Portalrahmen.
Bild 4.1. Statisches Schema der Rahmen. 43
Bild 4.2. RiegelStützen Verbindungen im Versuch. 44
Bild 4.3. Modell der SpannungsDehnungslinie. 45
Bild 4.4. Diskretisierung: Numerierung der Elemente. 46
Bild 4.5. Diskretisierung: Perspektive. 47
Bild 4.6. Instrumentierung der Versuche. (CRM ref (2)) 49
Büd 4.33. Folge der Hießgelenk Büdung. 52
I o Gleichzeitige Formierung von Fließgelenken in C und D;
2° Fließgelenk in B;
3° Fließgelenk in A.
Büd 4.7. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Vert Verschiebung zu vert. Last. 54
Büd 4.8. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Hor. Verschiebung zu vert. Last. 54
Büd 4.9. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Hor. Verschiebung zu vert. Last. 55
Büd 4.10. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 1 zu Last. 55
Büd 4.11. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 2 zu Last. 56
Büd 4.12. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 3 zu Last. 56
Büd 4.13. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 4 zu Last. 57
Büd 4.14. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 5 zu Last. 57
Büd 4.15. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 6 zu Last. 58
Büd 4.16. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 7 zu Last. 58
Büd 4.17. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 8 zu Last. 59
Büd 4.18. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 9 zu Last. 59
Büd 4.19. Simulieren Ρ 1, Ρ 2, Ρ 3, Ρ 4. Dehnungsmesstreifen 10 zu Last. 60
Bild 4.20. Simulieren Ρ 1, Ρ 2, Ρ 5, Ρ 7, Ρ 8. Vert. Verschiebung zu vert. Last. 60
Bild 4.21. Simuüeren Ρ 1, Ρ 2, Ρ 5, Ρ 7, Ρ 8. Hor. Verschiebung zu vert. Last. 61
Büd 4.22. Simulieren Ρ 9, Ρ 10. Vert. Verschiebung zu vert. Last 61
Büd 4.23. Simulieren Ρ 9, Ρ 10. Hor. Verschiebung zu vert. Last 62
Büd 4.24. Simulieren Ρ 11, Ρ 12. Vert. Verschiebung zu vert. Last 62
Büd 4.25. Simulieren Ρ 11, Ρ 12. Hor. Verschiebung zu vert. Last. 63
Büd 4.26. Simulieren Ρ 13, Ρ 14, Ρ 15. Vert. Verschiebung zu vert. Last 63
XV
Seite
Bild 4.27. Simulieren Ρ 13, Ρ 14, Ρ 15. Hor. Verschiebung zu vert. Last. 64
Bild 4.28. Simulieren Ρ 13. Plastifizierung. 64
Bild 4.29. Simulieren Ρ 16, Ρ 17. Vert. Verschiebung zu vert. Last 65
Bild 4.30. Simulieren Ρ 16, Ρ 17. Hor. Verschiebung zu vert. Last. 65
Bild 4.31. Simulieren Ρ 16, Ρ 17. Stützensfuß Rotation zu Last. 66
Bild 4.32. Simulieren Ρ 16, Ρ 17. Riegel Rotation in der Mitte zu Last. 66
5. Numerische Simulierung von fiktiven Stahlqualitäten.
Bild 5.1. Statisches Schema der Träger. 67
Bild 5.2. Geometrische Eigenschaften. 67
Bild 5.3. SpannungsDehnungslinie. 68
Bild 5.4. Allgemeine sinusoïdale Anfangsverformung. 69
Bild 5.5. Simulieren AA bis CC. 71
Bild 5.6. Plastische Verformung bei M=Mpl (abnehmender Teil). 72
Bild 5.7. Plastische Verformung bei M=Mpl (abnehmender Teil). 73
Bild 5.8. Plastische Verformung bei M=Mpl (abnehmender Teil). 74
Bild 5.9. R, Funktion von fy und fu/fy. 75
6. Schlußfolgerungen.
Bild 6.1. Nichtdimensionale Definition der Rotationskapazität. 77
Bild 6.2. Rotationskapazität der Träger (HEB 200) und der Portalrahmen. 77
XVI
List of the tables. p age
l.Introduction.
Table 1.1. Eurocode 3 requirements for plastic design (in part) 4-5
2. Experimental part
Table 2.1. η
Table 2.2. η
XVII
Liste des tables. Page
l.Introduction.
Table 1.1. Prescriptions de l'Eurocode 3 pour le calcul plastique. 4-5
2. Partie experimentale.
Table 2.1. 7
Table 2.2. 7
3. Simulations numériques des essais de flexion sur trois points.
Table 3.1. Simulations numériques. 24
Table 3.2. Résumé des figures 3.13. à 3.38. 26
XVIII
Tabellenliste. page
1.Einleitung.
Tabelle 1.1. Anforderungen des Eurocode 3 für plastisches Design. 4-5
2. Experimental Teil.
Tabelle 2.1. 7
Tabelle 2.2. 7
3. Numerische Simulierung der drei-punkt Biegebeanspruchung.
Tabelle 3.1. Numerische Simulierung. 24
Tabelle 3.2. Zusammenfassung der Bilder 3.13. bis 3.38. 26
4. Numerische Simulierung der Portalrahmen.
Tabelle 4.1. Numerische Simulierung. 48
Tabelle 4.2. Zusammenfassung der Bilder 4.7. bis 4.32. 50
5. Numerische Simulierung von fiktiven Stahlqualitäten.
Tabelle 5.1. Geometrische Eigenschaften. 68
Tabelle 5.2. Mechanische Eigenschaften. 68
Tabelle 5.3. Amplitude (mm) der Anfangsverformungen. 69
Tabelle 5.4. Zusatzwandstärke der Steg-Flansch Verbindungen. 70
Tabelle 5.5. Rotationskapazität. 70
XIX
List of symbols
Geometrical properties.
Ao Cross-section area
b Width of the flange
f Flange
H Height of the frame
h Height of the profile
L Length of the beam
R Radius of fillet
t Thickness
tf Thickness of the flange
tw Thickness of the web
Wpl Plastic section modulus
W Elastic section modulus
w Web
Material properties.
Ar Elongation at failure
E Young modulus
Est Strain hardening modulus
fy Yield strength
fu U ltimate strength
ε Strain
εγ Yield strain
est Strain corresponding to the end of the plateau
£p Plastic strain
eu Strain corresponding to the ultimate strength
ν Poisson's coefficient
σ Stress
XX
Deformations.
U x , Uy, U z Displacements in the χ, y and ζ directions
Oy, ûz, 6y, θ ζ Rotations around the y and ζ axes
R Rotation capacity
α Amplitude of global plane initial deformation
β,Υ Amplitude of global and local front initial deformation
δ Deflection, displacement
φ Rotation
<t>e Elastic rotation
Φΐ=Φρ1 Plastic rotation
Φ2 Rotation when the moment is equal to M p j in the decreasing part of the
moment-rotation curve
Units·
m Meter
cm Centimeter
mm Millimeter
kg K ilogram
t Ton
Ν Newton
kN K ilonewton
Pa Pascal
XXI
Liste des symboles
Propriétés géométriques.
A0 Surface de la section transversale
b Largeur de la semelle
f Semelle
H Hauteur du portique
h Hauteur du profilé
L Longueur de la poutre
R Rayon de congé
t Epaisseur
tf Epaisseur de la semelle
ty, Epaisseur de l'âme
WDi Module de flexion plastique
\V Module de flexion élastique
w Ame
Propriétés mécaniques.
Ar Elongation à la rupture
E Module d'Young
Est Module d'écrouissage
f C ontrainte élastique
fu C ontrainte ultime
ε Déformation
εν Déformation élastique
est Déformation correspondante à la fin du plateau
ε_ Déformation plastique
Ej, Déformation correspondante à la contrainte ultime
ν C oefficient de Poisson
σ C ontrainte
XXII
Déformations.
U x , Uy, UZ Déplacements dans les directions χ, y and ζ
*y» * z ' e y> θ
ζ Rotations autour des axes y and ζ
R Capacité de rotation
α Amplitude de la déformée initiale globale (vue en plan)
β,γ Amplitude de la déformée initiale globale et locale (vue de face)
δ Flèche, déplacement
Φ Rotation
Φβ Rotation élastique
Φΐ=Φρ1 Rotation plastique
Φ2 Rotation lorsque le moment est égal à Mp¡ dans la partie décroissante de
la courbe momentrotation
unités.
m Mètre
cm Centimètre
mm Millimètre
kg Kilogramme
t Tonne
Ν Newton
kN Kilonewton
Pa Pascal
XXIII
Verwendete Symbole
Geometrische Größen.
A
o Querschnittsfläche
b Profilbreite des Flansches
f Flansche
H Rahmenhöhe
h Profilhöhe
L Trägerlänge
R Rundungsradius
t Dicke
tf Flanschdicke
tW Stegdicke
Wpi Plastisches Widerstandsmoment
p
w Elastisches Widerstandsmoment
w Steg
Werkstoffkennwerte.
Ar Bruchdehnung
E Elastizitätsmodul
Est Verfestigungsmodul
fy Rechenwert der Streckgrenze
fu Rechenwert der Zugfestigkeit
ε Dehnung
£y Elastische Dehnung
Est Dehnung am Ende des Lüdersplateaus
Ep Plastische Dehnung
£u Dehnung am Ende der Zugfestigkeit
ν Querkontraktion
σ Spannung
Kräfte. Schnittgröften.
M Biegemoment
Me Elastische Biegemoment
Mp Plastische Biegemoment
Ρ, Ν Einzellast
ρ Streckenlast
Mim Plastische Last der Rahmen
XXIV
Verformungen.
U x , Uy, U z Verschiebung in die x, y und ζ Richtungen
Üy, # Z , e y , θ ζ Rotationen rundum die y und ζ Richtungen
R Rotationskapazität
α Globale Anfangsverformung in Draufsicht
β,Τ Globale Anfangsverformung in Ansicht
δ Durchbiegung, Verschiebung
Φ Verdrehung
Φε Elastische Verdrehung
Φΐ=Φρ1 Plastische Verdrehung
Φ2 Grenzverdrehung bei Wiedererreichen von M p l im abnehmenden Teil der
MomentRotationslinie
Einheiten.
m Meter
cm Zentimeter
mm Millimeter
kg Kilogramm
t Tonne
Ν Newton
kN Kilonewton
Pa Pascal
XXV
1. Introduction.
The plastic design theory of steel structures developed after the first world war has found its
first soar with the coming of numerical methods on computers. This theory is based on the
elastic perfectly plastic behaviour of the material, (seefigure1.1). In reality, the steels do not
follow exactly this representation. The FeE 235 steel (fy = 235 N/mm^) has a stress-strain
curve as indicated infigure12.
i
fu
^ \
fy
\E
est* ειι ε
Figure 12. Example of stress-strain curve for Steel FeE 235
The assumption of a steel with an elastic perfectly plastic behaviour implies the two following
fundamental concepts: the plastic moment and the plastic hinge. A bending beam in steel
elastic perfectly plastic has a moment rotation curve as indicated in the figure 1.3. At the
beginning of the loading, it exists an elastic domain and then, in function of the plastification in
the mid-span section, the curve becomes asymptotic to an horizontal level. The plastic design
theory is based on an elastic perfectly plastic moment rotation curve with unlimited rotation of
the beam. In practice, because of the strain hardening of the steel, the moment capacity of a
beam in class 1 (ref 1 §5.3) is higher than the theoretical plastic bending moment (see figure
1.4). The rotation φ 1 is, by definition, the rotation corresponding to the theoretical plastic
bending moment of the beam.
Rotation capacity
φ2-φ1
R=
φΐ
φβ φρ1=φ1
The rotation φ2 is the rotation of the beam when the moment reaches again the plastic moment
value in the decreasing part of the curve in consequence of local buckling in the plastic zone. By
definition, the difference between the rotation φ2 and φΐ divided by φ 1 is called the rotation
capacity of the beam. This ratio is a non dimensional value.
The example of plastic calculation mentioned at figure 1.5. shows that a plastic design can
lighten the weight of a structure. But this method requires materials which are able to undergo
sufficient plastic strains.
Ρ
ί i
5 b i 4, i v t
'— I
L
9
PL 2 vi?
bending moments 12
Ν. Β ^y 2
A
elastic domain! pL2
fy M fy
stress distribution
A , Β and C
fy ify/2 fy
bending moments Mp [ \ ^ Β /
fy \ Mp
A C
plastification without redistribution : \
g a i n l = WpI/W = 6 t o l 5 % fy fy
¡(function of the profile)
\
stress distribution
\
fy fy
bending moments Mp \ B
λ Mp
V Ài. ^ ^ / c
plastification with redistribution : Mp
gain 2 = 33.3% in this example
»J ' j ly
stress distribution
total gain = gain I + gain 2
= 39.3 to 48.3% fy fy fy
(2) At plastic hinge locations, the cross-section of the member which contains the
plastic-hinge shall have a rotation capacity of not less than the required
rotation at that plastic hinge location.
(4) For building structures in which therequiredrotations are not calculated, all
members containing plastic hinges shall have Class 1 cross-sections at the
1
E.C.C.S. = European Convention for Constructional Steelwork
plastic hinge location, (for the classification see Table 5.3.1 of Eurocode 3)
The aim of this research was to analyse the mechanical specifications for a plastic design given
by Eurocode 3.
Our partner in this research was the CRM2, which was dealing with the experimental part of the
work. ARBED has supplied the basic material for the tests. Five different steels (St 37-2, St 52-
3, FeE 460, FeE 235, FeE 420) have been tested by the CRM. All the used profiles were
HEB 200. For the five steels, the profile HEB 200 is classified in class 1 according to the
classification of the cross-section of Eurocode 3 (§5.3) (ref.l). That means that this profile has
a cross-section which can form plastic hinge with the rotation capacity required for plastic
analysis.
In order to generalise the results obtained in the practical tests, ARBED has carried out
numerical simulations by finite elements with the programme FINELG. After parametrising
this model and having verified the accuracy of the predictions with regard to the experimental
behaviour, the model has been applied to hypothetic steels, differing by their strain hardening
features and their fu/fy ratios. The aim of the simulations on hypothetic steel is to demonstrate
that the mechanical requirements given in Eurocode 3 are not necessary for all the plastic
design.
- Chapter 1 on "Introduction"
- Chapter 2 on "Experimental investigation" summarises the work carried out by CRM in the
ECSC research n°7210/SA-204. (ref 2). This chapter contains two parts: three-points bending
tests and portal frame tests.
- Chapter 3 on "Numerical simulations of three-point bending tests" presents the results of the
numerical simulations and the numerical model developed in order to simulate the behaviour of
an isostatic beam loaded at mid-span. These numerical simulations aims to analyse the influence
of geometrical and mechanical data on the plastic behaviour of HEB 200 beams.
- Chapter 4 on "Numerical simulations of portal frames" presents the results of the numerical
simulations of simple frame structures calibrated on practical tests in order to find the needed
requirements for plastic design of a structure.
2
CRM = Centre de Recherches Métallurgiques (Liège - Belgium)
- Chapter 5 on "Numerical simulations on hypothetic steels" presents the results of the
numerical simulations of isostatic beams in steel FeE235, FeE355 and FeE460 with ratios fu/fy
= 1.0,1.1 and 1.2.
- Chapter 6 gives the conclusions of these simulations and presents the requirements for a
plastic design.
2. Experimental part.
This chapter describes briefly the tests performed by the CRM (ref 2) in the frame of the
E.C.S.C. research n°7210SA/204.
tf
\
f s
r
tw.
^ R HEB 200 Beams
/
s1
V Λ
Table 2.1.
In order to simulate correctly the tests, the initial deformations and the residual stresses of the
profiles have been measured by the CRM. A description of the device for initial deformations
and of the cutting method for residual stresses can be found in the technical report n°2
(September 89) of the CRM research (ref 2).
The figures 2.1 and 2.2 describe the instrumentation of a beam submitted to a three-point
bending test
Rotation capacity
φ2-φ1
R=
J>L
<t>e φρΐ=φΐ φ2
straingages (10*/·)
straingages 127.)
The figure 2.3. and 2.4 show the static scheme of the portal frame test and a photo of a frame.
HEB 200
H = 2m
/////// Π ITU ι s
2m 2m
/- -/- -/
10
2.6. Results.
All the results of the tests are included in the diagrams with the numerical simulations in
chapters 3 and 4.
11
3. Numerical simulations of three-point pending tests.
The aim of this chapter is to describe the numerical model used to simulate the five three-points
bending tests, (performed by the CRM (ref 2)).
FINELG is a materially and geometrically non-linear finite element programme which has been
developed jointly at the University of Liège, Belgium, and at the Polytechnic Federal School of
Lausanne, Switzerland. It is used to solve problems such as:
- step-by-step structural response up to and beyond collapse;
- linear and non-linear instability with calculation of critical loads and instability modes;
- calculation of eigen frequencies and eigen modes, possibly taking into account the current
stress state.
Its library is composed of spatial truss bar, plane beam, spatial beam, membrane plate, thin and
thick shells, springs, linear constraints,...
The static scheme for the numerical simulations is indicated in figure 3.1.: three points bending
test - beams 3 meters long - 3 stiffeners.
12
3.2.3. Mechanical characteristics.
Type Ml : a four-linear law taking into account the tangent hardening modulus.
Type M2 : an eight linear law in order to try to follow more precisely the real behaviour of the
material during steel hardening.
"i i
/Type Ml
fu E = 205000 N/mm 2
/ ■
V = 0.285
1
fy ^Type M2
E
e
\
ey Est ερ eu Ar
The residual stresses have been simulated by a classical scheme for hot rolled H sections, (see
figure 3.3.). We have made simulations with and without residual stresses in order to analyse
their effects on the plastic behaviour.
13
profile following NBN B51-002
117.5 N/mm2
117.5 N/mm2
+100 (Mpa) measured at CRM
3.2.5. Discretisation.
The all structure, beam and stiffeners has been discretised in 600 thin shell finite elements, (see
figure 3.4).
In order to simulate the fillets, we have used an overthickness of the shell elements near the web
to flange connection, (see figure 3.5).
14
Figure 3.4. Discretisation : without simulation of the fillets : Type DI 1
15
For the overthickness of these elements, we can consider different possibilities. They are all
based upon an equivalence of area of the web toflangesjunction for the real hot rolled beam
and for the beam discretised intofiniteelements. We have supposed that thickness of elements
(web orflanges)are constant during thefirst1/5 of the fillet
Ρ il*
tw
Ar = tw. tf + 2.R. tf + R. tw + (4 π) .R2/2
First case : overthickness of the elements in the web and the flanges (recovering of shell
elements taken into account) : Type DI 2 A
tf + 2 . »1
>·
tw + 2 . i l
(47i)R twtf
S l = A r = > a l = ~24
2(jR+2tw +tf)
Second case : overthickness of the elements in the web and the flanges (without recovering of
shell elements taken into account): Type DI 2 Β
Y
^-
S2 = 2tf.R + tf.tw + tw.R + (24/5R + tw).a22.(a2) 2
16
24 ΠΑ
-(■jR +tw) + R+tw) 2 +4(47c)R 2
S2 = Ar = = > a2 =
Third case: overthickness of the shell elements in the flanges only (without recovering of shell
elements taken into account): Type DI 2 C
tf+ 2.·3
(4TI).RZ
S3=Ar=>a3 = Ï6~~
2(tw +5R)
3.2.6. Initial deformation.
17
2
I
ι si j
SI : compressed flange
S2 : tensile flange
| 4 S2
18
direction 1:
χ x* L = 3000 mm
^
direction 2:
0.8 sin ( π χ / 3 0 0 0 )
direction 3:
direction 4:
0.6sin(nx/100)
Figure 3.7. Example of initial deformation simulated with sinus curves (Steel D3): Type ID 1
19
b) In the second case, we have chosen a more simple general type of initial deformation, (see
figure 3.8).
Plane view
, α sin (π x/L)
^ ^ —
f Ν.
/
X
tz
J
Front view
X
Ν
ß
β sin (π x/L)
" " ~
c) In the third case, we have added to the global initial deformation mentioned in the point b, a
local initial imperfection (see figure 3.9 and 3.10). The aim of this initial imperfection is to
increase the speed of convergence of the iterating process during the calculation.
20
H-ysin(7ix/l)| v -ysin(nx/l)
Plan view of the
compressed flange: -Χ—ιέ.
t
ζ + g sin (πχ /1)
I 'b/2~'
21
Figure 3.10. Global and local initial deformation: Type ID 3
Different types of boundary conditions have been used for the numerical calculations to
simulate supports during the experimental tests.
a) Type Bl : (see figure 3.11) Nodes of the 2 extremities on the tensile flange fixed in y
direction (x=0 and x= 3000 mm).
Nodes of the 2 extremities and situated on the beam axesfixedin the Ζ direction.
Nodes in the middle of flange fixed in the X and Ζ directions.
b) Type B2 : same as Bl plus rotations around y axes (ûy) fixed for all the nodes situated in
the external sections (x=0 and x= L = 3000 mm)
c) Type B3 : (see figure 3.12) all the nodes of the tensile flange in external sections (y=0 for
x=0 and x=L=3000 mm) fixed in y and ζ direction and around y axes
(Uy,Uz,^y) plus one node fixed in χ direction to avoid rigid translation.
22
Figure 3.11. Boundary conditions: Type Bl
ιπτττί ιίπιηπ/ιπιπΐ7η/ΐπΐ7ΐΝ!ΐ " ITT 71Π ΙΤΠΤΠ ΓΓΤΠΤΙ ΓΓΠπ I Γ77Π1 ΓΠΤΤ7ΤΓΤΤ7ΤΠΙ Γ Π Τ Π
x=0 x=L
Figure 3.12. Type B3.
3.2.8. Loading.
The load at mid-span has been introduced following two different ways.
a) Type LI : the concentrated load is replaced by an equivalent model loads at mid-span on the
compressed flange.
b) Type L2 : we use one concentrated load and linear constraints to force the nodes situated at
mid-span on the compressedflangeto have the same displacement.
23
3.2.9. Data for numerical simulations.
The table 3.1 summarises the numerical simulations performed during this research on the five
beams tested by the CRM in order to find a good numerical model. These simulations are
classified in chronologic order: SI has been made at the beginning of the research and S i l at
the end. Before these simulation we have tried to simulate a test found in the literature (ref 3).
These first simulations have allowed to define the meshing of the numerical model. We have
beginning with 296 elements for the discretisation of the beam, then because of bad results, we
have increased this number of elements to 460 and finally to 600 elements. Due to the fact that
the number of elements and nodes is very high, the needed time for one calculation of a beam
discretised with 600 shell elements (nodes with 6 degrees of freedom) takes two or three weeks
on our micToVax Π computer. This time takes only into account the computer calculation and
not the preparation of the data file and the analysis of the results.
24
Discretisation: (see §3.2.5.)
DI 1 = no fillets
DI 2 A = fillets: overthickness in the web and in theflangeswithre-coveringtaking into account
DI 2 Β =fillets:overthickness in the web and in the flanges without re-covering
DI 3 C =fillets:overthickness in the flanges only without re-covering
Initial deformation: (see §3.2.6.)
ID 1 = real initial deformation
ID 2 = global initial deformation
ID 3 = global + local initial deformation
Boundary conditions: (see §3.2.7.)
Loading: (see §3.2.8.)
L 1 = equivalent model of loads at mid-span
L 2 = one concentrated load + linear constraints for vertical displacement of the compressed
flange
3.2.10. Numerical simulations results and comparison with experimental tests.
The table 3.2. gives a summary of the figures 3.13. to 3.38. All the figures are numbered in a
chronologic order. It is easier to explain the way that we have done in order to find a good
numerical model
General remarks:
a. The moment rotation curve is the total rotations φ of the beam extremities as a function of the
bending moment M. (M = Ρ L/4 where Ρ is the load and L is the span).
b. In all the moment rotation curves, the values of plastic moments are calculated with the
measured geometrical and mechanical properties of the tested beams.
c. The moment rotation curves called CRM are the results of the practical tests performed by the
CRM.
d. The figures showing the plastic deformation of the beams are obtained by removing the
elastic deformations from the total displacements.
25
Simulation Figures Contents
number
SI 3.13. Moment rotation curve
3.14. Plastic deformation at the maximum of the load
3.15. Plastic deformation after the maximum of the load
3.16. Plastification at the maximum of the load
3.17. Plastification after the maximum of the load
S I , S 2, S 3 3.18. Comparison between S 1, S 2 and S 3
Effect of the modélisation of the stress-strain curve
Effect of the residual stresses
Effect of the initial deformation
Effect of the boundary conditions
S4 3.19. Moment rotation curve
3.20. Plastic deformation at the maximum of the load
3.21. Plastic deformation after the maximum of the load
3.22. Plastification at the maximum of lhe load
3.23. Plastification after the maximum of the load
S 5, S 6, S 7 3.24. Moment rotation curve
S5 3.25. Plastic deformation after the maximum of the load
S5 3.26. Plastification after the maximum of the load
S8 3.27. Moment rotation curve
3.28. Plastic deformation after the maximum of the load
3.29. Plastification after the maximum of the load
S9 3.30. Moment rotation curve
3.31. Plastic deformation after the maximum of the load
'3.32. Plastification after the maximum of the load
S 10 3.33. Moment rotation curve
3.34. Plastic deformation after the maximum of the load
3.35. Plastification after the maximum of the load
SU 3.36. Moment rotation curve
3.37. Plastic deformation after the maximum of the load
3.38. Plastification after the maximum of the load
At the beginning of the research, the only beam tested by the CRM was the beam in steel D3
(FeE460). Thus, we have also begun our simulations with this type of steel with the meshing of
600 elements like explained in §3.2.9. Figure 3.13. gives the moment rotation curve of the
numerical simulations and of the test. We can see that the rotation capacity given by the
simulation is once more less than the experimental one. Nevertheless, the local buckling
phenomena appears in the same way than in the test, (see figure 3.14 and 3.15 and photo 3.1).
The figure 3.16 and 3.17 about the propagation of the plastification in the plastic hinge confirm
that the deformed shape is antimetricaL
26
photo 3.1. Antimetrical deformation shape of a test performed by CRM.
Due to the bad results of this first simulation, we have tried to analyse different parameters: the
stress-strain curve modélisation (§3.2.3.), the effects of the residual stresses (§3.2.4.), the
boundary conditions (§3.2.7.) and the way to apply the load at mid-span (§3.2.8.). The figure
3.18 gives three results of moment rotation curves in comparison with those of the test. We can
see that for all those three simulations, the rotation capacity is far from the experimental one and
that these different parameters can not explain the differences of behaviour between the
simulations and the test . As above concerning the local buckling and the shape of the
deformations, we have found the same conclusions for all the three simulations.
Looking the figure 3.18, we have thought that the problem of the numerical simulations was the
miss of rigidity between theflangesand the web because the fillets were not simulated. It would
be the reason why the local buckling would appear earlier in the simulations than during the
test. First of all, we have tried to simulate the existence of the fillets using suffer elements at the
web to flange connection. But the results of that numerical simulation were not good at all. The
response was too stiff. We have then decided to simulate the fillets using thicker shell elements
at the connection. The first overthickness for the simulation S 4 has been chosen arbitrarily
(new thickness = old thickness + 2. a with a = 1.7 mm) all along the beam and over a breadth
equal to 4/5 of the fillet (R=18 mm). Indeed, it is usually admitted that, in hot-rolled profiles,
thickness of elements (web of flanges) is constant until 1/5 R. (ref 4). For the simulation S 5, S
6 and S 7, we have tried to justify that overthickness making a calculation of area equivalence,
(see §3.2.5). Figure 3.19 shows the moment rotation curve for the simulation S 4 (steel D3).
27
Now, the correspondence between numerical and experimental results is quite good. But in this
case, the local instability of the compressedflangeoccurs only on one side of the beam (see
figure 3.20 and 3.21). This explains the plastic concentrations on the same side of the flange
after that the maximum of the load has been reached, (seefigures3.22 and 3.23). The figure
3.24 shows the moment rotation curves of simulation S 5, S 6 and S 7 (steel Dl = St37-2)
where different overthicknesses have been taken into account in order to simulate the fillets. In
the simulation S 5, we have an overthickness of 0.64 mm in the elements of theflangesand of
the web. (see §3.2.5 Type DI 2 A ). In the simulation S 6, the overthickness is equal to 1.42 mm
instead of 0.64 mm (see §3.2.5 Type DI 2 Β ). We see on figure 3.24 that the maximum
moment increases a little bit. In the simulation S 7, we have only thickened theflangesof the
beam of 2.07 mm (see §3.2.5 Type DI 2 C ) which, as we can see, does not give very different
results from the simulation S 6.
Anyway, though the overthicknesses used are not far from the one taken for the simulation S 4
(0.64 to 2.07 mm against 1,7 mm for S 4), we are to notice that, once again, the rotation capacity
obtained with numerical simulations is far from the experimental one. The figures 3.25 and 326
about the deformation and the plastification of the simulation S 5, show that the instability of the
compressedflangeand so the plastification appears antimetrically. It was not the case with the
good simulation S 4. But all the experimental tests have shown an antimetrical shape
deformation like in photo 3.1.
These simulations have shown that the initial deformation and especially the local initial
deformation in the plastic zone can influence enormously the rotation capacity of the beam.
Indeed, looking at the table 3.1, we see that, except numerical values for the stress strain curve,
very small differences in geometrical properties and in the choice of the overthickness of the
elements, the only great differences for the data of the three simulations S 5, S 6 and S 7 with
steel Dl and the simulation S 4 with steel D3 is the shape of initial deformation: no local
imperfection for simulation S 4.
We have then begun again the simulation of the steel Dl but without local imperfection
(simulation S 8). Thefigure3.27 shows that the moment rotation curve is better than those of
thefigure3.24 where a local imperfection was taken into account. But thefigures3.28 and 3.29
show that the buckling deformation appears on one side like in the simulation S 4. We have
simulated the beams in steel D2 (St52-3), D4 (FeE235) and D5 (FeE420) with the same
assumptions than the simulation S 8. Thefigures3.30 (simulation S 9), 3.33 (simulation S 10)
and 3.36 (simulation S i l ) show respectively the moment rotation curves of steel D2, D4 and
D5. Thefigures3.31,3.34 and 3.37 show that in this case the deformation shape is antimetrical
like in the tests. This is confirmed with the plastification after the maximum of the load, (see
figures 3.32,3.35,3.38).
In conclusion, the model using a global initial deformation without local deformation in the
plastic zone (in order to increase the convergence of the numerical calculation) and with fillets
28
taking into account by means of an overthickness of the web and the flanges gives a good
evaluation of the moment rotation curve of the five tested beams. In two cases (Steel Dl =
ST37-2 and D3 = FeE460) the buckling deformation appears only on one side of the beam
contrarily to the experimental tests where the buckling deformation is antimetrical. This
difference between the deformations at collapse in the tests and in the simulations depends on
the minimal energy needed to develop the plastic hinge.
29
Figure 3.13. Simulation S 1. Moment rotation curve
30
Figure 3.15. Simulation S 1. Plastic deformation after the maximum of the load
31
Figure 3.17. Simulation S 1. Plastification after the maximum of the load
Hp = 2 8 / f b m
ROTATION
TEST LABO
32
Ζ
40.00
J
33.00
30.00
33.00
[ i
ιΓ ι ι
i ι
ι
i\\. ac ι
!
»
\\ Λ 1 Ι Ι \ Ι Ι
20.00 7 ! ! ! 1 Λ !
15.00
l
J ' ' ¡ ! ! MOMENT — R O T A T I O N
10.00
S.OO / ! ! ! ! ! ! ..*. S4
33
Figure 3.21. Simulation S 4. Plastic deformation after the maximum of the load
34
Figure 3.23. Simulation S 4. Plastification after the maximum of the load
M ¡
'' ' ' i i
■.': 23.00
_20._00
Γ
f : i Κ' : ¡1 ;
13.00
10.00
S.00
35:
Figure 3.25. Simulation S 5. Plastic deformation after the maximum of the load
36
M
22.90
1 """^
: ι ινν ¡Ν « r - « *hm
V
ί . i ! Lij
10.00
«WENT ROTATION
s.oo
ι 3.00
!
¡1O0O
î
¡15.00
' i
¡20.00
! p7.io
¡23.00 30.00
---*--■ S 8
Figure 3.28. Simulation S 8. Plastic deformation after the maximum of the load
37
Figure 3.29. Simulation S 8. Plastification after the maximum of the load
Π
30.00
25.00
">*"^\ ttp=«/<Stm
if ! j ! X
SO. 00 Λ*
\ ! Ι ί .
10.00
MOMENT ROTATION
5.00
TEST LABO
---*-- S9
ρ■4.70
¡3.00 ¡10.00 ¡13.00 20. OO 23.00
DEG
38
Figure 3.31. Simulation S 9. Plastic deformation after the maximum of the load
39
46.¿ tir
ΜΟΙΈΝΤ - ROTBTION
TEST LABO
---*-- S 10
Figure 3.34. Simulation S 10. Plastic deformation after the maximum of the load
40
Figure 3.35. Simulation S 10. Plastification after the maximum of the load
Hp= 2¿/tm
MOPENT - ROTATION
TEST LABO
S 11
41
Figure 3.37. Simulation S i l . Plastic deformation after the maximum of the load
42
4. Numerical simulations of portal frames.
After having tested different beams in order to determine the rotation capacity of their section,
we have undertaken several tests on portal frames. These simple structures were done with the
same profiles as the beams, HEB 200. These tests are made in order to observe the rotations
that occur effectively in the frames, and, in particular, those rotations value at the time of the
formation of the plastic yield mechanism.
These results will allow us to show that the rotation necessary to reach the plastic collapse load
is less than the plastic rotation capacity of the structural elements.
The static scheme for the numerical simulations is indicated infigure4.1. : simple portal frame -
HEB 200 profiles - 1 horizontal load - 1 vertical load at mid-span of the beam - rigid column
bases - rigid connection between the beam and columns.
y <
HEB 200
H = 2m
ιι/τιιι 7777777
2m 2m
/- v- -/
In order to reach the maximum of the bearing capacity of the frames, it is important that the
beam to column joints don't create any weakness in the structure. This have led the CRM to
device very rigid connections, by welding and stiffening the joint in order to prevent any shear
43
deformation in the column web panel, (seefigure4.2)
We have taken the geometrical characteristics measured by the CRM at the time of the three-
points bending tests study .(see table 2.1)
The numerical simulations have been made with beam finite elements of the non-linear finite
elements programme FINELG. This beam finite element don't allow to define two stress-strain
curves in one section, it was impossible for us to simulate exactly the mechanical properties of
the beams. We have used a global four linear model for the stress-strain curve modélisation of
the whole section, (seefigure4.3).
44
σ
ii
fu E = 205000 N/mm2
/ V = 0.285
fy
I ey
E
Est ερ eu Ar
ε
We have also analysed the influence of rolling residual stresses, but not the welding residual
stresses that exist in the beam to column connections. The welding residual stresses may
influence the development order of the plastic hinges in the frame but do not change its plastic
capacity. The residual stresses used in the simulation are the same than for the bending tests.
Type RS 1 : with residual stresses (yes in table 4.1)
Type RS 2 : without residual stresses (no in table 4.1.)
45
4.1.5. Discretisation.
The structure has been discretised with 20 plane beams elements (see figure 4.4. and 4.5.);
elements 5, 6,15 and 16 simulate the beam to column joints. These elements have a greater
inertia than the other elements of the frame in order to simulate rigid joints.
ν 6 7 8 9 10 Π 12 13 ΜΙ 5
~ ^ ^ "™"^ΙΛ ι
^" 01
Ρ :
Ψ
m
α»
ι ι ιι
CM
<ο
> ι Ι
ro
777J777 77?fì7?
46
Figure 4.5. Discretisation: perspective view.
Due to the fact that the portal frame is loaded with an horizontal and vertical forces, it is not
necessary to put an initial deformation in the numerical simulations. Indeed, this loading system
is not symmetrical with respect to the structure and thus the second order effects can directly
appear during the loading.
The base of the column is simulated like a rigid support. We must note that in reality, it is very
difficult to realise a rigid support for an experimental test.
4.1.8. Loading.
The loading system (one horizontal and one vertical loads) is simulated with concentrated loads
applied on two nodes of the structure.
The table 4.1 summarises the numerical simulations performed on the portal frames. Like the
47
numerical simulations on bending tests, these simulations Ρ 1 to Ρ 17 are classified in
chronologic order.
3 MSM = Mécanique des structures, Stabilité des constructions, Mécanique des matériaux
48
££$XS>
p
\
Ρ hor. displ. left 7 i"
I
5 hor. displ. right
-1
Λ 4 3
vert. dispL
1 c i 2 D Straingages 10 c i 9
SXVOvV WVCvV
49
Simulation Figures Contents
number
P1,P2,P3,P4 4.7. vert, displ. versus vert load
P1,P2,P3,P4 4.8. hor. displ. versus vert, load
P1,P2,P3,P4 4.9. hor. displ. versus vert, load
P1,P2,P3,P4 4.10. straingage 1 versus load
P1,P2,P3,P4 4.11. straingage 2 versus load
P1,P2,P3,P4 4.12. straingage 3 versus load
P1,P2,P3,P4 4.13. straingage 4 versus load
P1,P2,P3,P4 4.14. straingage 5 versus load
P1,P2,P3,P4 4.15. straingage 6 versus load
P1,P2,P3,P4 4.16. straingage 7 versus load
P1,P2,P3,P4 4.17. straingage 8 versus load
P1,P2,P3,P4 4.18. straingage 9 versus load
P1,P2,P3,P4 4.19. straingage 10 versus load
P1,P2,P5,P7,P8 4.20. vert, displ. versus vert, load
P1,P2,P5,P7,P8 4.21. hor. displ. versus vert, load
Ρ 9, Ρ 10 4.22. vert, displ. versus vert, load
Ρ 9, Ρ 10 4.23. hor. displ. versus vert, load
Ρ 11, Ρ 12 4.24. vert, displ. versus vert, load
Ρ 11, Ρ 12 4.25. hor. displ. versus vert, load
Ρ 13, Ρ 14, Ρ 15 4.26. vert, displ. versus vert, load
Ρ13, Ρ 14, Ρ 15 4.27. hor. displ. versus vert, load
Ρ 13 4.28. plastification
Ρ 16, Ρ 17 4.29. vert, displ. versus vert, load
Ρ 16, Ρ17 4.30. hor. displ. versus vert, load
Ρ 16, Ρ 17 4.31. column base rotation versus load
Ρ 16, Ρ17 4.32. beam rotation at mid span versus load
Figures 4.7 to 4.19 show, as function of the load Ρ applied on the frame, the evolution of the
vertical displacement at mid-span, the horizontal displacement of the top of the column at the
side of the jack or at the other side and the evolution of the deformations that occur where the
straingages have been set during the tests in laboratory (seefigure4.6.) The curves called Labo
MSM or test Labo represent the experimental curves registered during the tests.
Thefigures4.10 to 4.19 show that rolling residual stresses modify the initial behaviour of the
structure without changing its plastic capacity.
The strain hardening does not influence plastic collapse but allows us to get nearer to the real
behaviour of the structure, (seefigure4.7 to 4.10, curves Labo MSM, Ρ 1 and Ρ 4). That
means that, for this type of frame structure, it is not necessary to have a strain hardening in
order to develop the plastic hinge and to reach the theoretical plastic load. Indeed, when the last
plastic hinge is formed, the maximum of strain in the structure (located at the first plastic hinge)
50
is lower than Est. (seefigure1.2)
The plastic load is also much influenced by the choice of mechanical characteristics (web,
flanges or mean values) of the section. Whatever are the hypotheses, the experimental plastic
load is always higher than that found by the simulations for the steel D4. (P 1 to Ρ 8).
On the contrary, infigures4.22 to 4.25 for simulations of steel Dl and steel D2 (P 9 to Ρ12),
the experimental curve is always between the two curves derived from the simulation. It is to be
noted that the sudden end of the experimental curves are due to a premature crack of the
weldings at the beam to column connection on the opposite side of horizontal jack. This has led
the CRM to reconsider the conception of the welding connections. A little difference of initial
rigidity between experience and simulations can be explained by the fact that it is not easy to
realise a complete fixed end for the columns during the test in a laboratory.
Concerning the experimental test on steel D3 (P 13 to Ρ 15), it had to be stopped because the
capacity of the hydraulic jacks had been reached. The curves displacements versus load (see
figure 4.26 and 4.27) have the same look as those given for the others types of steel. Anyway,
those calculation results are safe and the curve corresponding with the equivalent mechanical
data (Type Me §4.1.3) is located between those that consider flanges or web mechanical
properties. On the figures 4.26 and 4.27, we have mentioned the first order plastic load
calculated with a first order method. Plim(f), Plim(w) and Plim(e) represents respectively the
plastic load of the frame made in steel with the mechanical characteristics of the flanges, the web
and equivalent like explained in §4.1.3.
Thefigure4.28 shows the plastification and the position of the four plastic hinges. Wefindfor
all the simulations the same plastic hinges formation order than in the test, whatever are the steel
type or the hypotheses considered, (seefigure4.33 on next page).
51
Figure 4.33. Plastic hinges formation order.
I o - simultaneous formation of plastic hinges in C and D;
2° - plastic hinge in B;
3° - plastic hinge in A.
The figures 4.29. to 4.32. show the last results for the simulation of steel D5 (P 16 and Ρ 17).
Two portal frames made of steel D5 have been tested. They are supposed to be identical. The
two experimental curves are called Test Labo D5 and Test Labo D5bis. The rotations have also
been measured during the tests, (see figures 4.31 and 4.32). On figure 4.31., we can see the
rotation measured 400 mm higher than the base of the column as a function of the load for the
experimental tests and for the two simulations. During the test D5bis, the rotation of the base of
the column has also been recorded. (Test D5bis base, figure 4.31). This shows us that, as
explained before, the so called fixed ends of the columns are not in fact completely rigid in
rotation during the test The figure 4.32 gives the rotation that occurs near the plastic hinge in
the middle of the beam during experience and numerical calculation. On these figures, we have
also put the first order plastic load Plim(f) and Plim(w) calculated with the flanges and the web
characteristics.
- we can observe a good correspondence between the experimental and simulations results for
- ultimate load
- vertical displacement at mid-span
- rotation near the middle of the beam
- experimental results are greater than simulation ones concerning:
- horizontal displacements
52
- rotations of columns
- the residual stresses influence the frame deformation and the formation of the different plastic
hinges. It is in fact well known that any self-equilibrate stress state, if it does not change the
plastic load, can influence the way followed to reach that load.
- for all the frames, we can reach the plastic load before than the strains reach the end of the
plateau (est). That means that it is not necessary for this type of structure to have a strain
hardening in order to allow a redistribution of the bending moments inside the frames and thus
to develop completely the four plastic hinges. It is due to the fact that the required strains in
order to develop the plastic hinges are lower than est for this type of structure and this type of
loading. We can see that for very simple structures, it is possible to reach the theoretical plastic
load even when the ratio fu/fy is lower than 12\
53
s
DEP.VERT-CHAR.VERT-
LABOMSM
_-,—. PI
—-U— P3
—D— P2
--»— P4
y.i lM
OEP.H.VER-CHAR.VERT.
LABOMSM
__. PI
-.*_. P3
—c— P2
—■ P4
54
DEP.H.0P.VER-CH4R.V.
LABOMSM
_„—. PI
.-.-».-. P3
—°— P2
JAUGE 1 -CHAR.VERT.
LABOMSM
--,--. PI
--*— P3
—π— P2
_.„._.. P4
55
Figure 4.11. Simulation Ρ 1 , Ρ 2, Ρ 3, Ρ 4. Straingage 2 versus load
*'.ȃ3
56
r
LABOMSM
PI
■*— P3
-«■— P2
P4
iL*
JAUGE 5 - C HAR.VERT.
LABOMSM
PI
57
^r
~-'~~"ύ
iL
-JAUGE 6 - CHARTER.
LABOMSM
—■*— PI
■ -•-6— P3
—D— P2
—->— P4
s£_ ^ρ==·
_JL
-JAUGE 7 - CHAR.VER-
LABOMSM
—«—. PI
.-.„._.. P3
—D— P2
.„.-.. p4
58
JAUGE 8 - C HAR.VERT.
LABOMSM
PI
>— P3
P2
8.00 ţ f
E-06 P4
— JAUGE 9 CHAR.VER.
LABOMSM
— —*__ pi
3 .-.-».— P3
P2
28.00 f ^
e-os —o_.
P4
59
Figure 4.19. Simulation Ρ 1, Ρ 2, Ρ 3, Ρ 4. Straingage 10 versus load
35.00
1
I ) I I ! I
33-30 l i l i l í
r
I J-" I J^*++ *\ \ / Irí-** I J"
30.00
25.00 JïKjp»
¡4' *
*******
ι '
l
ι
i
ι
l i
ι
il
s
20-00
IS.00
Ι
DEP.VERT-CHAR.VERT.
10.00
LABOMSM
— M— PI
. 5-00
_.,.._ P5
J
— o — P8
*
-■*>·-■ P2
¡ :.00 j 4.00 ¡6.00 ¡8.00 ¡10.00 j12.00 *f
ι ι ι ι ι r MM O
' — P7
60
i, |M
iL
DEP.H.VERCHAR.VERT.
LABOMSM
_ . , . . . PI
_ . , » . - . P5
π P8
_..<,._.. P2
. P7
IUÆP
s
I D EP.VERCHAR.VERT.
LABOMSM
__«__. P9
* — PIO
61
=5:=;.·
DEP.H.VER-CHAR.VERT.
LABOMSM
__„._ P9
. .*— PIO
—
'-ni
43-0
Ί
«·ιο '
40.0
fJjfOi*tjf**-»*-'<-*-* «Η Ι Ι Ι
3 5 . OC
m Τ ! Γ Τ ^ ί
ii Ι ι ι ι ι ί ι
3 0 . OC
fí\ ! ! 1 ! ! }
25.00
1 '-^L·*
JL
20.00
¡Ι Ι Ι Ι Ι | | ι
15.00
DEP. VER-CHARTERT.
10.00
LABOMSM
Figure 4.24. Simulation Ρ 11, Ρ 12. Vert, displ. versus vert, load
62
4 ' 1 1 1 1 I 1
IO 1 1 1 1 1 1 1
45.01
L 1 J L i J _!_
42.10 ! ! ! ! ] ! !
ι rfUU.*l¿ΔΑ* Α-Δ—ώ— t ¡
40.OC
I */■ 1 l i l i !
/ / ¡ ¡ ' * ' 1 l
: /*ûK»***■-+—KUit!»·«* 1 | |
35. OC
30.00
20.00
'/ i ! i i · ! i ! _JL
« _, , Γ r — r _ _ _ _ T
15.00 // ί ! : ί ! ! !
.tí ι ι ι ι 1 1 1
OFP.H.VPRCHAR.VERT.
10.00
LABOMSM
5.00
—*__. PU
x ...-.._. P12
í í ! ! ί í '
! O
9
ι ι ! i í5a í !
,2.00 ,4.00 ¡6-00 ¡8.00 jtO.OO ¡12Ό0 ¡14.00 _T
Figure 4.25. Simulation Ρ11, Ρ12. Hor. displ. versus vert. load
ífaJSLfwj)
Figure 4.26. Simulation Ρ13, Ρ14, Ρ15. Vert, displ. versus vert, load
63
Figure 4.27. Simulation Ρ 13, Ρ14, Ρ15. Hor. displ. versus vert, load
64
Jrumllu}
LOAOVERT.DISP.
T E S T LABO OS
ώ — P16
—Q— P17
100.0
MM
Figure 4.29. Simulation Ρ16, Ρ 17. Vert, displ. versus vert, load
■"Pu» φ
LOAD-HOR.OISP.JACK
TEST LABO OS
T E S T LABO D SBIS
. * . Ρ16
_ — Ο ΡΗ
Figure 4.30. Simulation Ρ 16, Ρ 17. Hor. displ. versus vert load
65
L O A D - R O T . COI l ACK
TEST D 5
TEST D3BIS
_ .-a-Pie
. . . β - PH
Figure 4.31. Simulation Ρ 16, Ρ17. Column base rotation versus load
TEST D5
TEST D5SIS
- A - P16
_ ~ B ~ P17
Figure 4.32. Simulation Ρ 16, Ρ 17. Beam rotation at mid span versus load
66
5. Numerical simulations on hypothetic steels.
With the numerical model developed and calibrated on experimental three points bending tests,
(see §3) we have decided to simulate nine beams made with hypothetic steels. The aims of these
simulations are to analyse the effect of the yield point "fy" and the effect on the rotation
capacity of a beam of the ratio ultimate strength "fu" on the yield point "fy".
The data for the finite element simulations are given below:
67
Dimensions (mm)
h 200-15 = 185
b 200
R 18
tw 9
tf 15
Table 5.1. Geometrical properties.
Thickness of stiffeners: 15 mm
5.3. Mechanical Characteristics.
Steel fy fu εγ ερ fu/fy
(N/mm2) (N/mm2)
AA 235 235 0.00115 - 1
AB 235 258.5 0.00115 0.03073 1.1
AC 235 282 0.00115 0.03646 1.2
BA 355 355 0.00173 - 1
BB 355 390.5 0.00173 0.03366 1.1
BC 355 426 0.00173 0.04232 1.2
CA 460 460 0.00224 - 1
CB 460 506 0.00224 0.03622 1.1
CC 460 552 0.00224 0.04744 1.2
est = 2.5%
ερ < 10%
— ► ε
ey est ερ eu Ar
68
5.4. Residual stresses.
A global sinusoidal type of initial imperfection has been introduced for all the simulations.
AAtoCC (mm)
α 1.4
ß 0.7
Plane view
α sin (π x/L)
α
/
->'-
y
l
Front view
\ /
β sin (π x/L)
5.6. Discretisation.
All the beams have been discretised with the same mesh, using about 600 thin shell elements
with smaller elements near the webtoflanges junction to take fillets into account using thicker
elements in that region. (see figure 3.5)
69
where toid = thickness of the element (web or flange - see geometrical properties)
tnew = thickness of the elements of the web-toflangejunctions
a.2 = overwidth given here below (see table 5.4)
AAtoCC (mm)
a2 1.42
Supports have been simulated with the following conditions: all the nodes of the tensile flange
in external sections (x = 0 and χ = 3m) fixed in y and ζ direction and around y axes (Uy, Uz,
6y), plus one node fixed in χ direction to avoid rigid translation (see figure 3.12 Type B3).
5.8. Loading.
The beams have been loaded by a concentrated load in midwidth of the compressed flange at
mid-span. Linear constraints have been used to force the nodes situated at mid-span on the
compressed flange to have the same vertical displacement.
The figure 5.5 gives the moment - rotation curves of the 9 simulations AA to CC.
Thefigures5.6 to 5.8 show the plastic deformations (the elastic deformation is removed) at
M=Mpl after the maximum of the load (decreasing part)
The table 5.5.gives the rotation capacity R of the nines simulations.
Ρ
simulation fy fu/fy φ2-φ1
R
(N/mm2) - Φ1
AA 235 1 11.32
AB 235 1.1 15.48
AC 235 1.2 20.46
BA 355 1 7.37
BB 355 1.1 9.68
BC 355 1.2 11.84
CA 460 1 5.08
CB 460 1.1 6.94 Rotation
CC 460 1.2 7.61
70
black points are the tests results of the five tests performed by the CRM. For the tests, the yield
point fy and the ratio fu/fy are those of the flanges. (In the numerical simulations, the yield
point is the same in the flanges and in the web.)
Moment (kN.m)
400 -r
Rotation (Deg)
25 30
71
fy= 235 N/mm2 - HEB 200 - L=3 m
Simulation AA
fu/fy = 1.0
Simulation AB
fu/fy = l.l
72
fy= 355 N/mm2 - HEB 200 - L=3 m
Simulation BA
fu/fy = 1.0
Simulation BB
fu/fy = 1.1
Simulation BC
fu/fy = 1.2
73
fy= 460 N/mm2 - HEB 200 - L=3 m
Simulation CA
fu/fy = 1.0
Simulation CB
fu/fy = 1.1
Simulation CC
fu/fy = 1.2
74
ρ _ Φ2·φ1
RΛ φΐ
30 —ν-
ΙΟ - r
10 Η ^ÊÊSt TS^
iffsw >k5* ^^^^^^y^,
f ^ l Τ*^Γ>ί» -^'^
200 ^^^^^^^^Β< ^ > < Γ ^ 1.6
><
3^Ρ ^<. [ Î ^ ^ ^ C> f J ^ ^ ^> 1.4
400^ ^CB** ^><^' 1.2
¿><î' o = simulation
2
fy (N/mm ) • = test
75
6. Conclusions.
The plastic design based on a elasto-plastic behaviour of the steel is governed by the mechanical
and geometrical requirements given in Eurocode 3. These requirements are very severe
concerning the ultimate strength on yield strength ratio that must be higher than 1.2 (fu/fy ^
1.2). Nevertheless, the Eurocode 3 gives the possibility to use any steel in a plastic design if we
can demonstrate that : "At plastic hinge locations, the cross-section of the member which
contains the plastic hinge has a rotation capacity of not less than the required rotation at
the plastic hinge location.". "To satisfy the above requirement, it should be demonstrated that a
cross-section has a rotation capacity not less than the actual rotation required, "(ref 1 §5.5.3.
(2) and (3)).
A first remark concerning these requirements is the definition of the rotation capacity and the
actual rotation. In this report, we have used the nondimensional definition of the rotation
capacity (see figure 1.4). {R = (φ2 - φΐ) / φΐ}. In the Eurocode 3, the rotation capacity is an
angle, (φ2 - φΐ) or φ2. This angle must be compared with the rotations of the plastic hinges
calculated by a plastic design giving the deformations at collapse. If these rotations which are
also angles are the total rotations (elastic and plastic), they must be compared with φ2 of the
simple beam. If they represent only the plastic rotation of the cross-section, they must be
compared with (φ2 - φΐ).
The nondimensional definition of the rotation capacity allow to quantify the required rotation
capacity of a given structure whatever the steel grade. Indeed, the figure 6.1. shows that in order
to develop the plastic hinge in a cross-section, the deflection and the rotation of the beam are
proportional to the steel grade. The rotation φΐ = φρί is also proportional to the yield point.
Thus the required rotation capacity of a given beam is independent of the steel grade. A given
structure has also a given rotation capacity whatever the steel grade. It is possible to establish
different tables given the required rotation capacity of different structures.
On the figure 6.2., we have reported the nondimensional rotation capacity of the five tested
beams (see § 2.2) in function of the steel grade. We have also put the required rotation capacity
measured during the portal frame test which is equal to 2.3. As we can see, for this example of
frame, the required rotation capacity is very low. All the tested steels can be used because their
available rotation capacity is greater than 9. The steel FeE420 which have a ratio fu/fy equal to
1.16 (< 1.2) is also good for this plastic design. If the structure is more hyperstatic (see the dot
lines in figure 6.2), in order to develop all the plastic hinges, the rotations will be normally
greater, and thus the required rotation capacity will also increase. At this moment, the plastic
design could not be allowed anymore for the high strength steels.
76
f*y=2.fy=460N/mm2
M*=2M
φ*=2φ
=> R = R*
φ2φ1
R= —
φ1 HEB 200
in
JU -
r " ·—·*' ' ι
25« ^ ^ l a v a i able
20
i J. j-
15
10
5
L -L R required
2
~> 6 322 398 417 45 3 fy
Li53 1.33 1.36 1.16 U 11 fu/fy
Figure 6.2. Rotation capacity of beams (HEB 200) and portal frames.
77
A second remark concerns the requirement fu/fy ^ 1.2. The numerical simulations on
hypothetic steels (see figure 5.5) have shown that it is possible to develop a rotation
capacity of a beam in steel with a ratio fu/fy of 1.0. We could think that the plastic design
theory says that it is impossible to go over the theoretical plastic moment when the ratio is 1.0.
But the theory is based on the famous Bernoulli's assumption. "After deformation, a bending
right section stays in plane and perpendicular to the neutral axis". In reality, the stresses state in
the plastic zone is not a plane state of stresses and thus due to the Poisson's coefficient and the
combination of stresses, it is possible to have a load higher than the theoretical one based on the
Bernoulli's assumption. The stiffener brings also a contribution to the stiffness of the bending
beam which is not considering in the calculation of the plastic moment. The value 1.2 is thus
very conservative and could be decreased in the Eurocode 3 at a lower value, 1.1 for example. In
fact, this ratio is not important for a plastic design. Indeed, the rotations are directly linked to the
strain of the steel and in order to allow a plastic redistribution of the bending moments in a
structure, the steel needs only sufficient elongations before cracking whatever the fu/fy ratio.
But we must confess that without strain hardening, the required strains could become very large
in order to have a redistribution of the bending moments in a complex structure which is very
hyperstatic. It is the reason why it will be very interesting in the future to perform numerical
simulations of hyperstatic structures made in high strength steel with low ratio fu/fy (< 1.2) in
order to find the limitation of this ratio in function of their maximum elongation.
The safe global value of Rrequired (n°n dimensional rotation capacity independant of the steel
grades) for different structures will be given in different tables. Another possibility will be to
determine exactly the Rrequired in all plastic hinges with a programme for plastic design giving
the values of the rotations of all plastic hinges. The Ravailable (non dimensional rotation
capacity of the beams) will be given directly in the sales programme for the different steels.
In order to make this criterion available, the following work must be done in the next future:
1. To determine the required rotation capacity of commonly used structures, (by numerical
simulations).
2. To determine the available rotation capacity of all the sales programmes, (by tests and
numerical simulations).
3. To analyse the influence of the type of loading and of the span on the rotation capacity, (by
tests and numerical simulations).
4. To analyse the influence of the defects in combination with the material toughness, (by tests
and numerical simulations).
78
5. To analyse the influence of the welding technics on the rotation capacity, (by tests and
numerical simulations).
79
References»
1 : Eurocode 3 Part 1.1.: Design of Steel Structures, General rules and rule for buildings ENV
1993-1-1, February 1992.
2 : Comportement Elastoplastique des Constructions Métalliques. Interaction entre Résistance
et Ductilité des Aciers. Final report November 1992. E.C.S.C. research n°7210-SA/204.
3 : A.F. LUCKEY, P.F. ADAMS, "Rotation capacity of beams under moment gradiant",
Journal of the Structural Division, Vol. 95, N° ST6, Paper 6599, June 1969.
4 : Eurocode 3 Part 1.1.: Design of Steel Structures, General rules and rule for buildings ENV
1993-1-1, February 1992; Chapter 6; annex J
80
:
(( ' — — ■ ' ' —
The Communities research and development
information service
CORDIS
A vital part of your programme's
dissemination strategy
CORDIS is the information service set up under the VALUE programme to give quick and easy access
to information on European Community research programmes. It is available free-of-charge online via
the European Commission host organization (ECHO), and now also on a newly released CD-ROM.
Make sure your programme gains the maximum benefit from CORDIS
— Inform the CORDIS unit of your programme initiatives,
— contribute information regularly to CORDIS databases such as R&TD-news, R&TD-publications and
R&TD-programmes,
— use CORDIS databases, such as R&TD-partners, in the implementation of your programme,
— consult CORDIS for up-to-date information on other programmes relevant to your activities,
— inform your programme participants about CORDIS and the importance of their contribution to the
service as well as the benefits which they will derive from it,
— contribute to the evolution of CORDIS by sending your comments on the service to the CORDIS
Unit
If you are already an ECHO user, please mention your customer number.
European Commission
ISBN 92-827-6457-5
This research has shown that the requirements of Eurocode 3 for the plastic
design of a frame structure are not representative of the real behaviour of
the plastic hinge formation.
In the first part of this research, five three-point bending tests of an HEB 200
profile have been simulated with the non-linear finite elements program
Finelg. These simulations have allowed the determination of the influence
of the steel grade, the initial global and local deformations and the taking
into account of the fillets in the numerical model on the rotation capacity of
this beam. The numerical simulations have been calibrated on practical
tests performed by the CRM (Liège).
In the second part of this research, five simple frame structures have been
simulated with the non-linear finite elements program Finelg in order to
determine the required rotation capacity of a structure. The numerical
simulations have been calibrated with practical tests performed by the
CRM.
In the third part of this research, a parametrical study on hypothetic steels
has been performed in order to demonstrate that the requirements given in
Eurocode 3 are very conservative.
In conclusion, a new method is proposed to check the possibility of a plastic
design of a structure.
Venta · Salg · Verkauf · Πωλήσεις · Sales · Vente · Vendita · Verkoop · Venda · Myyntl · Försäljning
BELGIQUE/BELGIE IRELAND NORGE ISRAEL
Moniteur belge/ Government Supplies Agency NIC Info a/s Roy International
Belgisch Staatsblad 4-5 Harcourt Road Boks 6512 Etterstad 17, Shimon Hatarssi Street
Rue de Louvain 42/Leuvenseweg 42 Dublin 2 0606 Oslo P.O.B. 13056
B-1000 Bruxelles/B-1000 Brussel Tel. (1)66 13 111 Tel. (22) 57 33 34 61130 Tel Aviv
Tél. (02)512 00 26 Fax (1)47 52 760 Fax (22) 68 19 01 Tel (3) 546 14 23
Fax(02) 511 01 84 Fax (3) 546 14 42
ITALIA SCHWEIZ/SUISSE/SVIZZERA
Jean De Lannoy Sub-agent for the Palestinian Authority:
Avenue du Roi 202/Koningslaan 202 Licosa SpA OSEC
B-1060 Bruxelles/B-1060 Brussel Stampfenbachstraße 85 INDEX Information Services
Via Duca di Calabria 1/1
Tél. (02) 538 51 69 Casella postale 552 CH-8035 Zürich PO Box 19502
Fax (02) 538 08 41 1-50125 Firenze Tel. (01)365 54 49 Jerusalem
Tel.(055)64 5415 Fax (01) 365 54 11 Tel. (2)27 16 34
Autres distributeurs/ Fax 64 12 57 Fax (2) 27 12 19
Overige verkooppunten: BĂLGARIJA
Librairie européenne/ GRAND-DUCHÉ DE LUXEMBOURG Europress Klasslca BK Ltd EGYPT/
Europese boekhandel MIDDLE EAST
Messageries du livre 66, bd Vitosha
Rue de la Loi 244/Wetstraat 244 5, rue Raiffeisen BG-1463 Sofia Middle East Observer
B-1040 Bruxelles/B-1040 Brussel L-2411 Luxembourg Tel/Fax (2) 52 74 75
Tél. (02) 231 04 35 41 Sherif St.
Tél. 40 10 20 Cairo
Fax (02) 735 08 60 Fax 49 06 61 CESKA REPUBLIKA Tel/Fax (2) 393 97 32
Document delivery: NIS ČR
NEDERLAND UNITED STATES OF AMERICA/
C red oc Havelkova 22
SDU Servicecentrum Uitgeverijen CZ-130 00Praha 3 CANADA
Rue de la Montagne 34/Bergstraat 34 Tel/Fax (2) 24 22 94 33
Boîte 11 /Bus 11 Postbus 20014 UNIPUB
B-1000 Bruxelles/B-1000 Brussel 2500 EA 's-Gravenhage
Tél. (02) 511 69 41 Tel. (070) 37 89 880 HRVATSKA 4611 -F Assembly Drive
Fax (02) 513 31 95 Fax (070) 37 89 783 Lanham, MD 20706-4391
Mediatrade Tel. Toll Free (800) 274 48 88
P. Hatza 1 Fax (301) 459 00 56
ÖSTERREICH
DANMARK HR-4100 Zagreb
Manz'sche Verlags Tel/Fax (041) 43 03 92
J. H. Schultz Information A/S und Universitätsbuchhandlung CANADA
Herstedvang 10-12 Kohlmarkt 16 MAGYARORSZAG Subscriptions only
DK-2620 Albertslund A-1014Wien Uniquement abonnements
TH. 43 63 23 00 Tel. (1)531 610 EuroInfoService
Fax (Sales) 43 63 19 69 Fax (1)531 61-181 Europá Ház Renouf Publishing Co. Ltd
Fax (Management) 43 63 19 49 Margitsziget
Document delivery: 1294 Algoma Road
H-1138 Budapest Ottawa, Ontario K1B 3W8
Tel./Fax(1) 111 60 61, (1)111 62 16 Tel. (613)741 43 33
DEUTSCHLAND Wirtschaftskammer
Wiedner Hauptstraße Fax (613) 741 54 39
Bundesanzeiger Verlag A-1045 Wien POLSKA
Postfach 10 05 34 Tel. (0222) 50105-4356 Business Foundation
Fax (0222) 50206-297 AUSTRALIA
D-50445 Köln ul. Krucza 38/42
Tel.(02 21)20 29-0 PL-00-512 Warszawa Hunter Publications
Fax (02 21) 2 02 92 78 PORTUGAL Tel. (2) 621 99 93, 628 28 82 58A Gipps Street
International Fax&Phone (0-39) 12 00 77 Coliingwood
Imprensa Nacional—Casa da Moeda, EP Victoria 3066
GREECE/ΕΛΛΑΔΑ Rua Marquês Sá da Bandeira, 16-A Tel. (3)9417 53 61
P-1099 Lisboa Codex ROMANIA Fax (3 9419 71 54
G.C. Eleftheroudakis SA Tel. (01)353 03 99 Euromedia
International Bookstore Fax (01) 353 02 94/384 01 32
Nikis Street 4 65, Strada Dionisie Lupu JAPAN
GR-10563 Athens Distribuidora de Livros RO-70184 Bucuresti
Tel. (01)322 63 23 Bertrand, Ld.' Tel/Fax 1-31 29 646 Procurement Services InL (PSIJapan)
Fax 323 98 21 Grupo Bertrand, SA Kyoku Dome Postal Code 102
RUSSIA Tokyo Kojimachi Post Office
Rua das Terras dos Vales, 4-A Tel. (03) 32 34 69 21
ESPANA Apartado 37 CCEC Fax (03) 32 34 69 15
P-2700 Amadora Codex 9,60-letiya Oktyabrya Avenue
MundiPrensa Libros, SA Tel. (01)49 59 050 117312 Moscow
Fax 49 60 255 Sub-agent:
Castelló, 37 Tel/Fax (095) 135 52 27
E-28001 Madrid Kinokunlya Company Ltd
Tel. (91)431 33 99 (Libros) SUOMI/FINLAND Journal Department
SLOVAKIA
431 32 22 (Suscripciones) PO Box 55 Chitose
435 36 37 (Dirección) Akateeminen Kirjakauppa Slovak Technical Tokyo 156
Fax (91) 575 39 98 Akademiska Bokhandeln Library Tel. (03)34 39-0124
Pohjoisesplanadi 39 / Norra esplanaden 39 Nàm. slobody 19
Boletin Oficial del Estado PL /PB 128 SLO-812 23 Bratislava 1
Trafalgar, 27-29 FIN-00101 Helsinki / Helsingfors Tel. (7) 52 204 52 SOUTH and EAST ASIA
E-28071 Madrid Tel. (90) 121 4322 Fax (7) 52 957 85
Tel. (91)538 22 95 Fax (90) 121 44 35 Legal Library Services Ltd
Fax(91) 538 23 49 Orchard
CYPRUS
SVERIGE PO Box 0523
Sucursal: Cyprus Chamber of Commerce Singapore 9123
BTJAB and Industry Tel. 243 24 98
Librería Internacional AEDOS Fax 243 24 79
Traktorvägen 11 Chamber Building
Consejo de Ciento, 391 Box 200 38 Grivas Dhigenis Ave
E-08009 Barcelona S-221 00 Lund 3 Deligiorgis Street
Tel. (93) 488 34 92 Tel.(046)18 00 00 PO Box 1455 SOUTH AFRICA
Fax (93) 487 76 59 Fax (046) 18 01 25 Nicosia Safto
Tel. (2)44 95 00, 46 23 12
Librería de la Generalitat Fax (2) 36 10 44 5th Floor, Export House
de Catalunya UNITED KINGDOM Cnr Maude & West Streets
Rambla deis Estudis, 118 (Palau Moja) HMSO Books (Agency section) Sandton2146
E-08002 Barcelona MALTA Tel.(011)883-3737
Tel. (93) 302 68 35 HMSO Publications Centre Fax (011)883-6569
51 Nine Elms Lane Miller Distributors Ltd
Tel. (93) 302 64 62 London SW8 5DR
Fax(93) 302 12 99 PO Box 25
Tel.(0171)873 9090 Malta International Airport LOA 05 Malta ANDERE LÄNDER
Fax (0171) 873 8463 Tel. 66 44 88 OTHER COUNTRIES
Fax 67 67 99 AUTRES PAYS
FRANCE
ICELAND Office des publications officielles
Journal officiel TÛRKIYE
Service des publications BOKABUD des Communautés européennes
des Communautés européennes LARUSAR BLÖNDAL Pres AS 2, rue Mercier
26, rue Desalx Skólavördustig, 2 Dünya Infotel L-2985 Luxembourg
F-75727 Paris Cedex 15 IS-101 Reykjavik TR-80050 Tünel-Istanbul Tél. 29 29-1
Tél. (1)40 58 77 01/31 Tel. 551 56 50 Tel. (1)251 91 90/251 96 96 TélexPUBOFLU1324b
Fax 552 55 60 Fax (1)251 91 97 Fax 48 85 73, 48 68 17
Fax (1)40 58 77 00
NOTICE TO THE READER
All scientific and technical reports published by the European Commission are announced in the
monthly periodical 'euro abstracts'. For subscription (1 year: ECU 63) please write to the address
o
below.
Ol
σ>
IV)
m
O