sechstemann
Mechanical
Hello guys,
I'm no doing a nonlinear analysis for a wind turbine blade in Abaqus and encounter a convergence problem. My wind turbine model consists of shell and two webs. Slots are made at the back bottom part of the blade skin. Model as the following pictures showed:
conventional shell element is chosen and for the material i used simply glass fabric for the skin and left web (only one layup), the right web consists of three layups. All the three layups have the same orientation angle, the outer two layups are made of glass fabric while the middle one is made of CFRP-UD. The thickness in all parts except the right web is 10cm. The thickness of the right web varies from front to back from 10mm to 5mm and finally to 1mm. The geometric parameters for the wind turbine are taken from NREL 5MW in QBlade. The total length of the model is 61.5m.
the aerodynamic part is calculated using BEM method and then a pressure field is applied on the model in Abaqus.
My research purpose is to introduce buckling in the right web to see the buckling induced bending-twist coupling of the total structure.
Now when eigenvalue analysis and linear analysis are carried out, everything goes smoothly while by nonlinear analysis (NLGEOM ON) with initial imperfection the results cannot converge.
The error message says that system matrix has negative eigenvalues and that means the system matrix is not positive definite. For example, a bifurcation (buckling) load may have been exceeded.
I found that message helps nothing for the debugging of the problem since buckling is intentionally introduced in this structure thus it naturally exists in the analysis, that should not affect the convergence.
Totally no idea for solving this problem. Could the problem be caused by the design of the materials for example layup orientations, the thickness of the layups or any other reasons? I really appreciate that if you guys especially those who has experience in nonlinear analysis of thin walled structural model with composite materials could offer some help.
Best Regards,
Yijun
I'm no doing a nonlinear analysis for a wind turbine blade in Abaqus and encounter a convergence problem. My wind turbine model consists of shell and two webs. Slots are made at the back bottom part of the blade skin. Model as the following pictures showed:
conventional shell element is chosen and for the material i used simply glass fabric for the skin and left web (only one layup), the right web consists of three layups. All the three layups have the same orientation angle, the outer two layups are made of glass fabric while the middle one is made of CFRP-UD. The thickness in all parts except the right web is 10cm. The thickness of the right web varies from front to back from 10mm to 5mm and finally to 1mm. The geometric parameters for the wind turbine are taken from NREL 5MW in QBlade. The total length of the model is 61.5m.
the aerodynamic part is calculated using BEM method and then a pressure field is applied on the model in Abaqus.
My research purpose is to introduce buckling in the right web to see the buckling induced bending-twist coupling of the total structure.
Now when eigenvalue analysis and linear analysis are carried out, everything goes smoothly while by nonlinear analysis (NLGEOM ON) with initial imperfection the results cannot converge.
The error message says that system matrix has negative eigenvalues and that means the system matrix is not positive definite. For example, a bifurcation (buckling) load may have been exceeded.
I found that message helps nothing for the debugging of the problem since buckling is intentionally introduced in this structure thus it naturally exists in the analysis, that should not affect the convergence.
Totally no idea for solving this problem. Could the problem be caused by the design of the materials for example layup orientations, the thickness of the layups or any other reasons? I really appreciate that if you guys especially those who has experience in nonlinear analysis of thin walled structural model with composite materials could offer some help.
Best Regards,
Yijun
