Initial stress stiffness matrix
Initial stress stiffness matrix: Theoretical Foundations
Initial Stress Stiffness Matrix
Professor, are the initial stress stiffness matrix and the geometric stiffness matrix the same thing?
They are different names for the same concept. $[K_\sigma]$ is also called "initial stress stiffness," "geometric stiffness," or "stress stiffness." It is used in buckling analysis and vibration analysis.
Formulation
$[K_\sigma]$ is formulated from the current stress state $\{\sigma\}$. For beam elements, the axial force $N$ is used; for shell/solid elements, membrane stresses $N_{xx}, N_{yy}, N_{xy}$ are used.
$[G]$ is the displacement gradient matrix, $[S]$ is the stress matrix.
Summary
Euler Buckling and the Discovery of the Initial Stress Stiffness Matrix
Formulating buckling under compressive load as a linear algebra problem was the achievement of Leonhard Euler in 1744. The concept of "critical load" he derived actually corresponds to the modern expression "the initial stress stiffness matrix loses positive definiteness (eigenvalue becomes zero)." In the 1960s, Turner et al. introduced the initial stress stiffness matrix into FEM, making buckling analysis of arbitrary shapes possible for the first time.
Computational Methods for Initial stress stiffness matrix
Calculation of $[K_\sigma]$
Calculation in FEM:
1. Obtain stress distribution from static analysis (or nonlinear analysis)
2. Calculate $[K_\sigma]_e$ from the stress of each element
3. Assemble into global $[K_\sigma]$
Automatically calculated in all solvers (buckling analysis, prestress modal). Also used internally in NLGEOM nonlinear analysis.
Summary
How to Assemble the Initial Stress Stiffness Matrix
The initial stress stiffness matrix Kg is calculated by the integral ∫[G]ᵀ[σ][G]dV of the stress σij within the element and the shape function derivative matrix. When compressive stress dominates, Kg has negative definite components, and the load where the determinant of (K+Kg) becomes 0 is the linear buckling load Pcr. FEM linear buckling analysis reduces to solving the generalized eigenvalue problem (K+λKg)φ=0, where the minimum eigenvalue λmin represents the safety factor (safe if λmin>1).
Initial stress stiffness matrix in Practice
Practical Checklist
Geometric Nonlinear Design of Steel Frame Structures
AISC 360-10 recommends using the "Direct Analysis Method (DAM)" that considers initial stress for the design of high-rise steel buildings. DAM, which automatically accounts for the reduction in horizontal stiffness of columns under compressive load (P-δ effect) using the initial stress stiffness matrix, is 3-5% more conservative but more accurate than the "Effective Length Method" which uses a conversion for a virtual slenderness ratio of 0.2. For buildings over 400m, DAM is used as the standard analysis method.
Initial stress stiffness matrix: Software & Solver Comparison
Tools
Supported by all solvers. Concept is common but names differ.
- Nastran: Stress stiffness matrix (KGGT)
- Abaqus: Geometric stiffness (inside NLGEOM)
- Ansys: Stress stiffness (PSTRES)
SAP2000 Direct Analysis Method Settings
In CSI SAP2000's Direct Analysis Method (DAM) settings, it can automatically apply the initial stress stiffness matrix and column stiffness reduction factor (0.8×EI) to execute nonlinear analysis considering P-δ and P-Δ effects. As an AISC 360 certified software, its adoption rate is high among US structural design firms, with standard DAM settings established especially for wind/earthquake response analysis of super high-rise buildings.
Advanced Technology
Advanced Research
Initial Stress Stiffness: Prestressed Concrete Design
The concept of initial stress stiffness was put into practical use by prestressed concrete engineer Freyssinet in bridge design in the 1940s. The principle where tensile prestress in PC steel wires (typically 1,200~1,400 MPa) gives initial compressive stress stiffness to concrete, doubling load-bearing capacity, is inherited in modern CAE analysis as the `PRESTRESS` option in each solver.
Initial stress stiffness matrix: Common Issues & Debugging
Troubles
Unreasonable Results Where Linear Buckling Eigenvalue is Less Than 1
Even if linear buckling analysis yields a result with eigenvalue λ<1 (already buckling at design load), buckling may not actually occur in reality. Since linear buckling assumes a "perfect shape" without initial imperfections, it becomes non-conservative for structures with large geometric stiffening effects like membrane pressure vessels. For shells, it is recommended to confirm with nonlinear buckling analysis that adds initial imperfections (geometric errors of 0.1~1 times the plate thickness).
Related Topics
Experience the theory firsthand with the interactive simulator for this field
All Simulators