Hello,
I have recently attempted to demonstrate protection against plastic collapse using the Limit Load Analysis (LLA) method (Section 5.2.3) from ASME BPVC Section VIII, Division 2, Part 5 (2023 Edition).
My understanding of the LLA method is that an elastic-perfectly plastic material is modeled and subjected to the load case combinations in Table 5.4.
Since the material model relieves any post-yield stress as plastic strain, I presumed a strain limit should be stipulated. (For example, I have seen a 5% plastic strain limit used for similar elastic-perfectly plastic capacity designs in accordance with EU codes). However, Section 5.2.3.1(b) states that the displacements and strains from a limit analysis have no physical meaning. I understand the rationale—the limit load assessment is not intended to be a literal prediction but rather a conservative method. From this, I gather that convergence of the LLA solution is the sole criterion for acceptance, while non-convergence indicates that the plastic collapse load has been exceeded.
My LLA model has successfully converged. However, upon reviewing the subsequent section, 5.3 (Protection Against Local Failure), it appears that only two methods are permitted for this check: either 5.3.2 (Elastic Analysis - Triaxial Stress Limit) or 5.3.3 (Elastic-Plastic Analysis - Local Strain Limit).
I am a little puzzled by this. I initially chose the LLA method because I found stress linearization (required for an elastic analysis) too ambiguous for my model. Now, it seems my only option is to proceed with the Elastic-Plastic Analysis for local failure (5.3.3) and I am worried that considering non-linear geometry effects will require significantly more computational time. This makes my decision to perform the LLA (5.2.3) seem redundant, as I now need an analysis that complies with the Elastic-Plastic method (5.2.4) to satisfy 5.3.3.
In hindsight, it seems I would've saved more time carrying out the 5.2.4 method to begin with. Am I missing something?
On a side note, I am curious what metrics other designers use for mesh convergence studies of LLA or Elastic-Plastic Analysis (EPA) models. Currently, I am visually comparing averaged versus unaveraged equivalent plastic strain plots, and I am also considering comparing the percentage difference in key results across different mesh densities.
Any insight would be greatly appreciated.
I have recently attempted to demonstrate protection against plastic collapse using the Limit Load Analysis (LLA) method (Section 5.2.3) from ASME BPVC Section VIII, Division 2, Part 5 (2023 Edition).
My understanding of the LLA method is that an elastic-perfectly plastic material is modeled and subjected to the load case combinations in Table 5.4.
Since the material model relieves any post-yield stress as plastic strain, I presumed a strain limit should be stipulated. (For example, I have seen a 5% plastic strain limit used for similar elastic-perfectly plastic capacity designs in accordance with EU codes). However, Section 5.2.3.1(b) states that the displacements and strains from a limit analysis have no physical meaning. I understand the rationale—the limit load assessment is not intended to be a literal prediction but rather a conservative method. From this, I gather that convergence of the LLA solution is the sole criterion for acceptance, while non-convergence indicates that the plastic collapse load has been exceeded.
My LLA model has successfully converged. However, upon reviewing the subsequent section, 5.3 (Protection Against Local Failure), it appears that only two methods are permitted for this check: either 5.3.2 (Elastic Analysis - Triaxial Stress Limit) or 5.3.3 (Elastic-Plastic Analysis - Local Strain Limit).
I am a little puzzled by this. I initially chose the LLA method because I found stress linearization (required for an elastic analysis) too ambiguous for my model. Now, it seems my only option is to proceed with the Elastic-Plastic Analysis for local failure (5.3.3) and I am worried that considering non-linear geometry effects will require significantly more computational time. This makes my decision to perform the LLA (5.2.3) seem redundant, as I now need an analysis that complies with the Elastic-Plastic method (5.2.4) to satisfy 5.3.3.
In hindsight, it seems I would've saved more time carrying out the 5.2.4 method to begin with. Am I missing something?
On a side note, I am curious what metrics other designers use for mesh convergence studies of LLA or Elastic-Plastic Analysis (EPA) models. Currently, I am visually comparing averaged versus unaveraged equivalent plastic strain plots, and I am also considering comparing the percentage difference in key results across different mesh densities.
Any insight would be greatly appreciated.
