Statistical evaluation of a new resistance model for cold-formed stainless steel cross-sections subjected to web crippling

Abstract

This paper presents a statistical evaluation according to Annex D of EN 1990(2002 of a new resistance function for web crippling design of cold-formed stainless steel crosssections. This resistance function was derived in Bock et al. (2013) through the use of carefully validated numerical models with the aim to propose a design expression for stainless steel sections, which are currently designed following the provisions for coldformed carbon steel sections given in EN 1993EN -1-3 (2006. Although it was shown that the proposed design equation is appropriate for application to various stainless steels, the statistical uncertainties in material properties that the different types of stainless steels exhibit require an assessment of various partial safety factors. The statistical assessment showed that the proposed resistance function by Bock et al. (2013) requires adjustment to satisfy the safety level set out in EN 1993-1-4 (2006); A recalibration is performed herein. The web crippling design provisions given in EN 1993EN -1-3 (2006 and SEI/ASCE 8-02 (2002) American standard for application to stainless steel are also statistically evaluated herein. Comparison with test and numerical data showed that the predictions of the recalibrated resistance function are better suited and consistent than existing design provisions.

Design of prestressed cold-formed steel beams

“…As performed in similar studies [34,35], to account for the fact that the reliability analysis is based on numerical results, an additional term C FEM V 2 FEM was included in the determination of the combined coefficient of variation of the required resistance and thus within the square root term in Eq. (17).…”

Strength curves for web crippling design of cold-formed stainless steel hat sections

“…(22) For the statistical evaluation of the proposed design approach (resistance model), the database was split into two sub-sets based on their material grade to consider the difference in over-strength ratio (measured/minimum specified strength) following recommendations of [33]. Details of the procedure to statistically validate a resistance model are given in [34]. A summary of key statistical parameters is presented in Table 8 where n is the population of the data under consideration, b is the mean value of numerical data to predicted resistance ratio, V δ is coefficient of variation of the numerical data relative to the resistance model (error of the model) and V r is combined coefficient of variation making allowance for the error of the model V δ , including the basic variables V xi and the FE model V FEM [35].…”

Standardised material properties for numerical parametric studies of stainless steel structures and buckling curves for tubular columns

“…For the variability of the geometric properties, a COV value of 0.05 was used [10]. In order to add artificial variability to the numerical results, an additional variability term with coefficient of variation VFEM set to 0.03, which was determined by considering the deviation between the experimental and numerical results, as used in similar studies in [48], was also incorporated in the reliability analysis. For the purpose of the reliability analyses performed herein, the design resistance equations for flexural buckling resistance set out in Clause 5.4.2 of EN 1993-1-4, as given by Eq.…”