Cedric Xia

Sequential Optimization and Reliability Assessment Method for Metal Forming Processes

Uncertainty is inevitable in any design process. The uncertainty could be due to the variations in geometry of the part, material properties or due to the lack of knowledge about the phenomena being modeled itself. Deterministic design optimization does not take uncertainty into account and worst case scenario assumptions lead to vastly over conservative design. Probabilistic design, such as reliability-b ased design and robust design, offers tools for making robust and reliable decisions under the presence of uncertainty in the design process. Probabilistic design optimization often involves double-loop procedure for optimization and iterative probabilistic assessment. This results in high computational demand. The high computational demand can be reduced by replacing computationally intensive simulation models with less costly surrogate models and by employing Sequential Optimization and reliability assessment (SORA) method. The SORA method uses a single-loop strategy with a series of cycles of deterministic optimization and reliability assessment. The deterministic optimization and reliability assessment is decoupled in each cycle. This leads to quick improvement of design from one cycle to other and increase in computational efficiency. This paper demonstrates the effectiveness of Sequential Optimization and Reliability Assessment (SORA) method when applied to designing a sheet metal flanging process. Surrogate models are used as less costly approximations to the computationally expensive Finite Element simulations.

A New Method to Calculate Threshold Values of Ductile Fracture Criteria for Advanced High-Strength Sheet Blanking

A new approach is presented in this paper to calculate the critical threshold value of fracture initiation. It is based on the experimental data for forming limit curves and fracture forming limit curves. The deformation path for finally a fractured material point is assumed as two-stage proportional loading: biaxial loading from the beginning to the onset of incipient necking, followed plane strain deformation within the incipient neck until the final fracture. The fracture threshold value is determined by analytical integration and validated by numerical simulation. Four phenomenological models for ductile fracture are selected in this study, i.e., Brozzo, McClintock, Rice-Tracey, and Oyane models. The threshold value for each model is obtained through best-fitting of experimental data. The results are compared with each other and test data. These fracture criteria are implemented in ABAQUS/EXPLICIT through user subroutine VUMAT to simulate the blanking process of advanced high-strength steels. The simulated fracture surfaces are examined to determine the initiation of ductile fracture during the process, and compared with experimental results for DP780 sheet steel blanking. The comparisons between FE simulated results coupled with different fracture models and experimental one show good agreements on punching edge quality. The study demonstrates that the proposed approach to calculate threshold values of fracture models is efficient and reliable. The results also suggest that the McClintock and Oyane fracture models are more accurate than the Rice-Tracey or Brozzo models in predicting load-stroke curves. However, the predicted blanking edge quality does not have appreciable differences.

Experimental investigations on wear resistance characteristics of different die materials for advanced high-strength steel blanking in close section

Advanced high-strength steels are being used widely in automotive industry to achieve lightweight construction and fuel efficiency. During advanced high-strength steel blanking, the die components have to sustain higher pressure, and this may result in some problems such as chipping, cracking, galling and severe die wear, which has a great effect on the die life. This study aims to investigate the wear resistance performances of five die materials (carburized 4Cr13, conventional SKD11, Caldie, A2 and tungsten steel D60) using a featured blanking die setup combining five different die materials. The surface topography and microstructure of punching die materials are measured by optical profilometer and microscope. Based on the measured results, the specified wear rate and worn profiles of die inserts are obtained and compared. It is demonstrated that Caldie and A2 have higher wear resistance performance than SKD11 and 4Cr13. Moreover, the experimental results display that the most severe wear locations occur at the interfaces of straight lines and circular arcs.

Effect of micro-manufacturing processes on surface behavior

Surface texturing has become a valuable technique for reducing friction and wear in contacting parts; laser surface texturing is one such method used to create micro-dimples on the interface surface. This work investigates the surface material property variation caused by laser surface texturing. The hardness and modulus of elasticity of a steel laser surface texture sample were evaluated near the dimples and away from the dimpled zone through nano-indentation. Resulting data shows that no significant difference exists between the material properties from the two positions. An alternate technique for surface texture generation was also explored, involving the use of micro-punches to create surface features in a metal sample. Computational simulations were performed using a second material underneath a thin copper sheet. The second material was present to serve as a support and to allow extensive deformation of the top material. The choice of the support material and ratio of material thicknesses was optimized to minimize pile up. Trials were conducted for three base supporting materials: PTFE, PMMA, and aluminum. Results show that PMMA performed better than the other materials. Positive deflection was minimized when the PMMA thickness was at least fifteen times that of the copper sheet. Physical experiments were completed with a thin copper sheet to verify the results. An array of micro-indentations was also created in a bulk steel sample. In order to assess the effect of dimpling via micro-forming, nano-indentation was performed near and far from the deformed material of the dimples. Similar to the laser textured sample, no significant differences were found between the two locations.Copyright

Background and Tryout Report for BM2: Underbody Cross Member

An automotive underbody cross member was selected for one of the NUMISHEET’05 industrial benchmark to assess springback prediction capability of engineers around the world using various software. Binder and addendum were generated according to production intent process. Iterative design and draw simulation were performed on the part and addendum geometry to remove wrinkles and splits. Castings were poured and machined. Six different types of materials ranging from A15182‐O to DP965 were used in the production of the benchmark panels and three of these materials were included in the official benchmark data release. Draw panels were trimmed on a trimming fixture using laser and scanned with a whitelight optical device. Springback shapes at selected cross sections were recovered on the scanned data and original CAD data. In addition, major/minor and thickness strains were also measured at these sections.

Forming limits of anisotropic sheets with non-power-law hardening

For aluminum alloys and some advanced high-strength steels, the tensile flow curve exhibits a tendency to saturate. The suitability of constitutive equations was analyzed for aluminum 6111-T4 and a general non-power-law hardening model was adopted in the derivation of forming limits incorporated material anisotropy with varying R-values. A bifurcation analysis was conducted for the left-hand-side FLD under the assumption of proportional loading and zero-extension necking orientation. Analytical results showed good correlation with Nakajima forming limit test data for aluminum sheets following Voce hardening law.

Benchmark 2 - Springback of a draw / re-draw panel: Part C: Benchmark analysis

Benchmark analysis is summarized for DP600 and AA 5182-O. Nine simulation results submitted for this benchmark study are compared to the physical measurement results. The details on the codes, friction parameters, mesh technology, CPU, and material models are also summarized at the end of this report with the participant information details.

Elasto-Plastic Rough Surface Contact Analysis for the Effects of Material Properties, Topographical Characteristics and Load

Interaction of nominally flat engineering surfaces that lead to a large apparent contact area exists in many mechanical systems. Such an interaction can be modeled based on the periodic similarity of the surface topography and a numerical three dimensional elasto-plastic contact model [1] with the assistance of the continuous convolution and Fourier transform (CC-FT) algorithm. In this model, frequency response functions (FRF) are built to link the pressure excitation with the deformation response of materials. This model takes into account asperity interactions and work hardening caused by plastic strain. Contact pressure, real contact area, surface clearance, and resultant subsurface plastic strain and stress fields can be calculated. Following the two-dimensional digital filter technique developed by Hu and Tonder [2], rough surfaces can be generated numerically with specified autocorrelation length ratio λ *=β y /β x RMS roughness R q , skewness Sk, and kurtosis K (Fig. 1). A group of contact simulations has been performed for surfaces with various statistical characteristics, material properties, and loads. The behaviors of nominally flat contact, such as the contact area ratio, Λ=A c /A n , the average gap, \( \Gamma = \bar h/R_q \) and the volume containing plastic deformation, Ω=V p /A n R q , are obtained as functions of the average contact pressure (Fig. 2). The effects of topographic, material, and strain hardening parameters on the contact behaviors of rough surfaces are discussed in detail in this paper. Open image in new window Figure 1 A rough surface generated by computer Open image in new window Figure 2 Contact behaviors versus average pressure

Elasto-Plastic Contact Behaviors of Nominally Flat Surfaces: Modeling and Parametric Study

Interactions of nominally flat surfaces can be modeled based on the periodic similarity of surface topography and a numerical three-dimensional elasto-plastic contact model with the assistance of the continuous convolution and Fourier transform (CC-FT) algorithm. The rough surfaces were generated by a digital filtration technology with a wide range of topographical parameters. A group of contact simulations were conducted to investigate the effects of surface geometrical characteristics (including RMS roughness, correlation length ratio, skewness and kurtosis), material properties, and load on the elasto-plastic contact performance of materials.Copyright