Computer Aided Simulation and Optimisation of Material Forming and Shaping Processes
Simulation and Optimisation of Metal Forming Processes
Keywords: Computational Mechanics, Numerical Simulations, FEA (Finite Element Analysis), Adaptive Remeshing Concepts (ALE: Arbitrary Lagrangian Eulerian), Metal Forming Processes, Rolling, Forging, Elasto-Viscoplastic Constitutive Laws, Contact and Friction, Thermally Coupled Simulations, Rigid and Elastic Forming Tools, Process Validation, Tool Design, Manufacturing Feasibility Studies, Control Systems, Process Automation.
Introduction and Survey
In metal forming, a piece of material is plastically deformed between tools to obtain the desired product. Before the digital revolution, only few analytical tools were available to judge the manufacturing feasibility. Today, computer based metal forming simulation tools enable the validation of the tool and machine designs for production and an estimate of the final work-piece properties to be expected. The manner in which the material is worked directly determines the final product quality.
For metal production and treatment, the immensely complex and interrelated relevant process parameters require the application of highly sophisticated mathematical models, verified by actual operational experience. This is the basis and a vital precondition for producing high quality products satisfying exact tolerance demands. Mechatronics plays a dominant role in the process development and improvement process, where nearly all of its advanced methods are applied. It covers sensor-, actuator- and controller design and technology, conceptual feasibility studies, systematic designing-, manufacturing- and process automation methods.
By applying the FEA (Finite Element Analysis) method for the simulation and analysis of metal forming processes, the designer can validate and optimize his tool and machine design at an early stage before the tools (e.g. roll, blank holder, tie, punch, etc.) are manufactured to ensure that the final product meets the particular needs. A precise determination of the occurring final shapes and residual stress distributions, as well as the resultant strain distributions and forming limit curves, can be considered as key points to produce top quality products that satisfy the most demanding tolerances. The higher and higher demands from customers concerning product quality are the reason, why improved formalisms are of utmost importance to attain a better understanding of the underlying process details. By performing systematic regression methods, sets of characteristic curves are obtained. These serve as input basis for feedforward and feedback control systems and for the process automation.
We are particularly interested in the simulation of thermodynamically coupled hot and cold massive forming processes, which includes flat rolling and section rolling, die forging and hammer forging coupled with heat treatment. Besides, sheet metal forming processes, where the thickness of the material is small compared to other dimensions, are of interest as well. FEA simulations are performed both for axially symmetric, two-dimensional plane strain and plane stress scenarios, and for fully three-dimensional problems. For many problems under consideration, such as hot rolling and hot forging, the processes are thermally activated from a metallurgical point of view, requiring control within definite temperature bands and cooling rates in order to hit target properties.
In most forming processes under consideration, large material deformations and contact phenomena occur. Updated Lagrangian formulations are well suited for problems regarding path dependent material properties and free surfaces, but lead to problems with evolving contacts and suffer from numerical problems, when the mesh is distorted heavily. Eulerian formulations are able to cope with large material deformations, but are less suited for the description of free surfaces. Therefore, sophisticated mesh update algorithms like ALE (Arbitrary Lagrangian Eulerian) have become very popular. Moreover, intermediate Eulerian Lagrangian hybrid methods were developed, which combine the advantages of the Eulerian field representation by elimination of the explicit time dependence for steady-state processes, and the individual Lagrangian description with a well known reference configuration. It requires a systematic transformation (non-linear mapping) of the distorted geometry, the stress-, strain- and strain-rate tensors, the balance equations and variational principles of continuum mechanics into this fictitious intermediate reference system.
To obtain praxis-relevant results, the FE - models have to be calibrated and compared with measured plant data and pass schedule values. Many process- and modelling parameters, such as frictional and heat transfer coefficients and yield stress parameters, are not known exactly in most cases. Therefore, systematic parameter identifications have to be performed, leading to inverse problems. Such identification procedures often turn out to be “ill-posed problems”, which is due to the severe non-linearity of the models under consideration. The non-linearity results from large plastic deformations, the frictional contact between the work-piece and tools, and also from the non-linear elasto-viscoplastic path dependent constitutive laws, where the yield stress curve depends on the strain, strain rate and temperature. Besides, for many practical cases of interest (e.g. profile and flatness analysis of the strip for flat rolling processes), the tools cannot be modelled as rigid bodies, but elastic deformations and thermal crown effects have to be taken into account to obtain praxis relevant predictions.
In order to accomplish the development of customized metal forming simulation packages, commercial finite element simulation software packages such as Abaqus Standard and Explicit, Deform, Femlab etc. are utilized and supplemented by self-developed user subroutines (written in Fortran 90, C++ and script languages like Python, Tcl) for advanced pre- and post-processing purposes. This enables one to perform systematic parameter studies, to obtain sets of characteristic curves and to automatise the simulation and evaluation of coupled forging or rolling passes.
Simulation and Optimisation of Non-Metal Forming Processes
Keywords: Computational Mechanics, Thermoplastics, Numerical Simulations, FEA (Finite Element Analysis), Material Flow, Residual Stresses, Shrinkage, Distortion, Heat Transfer, Conductivity, Cooling, Phase Transition, Solidification.
Introduction and Survey
Beside metal forming processes, the treatment of forming and shaping processes of thermoplastics and other synthetics becomes more and more important from a designing and manufacturing point of view. The simulation and analysis of such material forming and shaping processes often uses the same physical basic investigations and requires similar tools, which are established and proven for metal forming simulations. Therefore, existing models and experiences from such manufacturing processes (e.g. metal die forging, bending, rolling, etc.) can be exploited for non-metal processes leading to valuable synergy effects.
Our main objective of the investigation of non-metal forming processes is the improvement of existing and development of new and highly sophisticated mathematical models, theories and optimisation tools in order to improve the insight into such processes. By performing systematic process parameter studies, parameter identification procedures and regression analysis methods, sets of characteristic curves can be obtained, which serve as input for the improvement of the process and for control and automation purposes. Of considerable practical importance are reliable predictions for convection and heat transfer coefficients, e.g. for the treatment of mould cooling processes.
Mag. DI Dr. Alexander Kainz