Inmetal forming simulation, the forming of sheet metal is simulated on the computer with the help of special software. Simulation makes it possible to detect errors and problems, such as wrinkles or splits in parts, on the computer at an early stage in forming. In this way, it is not necessary to produce real tools to run practical tests. Forming simulation has become established in the automotive industry since it is used to develop and optimize every sheet metal part.
To illustrate the metal forming process, there must be a model of the real process. This is calculated in the software using the finite element method based on implicit or explicit incremental techniques. The parameters of the model must describe the real process as accurately as possible so that the results of the simulation are realistic.
The typical parameters for forming simulation are, for example, part and tool geometry, material properties, press forces and friction. The simulation calculates stresses and strains during the forming process. In addition, simulations allow for the recognition of errors and problems (e.g. wrinkles or splits) as well as results (e.g. strength and material thinning). Even springback, the elastic behavior of material after forming, can be predicted in advance. Forming simulation also provides valuable information about the influence of process variations on stamping robustness.
Forming simulations are used throughout the entire process chain of sheet metal forming. The simulation allows a part designer to estimate the formability of a sheet metal part already during the design phase, which results in the design of a part which is easy to produce. A process engineer can already assess the process during the planning phase and optimize various alternatives using the simulation, which can subsequently reduce the fine tuning of a forming tool. Finally, regarding the fine tuning of a forming tool, simulation can provide useful information on how an existing, not yet fully functioning tool must be adjusted. It is also possible to see how the process parameters must be adjusted in order to guarantee optimal drawing results.
Metal forming simulation enables the fast review of several alternative concepts for quality and cost improvements, which results in huge cost and time savings. Furthermore, simulating the forming process improves development and planning reliability. The number of tool tryouts is reduced and tryout time is shortened. Metal forming simulation leads to the highest quality in part and tool design as well as maximum reliability in production.
Inspire Form is a complete stamping simulation environment that can effectively be used by product designers and process engineers to optimize designs, simulate robust manufacturing and reduce material costs.
With the fast and easy feasibility module, users can analyze parts in seconds to predict formability early in the product development cycle. The automated blank nesting proposes an efficient layout of the flattened blank on the sheet coil to maximize material utilization.
The tryout module includes a highly scalable incremental solver, helping users to iterate and simulate multi-stage forming, trimming and springback in a modern and intuitive user interface, reducing complexity and making the production of high quality parts more economical.
Inspire Form enables users to quickly and reliably check the formability of apart early in the product design cycle. With Inspire Form, users can visualize potential defects such as splits or wrinkles, and then modify to eliminate defects and improve overall design.
Inspire Form has a simple and highly intuitive workflow that is easy to learn and apply. Standard training sessions last only 4-6 hours, although most users can learn Inspire Form applications with no formal training at all.
Today the metal forming industry is making increasing use of simulation to evaluate the performing of dies, processes and blanks prior to building try-out tooling. Finite element analysis (FEA) is the most common method of simulating sheet metal forming operations to determine whether a proposed design will produce parts free of defects such as fracture or wrinkling.[1]
The most painful and most frequent defects are wrinkles, thinning, springback and splits or cracks. Few methods are being used around the industry to cope with the main defects, based on the experience of the technicians. However, the correct process is the most vital, since it involves the correct geometry followed by number of steps to reach at final geometry. Which demands for specific experience or higher number of iterations.[2]
Deformation of the blank is typically limited by buckling, wrinkling, tearing, and other negative characteristics which makes it impossible to meet quality requirements or makes it necessary to run at a slower than desirable rate.
Wrinkling in a draw are series of ridges form radially in the drawn wall due to compressive buckling. Practically these are duo to low blank holder pressure due to which material slips and wrinkles formed. The optimum blank holding pressure is the key, however in certain cases it doesn't work. Then draw beads are the solutions, the location and shape of draw bead is the challenge, which can be analysed with FEA during design stage prior to tool manufacturing.[2]
Crack in the vertical wall due to high tensile stresses, some small radius block the material flow and results in excessive thinning at that point usually more than 40% of the sheet thk. result in cracks. In some cases it may happen due to excessive blank holder pressure, which restrict the metal flow. Somewhere it might be due to wrong process design, like try to make a more deep draws in a single stage, which otherwise feasible only in two stages.[2]
Thinning is a Excessive Stretching in the vertical wall due to high tensile stresses cause thickness reduction specifically on the small radius in the metal parts, however up to 20% thinning is allowed due to process limitations.[2]
Springback is a particularly critical aspect of sheet metal forming. Even relatively small amounts of springback in structures that are formed to a significant depth may cause the blank to distort to the point that tolerances cannot be held. New materials such as high strength steel, aluminum and magnesium are particularly prone to springback.[3]
The traditional approach to designing the punch and die to produce parts successfully is to build try-out tools to check the ability of a certain tool design to produce parts of the required quality. Try-out tools are typically made of less expensive materials to reduce try-out costs yet this method is still costly and time-consuming.[4]
The first effort at simulating metalforming was made using the finite difference method in the 1960s to better understand the deep drawing process. Simulation accuracy was later increased by applying nonlinear finite element analysis in the 1980s but computing time was too long at this time to apply simulation to industrial problems.[citation needed]
Rapid improvements over the past few decades in computer hardware have made the finite element analysis method practical for resolving real-world metal forming problems. A new class of FEA codes based on explicit time integration was developed that reduced computational time and memory requirements. The dynamic explicit FEA approach uses a central different explicit scheme to integrate the equations of motion. This approach uses lumped mass matrices and a typical time step on order of millionths of seconds. The method has proved to be robust and efficient for typical industrial problems.[citation needed]
As computer hardware and operating systems have evolved, memory limitations that prevented the practical use of Implicit Finite Element Methods had been overcome.[5] Using the implicit method time steps are computed based on the predicted amount of deformation occurring at a given moment in the simulation, thus preventing unnecessary computational inefficiency caused by computing too small time steps when nothing is happening or too large a time step when high amounts of deformation are occurring.
Inverse One-step methods compute the deformation potential of a finished part geometry to the flattened blank. Mesh initially with the shape and material characteristics of the finished geometry is deformed to the flat pattern blank. The strain computed in this inverse forming operation is then inverted to predict the deformation potential of the flat blank being deformed into the final part shape. All the deformation is assumed to happen in one increment or step and is the inverse of the process which the simulation is meant to represent, thus the name Inverse One-Step.
Incremental Analysis methods start with the mesh of the flat blank and simulate the deformation of the blank inside of tools modeled to represent a proposed manufacturing process. This incremental forming is computed "forward" from initial shape to final, and is calculated over a number of time increments for start to finish. The time increments can be either explicitly or implicitly defined depending on the finite element software being applied. As the incremental methods include the model of the tooling and allow for the definition of boundary conditions which more fully replicate the manufacturing proposal, incremental methods are more commonly used for process validation. Inverse One-step with its lack of tooling and therefore poor representation of process is limited to geometry based feasibility checks.[6]
Incremental analysis has filled the role previously completed through the use of proof tools or prototype tools. Proof tools in the past were short run dies made of softer than normal material, which were used to plan and test the metal forming operations. This process was very time consuming and did not always yield beneficial results, as the soft tools were very different in their behavior than the longer running production tools. Lessons learned on the soft tools did not transfer to the hard tool designs. Simulation has for the most part displaced this old method. Simulation used as a virtual tryout is a metal forming simulation based on a specific set of input variables, sometimes nominal, best case, worst case, etc. However, any simulation is only as good as the data used to generate the predictions. When a simulation is seen as a "passing result" manufacturing of the tool will often begin in earnest. But if the simulation results are based on an unrealistic set of production inputs then its value as an engineering tool is suspect.
3a8082e126