Intelligente Prozessführung beim Presshärten

Within the project, a concept for a feedback controlled press hardening process was developed and implemented. The information for process optimization were taken from a database, the so-called knowledge base, which is a mathematical model of the press hardening process of the demonstrator geometry side sill. By entering target values, for example target work piece hardness, the knowledge base calculates under consideration of the current process conditions one or multiple control variables. Changing the process conditions due to the control variables leads to a variation of the work piece properties (e.g. open the tool when a certain calculated work piece temperature is reached). In work package 1 the significant parameters for the press hardening process were deter-mined. Furthermore, the flow curves of the investigated material 22MnB5 were determined with a dilatometer at different temperatures and compared with the state of the art. In the project, a sensitivity analysis of the process parameters was performed using FEM. The programs AutoForm and LS Dyna were used. Finally, a comparison of the FE simulations and the experiments was carried out. The results of the numerical sensitivity analysis were used for the design of the experiments. The variation of the process parameters was carried out in a small area, such as may occur in series production. To carry out the experimental tests, an automated press hardening set up was built, which was controlled by a PLC. The work pieces were heated in a chamber furnace and inserted with a robotic gripper system in a press hardening tool, which was located in a servo press. After the press hardening the work piece was removed out of the tool and dropped off, after which a new cycle could start. To capture the relevant process parameters the test set up was equipped with various sensors for measuring the work piece and the tool temperature, the forces and speeds. All sensors were synchronized via the PLC. The evaluation of the work piece quality was carried out at defined points regarding the hardness, the thinning and the springback. The hardness was measured with conventional Vickers hardness devices. The geometric data were measured with a GOM ATOS system. Taking all this data into account, a meta-model of the press hardening process was created, which is the basis for the knowledge base. Simplified, it is a table which establishes the correct relationships between input variables (e.g. work piece temperature at the beginning of the forming) and output variables (e. g. thinning in defined point X). 3 models have been developed for the planning of a process (offline use of the knowledge base) and the feedback control during a process (online use the knowledge base). The knowledge base with its control algorithms has been connected with the PLC and the sensor systems of the test set up. For testing the knowledge base, quality criteria of the aimed work piece were randomly entered into the knowledge base. During the press hardening process, the knowledge base processed the available process data and calculated the necessary value of a control variable, which was passed to the PLC. The subsequent analysis of the components indicated that the aimed component properties could be adjusted. Finally, in an economic evaluation the production costs of a work piece, produced in a knowledge based optimized press hardening process, were calculated