Abstract
The Chaboche model is a well-validated non-linear kinematic hardening material model. This material model, like many models, depends on a set of material constants that must be calibrated for it to match the experimental data. Due to the challenge of calibrating these constants, the Chaboche model is often disregarded. The challenge with calibrating the Chaboche constants is that the most reliable method for doing the calibration is a brute force approach, which tests thousands of combinations of constants. Different sampling techniques and optimization schemes can be used to select different combinations of these constants, but ultimately, they all rely on iteratively selecting values and running simulations for each selected set. In the experience of the authors, such brute force methods require roughly 2,500 combinations to be evaluated in order to have confidence that a reasonable solution is found. This process is not efficient. It is time-intensive and labor-intensive. It requires long simulation times, and it requires significant effort to develop the accompanying scripts and algorithms that are used to iterate through combinations of constants and to calculate agreement. A better, more automated method exists for calibrating the Chaboche material constants.
In this paper, the authors describe a more efficient, automated method for calibrating Chaboche constants. The method is validated by using it to calibrate Chaboche constants for an IN792 single-crystal material and a CM247 directionally-solidified material. The calibration results using the automated approach were compared to calibration results obtained using a brute force approach. It was determined that the automated method achieves agreeable results that are equivalent to, or supersede, results obtained using the conventional brute force method.
After validating the method for cases that only consider a single material orientation, the automated method was extended to multiple off-axis calibrations. The Chaboche model that is available in commercial software, such as ANSYS, will only accept a single set of Chaboche constants for a given temperature. There is no published method for calibrating Chaboche constants that considers multiple material orientations. Therefore, the approach outlined in this paper was extended to include multiple material orientations in a single calibration scheme. The authors concluded that the automated approach can be used to successfully, accurately, and efficiently calibrate multiple material directions. The approach is especially well-suited when off-axis calibration must be considered concomitantly with longitudinal calibration. Overall, the automated Chaboche calibration method yielded results that agreed well with experimental data. Thus, the method can be used with confidence to efficiently and accurately calibrate the Chaboche non-linear kinematic hardening material model.
LHC physicists embrace brute-force approach to particle hunt
