Abstract Scope |
Every prediction model of any physical phenomenon is judged according to its accuracy. This is a reasonable practice, however, a race towards perfect prediction is never reasonable and one important aspect of modeling is very often overlooked. This issue is the variability of a considered phenomenon. In this work, the variability of large radius air bending is discussed in detail. Since large radius bending is applied mostly to high-strength steels, Strenx 700 MC and Strenx 1300 have been considered. Based on previously performed modeling work, the thorough analysis of several prediction methods has been done. The following common modeling techniques have been analyzed: regression prediction, finite-element method and analytical approach. For these methods, the modeling limits have been estimated and the trade-off between effort and accuracy has been discussed. All prediction results are compared to demanding requests from industry based on the experimental campaign of more than 150 tests. |