Non-deterministic Calibration
179
visco-plastic self-consistent (VPSC) model [23] can provide accurate results by
accounting for the effect of grain shape while maintaining a tractable computational
effort. Additionally, while not exercised in this chapter, it is important to note that
stress-strain curves could be extracted from a variety of directions with respect to
the bulk material texture. This additional data would serve to improve the global
calibration approach.
4.3 Global-Local Methods
Like purely global methods, hybrid global-local methods use homogenized stress
as a target for the calibration method. A major difference, and improvement, comes
from the acquiring and integration of full-field displacement or strain data from
DIC. This additional full-field data fundamentally changes the numerical aspects
of the calibration. Instead of fitting a relatively simple (approximately a secondorder polynomial) global stress-strain curve using many (often greater than 5) CP
parameters, the DIC dataset helps alleviate the issue of uniqueness that plagues
global methods. Because of this, global-local methods are an improvement over
global methods.
4.3.1 Data Flow
In a global-local calibration method, the measured global force is combined with
local DIC data; see Fig. 6. Consequently, compared to the standard global methods,
global-local methods require additional DIC hardware and software to acquire and
process the acquired images. To inform the CP model, it is ideal to measure the
particular microstructure throughout the test coupon. Most commonly, this data
is acquired before mechanical testing using EBSD. While acquiring EBSD data
provides data beneficial to the calibration process, it is incomplete in the sense that
only surface microstructure is measured, leaving uncertainty about the underlying
microstructure. This not only means that subsurface grain orientations are unknown
but further means that subsurface defects could also influence the acquired surface
strain data. While volumetric methods for measuring microstructure are possible,
their availability is currently lacking in general common usage and less commonly
used for calibration.
4.3.2 Computational Model
Because the DIC data acquired in this method is local, local values for displacement
or strain must be computed using a FE model. To most closely match the acquired
DIC data, the FE model should be constrained on its boundaries with measured DIC
displacement data within a region of interest (ROI). Rokoš et al. [32] have recently
179
visco-plastic self-consistent (VPSC) model [23] can provide accurate results by
accounting for the effect of grain shape while maintaining a tractable computational
effort. Additionally, while not exercised in this chapter, it is important to note that
stress-strain curves could be extracted from a variety of directions with respect to
the bulk material texture. This additional data would serve to improve the global
calibration approach.
4.3 Global-Local Methods
Like purely global methods, hybrid global-local methods use homogenized stress
as a target for the calibration method. A major difference, and improvement, comes
from the acquiring and integration of full-field displacement or strain data from
DIC. This additional full-field data fundamentally changes the numerical aspects
of the calibration. Instead of fitting a relatively simple (approximately a secondorder polynomial) global stress-strain curve using many (often greater than 5) CP
parameters, the DIC dataset helps alleviate the issue of uniqueness that plagues
global methods. Because of this, global-local methods are an improvement over
global methods.
4.3.1 Data Flow
In a global-local calibration method, the measured global force is combined with
local DIC data; see Fig. 6. Consequently, compared to the standard global methods,
global-local methods require additional DIC hardware and software to acquire and
process the acquired images. To inform the CP model, it is ideal to measure the
particular microstructure throughout the test coupon. Most commonly, this data
is acquired before mechanical testing using EBSD. While acquiring EBSD data
provides data beneficial to the calibration process, it is incomplete in the sense that
only surface microstructure is measured, leaving uncertainty about the underlying
microstructure. This not only means that subsurface grain orientations are unknown
but further means that subsurface defects could also influence the acquired surface
strain data. While volumetric methods for measuring microstructure are possible,
their availability is currently lacking in general common usage and less commonly
used for calibration.
4.3.2 Computational Model
Because the DIC data acquired in this method is local, local values for displacement
or strain must be computed using a FE model. To most closely match the acquired
DIC data, the FE model should be constrained on its boundaries with measured DIC
displacement data within a region of interest (ROI). Rokoš et al. [32] have recently