PSI - Issue 42

Sebastian Henschel et al. / Procedia Structural Integrity 42 (2022) 110–117

112

S. Henschel and L. Kru¨ger / Procedia Structural Integrity 00 (2019) 000–000

3

and attenuation” leads to a di ff erence between the introduced and the measured pulses. Consequently, K ε eliminates these di ff erences. In the second step of calibration (‘bars together’), a certain loss of momentum during the transfer from the incident bar to the transmitted bar is considered possible. Gama et al. (2004) argue that an imperfect contact, e.g. due to misalignment, leads to this incomplete pulse transmission. A misaligned system can be recognized by deformed pulses and / or a small reflected pulse during the ‘bars together’ calibration. The loss of momentum is then incorporated in the coe ffi cient of pulse transmission ( K σ ). Hence, the di ff erence between ε I and ε T is eliminated. In both steps of calibration, the measured amplitudes of the pulses are adjusted to the respective theoretical pulse amplitudes. This procedure does not di ff erentiate between actual and apparent losses of momentum. An actual loss results from bar misalignment, whereas an apparent loss results from inaccurate strain gauge instrumentation (e.g. bonding, alignment). The apparent loss of momentum can be corrected by the above mentioned calibration steps. However, it depends on the user’s experience and the aimed level of precision if an actual loss is incorporated in the calibration or if it is reduced, e.g. by thorough realignment of the bars. There is a second common characteristic of these calibration steps. The actual loading of the specimen is not analyzed. Only the pulse introduced at one end and the pulse measured at some hundred millimeters away from the position of interest, i.e. the interface bar / specimen, are considered. In the present paper, a new method for dynamic calibration of the incident and transmitted bar’s strain instru mentation is proposed. This method does neither require the knowledge of the striker velocity nor the assumption of momentum conservation. The wave velocity of the bars is also not needed for calculation. It will be shown that the shape of the pulse does not a ff ect the calibration. An optical extensometer is applied to measure the displacements of the interfaces between the incident and transmitted bars and the specimen. The measured displacements are then com pared with the calculated displacements from the instrumentation of the bars, and the calibration factors are obtained. Moreover, this new method of dynamic calibration can be performed during an actual test. Separate calibrations steps are omitted.

2. Material and methods

Fig. 1 shows the setup of the SHPB. The bars are made of high-strength maraging steel and have a common diameter of 19.8 mm. The axial strain is measured with foil strain gauges (gauge length 1.5 mm) that are part of a Wheatstone bridge. The distances between the strain gauges and the positions 1 and 2 are given by x I and x T . The amplified voltage (cuto ff frequency 1MHz) is digitized in a transient recorder (sample rate 20MS / s). The velocity of the striker bar is measured by two light barriers with a distance of 100 mm. In addition to the strain measurements at the bars, the displacements at positions 1 and 2 are measured optically. To this end, an optical extensometer (type 200XR with objective lens 200XR-10, Rudolph, Germany) is applied, see Fig. 1. First, this extensometer detects two black-and-white edges, one for each channel. These edges are printed on stickers and are positioned on the bars, see Fig. 2a.

strain gauges specimen

v S

1 2

x I

x T

striker bar

incident bar (IB)

transmitted bar (TB)

momentum trap

light barriers

Tr

optical extensometer

data acquisition

Ch1 Ch2

Fig. 1. SHPB setup with strain instrumentation on incident and transmitted bars, and optical extensometer. Striker velocity v S . The right light barrier is used to trigger (Tr, pre-trigger) the data acquisition.