PSI - Issue 78

Han Liu et al. / Procedia Structural Integrity 78 (2026) 1759–1766

1764

Fig. 3: (a) Overall experimental configuration; and (b) close-up view of the flexural bending setup.

4. Results and Discussion

Figure 4(a) presents 28-day three-specimen averaged resistivity ρ computed using Eq. 1 as a function of CMF inclusion level, ranging from 0 wt% to 0.25 wt%, with error bars indicating the minimum-to-maximum range of resistivity measured across three specimens. The inset presents a similar analysis for varying G from 0 wt% to 16 wt%. It can be observed that the resistivity of the Sikacrete specimens significantly decreased with increasing G content, particularly at lower percentages from 0 wt% to 3 wt%. Percolation behavior for both MCMF and CCMF exhibited a similar trend of rapidly decreasing resistivity at low concentrations, with the most notable reduction occurring both within the range of 0.032 wt% to 0.125 wt%. Following these results, a hybrid composition consisting of 2 wt% G, 0.25 wt% MCMF, and 0.062 wt% CCMF, referred to as 2G250M62CCMF, was ultimately selected as the mix design for fabricating the conductive Sikacrete composite. Figures 4(c) and (d) present time-series plots of the FCR of specimens measured between AB, BC, CD, DE, and EF electrodes, respectively, compared against the strain input during the dynamic test. Data presented here was post processed using a detrend filter to eliminate intrinsic polarization drift caused by the dielectric nature Wen and Chung (2002); Bekzhanova et al. (2021) and / or direct piezoelectric e ff ect Sun et al. (2000); Chung (2002), and the bending strain was computed using Eq. 4. The results show a close match between the electrical signal and strain time histories, indicating that the beam specimen is capable of tracking the strain input. Notably, the FCR amplitudes measured between electrodes AB and EF were smaller than those measured at the central electrode pairs. This observation is consistent with the strain distribution along the beam, as bending strain is reduced near the ends and maximized toward the midspan. Additionally, all resistance measurements exhibited a relatively high level of signal noise, likely due to the contact resistance fluctuations at the embedded electrode interfaces.