PSI - Issue 64

Niloofar Heirani et al. / Procedia Structural Integrity 64 (2024) 6–13 Author name / Structural Integrity Procedia 00 (2019) 000–000

10

5

3.2. Fiber dispersion Figures 6A and 6B show optical microscope and backscattered SEM images, respectively, of a representative cured sample (10% CF, GU cement, centrifugal mixer, method A). Note that the images are not taken from the same spot and have different scales. These and similar images of all samples were used to analyse the fiber dispersion by measuring the inter-fiber distances as described in (Heirani, 2021). The smaller black circles in the backscattered image represent CFs. The larger black circles are air voids inside the specimen. For a given CF content, two main factors affected fiber dispersion: the addition of supplementary cementitious materials and the type of mixer. Fibers were closer to each other on average in the presence of supplementary cementitious materials, with less variability in distances. In samples mixed with the Centrifugal mixer, air voids were smaller on average, and CF were closer to each other, with a smaller variability in distance. Figures 7A and 7B present representative results that reflect these findings.

Fig. 6. Optical images of 10%CF samples mixed with (A) Hobart mixer and (B) Centrifugal mixer

Fig. 7. CF dispersion reflected by nearest fiber distance; representative results showing A) effect of silica fume (SF) on 5% CF samples, and B) effect of mixer type on 10% CF samples. 3.3. Sensing properties The electrical resistivity of the CFCP sensor samples decreased while loading and increased when unloading, confirming their piezoresistive behavior. Their strain sensitivity was calculated using ∆ / 0 vs. strain plots for the loading portion of 5 cycles. All sensors had relatively high sensitivity, with gauge factors (GFs) varying from of ~87 to ~1025, depending on fabrication factors (Heirani, 2021). In this study, special attention was also given to the repeatability of the sensor response as an important sensing characteristic for practical applications. Figure 8 shows the response of a 15% CF sample over 30 loading cycles. 15% CF sensors had the densest fiber network as the fiber content was much higher than the percolation threshold (found to be ~0.5% (Heirani, 2021)). This enhanced the repeatability of the sensing response. Here, the difference between GFs of the first and last compressive loading cycles (depicted in black and red, respectively, in Figure 8) was taken as a simplified measure of repeatability; a smaller difference indicated better repeatability. As 30 cycles of loading was extremely time-consuming (60 hours), the same procedure was applied with 5 loading cycles for all specimens.