PSI - Issue 64

Sizhe Wang et al. / Procedia Structural Integrity 64 (2024) 2075–2082 Author name / Structural Integrity Procedia 00 (2019) 000–000

2079

5

were 32‒68 ℃, and the maximum temperatures of the Fe-SMA anchorage parts across all the specimens were 36‒ 45 ℃, lower than the adhesive glass transition temperature =67 ℃.

A-250

A-150-a

Temperature (℃)

(b)

Time (s)

Fig. 3. Temperature history during activation. (a) Specimen A-150-a. (b) Specimen A-250.

Fig. 4a shows the typical strain history of specimen A-150-a during the cooling process. For safety reasons, strain measurements started after the electricity was cut off (i.e., the recorded strain history only corresponded to the cooling process). The strain history indicated the mechanical strain evolution under the combined effects of both thermal expansions and Fe-SMA recovery stress (the strain gauges were self-temperature-compensated). At an early stage (earlier than 500 s), the effect from thermal expansions was predominant, such that some strain gauges recorded tensile strains. The strain readings of the four stain gauges differed from each other because the effect either from thermal expansion or Fe-SMA prestress varied along the steel plate mid-section. With the specimens cooling to the room temperature, the strain readings gradually stabilized. The final strain readings were the strains of steel plate caused by the Fe-SMA prestress. Fig. 4b illustrates the final strain readings along steel plate mid-section width at the end of activation process. Compressive strains were measured across all specimens, indicating the presence of Fe-SMA prestress. Fig. 4c summarizes the average values of the final strains after activation against the activation length. The average strains , were ranging from approximately -40 to -120 µm/m. Overall, the average compressive strain increased with the increasing activation length. Equation (1) provides a simple way to estimate the Fe-SMA prestress from the average strain of steel plate, where is the elastic modulus of steel ( =205 GPa), is the section area of steel plate ( =1500 mm 2 ), and is the section area of Fe-SMA strip ( =75 mm 2 ). = ∙ , ∙ 2∙ (1) However, Equation (1) is only suitable for specimens with a sufficiently long activation length, but it leads to a significant error when applied to specimens with short activation lengths (e.g., A-50-a/A-50-b). It is because the strain distribution at the steel plate mid-section caused by the Fe-SMA prestress is non-uniform, which is more significant for shorter Fe-SMA strips, as shown in Fig. 4b. According to Equation (1), the calculated Fe-SMA prestress was ranging from 82 to 246 MPa. In contrast, finite element models were developed to take the non-uniform strain distribution into account, based on which the Fe SMA prestress was estimated to be ranging from 154 to 249 MPa. In particular, for the activation of the 100 mm long Fe-SMA strips, two concerns existed: (1) During the activation, the high temperature of activation part might transfer to the anchorage parts and cause softening of the adhesive of anchorage parts, resulting in anchorage failure; (2) With the prestress developing, the short bond length might have insufficient capacity to hold the prestress, such that premature debonding occurs. These two concerns were addressed