PSI - Issue 62

Alessandro Bellini et al. / Procedia Structural Integrity 62 (2024) 315–322 A. Bellini et al. / Structural Integrity Procedia 00 (2019) 000 – 000

319

5

4. Experimental results Results of prestress release tests will be presented and discussed in terms of strain vs time curves, pointing out the typical behavior of each test type and the accuracy to be expected during the application of the procedure. Examples of strain vs time curves measured during the experimental campaign are presented in Fig. 4: in detail, Fig. 4a shows the evolution of strain detected during a prestress release tests carried out making the saw-cuts first parallel (P) and the orthogonal (O) to the expected stress field, whereas Fig. 4b shows the results obtained by applying exactly the opposite procedure (O and then P saw-cuts). For both cases, 4 markers identify the starting time of each saw-cut with a letter (A-D). After completion of the cutting phase, the full detachment of the sample (letter R in Fig. 4a, b) was promoted by gently using a chisel. From a qualitative point of view, the two procedures show a common behavior: when the saw-cuts are parallel to the applied vertical stress (points A, B in Fig. 4a, b), their contribution to the overall prestress release is almost negligible, whereas saw-cuts orthogonal to the stress field (points C, D in Fig. 4a, b) offer the most significant release contribution. Comparing the effect of the orthogonal saw-cuts (O) from both tests, if the cutting depth is sufficient, the first one (D) shows, in general, the predominant contribution irrespective of previous or following P saw-cuts. Considerations about the accuracy of prestress release test can be drawn by observing the results reported in Table 2, where the residual error, for each group of tests, is calculated as the ratio between the residual strain (upon detachment of the sample) and the maximum deformation (after completing the load application), expressed in percent. Tests with the same applied stress, irrespective of the order of the saw-cuts, are joined together in the same group. Table 2 shows also the same percentage error considering the compensation by subtracting the effects of differential drying shrinkage, which can be significant for laboratory specimens with reduced curing time (Romano and Mazzotti, 2022). All the residual errors (considering or not considering the effects of the differential shrinkage) are included in the 6.0÷8.6% range, showing a good accuracy of the method.

C

A

B

D

(a) (c) Fig. 3. Detail of saw-cutting procedure: (a) use of angle grinder; (b) concrete surface after sample extraction; (c) example of a sample extracted with a truncated pyramidal shape. (b)

50

50

0

R

0

100

200

300

400

500

600

R

0

A

B

-50

0

100

200

300

400

500

600

C

-100

C

-100 Strain (µ ε ) -50

-250 Strain (µ ε ) -200 -150

C

C

B

A

B

-300

A

A

B

-150

-350

D

D

D

D

-200

-400

Time (s)

Time (s)

(a) (b) Fig. 4. Examples of prestress release tests with a different order during the execution of saw-cuts: (a) parallel (P) and then orthogonal (O); (b) orthogonal (O) and then parallel (P).