UHPC and NSFRC in Severe Environmental Conditions

Structure and properties of cement composite are time-varying characteristics, depending among others on environmental conditions. The key idea is a struggle for complex research of joint effect of physical, chemical and dynamic loads on the internal structure of cement composite and understanding the correlation between changes in microstructure and macro-scale properties. During the experimental program, specimens will be exposed to combined influence of freeze-thaw cycles, aggressive chemical agents and dynamic loading. The aim is to create a theoretical basis for design of effective cement composites meant to be used in severe environmental conditions.


Introduction
The study of the project was to evaluate the combined effect of severe environmental conditions (freezethaw cycles, aggressive chemical agents and dynamic loading [11]) on the internal structure and mechanical properties of cement composites by experimental and analytical methods. The intention of this work is to establish theoretical basis for design of engineered concrete mixes whose characteristics would allow more efficient use of this material in severe environmental conditions. The areas of civil engineering that will benefit from the outcomes most heavily are transportation and industrial structures.

Mix Design and Specimens
Two different concrete mix designs were used in the project: fibre reinforced concrete (NSFRC -Normal-Strength Fibre reinforced Concrete) used in load-bearing parts of bridges in the Czech Republic (C35/45 -XC4, XD1, XF2 -Cl 0,20 could be used as a good representative based on consultations with designers) and ultra-high performance concrete (UHPC -Ultra-High Performance Concrete) developed in the Department of Concrete and Masonry Structures, Faculty of Civil Engineering (FCE), Czech Technical University in Prague (CTU) in recent years by the team under supervision of prof. Kohoutkova [9].
 Two different mix designs, with different type of components are given in the Table 1

Experimental program
In this experimental program two groups of prismatic specimens (100100400 mm) were preparedfibre reinforced concrete (NSFRC) samples and ultra-high performance concrete (UHPC) samples. The specimens for loading went through a special loading program. This consisted of three different types of loading:  Freeze-thaw cycling test according to ČSN 73 13 22 [4],  Resistance against aggressive chemical agents and deicing chemicals according to ČSN 73 1326 [7].  Dynamic loading (four point bending test).
Specimens were subjected to physical, chemical and dynamic loads, while the rest were stored aside as reference samples.

Resistance to cyclic loads, Group A1
Resistance to cyclic loads was studied on prismatic specimens (100100400 mm). The specimens were subjected to cyclic bending. The frequency of the applied load was 25 Hz for UHPC specimens and NSFRC specimens. Maximum applied force was determined for each material as the resistance obtained from static four-point bending tests -24.0 kN for NSFRC (8 specimens) and 40.0 kN for UHPC (8 specimens). The loading schedule of the cyclic loading was: 1. Peak force = 80 % of the resistance, 500 000 cycles. 2. Peak force = 85 % of the resistance, 80 000 cycles. 3. Peak force = 90 % of the resistance, 80 000 cycles. 4. Peak force continuously increased until the failure of the specimen.
In case of NSFRC specimens, the results were somewhat unbalanced. Out of the 25 specimens tested, 28 % passed the three basic loading levels (average force at failure 25.8 kN), 8 % failed in the 3 rd level, 8 % failed in the 2 nd level and remaining 52 % did not pass the 1 st testing level. The performance of UHPC was significantly better. In total, 18 specimens were tested. One of them failed in the 3 rd level, all the others passed all three basic levels, reaching average force at failure of 45.9 kN. The results have proven excellent resistance of UHPC to cyclic loading compared to NSFRC. This can be attributed mainly to higher amount of fibers effectively eliminating propagation of microcracks in cement matrix. More compact structure and higher strength of cement matrix also provided better bond between the matrix and the fibers. The test arrangement is presented in Figure 1. The MTS 500 kN, B-262 machine was used for the tests.

Resistance to freeze-thaw cycles, Group A2
Freeze-thaw cycling tests (according to ČSN 73 1322 standard) of both UHPC and NSFRC were conducted, 100100400 mm prism were used for the tests. The tests were evaluated after each 25 cycles in the range of 0 -200 cycles, a total of 56 specimens were produced for each material. As a result, it can be stated that UHPC is slightly more sensitive to freeze-thaw effects at higher loading levels than NSFRC. This is probably caused by the brittleness of the extremely dense structure of UHPC. Porous pressures created by freezing water in the microscopic cracks within the matrix are more harmful for UHPC than in case of NSFRC, where the wider cracks and higher amount of internal pores create an additional space for compensation of increased volume of ice. Nevertheless, the strengths of UHPC were still more than 60 % higher than the strengths of NSFRC even after 200 cycles. Test specimens before and after the test are shown in Figure 2.

Resistance to combination of freeze-thaw cycles and water with deicing chemicals, Group A3
In the combined test, 12 NSFRC prism and 4 UHPC prism (100100400 mm) were tested. The specimens were first subjected to 200 freeze-thaw cycles (FTC, according to ČSN 73 1322 standard) and then to 500 cycles of water with deicing chemicals (CWDC, according to ČSN 73 1326 standard, method A). Amounts of waste were recorded and dynamic elastic modulus was measured after each 100 cycles. In the end, the specimens were subjected to four-point bending test to determine their loadbearing capacity after long-time exposure to environmental loads.
No waste was detected after 200 FTC. In case of NSFRC, the first disturbance of the surface was observed after 100 CWDC (waste up to 20 g/m 2 ). After that, the waste steadily increased, reaching an average of 230 g/m 2 after 500 CWDC (700 cycles in total). In case of UHPC, no waste was recorded up to 200 CWDC and after 500 CWDC (700 cycles in total), the average waste was 38 g/m 2 .
Dynamic modulus measurements showed virtually no change during the whole experimental program for both UHPC and NSFRC. There was constant mild increase of the modulus, reaching approximately 6 % after the 700th cycle. This can be attributed to aging of the material during the long-term tests. The results indicate that the environmental loads did not cause any internal damage of the materials.
After cycling, the specimens were subjected to four-point bending test to determine their load-bearing capacity. Average force at failure was 26.6 kN for NSFRC samples and 41.4 kN for UHPC samples. The forces obtained on samples not exposed to cycling were 24.0 kN for NSFRC and 40.0 kN for UHPC. This confirms the conclusion of the dynamic modulus measurements.
The results proved outstanding performance of both NSFRC and UHPC to combined environmental loads -even after 700 loading cycles, the surface was disturbed in very limited extent, no cracks occurred inside the specimens and the resistance was not affected. Test specimens before and after the test are shown in Figure 3.

Resistance to combination of cyclic loading (1st) and water with deicing chemicals (2nd), Group A4
Resistance to water with deicing chemicals was tested (according to ČSN 73 1326 standard, method A) on the remains of NSFRC beams from A1. The surface of all the specimens was undisturbed after 150 cycles, the amount of waste was up to 1 g/m 2 . The tests of UHPC samples were not performed, it was legitimately assumed that the waste would have been even smaller. When these results are compared with A3, it is clear that the dynamic cyclic loading has no effect on the rate of disturbance of the surface of the material exposed to environmental loads.
Penetration of samples by chlorides (NaCl) was measured by chloride ponding test on UHPC and NSFRC samples immersed in 3 % NaCl solution for 105 and 148 days. No substantial difference between concentrations in UHPC and NSFRC samples not exposed to dynamic cycling was found. The chlorides penetrated to the depth of max. 20 mm. This means that the risk of corrosion of the reinforcement in both materials is very low if the material is not exposed to dynamic cycling. In case of 5

1234567890
IMST 2017 IOP Publishing IOP Conf. Series: Materials Science and Engineering 251 (2017) 012031 doi:10.1088/1757-899X/251/1/012031 samples exposed to 660 000 loading cycles, NaCl concentrations were increased by 60 -120 % in both materials. This means that the microcracks formed during the dynamic loading play crucial role in the rate of penetration of chlorides in surface layers. However, the depth of penetration did not increase even after the dynamic loading.
3.5. A5: resistance to combination of water with deicing chemicals (1st) and cyclic loading (2nd) A set of six NSFRC samples was subjected to 100 CWDC. The surface of all the specimens was undisturbed after the test, the amount of waste was up to 0.5 g/m 2 . After that, they were exposed to dynamic cycling using the scheme from A1. Two beams passed the three basic levels with the average resistance of 25.8 kN in the last level, one beam failed in the 3 rd level, one in the 2 nd level, remaining two specimens did not pass the 1 st level.
The difference compared to A1 is negligible, therefore it can be said that up to 100 cycles, there is no significant variance in the behavior of NSFRC exposed to dynamic cycling whether it is exposed to CWDC or not. With respect to this finding, the tests for UHPC were not performed as it was legitimately expected based on the results of other tests that the outcomes would be similar.

Summary
The main findings of the experimental program can be summarized as follows: 1. Resistance of both NSFRC and UHPC samples to environmental loads was excellent. After 700 cycles of combined exposure to freeze-thaw cycles and water with deicing chemicals, surface of the specimens was just slightly disturbed, mechanical properties were unaffected. 2. UHPC was slightly more sensitive to effects of freeze-thaw cycles than NSFRC. Earlier beginning and higher rate of decrease of tensile strength was observed. This was probably caused by the brittleness of the extremely dense matrix of UHPC. 3. When subjected to dynamic cycling, the behavior of UHPC was significantly better than the behavior of NSFRC. The main difference was that the results of UHPC tests were steady, while the results of NSFRC showed significant variance. This can be attributed mainly to higher amount of fibers in UHPC effectively eliminating propagation of microcracks in cement matrix. 4. When dynamic cycling was combined with environmental loads, the rate of penetration of chloride ions into cement matrix was increased. However, no effect on surface disturbance and load-bearing resistance was observed in either NSFRC or UHPC specimens.