Effects of Interaction of Static Load and Frost on Damage Mechanism of Concrete Elements

Frost damage is typical material deterioration in concrete structures subjected to external environmental conditions. However, the weather conditions do not make the sufficient factor causing the worsening of the concrete properties. Concrete structures with frost damage in service are subjected to loadings. The investigation was carried out with the primary objective to assess the influence of interaction of mechanical load and freeze-thaw cycles on damage process of concrete. The salt scaling process was observed for beam specimens subjected to cyclic freezing and thawing in third-point loading condition. The damage development due to internal cracking was monitored using fracture mechanics parameters.


Introduction
Under severe freezing climate conditions, frost action is probably the most important cause of deterioration of exposed concrete structures.Even in structures with airentrained concrete, frost damage can occur under certain circumstances.Usually, the weather conditions do not make the sufficient factor causing the worsening of the concrete properties.Damage evolution as well as the concrete resistance to freezing and thawing have been of great concern of researchers for many years (Pigeon et al. 1996, Jana 2004, Pentala 2006, Valenza and Scherer 2007).
Two types of damage to concrete may occur as a result of cyclic freezing and thawing: ▪ surface scaling, ▪ internal cracking.Surface scaling is generally found where the surface of concrete is subjected to weak solutions of salt, typically used for de-icing purposes.Internal freeze-thaw damage results from expansive stresses generated by water on freezing when the pore structure of the concrete is saturated above a critical value, and leads to internal microcracking.Internal damage, which is likely to occur in concrete subjected to long-term wet/saturated conditions (Fagerlund 2002), is manifested macroscopically by irreversible tensile deformation and randomly oriented microcracking (Pigeon et al. 1996).Hence, the freezing and thawing action can be looked upon as a very complex fatigue crack propagation process (Hasan et al. 2008).
The most of previous works on the freeze-thaw durability of concrete have been focused on the mechanical property degradation (e. g. modulus and strength), weight change, length change, microstructural change or ultrasonic signature change after different numbers of freeze-thaw cycles, sometimes in the presence of salt solution (Cao and Chung 2002).Performance deterioration caused by a monodamaging process, such as freeze-thaw, is not consistent with real conditions to which concrete structures are actually exposed.It has been found that the deterioration of concrete could be accelerated when subjected to multidamaging processes, e. g. simultaneously exposed to external load, freeze-thaw cycles and chloride or sulphate attack.It is necessary to understand how the resulting stresses influence the concrete resistance to freezing and thawing.However, there are very few test results considering the effects of interaction of mechanical loading and cyclic freezing and thawing, and in large part, they concern the effect of internal damage on concrete properties (Zhou et al. 1994, Sun et al. 1999, Yu et al. 2008).
The analysis of research results, described in literature, has showed there is the lack of information concerning the influence of the tensile stress on the typical process of surface scaling due to freezing liquid contact with selected surface of element tested, which is the frequent situation in service life of concrete and reinforced concrete structures (Şahmaran and Li 2007).Since salt scaling is superficial, it does not affect mechanical integrity of a concrete body.However, this damage renders material susceptible to ingress of moisture and aggressive species that threaten durability (Valenza II and Scherer 2007).
The concrete is a heterogeneous material of high compressive strength, but its resistance to cracking is http://dx.doi.org/10.5755/j01.sace.1.1.2616low.The destruction of concrete under the influence of external loads is affected, among other things, by material discontinuities, disruptions, and local differences in mechanical properties of material.The local accumulations of stress caused by external pressures occur in the vicinity of concrete defects (Gettu et al. 1990, Jenq andShah 1991).They can cause violent propagation of damage, and finally lead to the destruction of the entire element.The most dangerous concentrators of stress are the tips of cracks where the greatest stress values are achieved (Schlangen and Garboczi 1997).
In the progress of freezing and thawing, recurrent frost expanding and penetration compressive stress act on concrete, each cycle produce freeze-thaw inner stress in concrete interior, the stress causes inner flaws of concrete expand, accumulate and form new damage.Freeze-thaw cycles and loads are both repetitive actions that cause accumulated physical damages in concrete.As the essential development and important supplement of fracture mechanics, damage mechanics is the part of material structure distortion and wreck theory, it emphasize the influence of material damage to mechanical properties, also the evolutive process and rules of materials or structure damage.Fracture mechanics can help analyze the response of microstructure to external load (Shah et al. 1995, Bažant 2002, Wang et al. 2010).So, it is reasonable and feasible to study the freeze-thaw damage rules of concrete by using fracture mechanics parameters.The critical stress intensity factor and the critical crack tip opening displacement, along with the Young's modulus, are sufficient to characterize the fracture resistance of concrete.The methods of fracture mechanics were approved in structural design by the latest proposal of CEB-FIP Model Code (2010).
Two different approaches to the estimation of cyclic freezing and thawing influence on concrete properties were presented in the paper.The analysis of interaction of load and freeze-thaw cycles with chloride exposure regime on surface scaling process of concrete was performed.The fracture parameters were used to assess the internal cracking progress in concrete.The investigation was not intended to perfectly simulate loading and exposure conditions, but to begin to make the preliminary understanding about the complex deterioration mechanisms of concrete in real service life conditions.

Specimen preparation
The tests were carried on for non-air-entrained concrete as well as for air-entrained concrete.The cement (CEM I 42,5) content in concretes tested was constant -350 kg/m 3 , and water to cement ratio was equal to 0,40.The natural aggregate with maximum diameter of 8 mm was used.The air-entraining agent content was 0,10% related to cement mass.After demoulding the specimens were stored in water with temperature 20 ± 2 ° C. The compressive strength, tested after 28 days of curing, was equal respectively 59,7 MPa for non-air-entrained concrete and 55,2 MPa for concrete with AEA.

Estimation of scaling resistance
The specimens sizes were 80×120×1100 mm.Every series was composed of 3 replicates.The concrete resistance to surface scaling due to cyclic freezing and thawing with de-icing salt saturation (3% NaCl solution) was determined using the procedure described in PKN-CEN TC 12390-9:2007.In order to realize the interaction of freezethaw cycles and load the beam specimens were tested in third-point loading condition.The load was realized using lever gears.The tensile stress ratio c was 0,0 (control specimens); 0,17 and 0,50 with respect to the failure stress.The arrangement of test stand was presented in Fig. 1.
The scaled material was collected from the top surface of specimen, which was subjected to tensile stress.The deicing salt solution was kept on the top of specimen thanks to the rubber sheet glued to all surfaces of the specimen except the test surface.The edge of the rubber sheet reached 20 mm above the test surface.Top specimen surface (total area for every specimen was equal to 45000 mm 2 ) was saturated with demineralized water during 72 hours.Immediately before the specimens were placed in the freezing chamber, the demineralized water was replaced with 3% NaCl solution.The freezing medium was prevented from evaporating by applying a flat polyethylene sheet.The loading devices with specimens were placed in freezing chamber.C to -18 o C. Every 7 d solution was exchanged.The material that scaled from the test surface was collected and dried to weight.The amount of the scaled material per unit area after n cycles m n was evaluated for each m occasion and each specimen.The specimens were subjected to 56 freeze-thaw cycles, the number sug unmodified concrete.

Fracture parameters determination
The specimens for fracture parameters evaluation were subjected to cyclic freezing in air and th water.The temperature changed from -18 o C to 18 o C. The duration of single cycle was 8 hours and th period duration was 6 hours.The freezing and thawing process was finished 1 day before testing.
The critical stress intensity factor s Ic K and the critical crack tip opening displacement CTO determined using procedure described in RILEM draft recommendation (1990), based on the fractu elaborated by Jenq and Shah (1985).The fracture parameters were assessed in three-point bend test with initial notches.The specimen sizes were 100×100×400 mm, and the initial saw-cut notch depth to 30 mm and width was 3 mm.The geometry of specimen and the way of load were presented in The single cycle duration was 24 hours with temperature change from 20 °C to -18 °C.Every 7 days NaCl solution was exchanged.The material that scaled from the test surface was collected and dried to constant weight.The amount of the scaled material per unit area after n cycles m n was evaluated for each measuring occasion and each specimen.The specimens were subjected to 56 freezethaw cycles, the number suggested for unmodified concrete.

Fracture parameters determination
The specimens for fracture parameters evaluation were subjected to cyclic freezing in air and thawing in water.The temperature changed from -18 °C to 18 °C.The duration of single cycle was 8 hours and the freezing period duration was 6 hours.The freezing and thawing process was finished 1 day before testing.
The critical stress intensity factor The single cycle duration was 24 hours with temperature change from solution was exchanged.The material that scaled from the test surface w weight.The amount of the scaled material per unit area after n cycles m n occasion and each specimen.The specimens were subjected to 56 freeze-tha unmodified concrete.

Fracture parameters determination
The specimens for fracture parameters evaluation were subjected to cy water.The temperature changed from -18 o C to 18 o C. The duration of single period duration was 6 hours.The freezing and thawing process was finished 1 The critical stress intensity factor s Ic K and the critical crack tip op determined using procedure described in RILEM draft recommendation (1 elaborated by Jenq and Shah (1985).The fracture parameters were assessed with initial notches.The specimen sizes were 100×100×400 mm, and the in to 30 mm and width was 3 mm.The geometry of specimen and the way of lo series was composed of 4 replicates.

a)
and the critical crack tip opening displacement CTOD c were determined using procedure described in RILEM draft recommendation (1990), based on the fracture model elaborated by Jenq and Shah (1985).The fracture parameters were assessed in threepoint bend test on beams with initial notches.The specimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal to 30 mm and width was 3 mm.The geometry of specimen and the way of load were presented in Fig. 2a.Each series was composed of 4 replicates.

48
arameters determination mens for fracture parameters evaluation were subjected to cyclic freezing in air and thawing in perature changed from -18 o C to 18 o C. The duration of single cycle was 8 hours and the freezing was 6 hours.The freezing and thawing process was finished 1 day before testing.al stress intensity factor s Ic K and the critical crack tip opening displacement CTOD c were ng procedure described in RILEM draft recommendation (1990), based on the fracture model enq and Shah (1985).The fracture parameters were assessed in three-point bend test on beams ches.The specimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal idth was 3 mm.The geometry of specimen and the way of load were presented in Fig. 2a.Each posed of 4 replicates.was monotonically loaded up to the maximum load.The applied load was reduced after the load imum value and was at about 95% of the peak load.Then, the applied load was reduced to zero g was applied.The specimen was cyclically loaded up to failure.

Fig. 2. Fracture testing configuration and geometry of specimen: (a) the way of load; (b) the place of CMOD measurement
The closed-loop testing machine with crack mouth opening displacement (CMOD) as the feedback signal was used to achieve a stable failure.The crack mouth opening displacement and the applied load were recorded continuously during the test.The CMOD, indicated in Fig. 2b, was measured by means of clip gauge.To measure the crack mouth opening displacement (CMOD) a pair of knife edges was attached to two sides of a notch performed on the lower surface of the beam.The rate of loading was controlled by a constant rate of increment of CMOD so that the peak load was reached in 5 min.
The beam was monotonically loaded up to the maximum load.The applied load was reduced after the load passed the maximum value and was at about 95% of the peak load.Then, the applied load was reduced to zero and the reloading was applied.The specimen was cyclically loaded up to failure.
After the initial cycle, each loading and unloading cycle was finished in about 1 min.The test result is a load-CMOD curve with several loading-unloading cycles.Based on the load-CMOD relation, the fracture parameters and Young's modulus can be calculated.The P-CMOD curve was prepared for each specimen.Typical test result (obtained for control concrete without AEA) was presented in Fig. 3.
According to RILEM recommendations (1990) the Young's modulus E is calculated from the equation

Effects of Interaction of Static Load and Frost on Damage Mechanism of Concrete Elements
After the initial cycle, each loading and unloading cycle was finished in about 1 min.The test result is a load-CMOD curve with several loading-unloading cycles.Based on the load-CMOD relation, the fracture parameters and Young's modulus can be calculated.The P-CMOD curve was prepared for each specimen.Typical test result (obtained for control concrete without AEA) was presented in Fig. 3.

Fig. 3. Typical experimental load-CMOD plot
According to RILEM recommendations (1990) the Young's modulus E is calculated from the equation Where C i is the initial compliance calculated from load-CMOD plot (Fig. 3), S, a 0 , d, b are geometrical characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack growth at peak load) is determined from Equation (1) for calculated value of Young's modulus and the unloading compliance C u measured at the maximum load (Figure 3).Using an iteration process, the critical effective crack length a c is found when Equation ( 2) is satisfied: Where C i is the initial compliance calculated from load-CMOD plot (Fig. 3), S, a 0 , d, b are geometrical characteristics of specimen.

49
After the initial cycle, each loading and unloading cycle was finished in about 1 min.The te load-CMOD curve with several loading-unloading cycles.Based on the load-CMOD relation, parameters and Young's modulus can be calculated.The P-CMOD curve was prepared for eac Typical test result (obtained for control concrete without AEA) was presented in Fig. 3.

Fig. 3. Typical experimental load-CMOD plot
According to RILEM recommendations (1990) the Young's modulus E is calculated from the e ], /[ ) ( 6 Where C i is the initial compliance calculated from load-CMOD plot (Fig. 3), S, a 0 , d, b are characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack growth at peak load) is deter Equation (1) for calculated value of Young's modulus and the unloading compliance C u mea maximum load (Figure 3).Using an iteration process, the critical effective crack length a c is Equation ( 2) is satisfied: The critical stress intensity factor is calculated according to relationship The critical crack tip opening displacement is calculated as: In Equations ( 1), ( 2), ( 3), (4 ) are geometric functions, described i Shah et al. (1995).
Simultaneously, the changes in compressive strength due to frost action were monitored, specimens 100×100×100 mm, subjected to the same freeze-thaw regime as specimens for fracture testing.

Scaling resistance of concrete subjected to static load
The test results for both non-air-entrained and air-entrained concretes are presented in Fig. 4 an The analysis of test results showed the significant influence of considered range of stress on th mass of material scaled from the specimen surface subjected to cyclic freezing and thawing.Th susceptibility to surface scaling was observed for both concretes, with and without air-entraining The critical effective crack length a c (a c = a 0 + stable crack growth at peak load) is determined from Equation (1) for calculated value of Young's modulus and the unloading compliance C u measured at the maximum load (Figure 3).Using an iteration process, the critical effective crack length a c is found when Equation ( 2) is satisfied: Where C i is the initial compliance calculated from load-CMOD pl characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack gro Equation (1) for calculated value of Young's modulus and the unloa maximum load (Figure 3).Using an iteration process, the critical eff Equation ( 2) is satisfied: The critical stress intensity factor is calculated according to relations The critical crack tip opening displacement is calculated as: In Equations ( 1), ( 2), (3), (4 Shah et al. (1995).
Simultaneously, the changes in compressive strength due to fros specimens 100×100×100 mm, subjected to the same freeze-thaw regime testing.

Scaling resistance of concrete subjected to static load
The test results for both non-air-entrained and air-entrained concrete The analysis of test results showed the significant influence of consi mass of material scaled from the specimen surface subjected to cyclic susceptibility to surface scaling was observed for both concretes, with (2) The critical stress intensity factor is calculated according to relationship Where C i is the initial compliance calculated from load-CMOD pl characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack grow Equation (1) for calculated value of Young's modulus and the unloa maximum load (Figure 3).Using an iteration process, the critical effe Equation ( 2) is satisfied: The critical stress intensity factor is calculated according to relationsh The critical crack tip opening displacement is calculated as: In Equations ( 1), ( 2), (3), (4 Shah et al. (1995).
Simultaneously, the changes in compressive strength due to fros specimens 100×100×100 mm, subjected to the same freeze-thaw regime testing.

Scaling resistance of concrete subjected to static load
The test results for both non-air-entrained and air-entrained concretes The analysis of test results showed the significant influence of consi mass of material scaled from the specimen surface subjected to cyclic susceptibility to surface scaling was observed for both concretes, with Where C i is the initial compliance calculated from load-CMOD plo characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack grow Equation (1) for calculated value of Young's modulus and the unloa maximum load (Figure 3).Using an iteration process, the critical effe Equation ( 2) is satisfied: The critical stress intensity factor is calculated according to relationsh Shah et al. (1995).
Simultaneously, the changes in compressive strength due to fros specimens 100×100×100 mm, subjected to the same freeze-thaw regime testing.

Scaling resistance of concrete subjected to static load
The test results for both non-air-entrained and air-entrained concretes The analysis of test results showed the significant influence of consi mass of material scaled from the specimen surface subjected to cyclic susceptibility to surface scaling was observed for both concretes, with × ×

49
After the initial cycle, each loading and unloading cycle was finished in about 1 min.The test result is load-CMOD curve with several loading-unloading cycles.Based on the load-CMOD relation, the fractur parameters and Young's modulus can be calculated.The P-CMOD curve was prepared for each specimen Typical test result (obtained for control concrete without AEA) was presented in Fig. 3.

Fig. 3. Typical experimental load-CMOD plot
According to RILEM recommendations (1990) the Young's modulus E is calculated from the equation Where C i is the initial compliance calculated from load-CMOD plot (Fig. 3), S, a 0 , d, b are geometrica characteristics of specimen.
The critical effective crack length a c (a c = a 0 + stable crack growth at peak load) is determined from Equation (1) for calculated value of Young's modulus and the unloading compliance C u measured at th maximum load (Figure 3).Using an iteration process, the critical effective crack length a c is found whe Equation ( 2) is satisfied: The critical stress intensity factor is calculated according to relationship in which a a / = .

Results and Discussion
Scaling resistance of concrete subjected to static load The test results for both non-air-entrained and air-entrained concretes are presented in Fig. 4 and 5.The analysis of test results showed the significant influence of considered range of stress on the increase i mass of material scaled from the specimen surface subjected to cyclic freezing and thawing.The increase susceptibility to surface scaling was observed for both concretes, with and without air-entraining admixture In Equations ( 1), ( 2), ( 3 ) are geometric functions, described in details by Shah et al. (1995).
Simultaneously, the changes in compressive strength due to frost action were monitored, using cubic specimens 100×100×100 mm, subjected to the same freeze-thaw regime as specimens for fracture parameters testing.

Scaling resistance of concrete subjected to static load
The test results for both non-air-entrained and airentrained concretes are presented in Fig. 4 and 5.
The analysis of test results showed the significant influence of considered range of stress on the increase in (4) mass of material scaled from the specimen surface subjected to cyclic freezing and thawing.The increased susceptibility to surface scaling was observed for both concretes, with and without air-entraining admixture, although, the dosage of AEA assured very good scaling resistance for control unload concrete specimens.The influence of the tensile stress on scaling was observed after 14 cycles of the freezing and thawing.The differences in mass of material, scaled from specimens subjected to various stress levels, increased together with the number of cycles.Besides, for specimens loaded the greater scatter of measurements was noticed.
M. Kosior-Kazberuk osage of AEA assured very good scaling resistance for control unload concrete specimens.The e tensile stress on scaling was observed after 14 cycles of the freezing and thawing.The ass of material, scaled from specimens subjected to various stress levels, increased together with ycles.Besides, for specimens loaded the greater scatter of measurements was noticed.
mass of scaled material m vs. freeze-thaw cycles number n as well as stress level c for non-air-entrained concrete ass of scaled material m vs. freeze-thaw cycles number n as well as stress level c for air-entrained concrete the unloaded concrete specimens, after the initial rapid growth in mass of scaling, the slowdown as observed and then the mass of scaling accumulated gradually.For loaded concrete specimens, e scaling increase was almost linear together with the number of cycles, for both stress levels.
the mass of scaled material for stress level c = 0,17 was ca.30% greater, and for c = 0,50 twice parison to the scaling from unloaded concrete surface.After 56 cycles, the loss in material for el was twice greater and for higher stress -more than three times greater than the mass of scaling ncrete.the air-entrained concrete, the greater relative difference in mass of material scaled from concrete d was pointed out in comparison to control concrete.The tensile stress level c = 0,17 caused four and stress level c = 0,50 -five times increase in the amount of scaled material.However, the aironcrete influenced the limitation of the susceptibility for scaling in comparison to the concrete ure.In considered range of external load of specimens in third point bending test, the rate of ss increased with the increase in applied stress value.ng the mass of scaled material after n cycles as a measure of accumulated damages, it is possible number of cycles after that, the scaling achieves unacceptable volume, for assessed stress level.evel of scaling can be determined arbitrarily, regarding durability requirements, or on the basis of ifferent concrete elements.The dependence can be useful for predicting concrete ability to scaling e tensile stress level.
ture parameters of concrete due to cyclic freezing and thawing re parameters were determined on the basis of P-CMOD curves obtained for concrete specimens.freezing and thawing on concrete properties was referee to the results obtained for reference

Fig. 4. Mean mass of scaled material m vs. freeze-thaw cycles number n as well as stress level c for non-air-entrained concrete
M. Kosior-Kazberuk sage of AEA assured very good scaling resistance for control unload concrete specimens.The e tensile stress on scaling was observed after 14 cycles of the freezing and thawing.The ass of material, scaled from specimens subjected to various stress levels, increased together with ycles.Besides, for specimens loaded the greater scatter of measurements was noticed.
mass of scaled material m vs. freeze-thaw cycles number n as well as stress level c for non-air-entrained concrete ss of scaled material m vs. freeze-thaw cycles number n as well as stress level c for air-entrained concrete the unloaded concrete specimens, after the initial rapid growth in mass of scaling, the slowdown as observed and then the mass of scaling accumulated gradually.For loaded concrete specimens, scaling increase was almost linear together with the number of cycles, for both stress levels.the mass of scaled material for stress level c = 0,17 was ca.30% greater, and for c = 0,50 twice arison to the scaling from unloaded concrete surface.After 56 cycles, the loss in material for el was twice greater and for higher stress -more than three times greater than the mass of scaling ncrete.the air-entrained concrete, the greater relative difference in mass of material scaled from concrete d was pointed out in comparison to control concrete.The tensile stress level c = 0,17 caused four nd stress level c = 0,50 -five times increase in the amount of scaled material.However, the airncrete influenced the limitation of the susceptibility for scaling in comparison to the concrete ure.In considered range of external load of specimens in third point bending test, the rate of s increased with the increase in applied stress value.g the mass of scaled material after n cycles as a measure of accumulated damages, it is possible number of cycles after that, the scaling achieves unacceptable volume, for assessed stress level.vel of scaling can be determined arbitrarily, regarding durability requirements, or on the basis of fferent concrete elements.The dependence can be useful for predicting concrete ability to scaling tensile stress level.
ture parameters of concrete due to cyclic freezing and thawing re parameters were determined on the basis of P-CMOD curves obtained for concrete specimens.reezing and thawing on concrete properties was referee to the results obtained for reference In case of the unloaded concrete specimens, after the initial rapid growth in mass of scaling, the slowdown of the process was observed and then the mass of scaling accumulated gradually.For loaded concrete specimens, the mass of the scaling increase was almost linear together with the number of cycles, for both stress levels.After 28 cycles the mass of scaled material for stress level c = 0,17 was ca.30% greater, and for c = 0,50 twice greater in comparison to the scaling from unloaded concrete surface.After 56 cycles, the loss in material for lower stress level was twice greater and for higher stress -more than three times greater than the mass of scaling for unloaded concrete.
In case of the air-entrained concrete, the greater relative difference in mass of material scaled from concrete subjected to load was pointed out in comparison to control concrete.The tensile stress level c = 0,17 caused four times increase and stress level c = 0,50 -five times increase in the amount of scaled material.However, the air-entraining of concrete influenced the limitation of the susceptibility for scaling in comparison to the concrete without admixture.In considered range of external load of specimens in third point bending test, the rate of damage progress increased with the increase in applied stress value.
Considering the mass of scaled material after n cycles as a measure of accumulated damages, it is possible to evaluate the number of cycles after that, the scaling achieves unacceptable volume, for assessed stress level.Unacceptable level of scaling can be determined arbitrarily, regarding durability requirements, or on the basis of standards for different concrete elements.The dependence can be useful for predicting concrete ability to scaling according to the tensile stress level.

Changes in fracture parameters of concrete due to cyclic freezing and thawing
The fracture parameters were determined on the basis of P-CMOD curves obtained for concrete specimens.The effect of freezing and thawing on concrete properties was referee to the results obtained for reference specimens cured in water.The force P plotted versus CMOD measured for air-entrained concrete after 350 cycles and control concrete were presented in Fig. 6.From the P-CMOD plot, one can see that the initial part of the curve for reference concrete is alm and the strain of the notch tip under tension increases with increasing load.After the linear portion of curve, deviation from linear response is observed and the tension strain reaches the maximum va indicates the onset of crack initiation at the tip of the notch.After the point of maximum tension, exhibits increasing load until reaching the peak.Therefore, the load at which the tension reaches its value is the initial cracking load.For extremely damaged concrete, the linear portion of P-CMOD cu short, the maximum load is achieved quickly, and the strain softening is observed.In the process of d the concrete behaviour is more ductile in comparison to reference concrete but the maximum loa strongly limited.

Effects of Interaction of Static Load and Frost on Damage Mechanism of Concrete Elements
The fracture parameters s Ic K and CTOD c as well as the measured maximum load P max , the critica crack length a c related to depth of specimen d, Young's modulus E and compressive strength f cm dete concrete without AEA and for air-entrained concrete were presented in tables 1 and 2, respectively.T for frozen as well as reference specimens were given.From the P-CMOD plot, one can see that the initial part of the curve for reference concrete is almost linear and the strain of the notch tip under tension increases with increasing load.After the linear portion of P-CMOD curve, deviation from linear response is observed and the tension strain reaches the maximum value, which indicates the onset of crack initiation at the tip of the notch.After the point of maximum tension, the curve exhibits increasing load until reaching the peak.Therefore, the load at which the tension reaches its maximum value is the initial cracking load.For extremely damaged concrete, the linear portion of P-CMOD curve is very short, the maximum load is achieved quickly, and the strain softening is observed.In the process of degradation the concrete behaviour is more ductile in comparison to reference concrete but the maximum load (P max ) is strongly limited.
The fracture parameters and CTOD c as well as the measured maximum load P max , the critical effective crack length a c related to depth of specimen d, Young's modulus E and compressive strength f cm determined for concrete without AEA and for air-entrained concrete were presented in tables 1 and 2, respectively.The results for frozen as well as reference specimens were given.
The specimens of non-air-entrained concrete were subjected to 200 cycles of freezing and thawing.The fracture parameters undergo greater changes than compressive strength under frost attack.After initial slight improvement of properties, the fracture parameters were influenced by frost action.For frozen specimens the value of C to -18 o C. Every 7 days NaCl The material that scaled from the test surface was collected and dried to constant e scaled material per unit area after n cycles m n was evaluated for each measuring en.The specimens were subjected to 56 freeze-thaw cycles, the number suggested for determination acture parameters evaluation were subjected to cyclic freezing in air and thawing in anged from -18 o C to 18 o C. The duration of single cycle was 8 hours and the freezing rs.The freezing and thawing process was finished 1 day before testing.ntensity factor s Ic K and the critical crack tip opening displacement CTOD c were re described in RILEM draft recommendation (1990), based on the fracture model ah (1985).The fracture parameters were assessed in three-point bend test on beams pecimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal mm.The geometry of specimen and the way of load were presented in Fig. 2a.Each replicates.

b)
sting configuration and geometry of specimen: (a) the way of load; (b) the place of CMOD measurement ng machine with crack mouth opening displacement (CMOD) as the feedback signal le failure.The crack mouth opening displacement and the applied load were recorded est.The CMOD, indicated in Fig. 2b, was measured by means of clip gauge.To opening displacement (CMOD) a pair of knife edges was attached to two sides of a ower surface of the beam.The rate of loading was controlled by a constant rate of at the peak load was reached in 5 min.tonically loaded up to the maximum load.The applied load was reduced after the load e and was at about 95% of the peak load.Then, the applied load was reduced to zero lied.The specimen was cyclically loaded up to failure.
decreased and the CTOD c increased, which means that the fracture process (stable crack propagation) appeared for greater crack opening displacement.After 150 cycles the reference concrete was characterized by higher critical stress intensity factor and higher value of Young's modulus than frozen concrete, but the values of CTOD c were comparable for both of them.After 200 cycles significant decrease in the mechanical as well as fracture properties of concrete subjected to freezing and thawing was found.C. Every 7 days NaCl The material that scaled from the test surface was collected and dried to constant e scaled material per unit area after n cycles m n was evaluated for each measuring en.The specimens were subjected to 56 freeze-thaw cycles, the number suggested for determination racture parameters evaluation were subjected to cyclic freezing in air and thawing in anged from -18 o C to 18 o C. The duration of single cycle was 8 hours and the freezing rs.The freezing and thawing process was finished 1 day before testing.ntensity factor s Ic K and the critical crack tip opening displacement CTOD c were re described in RILEM draft recommendation (1990), based on the fracture model ah (1985).The fracture parameters were assessed in three-point bend test on beams pecimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal mm.The geometry of specimen and the way of load were presented in Fig. 2a The specimens of air-entrained concrete were subjected to 350 cycles.The results presented in table 2 show similar changes in fracture parameters to the concrete without AEA.The concrete microstructure deterioration due to freezing caused the decrease in the maximum load, critical stress intensity factor, Young's modulus and the increase in the critical crack tip opening displacement.
Generally, the air-entrained concrete is characterized by smaller value of C to -18 o C. Every 7 days NaCl was exchanged.The material that scaled from the test surface was collected and dried to constant The amount of the scaled material per unit area after n cycles m n was evaluated for each measuring n and each specimen.The specimens were subjected to 56 freeze-thaw cycles, the number suggested for fied concrete.
acture parameters determination e specimens for fracture parameters evaluation were subjected to cyclic freezing in air and thawing in he temperature changed from -18 o C to 18 o C. The duration of single cycle was 8 hours and the freezing uration was 6 hours.The freezing and thawing process was finished 1 day before testing.e critical stress intensity factor s Ic K and the critical crack tip opening displacement CTOD c were ned using procedure described in RILEM draft recommendation (1990), based on the fracture model ted by Jenq and Shah (1985).The fracture parameters were assessed in three-point bend test on beams tial notches.The specimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal m and width was 3 mm.The geometry of specimen and the way of load were presented in Fig. 2a.Each as composed of 4 replicates.

a)
than non-air-entrained concrete.C. Every 7 days NaCl he material that scaled from the test surface was collected and dried to constant scaled material per unit area after n cycles m n was evaluated for each measuring n.The specimens were subjected to 56 freeze-thaw cycles, the number suggested for determination acture parameters evaluation were subjected to cyclic freezing in air and thawing in nged from -18 o C to 18 o C. The duration of single cycle was 8 hours and the freezing s.The freezing and thawing process was finished 1 day before testing.tensity factor s Ic K and the critical crack tip opening displacement CTOD c were e described in RILEM draft recommendation (1990), based on the fracture model ah (1985).The fracture parameters were assessed in three-point bend test on beams ecimen sizes were 100×100×400 mm, and the initial saw-cut notch depth was equal mm.The geometry of specimen and the way of load were presented in Fig. 2a.Each eplicates.

b)
ting configuration and geometry of specimen: (a) the way of load; (b) the place of CMOD measurement g machine with crack mouth opening displacement (CMOD) as the feedback signal e failure.The crack mouth opening displacement and the applied load were recorded st.The CMOD, indicated in Fig. 2b, was measured by means of clip gauge.To pening displacement (CMOD) a pair of knife edges was attached to two sides of a wer surface of the beam.The rate of loading was controlled by a constant rate of t the peak load was reached in 5 min.onically loaded up to the maximum load.The applied load was reduced after the load and was at about 95% of the peak load.Then, the applied load was reduced to zero ed.The specimen was cyclically loaded up to failure.
[MN/m The air-void system resulting from air-entraining treatment forms additional pores in concrete microstructure, which are the stress concentrators responsible for fracture.
Even though the concrete is recognized to be quasibrittle material (Gettu et al. 1990), in the process of degradation the material showed more ductile characteristics than reference concrete.Through experimental study presented, it was found that steady crack propagation stage exist before unstable fracture.The longer critical effective crack length a c and greater value of CTOD c , needed for failure, is characteristic for damage concrete.Similar changes in concrete behavior were noticed by Hanjari et al. (2011) during bond properties examination and by Li et al. (2011) during testing the flexural fatigue influence on concrete frost resistance.
Conclusions Two types of concrete damage due to cyclic freezing and thawing were studied.The interaction of load and freeze-thaw cycles with chloride exposure regime on surface scaling process of concrete was analyzed.The internal cracking progress in concrete was characterized using fracture parameters.The tests were carried on for nonair-entrained as well as for air-entrained concretes.
As the results of investigations, it was found that interaction of load and cyclic freezing and thawing in the presence of de-icing salt accelerates the process of surface scaling of concrete.In considered range of stress subjected to beams in three-point bending test, the rate of damages accumulation increased with the increase in stress, but the rate of damage accumulation was changing during longterm test.
The complete P-CMOD curves were measured from fracture tests on both frozen and reference concrete specimens.The fracture parameters of concrete were strongly influenced by cyclic freezing and thawing.It was found that the material damaged, due to cyclic freezing and thawing, is more ductile than undamaged one.The critical stress intensity factor and crack tip opening displacement can be valuable measures described the concrete degradation due to accumulation of physical damages in its microstructure.

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point bending test

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point bending test

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point b

.
Fracture testing configuration and geometry of specimen: (a) the way of load; (b) the place of CMOD measurement d-loop testing machine with crack mouth opening displacement (CMOD) as the feedback signal ieve a stable failure.The crack mouth opening displacement and the applied load were recorded uring the test.The CMOD, indicated in Fig. 2b, was measured by means of clip gauge.To ack mouth opening displacement (CMOD) a pair of knife edges was attached to two sides of a d on the lower surface of the beam.The rate of loading was controlled by a constant rate of MOD so that the peak load was reached in 5 min.
P max -the measured maximum load, W 0 -self-weight of the beam.

Fig. 5 .
Fig. 5. Mean mass of scaled material m vs. freeze-thaw cycles number n as well as stress level c for air-entrained concrete

Fig. 6 .
Fig. 6.P-CMOD curves for air-entrained concrete after 350 cycles and reference concrete cured in wat

Fig. 6 .
Fig. 6.P-CMOD curves for air-entrained concrete after 350 cycles and reference concrete cured in water

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point bending test

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point bending test ation was 24 hours with temperature change from 20 oC to -18 o C. Every 7 days NaCl The material that scaled from the test surface was collected and dried to constant e scaled material per unit area after n cycles m n was evaluated for each measuring en.The specimens were subjected to 56 freeze-thaw cycles, the number suggested for

Fig. 1 .
Fig. 1.Sketch of specimen subjected to three-point bending test

Table 1 .
Properties of concrete without air-entraining agent

Table 1 .
Properties of concrete without air-entraining agent