- Email: [email protected]
Damage analysis of concrete members containing expansive agent by mechanical and acoustic methods
Engineering Failure Analysis 74 (2017) 95–106
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Damage analysis of concrete members containing expansive agent by mechanical and acoustic methods Qiang Xia a,b, Hua Li a,b, Anqun Lu a,b, Qian Tian a,b,⁎, Jiaping Liu c,⁎⁎ a b c
State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing, 210008, China Jiangsu Sobute New Materials Co., Ltd, Nanjing, 211103, China College of Materials Science and Engineering, Southeast University, Nanjing, 211189, China
a r t i c l e
i n f o
Article history: Received 24 August 2016 Received in revised form 22 November 2016 Accepted 29 December 2016 Available online 04 January 2017 Keywords: Cement Expansive agent Mechanical strength Nondestructive tests
a b s t r a c t Shrinkage compensating concrete (SCC) and Self-stressing concrete (SSC) technique have been employed for reducing early-age cracking and leakage while the addition of expansive agent would have a negative impact on mechanical properties and durability. The objective of the current research was to quantitatively assess the damage development in cementitious materials with expansive agent by both the strength tests and nondestructive acoustic tests including ultrasonic measurements and acoustic emission (AE) tests. The damage degree was defined based on strength as well as ultrasonic properties and a significant linear relationship was observed between the damage degree and autogenous strains. AE parameters such as AE amplitude, AE counts and AE energy were related to AE activity of the cement-expansive agent system. Crack mode identification was performed based on the relationship between average frequency and RA value (rise time/amplitude). A decreasing ratio of tensile cracks and an increasing ratio of shear cracks were observed which could be an indication of aggravated damage inside the materials. © 2017 Published by Elsevier Ltd.
1. Introduction Shrinkage compensating concrete (SCC) technique has been applied to a wide range of structures as an effective method for reducing cracking and leakage [1–3] in cement based materials due to all kinds of shrinkages at early ages. The special products used to prepare SCC would react with water and lead to expansion in the concrete to maintain the volume stability [4]. For example, the transformation of calcium oxide to calcium hydroxide would increase its volume by about 90% [5]. In addition, the socalled Self-stressing concrete (SSC) could be produced by adding the expansive agent to cement based composites reinforced by steel bars. Therefore the expansion is restrained by steel bars and the steel bars are actually in the tensile state in which case a compressive stress would be created, typically 3– 6 MPa [6]. In theory an appropriate amount of expansion would lead to a denser structure [7] and improved impermeability of concrete as presented by Sun et al. [8]. Nevertheless decreases of mechanical properties including the compressive and splitting tensile strengths as well as elastic modulus were also observed by Meddah et al. [9] using the combination of expansive agent and
⁎ Correspondence to: Q. Tian, State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing 210008, China. ⁎⁎ Correspondence to: J. Liu, School of Materials Science and Engineering, Southeast University, Nanjing 211189, China. E-mail addresses: [email protected] (Q. Tian), [email protected] (J. Liu).
http://dx.doi.org/10.1016/j.engfailanal.2016.12.020 1350-6307/© 2017 Published by Elsevier Ltd.
96
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
shrinkage reducing admixture. The mix design method of SCC/SSC is not standardized and the expansion performance of cement based materials depends on many factors such as the expansive agent dosage, water to cement ratio, sand ratio and curing condition. Improper expansion (e.g., a too much expansion) would lead to inappropriate residual expansive stress in the matrix and thus a higher possibility of micro-cracks in the interface between the hardened cement composites and the expansive products that forms later. The risk of possible damage from expansive agent should also be paid enough attention like expansion damage caused by delayed ettringite formation (DEF) [10] and alkali-silica reaction (ASR) [11]. Concerning damage assessment a series of modellings have been proposed and the damage degree could be estimated from the decrease in mechanical strength (e.g., compressive strength) as Eq. (1) [12–13]
dðt Þ ¼ 1−
f ðt Þ f ðt 0 Þ
ð1Þ
where d(t) is damage degree, f(t) was the mechanical strength at time t and f(t0) was the initial strength. The definition of damage degree d in Eq. (1) is fundamental as strength property is considered as one of the major mechanical properties in materials design. However strength test is usually destructive and the specimens prepared in the laboratory fail to reflect the properties in the real structure. Nondestructive tests such as ultrasonic pulse velocity (UPV) and acoustic emission (AE) which show promising application have been widely adopted for evaluation of material properties and damage condition in the last decades [14–16]. As a non-destructive test, ultrasonic pulse velocity has been widely applied for describing degradation and damage development [17–18] inside materials, particularly for metal materials. And it had also been adopted as a useful tool for determination of materials properties [19] and assessment of different kinds of damage [20–21] in concrete. AE tests could capture any elastic wave in materials due to micro cracking in real time monitoring. AE parameters like counts, energy, amplitude, and frequency are sensitive in detecting active crack evolution inside the structures [22–23]. This paper carried out a comprehensive investigation on the damage development induced by expansive agent. The cementexpansive agent system was adopted and different dosages of expansive agent were added to produce different expansion magnitude. Autogenous deformations were measured. Both the mechanical tests and non-destructive acoustic tests including ultrasonic measurements and AE tests were performed to provide a better understanding of damage development. 2. Material and methods 2.1. Materials and mix design Blended Portland cement with a strength class 42.5 conforming to the Chinese national standard GB 175-2007 was used in this study. The Bogue phase composition was 54.1% C3S, 18.6% C2S, 6.3% C3A, 9.4% C4AF and the detailed chemical compositions were shown in Table 1. A compound CaO based expansive agent was employed. The chemical compositions of expansive agent were given in Table 1 and the phase composition from X-ray diffraction analysis was shown in Fig. 1. It was observed that the main components of the expansive agent was calcium oxide (CaO), anhydrite (CaSO4) and ye'elimite (4CaO·3Al2O3·SO3). No superplasticizer was added during the mix program. A total of five types of cement pastes with different dosages of expansive agent were prepared with a water to binder ratio of 0.35 by mass. The amount of expansive agent was 0, 2%, 4%, 6%, 8% (by mass of total binder), respectively. 2.2. Autogenous deformation Three 40 × 40 × 160 mm prisms which had copper probes at two sides were used to measure the autogenous deformation for each mixture by using the universal projection length measuring instrument. The specimens were covered with plastic sheets and sealed completely using aluminum tape to prevent water evaporation immediately after demoulding at 1 day. The specimens were placed in a 20 ± 2 °C, 60 ± 5% RH environment. 2.3. Strength tests Flexural strength and compressive strength were tested for mechanical strength conforming to the Chinese national standard GB/T 17671. Three prisms with the dimensions of 40 × 40 × 160 mm were tested for each mixture. In order to make comparisons Table 1 Chemical compositions of cement and expansive agent. Material
w/% Al2O3
CaO
Fe2O3
K2O
MgO
Na2O
SO3
SiO2
LOSS
Cement Expansive agent
4.34 6.45
62.16 68.30
3.08 0.51
0.72 –
1.71 1.39
0.12 –
2.68 20.80
20.74 2.22
4.45 0.33
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
97
Fig. 1. XRD pattern of expansive agent in the experiment.
with other tests all specimens were sealed with plastic sheets after demoulding at 1 day and placed in the same environment (20 ± 2 °C, 60 ± 5% RH). Strengths at 3, 7, and 14 days were recorded (after the demoulding day). 2.4. Ultrasonic measurements During the experiment, ultrasonic pulse velocity instrumentation containing a transmitter generating P-wave and a receiving transducer was used to assess the ultrasonic velocity in the specific material [24]. In the direct transmission, transmitter and receiving transducer were fixed at two sides of the 40 × 40 × 160 mm prism (sealed with plastic sheets). The transmitter and the receiving transducer were 38 mm in diameter with a central frequency of 50 kHz. The driving voltage was 500 V and the sampling period was 0.4 μs. Silicon grease was used as a coupling agent in order to guarantee full contact between the sensors and the specimen (Fig. 2). 2.5. AE tests The acoustic emission test system in this experiment from Physical Acoustics Corporation (PAC) consisted of a PCI-2 acoustic card, piezoelectric based acoustic sensors, selectable pre-amplifiers and a high performance data acquisition system which contained six channels (Fig. 3). The sensors used were R6α resonant type sensors with the operating frequency range of 35– 100 kHz. When the pre-amplifier was selected as 40 dB (in a previous study a similar setup was used [25]) the measuring range was 0– 80 dB with the accuracy of 0.5 dB. The sampling rates of data acquisition system were as high as up to 1 MSPS (mega samples per second). During the experiment the prisms for test were sealed with plastic sheets in case of moisture loss. To capture the accurate signals generated inside the materials the AE sensors were fixed on the prisms at the center using silicon grease as a coupling agent
Fig. 2. Illustration of direct transmission in ultrasonic measurement.
98
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 3. AE test system in the experiment.
[26]. The measured signals were then filtered and processed by the software packages AEWin obtained from PAC. Noise tests were performed to choose an appropriate amplitude threshold for elimination of possible environment noise. After a series of careful adjustments the amplitude threshold was chosen as 28 dB. 3. Results and discussion 3.1. Autogenous deformation Strains of specimens with different dosages of expansive agent in sealed condition in the first fourteen days were measured and plotted in Fig. 4. Compared with autogenous shrinkage of the reference sample (without expansive agent, the same below) the specimens with expansive agent showed continuous expansion of different magnitudes. Most of expansion was observed in the first three days which could be due to fast hydration of expansive components like calcium oxide [27]. The length of specimens along the direction of testing showed a gradual decreasing trend at later ages. For instance, relative shrinkage (compared with the strain at the testing day before) was measured after one day for the specimen with 2% expansive agent. The turning points from expansion to relative shrinkage were the second day for 4%, the third day for 6% and the fifth day for 8%, respectively. It should be noted that the expansion was not strictly proportional to the amount of expansive agent added and for the specimen with 8% expansive agent, a much larger expansion was measured, nearly 93 times and 7 times as large as that with 4% and 6% expansive agent, respectively. 3.2. Strength properties Both the flexural strength and compressive strength tests of the specimens were performed and the results were plotted in Figs. 5 and 6. It was observed that although expansive agent had an advantage of compensating for shrinkage the inclusion of
Fig. 4. Autogenous deformation of specimens with different dosages of expansive agent.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
99
Fig. 5. Flexural strength results.
expansive agent would influence the strength development over the curing ages. A higher dosage of expansive agent caused larger decreases in both flexural strength and compressive strength and the largest decreases were observed for the specimen with 8% expansive agent at different ages. Compared with the reference sample the specimen with 8% expansive agent (surface crack as shown in Fig. 7) had decreases of 29%, 19% and 21% in flexural strength and 56%, 43%, and 46% in compressive strength at 3, 7, and 14 days, respectively. It was reported expansive products such as calcium hydroxide and ettringite [28] would be formed due to the hydration of expansive agent, which could have an impact of the cement hydration process since both the cement content and water content were changed [2–3]. Meanwhile the replacement of cement by expansive agent naturally caused a strength loss. In addition, expansive products formed would create residual expansive stress in the contact surface with the hardened cement composites, which could lead to weaker interfacial bonding and the possible micro-cracks along the interface. For easier quantification of strength damage compared with the reference sample, Eq. (1) could be revised by taking the reference sample strength at the corresponding age into consideration as in Eq. (2).
dðt Þ ¼ 1−
f ðt Þ f r ðt Þ
ð2Þ
where f(t) was the strength of the specific sample and fr(t) was the reference sample strength. The damage degree of both the flexural strength and compressive strength for all the specimens at 14 days were calculated by Eq. (2) and the relationship between the strength damage degree and autogenous strains was shown in Fig. 8. A significant linear relationship was observed between the damage degree and autogenous strains, with the correlation coefficients of 0.9964 for the flexural strength and 0.9725 for the compressive strength. Meanwhile different slops of the regression lines indicated that the damage degree of the compressive strength were more sensitive to autogenous deformation.
Fig. 6. Compressive strength results.
100
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 7. Surface crack of the specimen with 8% expansive agent.
3.3. Ultrasonic pulse velocity In the direct transmission test the ultrasonic travel time and the true length of the prisms along the direction of testing (obtained by a vernier caliper) were used to calculate the ultrasonic pulse (P-wave) velocity [29] in the specific specimen as shown in Fig. 9. It was observed that for the specimens with no more than 6% expansive agent (including the reference sample) the ultrasonic velocity increased with the curing age in the densification process with cement hydration. Most of increases of ultrasonic velocity were observed at early ages, namely the first three days. The ultrasonic velocity at 7 days was almost same around 3.7 km/s for the specimens with 0, 2% and 4% expansive agent, but the specimen with 6% expansive agent had relatively lower velocity, 3.6 km/s, which could be due to damage from larger expansion. In particular, the ultrasonic velocity for the specimen with 8% expansive agent was far lower than that of other specimens (3.1 km/s at 7 days). Meanwhile a decrease of ultrasonic velocity was observed at the first day when a huge expansion of 9127 με (Fig. 4) was measured, after which the ultrasonic velocity increased with the curing age. Tirupan et al. [30] reported that the velocity of P-wave increased with the increase of material density. It could be expected that hydration of both the cement and expansive agent would cause reduction of the porosity in the matrix and lead to a denser structure. At the same time, the generation of residual expansive stress from expansion would lead to degradation and microcracking in the composites. The ultrasonic velocity was influenced by the two kinds of opposite effects and it was observed
Fig. 8. The relationship between strength damage degree and autogenous strains.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
101
Fig. 9. Ultrasonic pulse velocity in the specimens.
from Fig. 8 that the enhancement effect played a dominant role in the specimen with 2%, 4% expansive agent while degradation effect behaved in the specimen with 6%, 8% expansive agent. The damage degree based on ultrasonic pulse velocity could also be defined as Eq. (3). 0
d ðt Þ ¼ 1−
vðt Þ vr ðt Þ
ð3Þ
where v(t) was the ultrasonic velocity of the specific sample and vr(t) was the ultrasonic velocity of the reference sample, respectively. The damage degree based on the ultrasonic velocity for all the specimens at 14 days were calculated by Eq. (3) and the relationship between the damage degree d′ and autogenous strains was shown in Fig. 10. A significant linear relationship was also observed between the damage degree and autogenous strains with a correlation coefficients of 0.9977. By taking the absolute value of regression line slope from damage degree based on ultrasonic properties as well as strength properties into consideration it could be found that compressive strength showed larger decreasing slope than ultrasonic velocity and flexural strength. 3.4. AE activity 3.4.1. AE parameters AE parameters indicating the AE activity of the cement-expansive agent system were tested, e.g., AE amplitude, AE counts, cumulative AE counts, and AE energy. The results were shown in Figs. 11–14. The amplitude threshold for measurement was set as 28 dB and it was observed AE amplitude fell in the range from 28 dB to 48 dB indicating the AE activity due to inner expansion had relatively lower energy [31]. AE counts could be used as a representation of a damage event that occurred within the materials [32]. In Fig. 12 AE counts were observed during the whole experiment for all specimens and the specimens with higher dosage of expansive agent had
Fig. 10. The relationship between damage degree d′ and autogenous strains.
102
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 11. AE amplitudes for the specimens.
relatively more AE counts (Fig. 13). It should also be noted that AE counts were mostly at early ages, namely in the first 50 h for higher dosages of expansive agent (6%, 8%) consistent with quick development of expansion. Fig. 13 showed cumulative AE counts for the specimens and it could be observed generally specimens with more expansive agent had more cumulative AE counts indicating higher AE activity. The 6% sample showed higher cumulative AE counts at early age (before 30 h) but lower cumulative AE counts at later age than the 4% sample. As the AE signals detected were in the form of voltage and the AE counts measured experimentally depends on many factors such as AE sensors coupling to cement pastes, sensor sensitivity, source-sensor distance, etc. [33]. Analysis based on only one parameter was inadequate and thus it was significantly necessary to take AE energy released [34] into consideration. The cumulative AE energy for all specimens was shown in Fig. 14. As the AE energy gave a reflection about the true energy of any AE event or hit captured by AE station and could be related to the energy released in microcracking process. It was observed that a higher amount of energy released was recorded when a larger dosage of expansive agent was added because of larger expansion. The possible contradiction between cumulative AE counts and cumulative AE energy (4% and 6%) could be due to that AE energy associated with these AE events was relatively lower.
3.4.2. Crack mode identification Crack mode identification is fundamental to better understanding about damage properties [35] in the materials during expansion process. The mechanism by analysis of AE parameters lies in different elastic wave characteristics induced by different types of movement of the crack tips [36]. In tensile crack mode, volumetric change would cause the opposing displacement at the crack tips during which the longitudinal wave was emitted. However in shear crack mode, movements of the crack tips were in parallel to the crack plane and would lead to a change in shape instead of volume change.
Fig. 12. AE counts for the specimens.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
103
Fig. 13. Cumulative AE counts for the specimens.
Propagation and reflection in a heterogeneous medium would cause distortion and change AE waveform parameters like duration and rise time, especially for large size structures with complicated geometry [37]. In laboratory scale, assuming that distortion and attenuation are limited due to smaller dimensions and short propagation distances, crack classification could be studied based on RA value and average frequency. Ohtsu et al. [38–39] proposed a method to classify cracks based on the relationship between average frequency faverage and RA value using AE parameters. It was reported that AE signals which had lower RA value and higher average frequency could be an indication of tensile crack. Meanwhile, for AE signals which had high RA value and low average frequency it could be said that shear crack occurred. The RA value and average frequency faverage could be calculated from AE parameters as defined in Eqs. (4) and (5). RA ¼
t rise V
f average ¼
ð4Þ
N t dur
ð5Þ
where trise is the rise time, ms; V is the maximum amplitude, V; N is the AE ring-down counts, tdur is the duration time, μs. In AE test the amplitude is defined as the peak voltage (V) of an event and then transformed to AdB by logarithm. As a preamplifier (40 dB) was used, the relationship between voltage and dB was shown in Eq. (6) AdB ¼ 20 lg
V −40 V ref
ð6Þ
Fig. 14. Cumulative AE energy for the specimens.
104
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 15. Crack classification (specimens with 0 expansive agent).
Fig. 16. Crack classification (specimens with 8% expansive agent).
where AdB is AE amplitude (dB), V is the peak voltage of the AE waveform and Vref is the reference voltage, namely 1 μV. The RA value and average frequency for all detected AE hits were calculated from Eqs. (4) and (5) according to JCMS-III B5706 [40] and the proportion of the RA value and the average frequency was tentatively likewise chosen as 1:80 as stated by Ohtsu [39] for relative comparisons. Figs. 15 and 16 showed the relationship between average frequency and RA value for the specimens
Fig. 17. Ratio of tensile cracks and shear cracks for all specimens.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
105
with 0 and 8% expansive agent. It was observed that 62.2% of cracks were identified as tensile cracks for the reference sample while 67.4% of cracks were identified as shear cracks for the specimen with 8% expansive agent. Fig. 17 showed the variations of ratio of tensile cracks and shear cracks for all specimens with different dosages of expansive agent during the experiment. A decreasing ratio of tensile cracks and an increasing ratio of shear cracks were observed. As expansive agent had a shrinkage compensation effect residual tensile stress induced by shrinkage was largely relaxed or even eliminated. In addition, expansive stress induced by expansive products in the interface could lead to micro cracks including both tensile cracks and shear cracks when the volume expansion was assumed to be isotropic [10]. Lots of studies indicated that for cementitious materials tensile crack mode was dominated at the early stage of failure and shear crack mode was mostly active at the later stage [35]. Therefore the increase of shear cracks could be an indication of aggravated damage inside the materials. 4. Conclusions In this paper the damage development induced by expansive agent was investigated for concrete members by both mechanical and acoustic measurements. The results indicated that the inclusion of expansive agent would have an influence on the strength development and ultrasonic velocity over time. The hydration of the cement and expansive agent would cause reduction of the porosity in the matrix and lead to a denser structure. On contrast, the generation of residual expansive stress from expansion would lead to degradation and microcracking in the composites. A significant linear relationship was observed between the damage degree and autogenous expansion strains which was not strictly proportional to the amount of expansive agent added. According to this feature, the balance of shrinkage compensation and property deterioration should be paid enough attention to meet practical requirements when using expansive agent. Ultrasonic measurement offered a useful tool for determination of materials properties by ultrasonic velocity in the specific materials. Meanwhile AE activity could be assessed by AE parameter analysis and related to inner micro-cracking evolution of the cement-expansive agent system. Crack mode could be identified based on the relationship between average frequency and RA value. The addition of expansive agent lead to a decreasing ratio of tensile cracks and an increasing ratio of shear cracks which could be an indication of aggravated damage. Such nondestructive methods showed a promising future in better understanding of the damage development. Acknowledgements This research was financially supported by National Basic Research Program of China (No. 2015CB655105), National Outstanding Youth Science Foundation Program of China (No.51225801) and the Provincial Science and Technology Cooperation Project of Jiangsu Province-Jiangsu-Guangxi Cooperation Project (BM2014050). References [1] R.G. Colin, Cement and concrete as an engineering material: an historic appraisal and case study analysis, Eng. Fail. Anal. 40 (2014) 114–140. [2] C. Valeria, N. Alessandro, D. Jacopo, The influence of expansive agent on the performance of fibre reinforced cement-based composites, Constr. Build. Mater. 91 (2015) 171–179. [3] H. Jianguo, J. Di, Y. Peiyu, Understanding the shrinkage compensating ability of type K expansive agent in concrete, Constr. Build. Mater. 116 (2016) 36–44. [4] M. Collepardi, R. Troli, M. Bressan, et al., Crack-free concrete for outside industrial floors in the absence of wet curing and contraction joints, Cem. Concr. Compos. 30 (10) (2008) 887–891. [5] S. Chatterji, Mechanism of expansion of concrete due to the presence of dead burnt CaO and MgO, Cem. Concr. Res. 25 (1995) 51–56. [6] H. He, W. Dong, Z. Wu, Study on long-term expansive deformation of selfstressing concrete with combined restrictions of steel fibers and steel bar, Key Eng. Mater. 452–3 (2011) 533–536. [7] R. Sahamitmongkol, T. Kishi, Tension stiffening effect and bonding characteristics of chemically prestressed concrete under tension, Mater. Struct. 44 (2011) 455–474. [8] S. Wei, C. Huisu, X. Luo, Q. Hongpin, The effect of hybrid fibers and expansive agent on the shrinkage and permeability of high-performance concrete, Cem. Concr. Res. 31 (2001) 595–601. [9] M.S. Meddah, M. Suzuki, R. Sato, Influence of a combination of expansive and shrinkage-reducing admixture on autogenous deformation and self-stress of silica fume high-performance concrete, Constr. Build. Mater. 25 (2011) 239–250. [10] B. Hassina, M. Stephane, S. Alain, et al., Effects of restraint on expansion due to delayed ettringite formation, Cem. Concr. Res. 42 (2012) 1024–1031. [11] R. Patrice, B. Gérard, Assessment of the expansion related to alkali-silica reaction by the Damage Rating Index method, Constr. Build. Mater. 19 (2) (2005) 83–90. [12] A. Sellier, E. Bourdarot, S. Multon, et al., Combination of structural monitoring and laboratory tests for the assessment of AAR-swelling - application to a gate structure dam, ACI Mater. J. 106 (3) (2009) 281–290. [13] B. Capra, A. Sellier, Orthotropic modelling of alkali-aggregate reaction in concrete structures: numerical simulations, Mech. Mater. 35 (2003) 817–830. [14] H.D. Yun, W.C. Choi, S.Y. Seo, Acoustic emission activities and damage evaluation of reinforced concrete beams strengthened with CFRP sheets, NDT & E Int. 43 (7) (2010) 615–628. [15] M. Ohtsu, H. Watanabe, Quantitative damage estimation of concrete by acoustic emission, Constr. Build. Mater. 15 (5–6) (2001) 217–224. [16] T.P. Philippidis, D.G. Aggelis, Experimental study of wave dispersion and attenuation in concrete, Ultrasonics 43 (2005) 584–595. [17] J.B. Leonard, F.K. William, M.F. Dan, Improved assessment of mass concrete dams using acoustic travel time tomography: part I-theory, Constr. Build. Mater. 14 (2000) 133–146. [18] F.K. William, J.B. Leonard, M.F. Dan, Improved assessment of mass concrete dams using acoustic travel time tomography: part II-application, Constr. Build. Mater. 14 (2000) 147–156. [19] A. Vary, in: R.S. Sharpe (Ed.), Ultrasonic Measurement of Material Properties, In Research Techniques in NDT, 4, Academic Press, London 1980, pp. 159–204. [20] J. Lemaitre, J.L. Chaboche, Mechanics of Solid Materials, Cambridge University Press, 1989. [21] A.A. Shah, Y. Ribakov, Non-destructive evaluation of concrete in damaged and undamaged states, Mater. Des. 30 (9) (2009) 3504–3511. [22] M. Ohtsu, M. Shigeishi, H. Iwase, et al., Determination of crack location, type and orientation in concrete structures by acoustic emission, Mag. Concr. Res. 43 (155) (1991) 127–134. [23] K.M. Holford, Acoustic emission in structural health monitoring, Key Eng. Mater. 413–4 (2009).
106
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
[24] ASTM C597–97, Standard test method for pulse velocity through concrete, Annual Book of ASTM Standards, American Society of Testing and Materials, West Conshohocken, PA, 2003. [25] D. Soulioti, N.M. Barkoula, A. Paipetis, et al., Acoustic emission behavior of steel fibre reinforced concrete under bending, Constr. Build. Mater. 23 (2009) 3532–3536. [26] D.Y. Hyun, C.C. Won, Y.S. Soo, Acoustic emission activities and damage evaluation of reinforced concrete beams strengthened with CFRP sheets, NDT&E Int. 43 (7) (2010) 615–628. [27] S. Maltese, C. Pistolesi, A. Lolli, et al., Combined effect of expansive and shrinkage reducing admixtures to obtain stable and durable mortars, Cem. Concr. Res. 35 (12) (2005) 2244–2251. [28] A. Telesca, M. Marroccoli, M.L. Pace, et al., A hydration study of various calcium sulfoaluminate cements, Cem. Concr. Compos. 53 (2014) 224–232. [29] S. Selleck, E.N. Landis, M.L. Peterson, et al., Ultrasonic investigation of concrete with distributed damage, ACI Mater. J. 95 (1998) 27–36. [30] M. Tirupan, M.T. James, B.E. Tuncer, Non-destructive testing of cementitiously stabilized materials using ultrasonic pulse velocity test, Transp. Geotech. 6 (2016) 97–107. [31] Carpinteri, G. Lacidogna, G. Niccolini, Damage analysis of reinforced concrete buildings by the acoustic emission technique, Struct. Control. Health Monit. 18 (6) (2011) 660–673. [32] D. Prem, J.C. Kapil, K. Ashavani, Acoustic emission characteristics and b-value estimate in relation to waveform analysis for damage response of snow, Cold Reg. Sci. Technol. 119 (2015) 170–182. [33] D.G. Aggelis, D.V. Soulioti, N. Sapouridis, et al., Acoustic emission characterization of the fracture process in fibre reinforced concrete, Constr. Build. Mater. 25 (2011) 4126–4131. [34] E.N. Landis, L. Baillon, Experiment to relate acoustic emission energy to fracture energy of concrete, J. Eng. Mech. 128 (6) (2002) 698–702. [35] D.G. Aggelis, A.C. Mpalaskas, T.E. Matikas, Investigation of different fracture modes in cement-based materials by acoustic emission, Cem. Concr. Res. 48 (2013) 1–8. [36] D.G. Aggelis, Classification of cracking mode in concrete by acoustic emission parameters, Mech. Res. Commun. 38 (2011) 153–157. [37] D.G. Aggelis, A.C. Mpalaskas, D. Ntalakas, et al., Effect of wave distortion on acoustic emission characterization of cementitious materials, Constr. Build. Mater. 35 (2012) 183–190. [38] M. Ohtsu, T. Isoda, Y. Tomoda, Acoustic emission techniques standardized for concrete structures, J. Acoust. Emission 25 (2007) 21–32. [39] K. Ohno, M. Ohtsu, Crack classification in concrete based on acoustic emission, Constr. Build. Mater. 24 (12) (2010) 2339–2346. [40] JCMS-III B5706, Monitoring Method for Active Cracks in Concrete by Acoustic Emission, Federation of Construction Materials Industries, Japan, 2003.
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Damage analysis of concrete members containing expansive agent by mechanical and acoustic methods Qiang Xia a,b, Hua Li a,b, Anqun Lu a,b, Qian Tian a,b,⁎, Jiaping Liu c,⁎⁎ a b c
State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing, 210008, China Jiangsu Sobute New Materials Co., Ltd, Nanjing, 211103, China College of Materials Science and Engineering, Southeast University, Nanjing, 211189, China
a r t i c l e
i n f o
Article history: Received 24 August 2016 Received in revised form 22 November 2016 Accepted 29 December 2016 Available online 04 January 2017 Keywords: Cement Expansive agent Mechanical strength Nondestructive tests
a b s t r a c t Shrinkage compensating concrete (SCC) and Self-stressing concrete (SSC) technique have been employed for reducing early-age cracking and leakage while the addition of expansive agent would have a negative impact on mechanical properties and durability. The objective of the current research was to quantitatively assess the damage development in cementitious materials with expansive agent by both the strength tests and nondestructive acoustic tests including ultrasonic measurements and acoustic emission (AE) tests. The damage degree was defined based on strength as well as ultrasonic properties and a significant linear relationship was observed between the damage degree and autogenous strains. AE parameters such as AE amplitude, AE counts and AE energy were related to AE activity of the cement-expansive agent system. Crack mode identification was performed based on the relationship between average frequency and RA value (rise time/amplitude). A decreasing ratio of tensile cracks and an increasing ratio of shear cracks were observed which could be an indication of aggravated damage inside the materials. © 2017 Published by Elsevier Ltd.
1. Introduction Shrinkage compensating concrete (SCC) technique has been applied to a wide range of structures as an effective method for reducing cracking and leakage [1–3] in cement based materials due to all kinds of shrinkages at early ages. The special products used to prepare SCC would react with water and lead to expansion in the concrete to maintain the volume stability [4]. For example, the transformation of calcium oxide to calcium hydroxide would increase its volume by about 90% [5]. In addition, the socalled Self-stressing concrete (SSC) could be produced by adding the expansive agent to cement based composites reinforced by steel bars. Therefore the expansion is restrained by steel bars and the steel bars are actually in the tensile state in which case a compressive stress would be created, typically 3– 6 MPa [6]. In theory an appropriate amount of expansion would lead to a denser structure [7] and improved impermeability of concrete as presented by Sun et al. [8]. Nevertheless decreases of mechanical properties including the compressive and splitting tensile strengths as well as elastic modulus were also observed by Meddah et al. [9] using the combination of expansive agent and
⁎ Correspondence to: Q. Tian, State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing 210008, China. ⁎⁎ Correspondence to: J. Liu, School of Materials Science and Engineering, Southeast University, Nanjing 211189, China. E-mail addresses: [email protected] (Q. Tian), [email protected] (J. Liu).
http://dx.doi.org/10.1016/j.engfailanal.2016.12.020 1350-6307/© 2017 Published by Elsevier Ltd.
96
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
shrinkage reducing admixture. The mix design method of SCC/SSC is not standardized and the expansion performance of cement based materials depends on many factors such as the expansive agent dosage, water to cement ratio, sand ratio and curing condition. Improper expansion (e.g., a too much expansion) would lead to inappropriate residual expansive stress in the matrix and thus a higher possibility of micro-cracks in the interface between the hardened cement composites and the expansive products that forms later. The risk of possible damage from expansive agent should also be paid enough attention like expansion damage caused by delayed ettringite formation (DEF) [10] and alkali-silica reaction (ASR) [11]. Concerning damage assessment a series of modellings have been proposed and the damage degree could be estimated from the decrease in mechanical strength (e.g., compressive strength) as Eq. (1) [12–13]
dðt Þ ¼ 1−
f ðt Þ f ðt 0 Þ
ð1Þ
where d(t) is damage degree, f(t) was the mechanical strength at time t and f(t0) was the initial strength. The definition of damage degree d in Eq. (1) is fundamental as strength property is considered as one of the major mechanical properties in materials design. However strength test is usually destructive and the specimens prepared in the laboratory fail to reflect the properties in the real structure. Nondestructive tests such as ultrasonic pulse velocity (UPV) and acoustic emission (AE) which show promising application have been widely adopted for evaluation of material properties and damage condition in the last decades [14–16]. As a non-destructive test, ultrasonic pulse velocity has been widely applied for describing degradation and damage development [17–18] inside materials, particularly for metal materials. And it had also been adopted as a useful tool for determination of materials properties [19] and assessment of different kinds of damage [20–21] in concrete. AE tests could capture any elastic wave in materials due to micro cracking in real time monitoring. AE parameters like counts, energy, amplitude, and frequency are sensitive in detecting active crack evolution inside the structures [22–23]. This paper carried out a comprehensive investigation on the damage development induced by expansive agent. The cementexpansive agent system was adopted and different dosages of expansive agent were added to produce different expansion magnitude. Autogenous deformations were measured. Both the mechanical tests and non-destructive acoustic tests including ultrasonic measurements and AE tests were performed to provide a better understanding of damage development. 2. Material and methods 2.1. Materials and mix design Blended Portland cement with a strength class 42.5 conforming to the Chinese national standard GB 175-2007 was used in this study. The Bogue phase composition was 54.1% C3S, 18.6% C2S, 6.3% C3A, 9.4% C4AF and the detailed chemical compositions were shown in Table 1. A compound CaO based expansive agent was employed. The chemical compositions of expansive agent were given in Table 1 and the phase composition from X-ray diffraction analysis was shown in Fig. 1. It was observed that the main components of the expansive agent was calcium oxide (CaO), anhydrite (CaSO4) and ye'elimite (4CaO·3Al2O3·SO3). No superplasticizer was added during the mix program. A total of five types of cement pastes with different dosages of expansive agent were prepared with a water to binder ratio of 0.35 by mass. The amount of expansive agent was 0, 2%, 4%, 6%, 8% (by mass of total binder), respectively. 2.2. Autogenous deformation Three 40 × 40 × 160 mm prisms which had copper probes at two sides were used to measure the autogenous deformation for each mixture by using the universal projection length measuring instrument. The specimens were covered with plastic sheets and sealed completely using aluminum tape to prevent water evaporation immediately after demoulding at 1 day. The specimens were placed in a 20 ± 2 °C, 60 ± 5% RH environment. 2.3. Strength tests Flexural strength and compressive strength were tested for mechanical strength conforming to the Chinese national standard GB/T 17671. Three prisms with the dimensions of 40 × 40 × 160 mm were tested for each mixture. In order to make comparisons Table 1 Chemical compositions of cement and expansive agent. Material
w/% Al2O3
CaO
Fe2O3
K2O
MgO
Na2O
SO3
SiO2
LOSS
Cement Expansive agent
4.34 6.45
62.16 68.30
3.08 0.51
0.72 –
1.71 1.39
0.12 –
2.68 20.80
20.74 2.22
4.45 0.33
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
97
Fig. 1. XRD pattern of expansive agent in the experiment.
with other tests all specimens were sealed with plastic sheets after demoulding at 1 day and placed in the same environment (20 ± 2 °C, 60 ± 5% RH). Strengths at 3, 7, and 14 days were recorded (after the demoulding day). 2.4. Ultrasonic measurements During the experiment, ultrasonic pulse velocity instrumentation containing a transmitter generating P-wave and a receiving transducer was used to assess the ultrasonic velocity in the specific material [24]. In the direct transmission, transmitter and receiving transducer were fixed at two sides of the 40 × 40 × 160 mm prism (sealed with plastic sheets). The transmitter and the receiving transducer were 38 mm in diameter with a central frequency of 50 kHz. The driving voltage was 500 V and the sampling period was 0.4 μs. Silicon grease was used as a coupling agent in order to guarantee full contact between the sensors and the specimen (Fig. 2). 2.5. AE tests The acoustic emission test system in this experiment from Physical Acoustics Corporation (PAC) consisted of a PCI-2 acoustic card, piezoelectric based acoustic sensors, selectable pre-amplifiers and a high performance data acquisition system which contained six channels (Fig. 3). The sensors used were R6α resonant type sensors with the operating frequency range of 35– 100 kHz. When the pre-amplifier was selected as 40 dB (in a previous study a similar setup was used [25]) the measuring range was 0– 80 dB with the accuracy of 0.5 dB. The sampling rates of data acquisition system were as high as up to 1 MSPS (mega samples per second). During the experiment the prisms for test were sealed with plastic sheets in case of moisture loss. To capture the accurate signals generated inside the materials the AE sensors were fixed on the prisms at the center using silicon grease as a coupling agent
Fig. 2. Illustration of direct transmission in ultrasonic measurement.
98
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 3. AE test system in the experiment.
[26]. The measured signals were then filtered and processed by the software packages AEWin obtained from PAC. Noise tests were performed to choose an appropriate amplitude threshold for elimination of possible environment noise. After a series of careful adjustments the amplitude threshold was chosen as 28 dB. 3. Results and discussion 3.1. Autogenous deformation Strains of specimens with different dosages of expansive agent in sealed condition in the first fourteen days were measured and plotted in Fig. 4. Compared with autogenous shrinkage of the reference sample (without expansive agent, the same below) the specimens with expansive agent showed continuous expansion of different magnitudes. Most of expansion was observed in the first three days which could be due to fast hydration of expansive components like calcium oxide [27]. The length of specimens along the direction of testing showed a gradual decreasing trend at later ages. For instance, relative shrinkage (compared with the strain at the testing day before) was measured after one day for the specimen with 2% expansive agent. The turning points from expansion to relative shrinkage were the second day for 4%, the third day for 6% and the fifth day for 8%, respectively. It should be noted that the expansion was not strictly proportional to the amount of expansive agent added and for the specimen with 8% expansive agent, a much larger expansion was measured, nearly 93 times and 7 times as large as that with 4% and 6% expansive agent, respectively. 3.2. Strength properties Both the flexural strength and compressive strength tests of the specimens were performed and the results were plotted in Figs. 5 and 6. It was observed that although expansive agent had an advantage of compensating for shrinkage the inclusion of
Fig. 4. Autogenous deformation of specimens with different dosages of expansive agent.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
99
Fig. 5. Flexural strength results.
expansive agent would influence the strength development over the curing ages. A higher dosage of expansive agent caused larger decreases in both flexural strength and compressive strength and the largest decreases were observed for the specimen with 8% expansive agent at different ages. Compared with the reference sample the specimen with 8% expansive agent (surface crack as shown in Fig. 7) had decreases of 29%, 19% and 21% in flexural strength and 56%, 43%, and 46% in compressive strength at 3, 7, and 14 days, respectively. It was reported expansive products such as calcium hydroxide and ettringite [28] would be formed due to the hydration of expansive agent, which could have an impact of the cement hydration process since both the cement content and water content were changed [2–3]. Meanwhile the replacement of cement by expansive agent naturally caused a strength loss. In addition, expansive products formed would create residual expansive stress in the contact surface with the hardened cement composites, which could lead to weaker interfacial bonding and the possible micro-cracks along the interface. For easier quantification of strength damage compared with the reference sample, Eq. (1) could be revised by taking the reference sample strength at the corresponding age into consideration as in Eq. (2).
dðt Þ ¼ 1−
f ðt Þ f r ðt Þ
ð2Þ
where f(t) was the strength of the specific sample and fr(t) was the reference sample strength. The damage degree of both the flexural strength and compressive strength for all the specimens at 14 days were calculated by Eq. (2) and the relationship between the strength damage degree and autogenous strains was shown in Fig. 8. A significant linear relationship was observed between the damage degree and autogenous strains, with the correlation coefficients of 0.9964 for the flexural strength and 0.9725 for the compressive strength. Meanwhile different slops of the regression lines indicated that the damage degree of the compressive strength were more sensitive to autogenous deformation.
Fig. 6. Compressive strength results.
100
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 7. Surface crack of the specimen with 8% expansive agent.
3.3. Ultrasonic pulse velocity In the direct transmission test the ultrasonic travel time and the true length of the prisms along the direction of testing (obtained by a vernier caliper) were used to calculate the ultrasonic pulse (P-wave) velocity [29] in the specific specimen as shown in Fig. 9. It was observed that for the specimens with no more than 6% expansive agent (including the reference sample) the ultrasonic velocity increased with the curing age in the densification process with cement hydration. Most of increases of ultrasonic velocity were observed at early ages, namely the first three days. The ultrasonic velocity at 7 days was almost same around 3.7 km/s for the specimens with 0, 2% and 4% expansive agent, but the specimen with 6% expansive agent had relatively lower velocity, 3.6 km/s, which could be due to damage from larger expansion. In particular, the ultrasonic velocity for the specimen with 8% expansive agent was far lower than that of other specimens (3.1 km/s at 7 days). Meanwhile a decrease of ultrasonic velocity was observed at the first day when a huge expansion of 9127 με (Fig. 4) was measured, after which the ultrasonic velocity increased with the curing age. Tirupan et al. [30] reported that the velocity of P-wave increased with the increase of material density. It could be expected that hydration of both the cement and expansive agent would cause reduction of the porosity in the matrix and lead to a denser structure. At the same time, the generation of residual expansive stress from expansion would lead to degradation and microcracking in the composites. The ultrasonic velocity was influenced by the two kinds of opposite effects and it was observed
Fig. 8. The relationship between strength damage degree and autogenous strains.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
101
Fig. 9. Ultrasonic pulse velocity in the specimens.
from Fig. 8 that the enhancement effect played a dominant role in the specimen with 2%, 4% expansive agent while degradation effect behaved in the specimen with 6%, 8% expansive agent. The damage degree based on ultrasonic pulse velocity could also be defined as Eq. (3). 0
d ðt Þ ¼ 1−
vðt Þ vr ðt Þ
ð3Þ
where v(t) was the ultrasonic velocity of the specific sample and vr(t) was the ultrasonic velocity of the reference sample, respectively. The damage degree based on the ultrasonic velocity for all the specimens at 14 days were calculated by Eq. (3) and the relationship between the damage degree d′ and autogenous strains was shown in Fig. 10. A significant linear relationship was also observed between the damage degree and autogenous strains with a correlation coefficients of 0.9977. By taking the absolute value of regression line slope from damage degree based on ultrasonic properties as well as strength properties into consideration it could be found that compressive strength showed larger decreasing slope than ultrasonic velocity and flexural strength. 3.4. AE activity 3.4.1. AE parameters AE parameters indicating the AE activity of the cement-expansive agent system were tested, e.g., AE amplitude, AE counts, cumulative AE counts, and AE energy. The results were shown in Figs. 11–14. The amplitude threshold for measurement was set as 28 dB and it was observed AE amplitude fell in the range from 28 dB to 48 dB indicating the AE activity due to inner expansion had relatively lower energy [31]. AE counts could be used as a representation of a damage event that occurred within the materials [32]. In Fig. 12 AE counts were observed during the whole experiment for all specimens and the specimens with higher dosage of expansive agent had
Fig. 10. The relationship between damage degree d′ and autogenous strains.
102
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 11. AE amplitudes for the specimens.
relatively more AE counts (Fig. 13). It should also be noted that AE counts were mostly at early ages, namely in the first 50 h for higher dosages of expansive agent (6%, 8%) consistent with quick development of expansion. Fig. 13 showed cumulative AE counts for the specimens and it could be observed generally specimens with more expansive agent had more cumulative AE counts indicating higher AE activity. The 6% sample showed higher cumulative AE counts at early age (before 30 h) but lower cumulative AE counts at later age than the 4% sample. As the AE signals detected were in the form of voltage and the AE counts measured experimentally depends on many factors such as AE sensors coupling to cement pastes, sensor sensitivity, source-sensor distance, etc. [33]. Analysis based on only one parameter was inadequate and thus it was significantly necessary to take AE energy released [34] into consideration. The cumulative AE energy for all specimens was shown in Fig. 14. As the AE energy gave a reflection about the true energy of any AE event or hit captured by AE station and could be related to the energy released in microcracking process. It was observed that a higher amount of energy released was recorded when a larger dosage of expansive agent was added because of larger expansion. The possible contradiction between cumulative AE counts and cumulative AE energy (4% and 6%) could be due to that AE energy associated with these AE events was relatively lower.
3.4.2. Crack mode identification Crack mode identification is fundamental to better understanding about damage properties [35] in the materials during expansion process. The mechanism by analysis of AE parameters lies in different elastic wave characteristics induced by different types of movement of the crack tips [36]. In tensile crack mode, volumetric change would cause the opposing displacement at the crack tips during which the longitudinal wave was emitted. However in shear crack mode, movements of the crack tips were in parallel to the crack plane and would lead to a change in shape instead of volume change.
Fig. 12. AE counts for the specimens.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
103
Fig. 13. Cumulative AE counts for the specimens.
Propagation and reflection in a heterogeneous medium would cause distortion and change AE waveform parameters like duration and rise time, especially for large size structures with complicated geometry [37]. In laboratory scale, assuming that distortion and attenuation are limited due to smaller dimensions and short propagation distances, crack classification could be studied based on RA value and average frequency. Ohtsu et al. [38–39] proposed a method to classify cracks based on the relationship between average frequency faverage and RA value using AE parameters. It was reported that AE signals which had lower RA value and higher average frequency could be an indication of tensile crack. Meanwhile, for AE signals which had high RA value and low average frequency it could be said that shear crack occurred. The RA value and average frequency faverage could be calculated from AE parameters as defined in Eqs. (4) and (5). RA ¼
t rise V
f average ¼
ð4Þ
N t dur
ð5Þ
where trise is the rise time, ms; V is the maximum amplitude, V; N is the AE ring-down counts, tdur is the duration time, μs. In AE test the amplitude is defined as the peak voltage (V) of an event and then transformed to AdB by logarithm. As a preamplifier (40 dB) was used, the relationship between voltage and dB was shown in Eq. (6) AdB ¼ 20 lg
V −40 V ref
ð6Þ
Fig. 14. Cumulative AE energy for the specimens.
104
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
Fig. 15. Crack classification (specimens with 0 expansive agent).
Fig. 16. Crack classification (specimens with 8% expansive agent).
where AdB is AE amplitude (dB), V is the peak voltage of the AE waveform and Vref is the reference voltage, namely 1 μV. The RA value and average frequency for all detected AE hits were calculated from Eqs. (4) and (5) according to JCMS-III B5706 [40] and the proportion of the RA value and the average frequency was tentatively likewise chosen as 1:80 as stated by Ohtsu [39] for relative comparisons. Figs. 15 and 16 showed the relationship between average frequency and RA value for the specimens
Fig. 17. Ratio of tensile cracks and shear cracks for all specimens.
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
105
with 0 and 8% expansive agent. It was observed that 62.2% of cracks were identified as tensile cracks for the reference sample while 67.4% of cracks were identified as shear cracks for the specimen with 8% expansive agent. Fig. 17 showed the variations of ratio of tensile cracks and shear cracks for all specimens with different dosages of expansive agent during the experiment. A decreasing ratio of tensile cracks and an increasing ratio of shear cracks were observed. As expansive agent had a shrinkage compensation effect residual tensile stress induced by shrinkage was largely relaxed or even eliminated. In addition, expansive stress induced by expansive products in the interface could lead to micro cracks including both tensile cracks and shear cracks when the volume expansion was assumed to be isotropic [10]. Lots of studies indicated that for cementitious materials tensile crack mode was dominated at the early stage of failure and shear crack mode was mostly active at the later stage [35]. Therefore the increase of shear cracks could be an indication of aggravated damage inside the materials. 4. Conclusions In this paper the damage development induced by expansive agent was investigated for concrete members by both mechanical and acoustic measurements. The results indicated that the inclusion of expansive agent would have an influence on the strength development and ultrasonic velocity over time. The hydration of the cement and expansive agent would cause reduction of the porosity in the matrix and lead to a denser structure. On contrast, the generation of residual expansive stress from expansion would lead to degradation and microcracking in the composites. A significant linear relationship was observed between the damage degree and autogenous expansion strains which was not strictly proportional to the amount of expansive agent added. According to this feature, the balance of shrinkage compensation and property deterioration should be paid enough attention to meet practical requirements when using expansive agent. Ultrasonic measurement offered a useful tool for determination of materials properties by ultrasonic velocity in the specific materials. Meanwhile AE activity could be assessed by AE parameter analysis and related to inner micro-cracking evolution of the cement-expansive agent system. Crack mode could be identified based on the relationship between average frequency and RA value. The addition of expansive agent lead to a decreasing ratio of tensile cracks and an increasing ratio of shear cracks which could be an indication of aggravated damage. Such nondestructive methods showed a promising future in better understanding of the damage development. Acknowledgements This research was financially supported by National Basic Research Program of China (No. 2015CB655105), National Outstanding Youth Science Foundation Program of China (No.51225801) and the Provincial Science and Technology Cooperation Project of Jiangsu Province-Jiangsu-Guangxi Cooperation Project (BM2014050). References [1] R.G. Colin, Cement and concrete as an engineering material: an historic appraisal and case study analysis, Eng. Fail. Anal. 40 (2014) 114–140. [2] C. Valeria, N. Alessandro, D. Jacopo, The influence of expansive agent on the performance of fibre reinforced cement-based composites, Constr. Build. Mater. 91 (2015) 171–179. [3] H. Jianguo, J. Di, Y. Peiyu, Understanding the shrinkage compensating ability of type K expansive agent in concrete, Constr. Build. Mater. 116 (2016) 36–44. [4] M. Collepardi, R. Troli, M. Bressan, et al., Crack-free concrete for outside industrial floors in the absence of wet curing and contraction joints, Cem. Concr. Compos. 30 (10) (2008) 887–891. [5] S. Chatterji, Mechanism of expansion of concrete due to the presence of dead burnt CaO and MgO, Cem. Concr. Res. 25 (1995) 51–56. [6] H. He, W. Dong, Z. Wu, Study on long-term expansive deformation of selfstressing concrete with combined restrictions of steel fibers and steel bar, Key Eng. Mater. 452–3 (2011) 533–536. [7] R. Sahamitmongkol, T. Kishi, Tension stiffening effect and bonding characteristics of chemically prestressed concrete under tension, Mater. Struct. 44 (2011) 455–474. [8] S. Wei, C. Huisu, X. Luo, Q. Hongpin, The effect of hybrid fibers and expansive agent on the shrinkage and permeability of high-performance concrete, Cem. Concr. Res. 31 (2001) 595–601. [9] M.S. Meddah, M. Suzuki, R. Sato, Influence of a combination of expansive and shrinkage-reducing admixture on autogenous deformation and self-stress of silica fume high-performance concrete, Constr. Build. Mater. 25 (2011) 239–250. [10] B. Hassina, M. Stephane, S. Alain, et al., Effects of restraint on expansion due to delayed ettringite formation, Cem. Concr. Res. 42 (2012) 1024–1031. [11] R. Patrice, B. Gérard, Assessment of the expansion related to alkali-silica reaction by the Damage Rating Index method, Constr. Build. Mater. 19 (2) (2005) 83–90. [12] A. Sellier, E. Bourdarot, S. Multon, et al., Combination of structural monitoring and laboratory tests for the assessment of AAR-swelling - application to a gate structure dam, ACI Mater. J. 106 (3) (2009) 281–290. [13] B. Capra, A. Sellier, Orthotropic modelling of alkali-aggregate reaction in concrete structures: numerical simulations, Mech. Mater. 35 (2003) 817–830. [14] H.D. Yun, W.C. Choi, S.Y. Seo, Acoustic emission activities and damage evaluation of reinforced concrete beams strengthened with CFRP sheets, NDT & E Int. 43 (7) (2010) 615–628. [15] M. Ohtsu, H. Watanabe, Quantitative damage estimation of concrete by acoustic emission, Constr. Build. Mater. 15 (5–6) (2001) 217–224. [16] T.P. Philippidis, D.G. Aggelis, Experimental study of wave dispersion and attenuation in concrete, Ultrasonics 43 (2005) 584–595. [17] J.B. Leonard, F.K. William, M.F. Dan, Improved assessment of mass concrete dams using acoustic travel time tomography: part I-theory, Constr. Build. Mater. 14 (2000) 133–146. [18] F.K. William, J.B. Leonard, M.F. Dan, Improved assessment of mass concrete dams using acoustic travel time tomography: part II-application, Constr. Build. Mater. 14 (2000) 147–156. [19] A. Vary, in: R.S. Sharpe (Ed.), Ultrasonic Measurement of Material Properties, In Research Techniques in NDT, 4, Academic Press, London 1980, pp. 159–204. [20] J. Lemaitre, J.L. Chaboche, Mechanics of Solid Materials, Cambridge University Press, 1989. [21] A.A. Shah, Y. Ribakov, Non-destructive evaluation of concrete in damaged and undamaged states, Mater. Des. 30 (9) (2009) 3504–3511. [22] M. Ohtsu, M. Shigeishi, H. Iwase, et al., Determination of crack location, type and orientation in concrete structures by acoustic emission, Mag. Concr. Res. 43 (155) (1991) 127–134. [23] K.M. Holford, Acoustic emission in structural health monitoring, Key Eng. Mater. 413–4 (2009).
106
Q. Xia et al. / Engineering Failure Analysis 74 (2017) 95–106
[24] ASTM C597–97, Standard test method for pulse velocity through concrete, Annual Book of ASTM Standards, American Society of Testing and Materials, West Conshohocken, PA, 2003. [25] D. Soulioti, N.M. Barkoula, A. Paipetis, et al., Acoustic emission behavior of steel fibre reinforced concrete under bending, Constr. Build. Mater. 23 (2009) 3532–3536. [26] D.Y. Hyun, C.C. Won, Y.S. Soo, Acoustic emission activities and damage evaluation of reinforced concrete beams strengthened with CFRP sheets, NDT&E Int. 43 (7) (2010) 615–628. [27] S. Maltese, C. Pistolesi, A. Lolli, et al., Combined effect of expansive and shrinkage reducing admixtures to obtain stable and durable mortars, Cem. Concr. Res. 35 (12) (2005) 2244–2251. [28] A. Telesca, M. Marroccoli, M.L. Pace, et al., A hydration study of various calcium sulfoaluminate cements, Cem. Concr. Compos. 53 (2014) 224–232. [29] S. Selleck, E.N. Landis, M.L. Peterson, et al., Ultrasonic investigation of concrete with distributed damage, ACI Mater. J. 95 (1998) 27–36. [30] M. Tirupan, M.T. James, B.E. Tuncer, Non-destructive testing of cementitiously stabilized materials using ultrasonic pulse velocity test, Transp. Geotech. 6 (2016) 97–107. [31] Carpinteri, G. Lacidogna, G. Niccolini, Damage analysis of reinforced concrete buildings by the acoustic emission technique, Struct. Control. Health Monit. 18 (6) (2011) 660–673. [32] D. Prem, J.C. Kapil, K. Ashavani, Acoustic emission characteristics and b-value estimate in relation to waveform analysis for damage response of snow, Cold Reg. Sci. Technol. 119 (2015) 170–182. [33] D.G. Aggelis, D.V. Soulioti, N. Sapouridis, et al., Acoustic emission characterization of the fracture process in fibre reinforced concrete, Constr. Build. Mater. 25 (2011) 4126–4131. [34] E.N. Landis, L. Baillon, Experiment to relate acoustic emission energy to fracture energy of concrete, J. Eng. Mech. 128 (6) (2002) 698–702. [35] D.G. Aggelis, A.C. Mpalaskas, T.E. Matikas, Investigation of different fracture modes in cement-based materials by acoustic emission, Cem. Concr. Res. 48 (2013) 1–8. [36] D.G. Aggelis, Classification of cracking mode in concrete by acoustic emission parameters, Mech. Res. Commun. 38 (2011) 153–157. [37] D.G. Aggelis, A.C. Mpalaskas, D. Ntalakas, et al., Effect of wave distortion on acoustic emission characterization of cementitious materials, Constr. Build. Mater. 35 (2012) 183–190. [38] M. Ohtsu, T. Isoda, Y. Tomoda, Acoustic emission techniques standardized for concrete structures, J. Acoust. Emission 25 (2007) 21–32. [39] K. Ohno, M. Ohtsu, Crack classification in concrete based on acoustic emission, Constr. Build. Mater. 24 (12) (2010) 2339–2346. [40] JCMS-III B5706, Monitoring Method for Active Cracks in Concrete by Acoustic Emission, Federation of Construction Materials Industries, Japan, 2003.