International Symposium (NDTCE 2003) NonDestructive Testing in Civil Engineering 2003  
Start > Contributions >Lectures > Materials 2: 
Nonlinear nondestructive methods for evaluating strength of concreteI. Shkolnik, Wayne State University, Detroit, Michigan, USA [Visiting Professor, skholnikj@juno.com]H. Aktan, Wayne State University, Detroit, Michigan USA [Professor and Director of Center for Structural Durability (http:/webpages.eng.wayne.edu/durabilitycenter/ ) Haluk.Aktan@wayne.edu] R. Birgul, Mugla University, Mugla, Turkey [Asst. Professor, rbirgul.mu.edu.tr] AbstractExperimental data obtained using various methods shows that when concrete is subjected to increasing uniaxial compression, there are ranges of stress, reflecting the reaction of concrete material structure to external force and indicating the nonlinearity of deformation. It is also an established fact that the acoustic properties of solids depend on the properties of their flaws. Utilizing the nonlinear force deformation behavior of concrete, new nonlinear acoustical methods are described for evaluating the strength of concrete. There are two main steps involved in this research. The first step is the description of the physical process using nonlinear approach and known relations representing the fracture of concrete, and establishment of a relationship for the nonlinear strength parameter. The second step is development of new nondestructive test methods for evaluating the nonlinear parameter of concrete. The nonlinear parameter is defined as change in the static or dynamic modulus of elasticity under loading. It will be shown that analytical solution of strength is in substantial agreement with the corresponding formula in ACI Building Code 318. It is proven that the nonlinear parameter depends on structural defects present in concrete, moisture content, age, mineralogical content of cement, and integral energy losses. Three methods for the measurement of nonlinear parameters of concrete that will be presented are;
The testing methods do not require special equipment, may be used when conventional acoustic methods are not applicable, and point out new perspectives for the use of ultrasonic and impactecho methods for evaluation of strength parameters of concrete insitu. As important examples, the application of the three methods will be demonstrated for use in predicting both the strength of concrete at different moisture states, as well as the strength of high performance concrete. IntroductionIn recent years, the emphasis in nonlinear ultrasonics has shifted towards the determination of properties that correlate strongly with mechanical properties of engineering materials such as adhesion and hardening of precipitates (Panraz and Arnold 1994, Cantrell and Yost 1996). However, to our knowledge, there are no nonlinear acoustical methods for predicting strength of concrete. At the present time, the conventional ultrasonic method for evaluating the strength of concrete does not consider the nonlinear behavior of concrete. Quick, simple, and widely acceptable procedures are needed for evaluation of the reaction of concrete's structure on the external loading, i.e. for measurement of the nonlinearity of concrete deformation. Based on the nonlinear behavior of concrete, this article discusses the fundamentals as well as nondestructive methods for evaluation of the strength of concrete. Basic StatementsWhen developing nonlinear methods for evaluating the strength of concrete, the most important condition is that within a wide range of strain (10^{6} 10^{4}), the relationship between stress ( s) and strain (e) of brittle materials can be described as parabolic (Bell 1984):
It is then necessary to relate the fracture of brittle materials to their corresponding modulus of elasticity and the nonlinear parameter (E_{0} is the initial modulus of elasticity). Carrying the analogy over the stressfree region from the line crack model, it is assumed that the stressfree region is a volume, which is proportional and within the order of magnitude equal to the initial porosity of concrete.Thus total energy (U) for the system under consideration can be written as (Sedov 1973; Shkolnik 1993):
where "p" is the relative volume of defects (p= NV_{n}/V=nV_{n}<1); n is the number of defects per unit volume; V_{n} is the volume of a spherical defect with the diameter D; G is the density of surface energy; K=(6^{2/3}/4)(3.14)^{1/3}= 1.2. It is assumed that for the nonlinear model (1), the porosity under loading as well as the modulus of elasticity depends on strain (stress). Thus, the values of strength are found to be the roots of the condition ¶U /¶s = 0 applied using Eq. (2). After simple manipulation and substitution of R for the root s _{2}, we have from Eq. (2):
Eqs. (2a and 2b) describe the relationship between compressive strength "R" of concrete, modulus of elasticity "E", (E_{0}~ 50GPa for concrete), and the nonlinear parameter ¶E / ¶s (Shkolnik 1993). From Eq. (2b), it follows that the modulus of elasticity should have a maximum value given by the product of strength and the nonlinear parameter. Such an approach explains existing experimental relationships between modulus of elasticity and strength of concrete in which modulus of elasticity decreases in spite of an increase in strength if the nonlinear parameter of concrete decreases simultaneously (Freudental, A.M.; Walker, S. 1959). In general, Eq. (2b) is in substantial agreement with the corresponding formula in ACI Building Code 318. Methods and ResultsStatic Loading.
Table 1 contains corresponding experimental data (Moskvin et al. 1974) for concrete of various water/cement ratios (w/c). Three mix proportions were prepared using the same fine and coarse aggregate and cement with mix proportions of 1:3.33:5.55 for w/c = 0.7; 1:2.17:4.50 for w/c = 0.5; 1:1.5:3.5 for w/c = 0.4. Various specimens were subjected to four methods of inundation and subsequent load testing as shown in Table 1.
Table 1 shows the critical stress values (s_{c}), and strain values (e_{c}), depicting the onset of a significant increase in bond cracking. For saturated specimens, the change in modulus at the critical stress can be approximated as ¶E / ¶s = EE_{c} / s_{c}. It is also seen in Table 1 that the nonlinear parameter, reflecting the response of moist concrete's structure to external load, increases with moisture content more significantly than the increase of the modulus of elasticity. Therefore, in spite of the increasing modulus of elasticity, the strength decreases, which displays agreement with Eq. (2c). As was pointed out above, for the nonlinear model (1), the modulus of elasticity depends on strain. To investigate that relationship, mix proportions of concrete with five watercement ratios (0.35, 0.40, 0.45, 0.5, 0.55) were batched. Young's Modulus, Poisson's ratio, and compressive strength tests were performed on designated specimens according to ASTM C 469 and ASTM 39. During the 1st cycle of loading, the ratio of normal stress to corresponding strain, i.e. modulus of elasticity, was measured within the strain range (101000) me . Then the nonlinear parameter dE/ds was been found for data with a correlation coefficient at least 0.9. Table 2 shows the comparison of experimental compressive strength with the corresponding prediction using Eq. (2 a) and Eq. (2c). Bold type represents the values of strength using Eq. (2a) or Eq. (2c) for dry and wet concrete respectively, leading to the minimal prediction error. It is seen in Table 2 that the average relative error is less than 15%. Excluding, with probability more than 90%, such errors as 75%, and then 71%, the average error can be reduced from 13% to 6.4%, as seen in the rightmost column.
The described nonlinear approach also allows investigation of the physical meaning of the coefficient recommended by ACI Building Code 318. Indeed, Table 3 illustrates the values of the coefficient between modulus of elasticity and strength for different watercement ratios according to ACI Building Code and by use of Eq. (2a). Corresponding relative error is within 1%. Therefore, the empirical coefficient reflects the nonlinear behavior of concrete and can be calculated using the nonlinear parameter dE/ds while holding E_{o}=50GPa as a constant._{ } It is possible (Table 2 and Table 3) simultaneously evaluate the strength of concrete, using only the 1^{st} cycle during the test according to ASTM C 469, and to reduce the quantity of specimens for compressive strength.
Dynamic Loading It is logical to suppose that nonlinearity of deformation, reflecting a reaction of the structure to the external force, may be measured by increase the accuracy of measurement Therefore new testing methods were developed: spectral, resonant frequency shift, and phase shift (Shkolnik 1993). In spectral method an ultrasonic wave of known frequency, f= 50KHz, was applied trough a concrete samples. The energy wasted by scattering when the wave "stumbles" upon inner defects is proportional the energy spend on forming the second harmonic, f=100KHz. Consequently, in spectral method the nonlinear parameter was defined by the ratio of the second harmonic amplitude to the first one. The ratio of the second harmonic to the first one, taking into account the transmission coefficients of measuring device, was within a few fractions to 1 percent (Shkolnik 1992). From a practical point of view, it is important to measure the nonlinear parameter in the field. For this purpose, two nondestructive test methods were developed based on the measurement of resonant frequency shift and phase dependence to dynamic loading. The first procedure is based on measurement of the resonant frequency shift of the fundamental mode, and the second procedure is based on the phase shift dependence on the excitation amplitude (Shkolnik 1985; Shkolnik 1992). The resonance frequency shift measurement allows the use of the wellknown impactecho method (Sansalone and Carino 1991) for measuring the nonlinear parameter of concrete in situ. In the phase dependence measurement, the nonlinear parameter can be determined from the phase change due to the changes in the dynamic modulus of elasticity. The phase difference, " D f " between two amplitudes of excitation was measured on several high performance concrete specimens. A linear regression of static strength with phase difference D f produced R_{1 }= 66.56.84D f , with a meansquare error of less than 10 % in the investigated range satisfying the practical requirements (Fig.1).
The nondestructive test procedures can be used when conventional acoustic methods are not applicable, and significantly reduce the effort required for evaluating the static, impact, and fatigue strengths of concrete. SummaryNonlinear parameters are integral structuresensitive characteristics that contain information on structural defects of the material, mineralogical content of Portland cement, and moisture content. Nonlinear parameters can be used for quantitative evaluation of concrete strength under quasistatic or dynamic loading. The nonlinear approach verifies the relationship between static elastic modulus of concrete and strength accepted by ACI 318. Although additional tests are needed for better statistical representation, data presented allows the evaluation of the strength of concrete specimens from standard compressive tests according to ASTM C 469. The findings also indicate that in situ nondestructive stress wave techniques can be applied for the prediction of strength and fracture of concrete. References

