NDT.net - February 2000, Vol. 5 No. 02 |
- International Symposium on NDT Contribution to the Infrastructure Safety Systems, 1999 NOV 22-26 Torres, published by UFSM, Santa Maria, RS, Brazil |
TABLE OF CONTENTS |
In order to improve the significance of the NDT x Concrete Strength correlation the coarse aggregate strength may be introduced as an independent variable. Another way of improving the concrete strength estimation consists in combining two different non destructive tests.
The aim of the present paper is to check the validity of pulloff, pin penetration and ultrasonic pulse velocity for assessing concrete strength, that are widely used in structural field. The NDT are compared on basis of sensitivity of measurements and scatter of results. The advantage deriving from the aggregate strength used as a variable is also examined. An extensive series of tests was carried out to obtain a correlation of the NDT results with compressive strength. Nine concrete mixes for each three different coarse aggregate strength, of 19 mm max. size, were tested.
Keywords: Concrete; NDT; Pulloff Test; Penetration Test; UPV
Nondestructive tests (NDT) provide indirect data that can be empirically related to compressive strength by calibration with strength measurements from a number of cast specimens. NDT measurement technique has been used for more than two decades for concrete quality evaluation and assessing concrete compressive strength (f_{c}). During this period, the factors influencing the test results have been widely reported.
According to Castro [1] age, mix proportion, water/cement ratio, cement type, aggregate type have an influence on the pulse velocity test results and its relationship to f_{c}. Using different concrete mixes, but with the same materials Castro [2] showed a comparison between NDT through their relationship with compressive strength of cast cylinder specimens and f_{c} by extracted cores. Structural inspection to repair or reinforce structures is probably the main application for combine NDT. Measurement of NDT at points on a regular grid on the surface of concrete structures provides a reliable method for assessing the homogeneity of the concrete as shown by Ludwig and Castro [3].
This paper shows the results of an experimental program on NDT for assessing f_{c}. The variable assessed were: (1) concrete strength at different mix proportion with a range of 15 to 40 MPa to duplicate strengths found in practice; and (2) coarse aggregate of different crushing strength. The effect of these parameters on the relation between compressive strength by cast specimens and the nondestructive tests results was measured.
Portland cement (CP-II F40), complying with the ABNT - NBR 11578 / 91 [4] , and natural sand were used in all mixes. Crushed coarse aggregate of three different origins were used in the research work. They were named after their crushed strength (kN) 95, 120 and 180. Nine series of mixes was carried out for each coarse aggregate origin. Therefore the mixes set used have the same: (1) Portland cement; (2) sand / aggregate ratio; (3) coarse aggregate maximum size of 19 mm; (4) and admixture proportion. The mixes used for the test series are shown in Table 1.
MixID | Mix Proportion | W/C Ratio | CACS* (kN) | fc** (MPa) | Prismatic Specimens (230x230x340 mm) | ||
Pulloff (N.m) | UPV (km/s) | Pin*** (mm) | |||||
1/95 | 1 : 2.53: 3.53 | 0.60 | 95 | 18.7 | 14.7 | 3.41 | 14.8 |
2/95 | 1 : 2.42: 3.64 | 0.60 | 26.3 | 16.3 | 3.71 | 19.0 | |
6/95 | 1 : 1.58: 2.67 | 0.42 | 28.7 | 18.2 | 4.00 | 25.7 | |
7/95 | 1 : 1.41: 2.50 | 0.42 | 32.6 | 23.3 | 4.12 | 26.6 | |
8/95 | 1 : 1.24: 2.35 | 0.39 | 38.2 | 22.3 | 4.09 | 29.4 | |
9/95 | 1 : 1.17: 2.21 | 0.37 | 31.8 | 23.2 | 4.07 | 27.9 | |
2/120 | 1 : 2.42: 3.64 | 0.60 | 120 | 26.8 | 16.3 | 4.14 | 23.0 |
3/120 | 1 : 2.20: 3.14 | 0.54 | 24.1 | 11.7 | 4.12 | 25.0 | |
4/120 | 1 : 1.95: 2.97 | 0.49 | 29.3 | 17.0 | 4.15 | 27.2 | |
5/120 | 1 : 1.75: 2.75 | 0.45 | 26.3 | 17.8 | 4.15 | 24.6 | |
7/120 | 1 : 1.41: 2.50 | 0.42 | 40.2 | 24.5 | 4.22 | 23.1 | |
8/120 | 1 : 1.24: 2.35 | 0.39 | 40.6 | 22.3 | 4.25 | 25.7 | |
9/120 | 1 : 1.17: 2.21 | 0.37 | 28.2 | 18.2 | 4.07 | 25.1 | |
1/180 | 1 : 2.53: 3.53 | 0.60 | 180 | 29.5 | 20.7 | 4.22 | 23.7 |
2/180 | 1 : 2.42: 3.64 | 0.60 | 35.3 | 14.5 | 4.29 | 33.9 | |
3/180 | 1 : 2.20: 3.14 | 0.54 | 24.7 | 14.8 | 4.19 | 28.2 | |
4/180 | 1 : 1.95: 2.97 | 0.49 | 32.8 | 18.3 | 4.28 | 30.8 | |
5/180 | 1 : 1.75 : 2.75 | 0.45 | 34.5 | 14.5 | 4.30 | 28.7 | |
6/180 | 1 : 1.58: 2.67 | 0.42 | 34.8 | 22.0 | 4.36 | 24.5 | |
7/180 | 1 : 1.41 : 2.50 | 0.42 | 41.9 | 19.7 | 4.30 | 35.5 | |
8/180 | 1 : 1.24 : 2.35 | 0.39 | 26.3 | 19.7 | 4.39 | 26.7 | |
Table 1: Mix Characteristics and test results |
The development of the pulloff test device is described in Miranda [5]. The pulloff test is a partial NDT. The steel plate of 75 mm diameter is fixed to the concrete surface by epoxy. Applying an axial pulloff force to the bolt by means of a torque wrench and bearing system results in a vertical force on the plate, which eventually leads to the fracture of the concrete. The penetration test was introduced in Brazil by Pontes Vieira [6]. Thus, pin penetration tests were carried out according to the Brazilian tradition: smooth 55 mm long pin of 6.4 mm diameter and a short cartridge 22, medium powder potential. Test results were measured by using a caliber and referred to millimeter. UPV (54 kHz) measurements by direct transmission arrangement were used in the present research program.
The tests results of the present investigation were previously analyzed to figure out which were the outlier results for each test. Whenever a mix proportion has more than two outlier results from different non destructive tests, all NDT results for that mixes were excluded. Therefore, the analysis and conclusions are based on tests results, shown in Table 1.
Statistical tests on NDT results were carried out. The first test was to verify whether the NDT results and compressive test result fitted to a Gauss Distribution. The Henry straight line test shows all test fitted to Gauss Distribution. Just as a frequency distribution can be described diagrammatically by a frequency polygon, so can a cumulative frequency distribution be represented by a curve called an ogive. The ogive curves for the experimental test results are shown in Figure 1.
Fig 1: Ogive for the test results of the experimental program |
The pulloff test results produced an ogive close to the concrete compressive test results. Actually the mechanical measurement of both tests results can be assumed as the main support for that close behavior. UPV and pin penetration test results produced an ogive significantly different from the compressive strength test results. This statement is supported by hypothesis test carried out to ogive area at 5% risk. The effect of the coarse aggregate crushed strength on the pin test result is assumed as a parameter for that behavior. The crush strength of the coarse aggregate is related to its density. Therefore, the effect of the coarse aggregate origin is assume to explain the UPV ogive behavior.
Regression analysis is an useful tool to evaluate the relationship between NDT and concrete tests. Statistical parameter are first used to evaluate the fit of the chosen regression curve. However, when comparing models obtained for different test types, the regression curves parameters should be carefully used. Table 2 shows regression models, correlation coefficient and standard errors obtained for the experimental program. The standard errors of these correlation is the statistical parameter used as basis of comparison between tests: it has the units of MPa. Although the correlation coefficient shows the extend to which the NDT results and the compressive strengths are related, the effectiveness of f_{c} evaluation cannot be obtained from this parameter. The residuals standard deviation is the measure of the accuracy in evaluating the f_{c} hence the measure of the model's effectiveness. Gross model violations, when present, are often exposed by an appropriate graph of the residual. The standardized residuals were plotted as the ordinate against the model value. The normal pattern fits into two parallel straight lines.
The procedure used for analyzing data included one variable statistic and regression analysis by least squares method. An analysis of residual variation was undertaken to evaluate the effect of the aggregate origin on the tests relationship.
The present NDT test results confirms that a crushed strength effect is present, i. e., coarse aggregate crushed strength has an influence on NDT - concrete strength relationship.
NDT | CACS (kN) | Regression Curves | Correlation Coefficient | Standard Error |
Pulloff Test | 95 | fc = 1.101 B^{1.100} | R^{2} = 0.792 | 3.660 |
120 | fc = 13.005 e^{0.046.B} | R^{2} = 0.856 | 3.021 | |
180 | fc = 1.763B | R^{2} = 0.694 | 5.934 | |
All results | fc = 0.049 B^{2} - 0.776 B + 27.849 | R^{2} = 0.379 | 4.961 | |
All with C | fc = 1.346 B^{0.667} C^{0.242} | R^{2} = 0.466 | 0.067 | |
UPV | 95 | fc = 0.435 V^{3.083} | R^{2} = 0.896 | 2.936 |
120 | fc = 94.644 V - 362.647 | R^{2} = 0.705 | 4.018 | |
180 | fc = 0.437 V^{2.948} | R^{2} = 0.069 | 5.804 | |
All results | fc = 1.304 V^{2.222} | R^{2} = 0.394 | 5.138 | |
All with C | fc = 1.415 V^{3.070} C^{-0.263} | R^{2} = 0.463 | 0.068 | |
Pin Penetration Test | 95 | fc = 1.825 P^{0.876} | R^{2} = 0.910 | 2.486 |
120 | fc = 0.674 P + 10.112 | R^{2} = 0.257 | 1.994 | |
180 | fc = 0.819 P + 8.713 | R^{2} = 0.386 | 4.689 | |
All results | fc = 12.782e^{0.032 P} | R2 = 0.635 | 3.493 | |
All with C | fc = 3.876 P^{0.771} C^{-0.091} | R^{2} = 0.437 | 0.069 | |
Table 2:The regression models |
© NDT.net - info@ndt.net | |Top| |