·Home ·Table of Contents ·Civil Engineering

Evaluation of Actual Compressive Strength of High Strength Concrete by NDT Giovanni Pascale, Antonio Di Leo, Roberto Carli DISTART - University of Bologna Viale Risorgimento, 2 - 40136 BOLOGNA - ITALY Virna Bonora DICASM - University of Bologna Viale Risorgimento, 2 - 40136 BOLOGNA - ITALY E-mail : giovanni.pascale@mail.ing.unibo.it antonio.di.leo@mail.ing.unibo.it virna.bonora@mail.ing.unibo.it roberto.carli@mail.ing.unibo.it Contact

Abstract

This paper deals with Non Destructive Testing (NDT) of High Strength Concrete (HSC).
An experimental research was carried out, involving both destructive and non destructive methods applied to different concrete mixes, with compressive strength varying from 30 up to 150 MPa. Both cubic and cylindrical standard specimens and bigger blocks were cast.
Correlation curves were derived for different methods employed, as pulse velocity, rebound hammer, surface hardness, pull out, probe penetration, microcoring and combined methods.
The results shows good behaviour for some methods, like pulse velocity, surface hardness and combined methods, also if new correlations are to be established in some cases, respect to the ones already established for normal concrete.
Some problems arose for other methods, like pull out and probe penetration, for which the available commercial testing systems appeared not adequate to high strength concrete.

Keywords : Concrete, HSC, Non destructive methods, Combined methods.

Introduction

The evaluation by non destructive methods of the actual compressive strength of concrete in existing structures is based on empirical relations between strength and non destructive parameters [1]. The most commonly used testing methods are rebound hammer, pulse velocity, pull-out, probe penetration test, microcoring and combined methods. The validity of the above mentioned relations is actually limited to normal strength concrete, up to 50 MPa.
HSC has been employed in recent years, with compressive strength up to 150 MPa and over. The relations used to evaluate the compressive strength of normal concrete by non destructive tests may be no longer valid for HSC.
This experimental research is aimed to verify the possibility of applying the known NDT methods to HSC, to state the limits of the testing equipment available and to extend the existing relations, or determine new ones between ND parameters and compressive strength of HSC.
In the next section an experimental program is presented, for a concrete with fixed components. Different mixes has been designed, in order to result in a compressive strength as wide as possible. Such a choice could affect the generalization of the results of the analysis of experimental data [2,3]: the same strength value can be obtained with different constituents and mix proportions.
This work is thus to be considered as a first approach to the analysis of experimental tests on HSC by non destructive methods.

Experimental program

Materials
The materials used in this investigation and their characteristics are here summarized.

Cement
High early-strength Portland cement CEM I 52.5R (Italcementi, Bergamo, Italy) conforming to italian standard UNI EN 197/1.

Supplementary cementitious materials
Silica fume, Starkzement 1 (ADDIMENT Italia S.p.A., Bergamo, Italy), in the form of powder with the characteristics showed in Table 1.

Fine aggregate
Commercial, locally available sand was used (Cave AGES, Bologna Italy). Based on petrographic analysis it was composed of 94% of limestone. The fineness modulus, density, on a saturated and surface-dried basis, and water absorption were 2.67, 2.70 and 0.23 respectively.

Coarse aggregate
Commercial, locally available crushed coarse aggregate (Cave AGES, Bologna Italy) with a nominal maximum aggregate size of 15.0 mm was used. Based on petrographic analysis it was composed of 95% of limestone. The fineness modulus, density, on a saturated and surface-dried basis, and water absorption were 6.78, 2.74 and 0.71 respectively.

Admixture
A new superplastifier based on carboxylic ether polymer with long side chains (Glenium 51, MAC S.p.A., Treviso, Italy) was used.

Chemical composition		Physical characteristics
SiO2	min. 85%	Fineness	grading 100% < 100 mm
CaO	ca. 1%		BET spec. surf. area 22 mq/g
MgO	max. 1%	Density	2.2 kg/dm3
Al2O3	max. 1%	Apparent density	0.2 kg/dm3
Fe2O3	ca. 2%
Na2O	0.50%
Carbon	ca. 2%
Loss of ignition	max. 5%
Moisture content	max. 1%
Table 1: Silica fume characteristics

Concrete and mix proportions
Ten concrete mixes were prepared with five different water/(cement + silica fume) ratios. 11% of silica fume by wt. of cement was added. The curing regime of 21 ± 2 °C and 72 ± 5% of relative humidity (R.U.) was adopted. The concrete mix proportions are summarised in Table 2.
The mix 1 was chosen as reference mixing [4] and the others were obtained changing the w/(cement + silica fume) ratio and, as a consequence, the nominal strength.
Three batches of 38 dm³ of concrete were cast for each mix to provide the necessary specimens for the investigation.

Mix	1/1 - 1/2	2/1 - 2/2	3/1 - 3/2	4/1 - 4/2	5/1 - 5/2
W/(cement+silica fume)	0.21	0.26	0.30	0.35	0.40
Water (w), kg/m3	126	126	126	126	126
Cement, kg/m3	540	436	378	324	284
Silica fume, kg/m3 (powder)	60	48	42	36	32
Fine aggregate, kg/m3	467	673	775	904	940
Coarse aggregate, kg/m3	1367	1273	1232	1190	1166
Admixture, l/m3	17.9	1405	12.5	10.7	9.4
Table 2: Concrete mix proportions

Test specimens and schedule
For each mix different specimens were done, as a function of the specific characterisation as follows:

• thirteen 150 mm cubes for density, pulse velocity, rebound hammer and standard compressive tests at different ages and curing conditions;
• one beam 600x200x200 mm and one beam 400x200x200 mm for pull out, Windsor probe and pulse velocity tests at different ages;
• one 200 mm cube used to drill small core samples of 28 and 50 mm diameter;
• two cylinders, 100 mm diameter and 200 mm high, for compressive tests.

The real dimensions of specimens were taken into consideration in calculations.
The following tests were done at each age and for each mix:
• compression tests on 2 cubic specimens;
• rebound index and pulse velocity on cubic specimens;
• compressive tests on 12 small core samples of 28 mm and 5 of 50 mm;
• 3 penetration tests on blocks;
• 1 pull-out test on blocks.

Cylindrical specimens were cast for another research program, not reported in this paper.

Batching, casting and curing
A horizontal rotary drum mixer with 50 dm³ capacity was used. The surface of the mixer was wetted before mixing to avoid water absorption and to insure the same mixing conditions for all mixes. First, coarse aggregate, fine aggregate, cement and silica fume were added and then, starting the mixing, water and admixture, at the end. Mixing was continued for about 5 min. Slump test was then done according with Italian standard UNI 9418. The cubic and beam specimens were cast in steel moulds and cylinders specimens in plastic moulds. All the samples were consolidated on an standard vibrating table for 2 min. All the moulds were covered with polyethylene sheets for 24 h, in a laboratory environment, until demoulding. After this time, the specimens were removed from the moulds and transferred to the controlled environment room for the curing. The tests were carried out at 1, 3, 7, 28, 90 days. Two 150 mm cubes were cured in water for standard compressive testing at 28 days.

Non destructive tests and equipment
The following equipment were used for non-destructive testing.

Pulse velocity tests
A microseismic analyzer was utilized, with the following characteristics:

• Piezoelectric transducers, either emitting or receiving, with a flat active surface (emitting or receiving) with a diameter of 30 mm .
• Two on-line commutators, allowing the operator to activate any emitting or receiving probe, among fixed multiple transducers applied to the stone specimen.
• Operator-adjustable pulse repetition rate, avoiding reverberation and resonance giving rise to disturbed measurements or to the impossibility of carrying them out.
• Pulse frequency: this was set at 70 KHz, being the resonance frequency of the emitting and receiving probes, tuned and applied to the specimen.
• Pulse energy: the instrumentation was designed to power the emitting probes at various energy levels from 0.25### J to 50 ###J, according to the absorbing characteristics of the specimen and trajectory lengths. The emitting probes were powered respectively with 2.5 ### J for the smaller specimens (prisms) and 250 ###J for the larger ones (blocks).
• Amplification of received signal: amplification varying between -16 and +80 dB, in 1-dB steps, was used for the electric signal of the receiving probes. The amplification value was both digitally displayed to the operator and available as electric output for recording equipment during measurements.
• Transit time and intensity variations: the CRT (Cathode Ray Tube) of an oscilloscope was used for continual display of time function vibrations, as received by the probe and suitably amplified.

The tests were done according to the italian standard UNI 9524.

Rebound tests
A standard Schmidt rebound hammer type N was used, and the tests were done according to the italian standard UNI 9189.

Pull-out tests
The commercial apparatus FISCHER TCP was used, with post-inserted, forced expansion inserts, according to the italian standard UNI 10157.

Probe penetration tests
The Windsor Probe System was used, according to ASTM C803M-97.

Microcoring
A Hilti DCM-1 drilling machine was used to core the cylinders. A REMET MT70/LS3 apparatus was used to cut and smooth the specimens. A standard laboratory testing machine was used to do the compressive tests (UNI 10766).

Experimental results and comments

All results, for all mixes and ages, were treated together in order to obtain correlation curves between destructive and non-destructive parameters. This was possible, because only the age and proportions were variable.

The regression curves were established according to the law:
y = a.x^b

where the coefficient a and the exponent b were determined by the least square method. The correlation coefficient r and the 95% confidence limits were determined for each method.
The values of the rebound index and the pulse velocity were correlated to the compressive strength separately, as well as in combination, according to the "SonReb" method. In the second case, the following form was adopted for the correlation curve:
z = a . x ^b . y^g

All the experimental results are plotted in Figures 1 to 6.

The correlation laws determined are reported in Table 3, where:

Rs	=	estimated cubic compressive strength (MPa);
Vl	=	Pulse velocity (m/s);
Ir	=	Rebound number;
P	=	Pull-out pressure (bar);
L	=	Exposed probe length (in);
Rc,28	=	Compressive strength of microcores Æ 28 mm;
Rc,50	=	Compressive strength of microcores Æ 50 mm;
r	=	Correlation coefficient.

Correlation law	r2
Rs = 10-28V18.1272	0.84
Rs = 0.000135 Ir 3.4424	0.81
Rs = 0.01714 P 1.6383	0.78
Rs = 4.95654 L 3.4383	0.83
Rs = 29.326 + 0.6144 Rc,28	0.86
Rs = 19.8938 + 0.7633 Rc,50	0.91
Table 3: correlation laws

Comparing the correlation laws found to the ones usually employed for normal concrete, we can observe that the exponents are generally higher. This is due to a lower sensitivity of non-destructive parameters to strength variations, when the strength is very high [1].
Correlations of Figures 1 and 2 can be favourably judged, with reference to the whole range of experimental results, according to the value of r calculated.
With reference to the other curves, it was not possible to use the results relative to the mixes with the highest strength.
The pull-out method (see Fig. 3) showed good performance for the entire strength range, but it was necessary to discard some results, because of anomalous cracking around the testing area.

Fig 1: Compressive strength vs pulse velocity.

Fig 2: Compressive strength vs rebound index.

Fig 3: Compressive strength vs pull-out pressure.

Fig 4: Compressive strength vs exposed probe length.

Fig 5: Compressive strength vs microcore compressive strength (Æ 28 mm).

Fig 6: Compressive strength vs microcore compressive strength (Æ 50 mm).

Fig 7: Compressive strength esitimated by SonReb method


The probe penetration method gave bad results for strength values greater than about 80 MPa, as it can be seen in Fig. 4. The probes couldn't penetrate correctly and in most cases were bent. The correlation law was determined rejecting these values.
With regard to microcoring, the experimental results for 1/1 and 2/1 mixes were rejected, because the apparatus used did not work correctly.
The combined SonReb method gave the best results. This confirm the results known for normal concrete.

Conclusions

In conclusion, we can observe that:
1. the working range of probe penetration method is limited to strength values of about 80 MPa, due to inadequacy of the employed configuration of the Windsor Probe System (silver probes and standard power);
2. the effectiveness of microcores depends on the specimen's preparation;
3. the combined rebound-pulse velocity method gives good results for any strength level;
4. the sensitivity become lower at high strength levels.

Acknowledgements

The authors gratefully acknowledge the components of the technical staff of the Experimental Laboratory for the Strength of Construction Materials at the University of Bologna, namely Mr. Davide Betti and Mr. Gregorio Bartolotta.
Graduate students Marco Ballestrazzi and Michele Colangelo are acknowledged for their collaboration and active participation in the experimental work.
The aggregates have been supplied by LIVABETON, Calcestruzzi Preconfezionati S.p.A. (Bologna, Italy), which is gratefully acknowledged by the authors.
This research was financially supported by the Italian Ministry of University and Scientific Research [MURST contract N°9708165746_006].

References

1. A. Di Leo, G. Pascale, E. Viola, Core sampling size in N.D.T. of concrete structures, ACI SP-82, Ed. Malhotra, Detroit, 984, pp. 459-477.
2. ACI Committee 26, Silica fume concrete. Preliminary report, ACI Materials Journal, 84(2), 1987, pp.158-166.
3. Ministero Lavori Pubblici, Guidelines on structural concrete, In concreto, 1996, pp. 19-23 (in italian).
4. Said Iravani, "Mechanical properties of high-performance concrete", ACI Materials Journal, vol.93 (5), p. 416-426, 1996

© AIPnD , created by NDT.net

|Home|    |Top|