Bundesanstalt für Materialforschung und -prüfung

International Symposium (NDT-CE 2003)

Non-Destructive Testing in Civil Engineering 2003
Start > Contributions >Posters > Ultrasonic: Print

Reference concrete for ultrasonic-testing and its creation by components with pre-analysed properties

Dipl.-Ing. André Glaubitt, Universität Dortmund, Dortmund (Germany)
Dipl.-Geophys. Sascha Bussat, Universität Dortmund, Dortmund (Germany)
Prof. Dr.-Ing. Jürgen Neisecke, Universität Dortmund, Dortmund (Germany)

1.Introduction

The aim of the project is to give prediction about the dynamic E-modulus of concrete by the use of the elastic properties (dynamic E-modulus, Poisson´s ratio and density) of its components (cement-matrix, aggregate).

These predictions should be made by inclusion of the aggregate (e.g. basalt, granite and sandstone), water/cement-ratio and different kinds of cement. However, in the past the acoustic properties could not be measured so exactly like it is possible nowadays by using new measuring methods.

Elastic properties of aggregate and cement-matrix are now defined by measuring of longitudinal and surface waves (Rayleigh wave) with an accuracy of +/- 0,5%.

By this it should be possible to examine existing prediction-models for dynamic E-modulus referring to their accuracy or if necessary create a new model. Using such models it even should be possible to predict ultrasonic wave speed VT, VL (transverse- and longitudinal-wave-velocity) in concrete with a required reliability by measuring of VT, VL of concrete´s separate components.

As a conclusion, it is absolutely necessary to know as much as possible about each component of cement concerning factors having influence on their elastic properties. Otherwise a creation of reference concrete with reproducibility will not be possible.

The predictions can be used for targeted production of reference concrete for ultrasonic-testing to compare measuring of different testers and testing methods. By this it is possible to give statements for example about their accuracy.

2.Determination of elastic properties (dynamic E-modulus Edyn, Poisson´s ratio mdyn)

To determine elastic properties (dynamic E-modulus Edyn, Poisson´s ratio m dyn) as exactly as possible, it is necessary to determine ultrasonic-wave-velocity of longitudinal and transverse wave (VL,VT) and density r.

2.1 Determination of longitudinal-wave-velocity
The determination of longitudinal-wave-velocity is combined with a calibration of the measuring equipment. Therefore several pieces with different length of one sample are cut and scanned by sound in usual way. The measured transit times are now placed against their length in a diagram (fig. 1). The increase of the straight line is the longitudinal-wave-velocity of the sample´s material and the y-axis intercept shows the delay of measuring equipment (see fig.1).

By this method a very exact determination of VL is possible, because several separate measurings are analysed for one result. The accuracy of this method is at least about +/- 0,5% and the determination of the delay is more exactly than by the "face-to-face"-method.

Fig 1: Determination of VL and device-dependent delay.

2.2Determination of transverse-wave-velocity
The determination of transverse-wave-velocity is done by using the measuring of Rayleigh waves (surface waves). This method has been chosen, because test with transverse-probes gave no exact information where the first onset of transverse-wave has to be identified, even by the use of hardening couplant.

To determine the transverse-wave-velocity exactly, the phase velocity VR of the surface wave is measured. If a material is homogeneous, phase velocity is equal to group velocity. A precondition is, that wave velocities VT, VL and density of the sample do not vary by depth. Otherwise an analysis of dispersion has to be made [1].

The measuring of Rayleigh-wave-velocity is done with two compressional probes. While the transmitter is fixed on the sample, the transceiver is placed in different distances to the transmitter. For both probes glycerin is used as couplant. Depending on the distance between transmitter and transceiver different A-scans are recorded, which are placed in parallel over the offset (distance between transmitter and transceiver). The last step to get the Rayleigh-wave-velocity is to define the increase of the straight line through the onsets of regististrated Rayleigh waves.

Fig 2: Principle of measuring Rayleigh-wave-velocity VR.

2.3Determination of modulus Edyn and Poisson´s ratio m dyn
The determination of modulus Edyn and Poisson´s ratio m dyn is now done by using longitudinal-wave-velocity VP, Rayleigh-wave-velocity VR and density r of the material. Therefore the transverse-wave-velocity is calculated. The formula used for velocity of Rayleigh wave is [2]:

Because of knowing VP and VR the formula can be written as followed:

By determing f(VT)=0 the velocity of transverse-wave VT is been given. For further determination of Edyn and m dyn the system of equations with

has to be solved.

2.4 Example
Using the measuring of Rayleigh-wave-velocity, it is uncomplicate to determine elastic properties (Edyn , m dyn) of approximately homogeneous samples. In the following there is an example of measuring and interpretation done on a prism of mortar. Some information about this prism can be seen from table 1.

Cement w/c-ratio CEM I 32,5 R 0,5
aggregate size 99% SiO2-content 0,063-1 mm
specification production under vacuum
Table 1: Information about examined prism of mortar.

Fig 3: Picture of configuration to measure Rayleigh-waves

As a result of the trough-transmission, there is a longitudinal-wave-speed VL = 4440 m/s. The configuration for determination of Rayleigh-wave-velocity can be seen in figure 3., the recorded A-scans with different distances to the transmitter are shown in figure 4. For VR the velocity was determined to 2500 m/s.

The result of the subsequent determination of transverse-wave-velocity was VT = 2655 m/s. The dynamic E-modulus was calculated to Edyn = 35,3 N/mm² and Poisson´s ratio to m dyn = 0,22.

Fig 4: Measuring of Rayleigh-wave-velocity (A-scans dependent on offset).

2.5 Advantages of determination of Rayleigh-waves
There are several advantages to determine transverse-wave-velocity by measuring Rayleigh-wave-velocity:

  • For this method it is sufficient that components are accessible only on one side, because determination of longitudinal-wave-velocity can also be done by use of direct wave.
  • A new acquisition of probes is not necessary, if dimensions of samples do not change.
  • There is no need to use special couplant.
  • Measuring on samples of almost all dimensions are possible. The minimal size of samples just depends on the size of probes and rayleigh-wave has to spread out undisturbed.

3.Prediction-models

The exact determination of longitudinal and transverse-wave-velocities VL and VT gives the possibility to determine Edyn and m dyn . The dynamic E-moduli and Poisson´s ratios m dyn are input data of several prediction-models. With these models the E-modulus of for example concrete should be calculated by the E-moduli and Poisson´s ratios of its components (cement, aggregate). These ideas of models are based upon the assumption that concrete is a two-material-system. Extensive analyses of different ideas of models were made in the past. There is a differentiation between basic-models and complex-models. The basic models do not include the influence of transverse elongation or it is disregarded. The complex models do exist with the influence of transverse elongation and also without [3].

Because of the exact determination of VL and VT and by this Edyn and m dyn , these prediction-models have to be controlled belonging to their accuracy. Therefore concrete´s two components cement and aggregate have to be examined in matters of Edyn and m dyn . Using prediction-models these results lead to a statement of the E-modulus and Poisson´s ratio of concrete produced with these two components.

As a conclusion, it is strongly necessary to examine the influence of different parameters of concrete´s production. By this it should be avoided that the mixing of the two components to one material will have influence on their E-moduli and Poisson´s ratios. If it is not possible to avoid such effects, at least it should be known in which dimension there is an influence.

Therefore one first step is the examination of cement-matrix. Prisms of cement made of CEM I 32,5 R and varying w/c-ratio under vacuum showed a linear variation of longitudinal-wave-velocity VL dependent on w/c-ratio. By the production under vacuum the air- and compaction-pores should be eliminated, so that concerning the pores only capillary pores should have an influence on longitudinal-wave-velocity VL . Theoretically there should not be any capillary-pores in cementstone with an w/c-ratio less than " 0,4. Therefore, if capillary-pores have a great influence on ultrasonic-wave-velocity, the linear dependence of VL on w/c-ratio should be different between w/c-ratio less than 0,4 and more than 0,4. However, this effect can not be seen in the examined test series (see fig. 5).

Fig 5: Longitudinal-wave-velocity of cement-stone (vacuum-production) depending on w/c-ratio.

An examination of cement-stone with a mercury porosimetry analysis will give more detailed information about pore-structure.

4.Outlook

If all important factors that have influence on ultrasonic-wave-velocity of cement-stone are known, it is possible to produce one or if necessary several "Standard-Ultrasonic-Cementstone(s)", which is (are) reproducible.

Afterwards there will be a further examination with diverse test series to discover the influence of aggregate in a model-test-block of "Standard-Ultrasonic-Cementstone". First the aggregate will be, concerning to the idea of a two-material-system, mono-mineral and later even a mixture of several minerals. These aggregates will vary in proportion to cementstone, particle shape, maximum particle size and grading curve.

With these results existing prediction-models will be controlled and tested concerning their accuracy. In the past the examination of such prediction-models was not so accurate, because it was not possible to give detailed information about the accuracy of measured ultrasonic-wave-velocities (there was no calibration of the equipment), nor the Poisson´s ratio was considered in the calculation of the E-moduli. If need be, a new prediction-model has to be developed with the extracted information of the test series.

With the above-named basic steps it is possible to develop one or several reference concrete(s), because fixed specifications about recipe and creation can be made, which will ensure a reproducibility. With these reference concretes it is possible to develop reference-samples for ultrasonic-testing. Such reference-samples are a precondition for comparability, transferability and reliability of measuring results, testing-methods and equipment for the non-destructive testing of cement-bonded mortar and concrete.

References

  1. FORBRIGER, Th. 2001. Inversion flachseismischer Wellenfeldspektren. Ph.D. thesis, Stuttgart University
  2. GLANGEAUD, F., MARI, J.-L., LACOUME, J.-L., MARS, J. & NARDIN, M. 1999 Diapersive seismic waves in geophysics. European Journal of Environmental and Engineering Geophysics, 3, 265-306.
  3. MANNS, W. 1970. über den Einfluß der elastischen Eigenschaften von Zementstein und Zuschlag auf die elastischen Eigenschaften von Mörtel und Beton. Forschungsberichte des Landes Nordrhein-Westfalen Nr.2112
STARTPublisher: DGfZPPrograming: NDT.net