1 Abstract

A lot of refere samples already exist in the domain of non-destructive testing of steel, whereas for concrete-testing a consisten reference sample is up tp now missing. Concrete is a complex medium, that comprises several materials so that the construction of reproducible samples is very difficult. However, reproducible samples are esential for multitude of investigations. Now it could be realised to create a reproducible testing-sample, that has nearly the characteristics of concrete. For that purpose the aggregate of normal concrete will be substitutted by a lot of cylinders (a few hundred) of glassware with equal orientation. Thereby a reproducible distribution of scattering aggregates inside the testing-sample is possible. The 3D-arrangment of aggregates will simplify to a 2D- arrangment of cylinders. The space between the cylinders will be filled with cement. This has to be carried out under vacuum, in order to minimise and control the compaction-pores inside the cement matrix. Ultrasonic-measurrements of longitudinal and transverse-wave-velocities at several locations on the prototype should in combination with porosity measurements show the reproducibility of this testing-sample. Furthermore, these measurements are very useful for the verification of prediction models for Young's modulus of concrete. Such a reproducible testing-sample allows now an accurate coordination between real ultrasonic measurements and software modelling the seismic wave propagation in concrete. Finally it is thinkable to develop this sample for specified problems to a standard-reference-sample for ultrasonic measurements on concrete.

2 Introduction

Standardized reference-sample are a deciding requirement in research and development as well as in practical application of ultrasonic-methods. This fact is reflected in a multitude of norms about kind and application of reference-samples in ultrasonic-testing of metallic, orgnic or ceramic material. Still such reference-samples are lacking for methods and instruments of ultrasonic-testing of concrete-units. The comparability of measurement results obtained by different users as well as the evaluation of significance of different apparatuses and/or methods is up to now difficult or impossible. The intention of this work is to provide a method, that allows everybody to produce comparable to ensure the reproducibility of these reference-samples. Thus it's a question of constructing a modell-sample with nearly the same properties like a conventional concrete.

3 Requirements of a reproduceable sample

For the production of a reproduceable sample some special requirements have to fulfilled. The most important requirements are:

• The materials-characteristics of all concrete parts (aggregate and cement-stone) have to be exactly known. For the wave-propagation these characteristics are the elastic properties; the velocity of logitudinal- and transverse-waves as well as the density.
• The scattering inside the sample, particularly the arrangement of the scattering bodies (aggregate) should be reproduceable. The accurate position of aggregate have to be know.
• The dispensation of pores inside the cement-stone should be well adjustable, because the pores have great effect on the elastic propeties of a concrete sample.
• In ideal case the bond between aggregate and cement-stone should be constant, but also with the possibility to vary.

4 Realisation of these requirements

Certainly it's not possible to accomplish the described requirements with a conventional concrete. Simplifications are essential to realise the reproducibility of such reference samples. On this account a concrete-modell-sample has been developed, that accomplishes the desceibed requirements while having nearly the same characteristics as a conventional concrete.

This concrete-modell-sample is just in development and testing-phase and will implement the described requirements as follows:

4.1 Properties of the wave-scattering
The different kind and position of aggregate inside a concrete-sample generate different propertiews at every repeated production. These properties are for example the scattering of wave-propagation and variations of traveltimes of transmission-measurements (Muller et al;1992)

Fig 1: Arrangement of the 709 cylinders inside the sample. There are 10 different diameters (from 3mm to 14mm) in use. The size of the sample is: 300mm width; 400mm length and 100mm height.

The construction of a reproduceable 3D-arrangement of aggregate is not realizable. For this reason the 3D- arrangement should be simplify to a 2D-arrangement. The aggregate-"balls" will be replaced with aggregate-cylinders. These cylinders will be stacked with the same orientation inside a box and can be filled afterwards.

The arrangement of the cylinders with different diameters is shown in figure 1 and is guided by ideal-grading-curve with a largest grain of 14mm diameter. With this method each other grading-curve is realizable (see also section 4.2)

A repeatable production with an accurate positioning of cylinders can be warranted with two high precise adjusting plates. That way a reproducibility of the position of the scattering-bodies is practicable. Figure 2 shows the hung in glassware-cylinders inside the adjusting plates before the filling.

Fig 2: The sample before filling the spaces with cement. The glassware-cylinders are positioned in adjusting planes.

4.2 Material
The aggregate-cylinders are made of AR-glassware (SCHOTT enterprise), which material characteristics are very accurate published from the manufacturer. For the production of the reference-sample glassware-sylinders with diameters of 3mm to 14mm (3,4,...10,12,14) are used. Every other combination is realisable as well unless the least diameter is greater than 3mm. All together there are 709 cylinders inside the reference-sample. For grain smaller than 3mm no glassware-cylinders are available so that sand could be useful for the production. But considering the small space between the cylinders that have to be filled and the demands on the reproducability, this was not brought in. In conclusion the free space will be filled only with water and cement. The cement is a CEM I 32,5 R. The ratio between water and cement is 0,375, so that undesirable effects by skrinkage can be excluded.

4.3 Pore space
Because porespace has a great effect on ultrasonic wave-propagation and consequently a great effect on elastic properties (Schubert,2000), the pore space will be minimised. In order to reduce the compaction pores the production takes place under vacuum.

As new studies show, the capillary pores have also a great effect on ultrasonic wave-velocities. On this account the accurate observance of water/ cement-ratio is decisive for the reproducibility.

The implementation of defined pore rate is possible by the application of pore builders. That way a defined porosity is realisable.

Further studies relating the analysis of pore spaces will show how far this application is really usable.

4.4 Bond between agregate and cement
In 0rder to vary cylinders roughness it is possible to cauterise the glassware with a composite of hydrofluoric acid (HF) and ammoniumhydrogendifluoride (NH₄HF₂). That way the glassware can be provided with a defined and consistent roughness. The reproducibility of this procedure is realisable. But this cauterisation is very complex in realization, ao that it will not be applied for the present.

4.5 Diamension of reference-sample
The dimension of the reference-sample were fixed on 400mm length, 300mm width and 100mm height.This dimension was selected,because a handling in a laboratory is unproblematic and an undisturbed wave-propagatiom inside the reference-sample is warranted using conventional probes for concrete-testing. Measurements of longitudinal transversal and rayleigh-waves are practicable.

5 Modellimg of elastic wave propogation

To test the undisturbed wave-propagation inside the reference-sample modelling of 2D elastic wave propagation was performed. Furthermore amplitude spectrums were calculated to check on characteristic frequencies, that were caused by eigenfrequencies of glassware-cylinders. This analysis shows no significant characteristics, nor does it while modelling with elliptical glassware. The modelling method used,is a finite difference application using the ratated staggered grid (Bohlen,1998) and (Saenger,2000) for handling the strong contrasts of velocities inside the reference-sample and at its surface. The figures 3 and 4 show snapshots of wave-propagation inside the reference-sample at t =20ms and t =35ms. Only the energy of longitudinal wave is displayed.

Fig 3: Wave propogation of longitudinal waves inside the reference sample at t=20ms.

Fig 4: Wave propogation of longitudinal waves inside the reference sample at t=35ms.

Thus the first database was created for a comparison of the output of modelling software with the material measurements.

6 Applications for reproducible reference-samples

The development and production of reproducible samples open up a multiplicity of favourable and necessary application types, that are exemplarily called below:

• For the first time a very good tuning of different modelling methods is possible on an existing test specimen. The simulation of wave propagations can thus be controlled and optimized.
• By means of the presented procedure many further investigations are conceivable, that are due to the defined mode of production of sample feasible now. Among these investigations rank:
  • the influence of the grading curve,
  • the influence of the roughness of the aggregate and
  • the distribution of pore-size of cement-stone
  on the ultrasonic propagation. A permanent alignment with modelled data would be conceivable.
• Due to the exactly defined characteristics of the sample forecast models can be examined at it. Such forecast models try to give the elastic characteristics of a finished concrete with elastic characteristics of the concrete components as input.
• Production of identical reference-samples at any place and any time.
• Measuring methods can be compared fast at a uniform reference-sample.
• Development to a standard-reference-sample is possible. Such a standard-reference sample provides users of ultrasonic-testing a possibility to check their apparatuses and ultrasonic methods (similary the reference-samples for testing steel). For this employment still defects have to be defined within such a reference-samples. Furthermore standard-reference-samples for different tasks (grading curves) can be produced.

7 Results

A procedure for the reproducible production of reference-samples for ultrasonic testing of concrete was developed. Thereby a model of a concrete with a 2D-arrangement of glassware-cylinders isused. The material properties of the components are exactly known and the pore space can be controlled by a production under vacuum. In order to compare real measurements on this reference-sample with numerically computed data first modellings were performed. An advancement to a standard-reference-sample for nondestructive testing of concrete is conceivable.

8 Outlook

The procedure for the production of reproducible sample for ultrasonic testing of concrete has to be tested moreover concerning on the reproductibility. Measurements of the longitudinal wave velocity at different position will be performed. Furthermore measurements of the trasversal wave velocity will be accomplished by means of measurements of the rayleigh wave velocity. The results of these measurements are useful to examine forecast models, that provide the elastic properties of the finished concrete on the basis of elastic properties of the concrete components.

The accomplished measurements will be compared with the synthetic data,to check the numerical methods of modelling software.

The applicability of this manufacturing process should be evaluated regarding to the development of standard-reference-samples for non-destructive ultrasonic testing of concrete.

References

1. BOHLEN, T.1998. Viskoelastische FD-Modellieung seismischer Wellen zur Interpretation gemessener Seismogramme. pH.D. thesis Kiel University.
2. BOHLEN,T. 2002 parallel 3-D viscoelastic finite-diffence seismic modelling. Computers & Geosciences, 28(8) 887-899.
3. MuLLER,G.,ROTH,M.,& KORN,M. 1992. seismic-wave traveltimes in random media. Geophysical Journal InternationaL, 110,29-41.
4. SAENGER,E.2000. Wave propagation in fractured media: theory and applications of the rotated staggered finite-difference grid. Ph,D. thesis, Karlsruhe University.
5. SCHUBERT,F 2000. Ausbreitungsverhalten von Ultraschallimpulsen in Beton und Schlussfolgerungen für die zerstörungsfreie Prüfung. ph.D. thesis, Dresden Universitat.