Keywords: Compressive Strength, Concrete, Standards, Ultrasound
This paper was presented at the International Symposium Non-DestructiveTesting in Civil Engineering (NDT-CE) 26.-28.09.1995 in Berlin. NDT-CE, Full Program or the Ultrasound Part

Table of contents
1. Introduction
2. The Method
3. Factors Influencing the Pulse Velocity
4. General Comparison of the Standards Velocity
5. Objections
6. Conclusions

Introduction

Nondestructive testing of concrete is rapidly gaining in importance due to the deteriorating infrastructure. Many test methods have been offered but none of them are satisfactory. Nevertheless, the method based on longitudinal pulse velocity has become popular because of its simplicity and cost effectiveness.

Most nations have standardized procedures for the performance of this test ( 1). Out of these standards, the following eight are analyzed in this paper:

• DIN/ISO 8047 (Entwurf) "Hardened Concrete - Determination of Ultrasonic Pulse Velocity" in German
• "Testing of Concrete - Recommendations and Commentary" by N. Burke in Deutscher Ausschuss fur Stahlbeton (DAfStb), Heft 422, 1991, as a supplement to DIN/ISO 1048 in German
• ASTM C 597-83 (91) "Standard Test Method for Pulse Velocity Through Concrete"
• BS 1881: Part 203: 1986 "Testing Concrete - Recommendations for measurement of velocity of ultrasonic pulses in concrete"
• RILEM/NDT 1 1972 "Testing of Concrete by the Ultrasonic Pulse Method"
• GOST 17624-87 "Concrete - Ultrasonic method for strength determination" in Russian
• STN 73 1371 "Method for ultrasonic pulse testing of concrete" in Slovak (Identical with the Czech CSN 73 1371)
• MI 07-3318-94 "Testing of concrete pavements and concrete structures by rebound hammer and by ultrasound" technical guidelines in Hungarian.

The comparison and analysis of the test methods are followed by a critical evaluation. This is necessarily subjective; nevertheless, it is hoped that it will help in the improved use of the ultrasonic pulse velocity method and contribute to the improvement of future specifications.

Several standards use the term "measurement" (Messung) or equivalent, of pulse velocity. This is not exactly correct because only the distance between the two transducers and the transit time are measured directly. (The transit time is the time taken for the onset of the pulse to pass through the concrete.) The pulse velocity is calculated from these two. Nevertheless this misnomer does not cause much confusion.
For the sake of clarity, the text dealing with specifications in the DIN/ISO stanard are written in italics, everything else is written as plain text.

The Method

The basis of the comparison is the DIN/ISO 8047 (Entwurf) standard. It is divided into 7 chapters and an Appendix. Each chapter contains several subchapters. These subchapters will be described briefly and compared to the pertinent chapters of the other standards. In the comparisons, differences between the test methods are emphasized. If there is no comparison of DIN to ASTM concerning a particular point, this means that the two standards are similar with regard to that point.

Goal and Use
The scope of DIN/ISO is restricted to the determination of the velocity of longitudinal ultrasonic waves in concrete. This so-called "pulse velocity" may be used for the assessment of the uniformity of concrete in a structure, measurement of layer thickness of a concrete of inferior quality, monitoring changes in oncrete with time, and detection of defects and anisotropy. It is also permissible to use it for the assessment of the strength of concrete if reliable calibration curves are available. It is pointed out, however, that this ultrasonic test is not an acceptable substitute for the standard, destructive strength determination. Determination of the elastic constants is not mentioned in the DIN.

Similar restrictions and uses are given in the other standards, especially in the ASTM and RILEM standards. Most of them, however, permit the assessment of the elastic constants from pulse velocity measurements. The BS also offers explanations for the various uses of pulse velocity.

Basic Principles of the Test Method
The method specified in all standards is based on the same principle. Pulses of longitudinal ultrasonic waves are generated by an electro-acoustical transducer that is held in contact with the surface of the concrete under test. After traversing through the concrete, the pulses are received and converted into electrical energy by a second transducer. The velocity v is calculated from the distance 1 between the two transducers, and the electronically measured transit time t of the pulse as v = l/t.

Apparatus
Generally, the apparatus consists of a pulse generator, a pair of transducers, an amplifier, and an electronic timing device for measuring the transit time. According to the DIN, the generator should have: a precision of +/-1% in time measurements, short rise time, a capacity for a low frequency generation, and field worthiness. For short path lengths the use of transducer of high frequency (60 to 200kHz) is recommended; for long path lengths, low frequency (10 to 40 kHz) is recommended Transducers with in the frequency range of 40 to 60 kHz are acceptable in most cases. The timing device should be sensitive enough to be triggered by low amplitude pulses.

The ASTM also specifies that

• the pulse generator should produce repetitive pulses at a rate of not less than 10 pulses per second nor more than 150 pulses per second;
• the time measurement should hay: the precision of 0.5%;
• the voltage generated by the transducer should be amplified as necessary to produce triggering pulses to the time measuring circuit;
• a calibration device should be provided for the purpose of checking the proper operation of the time-measuring circuit.

The British standard offers a mettod for checking the the precision of the transit it measurement. The Hungarian specification requires a 0.1µs precision of the time measurement.
According to the GOST, the limits If the permissible absolute error of the measurements of transit time of the standard specimens should not be more than delta = +/-(0.01t + 0.1), where t is the transit time in µs. Also, the deviation If the individual measurements of the transit tinge of a specimen from the average value of the measurements of the same specimen should not be more than 2%.

According to the STN, the precision of the test apparatus on the reference bars should be +0.01 As if the surroundingtemperatureranges from -10 to +45°C, and the humiditvis not more than 80%. RILEM provides details about transducer characteristics.

Procedure
The DIN describes three possible arrangementsfor the transducers for velocity determination. These are:

• the transducers are located directly opposite each other. This is the most sensitive arrangement, and is called direct transmission;
• the transducers are located diagonally with respect go each other, that is, the transducers are across corners. This is less sensitive than the direct transmission, and is called diagonal transmission;
• the transducers are attached to the same side surfaces This is the least sensitive arrangement and is called indirect transmission.

Whenever possible, the direct transmission arrangement should be used and on the surfaces that were in contact with the mold.
It is essential that there be adequate acoustical coupling between the concrete and the face of each transducer. For most concrete, the surfaces are usually smooth enough to secure a good transfer of ultrasound if a thin layer of an appropriate coupling agent is applied. The accuracy of the transit time measurements should be checked against a calibration device before very series of measurements. The distance between the two transducers should be measured vith the precision of +/-1%, and the transit time should be recorded to three significant digits.

The most detailed description of the measurements with any of the three transducer arrangements is presented in the BS. The details deal with calibration, accessories, such as cathode ray oscilloscope, digital instruments, etc. According to ASTM, repeat measurements should be made at the same location to minimize erroneous readings due to poor contact. It is suggested in the RILEM as well as in the Hungarian specification that before testing, the concrete surface be smoothed when it is rough. RILEM also provides details about the transit time measurements with an oscilloscope both by the maximum amplitude technique and by the fixed amplitude technique. Both the BS and the STN warn that the indirect transmission produces lower pulse velocities than the direct transmission method. The GOST specifies that the maximum depth and diameter of voids on the contact area must not exceed 3 mm and 6 mm respectively, and the maximum height of any proturbance should not be more than 0.5 mm.

Calculations
All standards specify that the pulse velocity v should be calculated as

v = l/t (1)

According to the Slovak standard, the pulse velocity determined in a one- or two dimensional specimen should be recalculated for the equivalent pulse velocity in a three-dimensonal-specimen, as follows

v_l3 = k₃ v_l1 (2)

v_l3 = (k ₃/ k₂) v_l2 (3)

where
v_l1 = pulse velocity in one-dimensional specimen, such as bar
v_l2 = pulse velocity in two-dimensional specimen, such as plate
v_l3 = pulse velocity in three-dimensional specimen.

The values of coefficients k₂ and k₃ are dependent on the value of dynamic Poisson's ratio pcu, and can be obtained, as follows:



The GOST permits the use of the transit time t, instead of the velocity when the value of 1 is kept constant.

Report
The DIN/ISO provides detailed instructionsfor the preparation of the test report. This includes: a description of the tested structure or specimen; specificationsfor the concrete; concrete composition; curing condition; and age; the testing apparatus and procedure; arrangement of the transducers; location of the reinforcement; properties of the concrete surface; estimated moisture content; path length; pulse velocity in various directions; and other meaningful information.

The requirements of other standards are shorter but cover essentially the same items for the report. The ASTM requires the measured transit time and also the adjusted transit time. The BS specifies the recording of the date, time and place of the investigation. The Hungarian specification requires the name of the client, the purpose of the testing, the names of the performers of the measurements, the apparatus used, visual observations, and details of the sampling.

Precision
The DIN/ISO states in the Appendix that the precision of the transit time should be checked. If this checking takes place with a calibration bar, the transit time should be known with the precision of +/-0.2s. Measured values should not differ more than +/-0.5% from the known value of the calibration bar.

According to the ASTM precision statement, tests involving three test instruments and five operators have indicated that, for path lengths from 0.3 m to 6 m through sound concrete, different operators using the same instrument, or one operator using different instruments, will achieve repeatability of time test results within 2%. In the case of deteriorated concrete, the variation of results are substantially increased. In such cases, however, calculated velocities will be sufficiently low as to indicate clearly the presence of distress in the concrete tested.

Factors Influencing the Pulse Velocity

Several factors are discussed in the DIN that can influence the measured transit time beside the quality of concrete. These are:
• temperature (effect is not significant within practical limits)
• too short path length (path should be longer than 100 - 150 mm for direct transmission, and longer for indirect transmission)
• microcracks (can reduce the velocity)
• moisture in concrete (can slightly increase the velocity).

The effects of two additional factors are discussed below.

Size and Shape of the Specimen
The dimension of the concrete specimen in the direction of pulse propagation should be at least 80 mm when tested with ultrasound of 40- to 60-kHz frequency. Smaller specimens should be used with caution.

The STN regulates the ultrasonic wavelength according to the shape and dimensions of the tested elements. The shapes are defined, as follows:

• a specimen is one-dimensional when a and b < 0.2 l_L (bars, prisms, cylinders, and beams),
• a specimen is two-dimensional (planar) when b < 0.2 l_L. These are thin slabs;
• otherwise, the specimen is considered three-dimensional (cubes, short prisms and cylinders and beams).
  In such cases the criterion is
  a > 2 l_L and b > 2 l_L

In the case of thick slabs, in which case the transducers are placed on opposing surfaces, the criterion is T > 0.9 l_L where

a, b = the dimensions of the cross-section normal to the directions of transmission.

T = the thickness of the slab,

l_L = is the wave length in the concrete determined from the relation l_L = v_L / f_u, where v_L is the pulse velocity of ultrasonic wave in the concrete and f_u is the frequency of the ultrasonic wave motion in the concrete.

The effect of the specimen shape on the pulse velocity should be taken into consideration by Eqs. 2 through 4.

Effect of Steel Reinforcement on Pulse Velocity
Steel reinforcement increases the measured pulse velocity, when it is in close proximity to the pulse path. This influence is especially strong when the reinforcement is parallel to the direction of pulse propagation. The increase, however, is negligible if the distance between the steel surJace and the path is more than a sixth of the measured length. The influence of the steel reinforcement perpendicular to the direction of the measurement is very small, with the exception of heavy reinforcement.

If it is not possible to avoid wave propagation paths which are parallel to the reinforcing bars, and the path is in the vicinity (a/I < 0.25) of a steel bar, the British Standard recommends the following correction of the measured pulse velocity:



where

v_c = corrected pulse velocity in the concrete, km/s

v_s = pulse velocity in the steel bar, km/s

a = distance from the surface of the steel bar to the line joining the nearest point of the two transducers, mm

t = transit time, ms

l = length of the direct path between transducers, mm.

Eq. 6 may be modified to give the following:

V_c = k v_m(7)

where
V_m = I/t = measured apparent pulse velocity, km/s
k = correction factor given by k = g + 2(a/l) (1-g²) in which g = v_c/v_s.

The effects of reinforcing bars which have axes perpendicular to the direction of wave propagation and with diameter less than 20 mm, may be ignored. The GOST also specifies that the measurements of the transit time should be made in the direction perpendicular to the direction of steel reinforcement. The concentration of the reinforcement along the path of the wave propagation should be less than 5%. Measurements along the path parallel to the direction of steel reinforcement are permissible if the distance between the path and the steel surface is more than a sixth of the measured length.

The STN also states that measurements perpendicular to the direction of the reinforcement are preferred. In this case, the effect of the steel bars is negligible, unless the steel concentration S is

	where di = diameters of the reinforcing bars n = number of reinforcing bars l = path length. When the steel concentration is greater, the effect of reinforcing bars perpendicular or inclined to the path length is expressed by the following equation:
	where vs = pulse velocity in the steel bar, km/s vcs = pulse velocity in the reinforced concrete by measuring in the direction perpendicular or inclined to the direction of the reinforcing bars, km/s.

The RILEM specification presents somewhat different formulas for the effects of parallel and perpendicular reinforcement.
The Slovak standard severely restricts the velocity measurement if the path would be parallel to the direction of the reinforcement. In such cases, the location of transducers should be outside of the area influenced by the reinforcement. The assumed influenced area is a cylindrical surface with an approximate diameter of l/6.

General Comparison of the Standards Velocity

Most of the examined ultrasonic standards were issued more than ten years ago. This may indicate a lack of progress in testing concrete ultrasonically.
Beyond the general similarities, there are less general similarities For instance, the DIN/ISO is closer to the ASTM than to the others. Similarly, the BS and RILEM closely resemble each other, and the STN and GOST are similar. The reasons for these similarities are probably geographic and/or political. The Hungarian specification of 1994, for instance, does not show any dependence on the Russian specifications any more.
The ASTM, DIN/ISO and Hungarian specification are quite compact. They concentrate more on the specification of measurement of the transit time. Other standards provide details also for the assessment of concrete strength from the pulse velocity, and for the assessments of other concrete properties, such as elastic constants, defect detection, and determination of concrete uniformity. Another reason for the considerable length of the British standard is that many formulas are presented and also explained, as in a textbook.

Objections

Many statements and specifications in the analyzed standards are supported by the literature, but not all of them. It would be prohibitively lengthy to discuss this in a comprehensive manner, therefore the illustration is restricted to previous works of the writers of this paper. For instance, it has been shown (2,3), that the assumption that pulse velocity is independent of the size and shape of the specimen, path length, frequency, and stresses in the concrete is acceptable but only as a first approximation.

Much more important from engineering point of view is the writers' main objection against the analyzed standards. This is that the standards do not warn the user about the pitfalls of the assessment of concrete properties from longitudinal pulse velocity. Most standards list about half a dozen possible applications of this ultrasonic test, such as assessment of strength, elastic constants, defect detection, etc., frequently supplemented with formulas However, none of the standards rate these applications according to their reliability. This is unfortunate because it gives the impression that the pulse velocity test is equally suitable for all these applications, which is not the case (4,5). In fact, the best, and perhaps the only reliable, applications of the longitudinal pulse velocity are for (a) checking the uniformity of concrete, and (b) monitoring the changes in a concrete over time. Strength estimation is possible only within a 20% accuracy, and even this can be achieved only under strict laboratory conditions with an established calibration curve. This low accuracy is not improved by supplementing the pulse velocity measurement with other tests, such as the rebound hammer test (6) The other suggested applications of pulse velocity (defect detection, crack depth measurement, etc.) are even less reliable.

The present state of ultrasonic concrete testing clearly needs improvement. The first step toward this may be a warning in the standards about the uncertainties of the use of the standardized longitudinal pulse velocity method. Further improvement should come from a better understanding of the theory of ultrasonic pulse propagation in concrete. This may lead to the use of surface and other guided waves, and advanced signal processing techniques (7,8). Unfortunately, these writers do not know of any standards dealing with such tests.

Conclusions

The eight standards and specifications analyzed show considerable similarities for the measurement of transit time of ultrasonic longitudinal pulses in concrete. Nevertheless. there are also differences. Some standards provide more details about the applications of the pulse velocity, such as strength assessment, defect detection, etc. It has been established, however that the accuracy of most of these applications, including the strength assessment, is unacceptably low. Therefore it is recommended that future standards rate the reliability of the applications.

It also follows that the present state of ultrasonic concrete testing needs improvement. Since further improvement may come from the use of surface and other guided waves, advanced signal processing techniques, etc., development of standards for these is timely.

Acknowledgment
This paper was partially sponsored by the U.S. - Slovak Science and Technology Program.

References

1. Teodoru, G., Zerstorungsfreie Betonprufungen (Nondestructive Testing of Concrete), Beton-Verlag, Dusseldorf, 1989. 158 p.
2. Popovics, S., Rose, J. L., and Popovics, J. S., "The Behavior of Ultrasonic Pulses in Concrete," Cement annul Concrete Research, Vol. 20, No. 2, 1990. pp. 259 - 270.
3. Popovics, S., Popovics, J. S., "Effect of stresses on the ultrasonic pulse velocity in concrete," Materials and Structures - Research and Testing, RILEM, Vol. 24, No. 139, Paris, January 1991. pp. 15 - 23.
4. Popovics, S., Popovics, J.S., "Misapplications of the Standard Ultrasonic Pulse Velocity Method for Testing Concrete," Structural Materials Technology - An NDT Conference, Scancella, R. J., Callahan, M. E., (Editors), Technomic, Atlantic City, New Jersey, February 23 - 25, 1994. pp. 241 - 246.
5. Popovics, S., and Popovics, J. S., "A Critique of the Ultrasonic Pulse Velocity Method for Testing Concrete," Nondestructive Testing of Concrete Elements and Structures, F. Ansari and S. Strue, Editors, Proc. ASCE, San Antonio, April 1992. pp. 94 - 103.
6. Popovics, S., "Stato attuale delta determinazione delta resistenza del calcestruzzo mediante la velocita degli impulsi in America" (Present State of the Determination of Concrete Strength by Pulse Velocity in America), II Cemento, Anno 83°, No. 3, July September 1986. pp. 1 17-128.
7. Popovics, S., and Popovics, J. S., "Potential ultrasonic Techniques Based on Surface Waves and Attenuation for Damage Evaluation in Concrete - A Review," Diagnosis of Concrete Structures, T. Javor, Editor, Proceeding of the International RILEM - IMEKO Conference, Experteentrum, Bratislava, 1991. pp. 101 - 104.
8. Popovics, J. S., "Are Advanced Ultrasonic Techniques Suitable for Concrete? - An Exploratory Investigation," Proceedings, Nondestructive Evaluation of Civil Structures and Materials, B. A. Suprenant, et al., Editors, University of Colorado, Boulder, Colorado, October 1990. pp. 327 - 339.