2nd International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, New Orleans May 2000.

TABLE OF CONTENTS

Introduction Why test concrete? NDT methods for testing concrete Examples of NDT methods NDT for location and inspection of pre-stressed cable ducts The tests The AGFA Strukturix DPS system The test objectives Results Conclusion - High Energy radiography and The AGFA Strukturix DPS-system

Introduction

Concrete structures are normally cast on site using locally available materials and in conditions, which may vary considerably throughout the construction process. The concrete material itself is usually prepared according to careful design and quality controlled by the manufacturer. The successful casting of large volumes of concrete in heavily reinforced and awkward structures can however not always be achieved at this most critical stage. The uniqueness of most large structures can introduce unpredictable factors with the risk of upsets during the pouring stage and intervals between pours. The use of innovative techniques intended to make the work more efficient, can lead to deviations from the intended quality and conformance of the finished product. Once started a large concrete pour is normally completed and the outward appearance of the concrete surface may not reveal internal defects.

Quality control of the finished product may be carried-out by visual inspection and in some cases by performance monitoring, for example by measurement of concrete strain under load or in the long term due to creep. The general behaviour and appearance of the structure is thus checked. It has been shown however that critical damage and deterioration to concrete structures is often less dependant on the general condition than on specific defects, particularly in the case of pre-/post-tensioned structures or those in a (potentially) aggressive environment.

The uncertainty regarding the actual condition of a finished concrete structure has not been generally addressed due partly to the lack of sufficiently reliable and effective inspection methods. Concrete has a heterogeneous nature; it is susceptible to change its uniformity and properties locally at the fresh stage and in the long-term due, for instance to the effects of moisture movement. Non-destructive seismic and radar testing methods are affected by these factors making the process of wave propagation prediction uncertain.

Defects in concrete must also be treated from the point of view of engineering significance and not purely on type, size and position. A non-destructive survey of a large concrete structure must therefore be made efficiently, using techniques with predictable behaviour, reliability and capability, as well as with understanding of defect significance in an engineering context. This requires harmony between structural engineering expertise, preferably with historic knowledge of the structure in question, and inspection personnel with access to non-destructive technology with proven performance for the case in question. Qualitative surveys of concrete are relatively simple to carry out, but usually leave many questions unanswered. Quantitative surveys are possible with the techniques available today, although a key factor is pre-information - in short it helps to know what you are looking for.

Significant developments in NDT-technology have recently been made which have extended the boundaries of application, reliability, detail and capability. Some of these methods are presented below with examples of performance on real structures and mock-ups with well-defined defects.

Why test concrete?

There are two main reasons for testing the condition of concrete structures:
To obtain information which can be used in the prediction of service life
To establish the actual conformance of the structure with design

The service life of a structure may be affected by the ingress of aggressive elements, which in time may cause corrosion of embedded steel. Long-term creep of concrete and relaxation of pre-stressed reinforcing can result in loss of compression forces in safety-related structures.

The actual conformance of a structure may be different from actual documented design due to a number of factors, for example the use of on-site detail changes which have not been recorded, or as is often the case, due to workmanship-related problems.

In the majority of cases the more ambitious inspection programs, some of them involving NDT, are initiated by observations of deterioration, i.e. at a late stage. It may also be necessary to establish parameters, which can be used in load-bearing capacity calculations or simply to determine the position of reinforcement prior to intrusive works.

Extensive cracking of concrete can lead to loss of reinforcement bond, as can settlement of the concrete at its plastic stage during casting. Voids can constitute weak zones and loss of anchorage capacity, and in some circumstances can lead to serious corrosion of embedded steel.

A particular problem and one which has received much attention due to catastrophic failure of some major civil engineering structures is that of corrosion to pre-/post-tensioned reinforcing cables or bars. These are normally placed inside ducts, which are either injected with a cement grout or filled with some other corrosion-inhibiting material such as oil or grease. In the former case there is no possibility of removing the cables for inspection. Voids inside cable ducts may occur due to unsuccessful grout injection thus increasing the risk of corrosion to pre-stressed steel. Inspection of these cables requires first location - usually behind one or more layers of reinforcing- detection of voids and inspection of the physical condition of the cables themselves.

In Table 1 a summary is given of potential defect/damage types and problem cases that are frequently encountered.

Item/Object of test	Description
1	Cracking perpendicular to surface concrete
2	Cracking (internal) parallel with concrete surface
3	Deterioration of concrete with time
4	Damaged (weak) concrete layers
5	Elastic properties of concrete (E-modulus)
6	Elastic properties of concrete (G-modulus)
7	Thickness of concrete member or layer
8	Voids and inhomogeneities
9	Reinforcement location
10	Reinforcement diameter
11	Damage to reinforcement
12	Location of pre-stressed cable ducts
13	Detection of voids in pre-stressed cable ducts
Table 1. Non-destructive testing of concrete - Point of interest / Damage type

The items in bold letters are those which are commonly encountered.

NDT methods for testing concrete

Some of the methods that are used for testing metals can also be used for concrete, the difference being that lower seismic frequencies are required (ultrasonics and related methods) and higher energies are required (radiography). In addition there are some methods such as radar, which can be applied to reinforced concrete but are of little use for testing highly conductive materials such as metal.
Concrete is unique in the sense that its properties usually vary for different strengths, on a micro- and macro-scale due to aggregate size and shape as well as moisture content and due to inherent inhomogeneities, not least as a result of its treatment during casting. Seismic and radar techniques must therefore be calibrated for a given concrete type and the environment. Wave velocities and reflected energy, as well as damping effects on these waves will differ significantly and can be unpredictable.

In Table 2 a list is given of commonly used NDT-techniques and their applications to reinforced concrete.

Test Method	Principle	Application (cf. Table 1)	Capability
UPE	Ultra-sonic P- or S-Wave transmission & reflection.	1,2,3,4,5,6, 7,8,9,12,13	Maximum thickness of concrete approx. 1000 mm. Can detect a 100 mm void at 300 mm or a 200 mm void at 500 mm.
UPV	Ultra-sonic wave velocity & attenuation	1,2,3,5,8	Wave travel time only. Interpretation demanding & usually requires a complementary technique.
Impact Echo	Transmission & reflection of transient low frequency P-waves.	1,2,3,5,7,8	Thick concrete members. Simple problem cases and simple geometries. Interpretation can be difficult.
SASW	Rayleigh Wave velocity with depth.	3,4,6,7,8	Defect must be relativley large in relation to depth from surface. Unlimited depth measurement provided access at surface is sufficient.
Radar	Transmission & reflection of electro-magnetic waves.	9,12	Maximum penetration in moist and heavily reinforced concrete approximately 150 to 300 mm.
HER	High energy radiography.	8,9,10,11,12,13	Maximum concrete thickness 1500 mm (7.5MeV Betatron)
Table 2. Non-destructive testing of concrete - Testing Methods

The applications shown in bold text are those, which are considered to be the most appropriate for the respective methods based on the experience of the author.

Examples of NDT methods

1) Impact Echo
This technique is based on the propagation and reflection of low-frequency transient P-waves which are generated by mechanical impact on the concrete surface. A wave front will travel through the concrete with a velocity of typically 4000 m/s to 4500 m/s depending on the quality and strength of the concrete. The wave will reflect at interfaces between materials of different acoustic impedance such as voids or cracks and a periodic wave motion is created between the test surface and the interface. The frequency of peak echo signals is determined from a frequency spectrum and the depth to reflecting surfaces can be calculated.

Fig 1: Impact Echo response from an 800 mm thick solid concrete slab using a light impactor. Upper diagram shows the wave response while the centre and lower diagrams show individual and average values of the frequency spectrum.

In Fig.1 an example is shown of a typical response from a concrete slab with no defects at the point of test. A bottom echo with frequency 2.69 kHz can be seen to the left, although this echo is not dominant. The test was made with a relatively light hammer, which produces a lower-energy, higher frequency wave packet. The peak echoes around 5 and 9 kHz originate from what is thought to be casting joints in the slab. This effect may be caused by the segregation of some aggregates or air-void concentrations at the joint interfaces, as no defects exist at these levels.

In Fig.2 can be seen the result of a test at the same location but with a larger impactor. This has caused lower-frequency signals to be created with a higher energy level. The bottom echo is now dominant and the effects of internal 'interfaces' become insignificant.

Fig 2: Impact Echo response of 800 mm thick slab with a large impact source. The bottom echo has a frequency of 2.64 kHz corresponding to a P-wave velocity of about 4300 m/s.

This simple example demonstrates the importance of understanding the response of a given concrete structure, as well as the correct choice of test parameters and input. Impact Echo will normally detect defects, which have dimensions perpendicular to the direction of wave propagation of the same order as the measuring depth. This method does not produce a focussed wave packet, which tends to restrict its use to relatively simple geometries. Depth estimation to internal defects and thickness measurement can usually be made with an accuracy of < 5 %.

2) Spectral Analysis of Surface Waves (SASW)
This method is based on the propagation of mechanically induced Rayleigh waves. By striking the concrete surface with a light hammer a transient stress wave is created, including surface or Rayleigh waves, which are registered by two transducers placed in line with the impact point on the concrete surface at fixed separations. The transducers, in our case small accelerometers, register the passage of the waves. The receiver outputs are plots of the phase difference between the two transducers as a function of frequency. A profile of Rayleigh wave velocity versus wavelength, or so-called dispersion curve, is calculated from the phase plot. The ratio of Rayleigh wave velocity to shear wave velocity is approximately 0.9:1; thus the shear wave velocity can be estimated. The shear stiffness (G) of the concrete can be calculated from the shear velocity if the material density is known. Thus we obtain a plot of concrete stiffness as a function of depth from the surface.

This principle is used to measure the thickness of concrete layers with different stiffness properties. It can also be used to locate and roughly delineate inhomogeneities such as voids.

Fig 3: SASW-result (black line) from a 200 mm thick concrete repair. The deviation in the curve from a velocity of around 2300 m/s to approximately 2000 m/s suggests that there is a lower strength material at about 120 mm from the surface. A model was calculated on the basis of this (blue line) and suggested that there was a weaker layer of concrete approximately 60 mm thick below 120 mm from the surface. This was confirmed by opening up the concrete - a weaker layer existed due to lack of aggregates in a layer of cement paste.

The above examples demonstrate some uses of these seismic testing techniques. Their usefulness as individual techniques is perhaps limited, but they can provide reliable qualitative and quantitative information. Reliability of results and defect characterisation possibilities are improved if two or more methods are combined.

NDT for location and inspection of pre-stressed cable ducts

The most useful techniques for this purpose at the moment are high-frequency radar and radiography. Seismic techniques such as described above and including ultra-sonic pulse velocity have been found to give ambiguous results. Newly developed UPE-equipment may have a good potential for locating cable ducts with greater depth penetration and resolution in the case of heavily reinforced and moist concrete compared to radar.

Radar
As mentioned above radar has limited penetration ability in concrete due to its relatively high moisture content. Also the near-surface reinforcing will tend to screen deeper lying objects such as cable ducts. None-the-less, high frequency radar has the advantage that no surface contact with the concrete is required, a scan can be make rapidly and with good resolution. The waves reflected from a steel duct in concrete are normally very clear as the steel duct has significantly different electro-magnetic impedance properties compared to concrete. This enables steel ducts to be located quickly and accurately to a maximum depth of between 150 and 300 mm. Radar is thus a good complement to more detailed techniques such as high-energy radiography. Radar waves are reflected from the steel duct and will not provide information about the condition of the injection grout or pre-stressed cables inside.

Ultrasonic Pulse Echo
Recently developed UPE-equipment has shown some promising results both for locating concrete inhomogeneities as well as location of cable ducts. UPE may be a useful complement in this type of investigation as the penetration depth of the signal is considerable (around 800 mm in high quality concrete). Also, ultrasonic pulses are less sensitive to the screening effects of near surface reinforcing, enabling deeper lying objects to be detected.

Fig 4: Result of UPE test of a reinforced concrete beam with a pre-stressed cable duct. The duct wall is recognised at approximately 80 mm from the surface (hyperbolic line with multiple reflections in the centre of the image).

High Energy Radiography
High energy radiography has been found to be the only reliable NDT-method that can locate pre-stressed cable ducts, determine the existence and size of voids in the grout filler inside the ducts, and also to enable inspection of the cables themselves. The method has not been fully exploited on site, possibly because of radiation safety considerations and possibly because of misunderstandings relating to cost.

A series of tests were carried-out to establish the capability of modern portable high energy radiographic equipment. In this case a Betatron of type PXB 7.5 MeV was used. This is a small accelerator, which produces pulsed radiation energy.

Fig 5:Radiographic image of 1200 mm thick concrete using a 7.5 MeV Betatron and Agfa Strukturix DPS image plate. The three holes in the centre have a diameter of 20 mm and depth 15, 20 and 40 mm (from top to bottom).

The tests

The object of the tests was primarily to determine the capability of certain NDE methods in detecting voids in cable ducts cast in concrete. Only high energy radiography is considered here. A 7.5 MeV Betatron was used as the source and comparisons were made between the AGFA Strukturix Digital Phosphor System and conventional film (AGFA FD8p and D7).

The majority of tests were made on a concrete block with encast reinforcing bars and a centrally placed pre-stressed cable duct. The duct was injected with cement grout and a number of square voids in the grout were made with dimensions 70mm down to 15 mm in intervals of 10 (and 5 mm), i.e. 7 voids. The concrete thickness was varied by placing various smaller blocks in front or behind this block to produce a maximum thickness of 1200 mm and a minimum thickness of 300 mm. In the examples shown here the duct with voids was placed closest to the film or image plate, i.e. the object-film distance (OFD) was on average 150 mm. The FFD used for all exposures was 1600 mm.

The AGFA Strukturix DPS system

The system consists of imaging plates, a scanner, a workstation for processing and evaluation and a printer for producing hard copies. The imaging plates contain phosphor crystals which when exposed to radiation are excited to a higher energy level depending on the accumulated radiation dose. The density of the object exposed thus determines the number and level of excitation at different locations on the imaging plate.

The exposed plates are then scanned with a laser and the equivalent trapped energy in the plates is converted to light. This light is then captured by a photo-multiplier and converted into digital data before being sent to the workstation for post processing. After the imaging plates have been scanned and erased they are re-useable.

The test objectives

• to determine the capability of detecting voids of various sizes in concrete with thickness from 300 mm to 1200 mm
• to determine the capability and accuracy of measuring reinforcing bar diameter and position in concrete up to 1200 mm thick
• to determine the capability of measuring reinforcing bar section reduction (corrosion)

Measurements were made through concrete with thickness 300, 600, 800, 1000 and 1200 mm. Comparisons were made of image quality, void detectibility and exposure times for the AGFA DPS IPC-cassettes, D7-film (maximum 300 mm thick concrete) and FD8p-film (maximum 1000 mm thick concrete).

Results

In Fig.5 can be seen the DPS image for 1200 mm thick concrete. The three dark spots in the centre are 20 mm diameter holes drilled in the backside of the block to depths 15, 20 and 40 mm. Several reinforcing bars cause the light area, while the other dark areas are caused by voids. The exposure time required for 1200 mm thick concrete was 4500 s, compared to 14400 s for FD8p-film.

Fig 6: Images of reinforced concrete of different thickness. The duct and cables can be seen in the middle of the images, together with three rectangular voids of 15, 20 and 30 mm dimension (left to right).

The images have been subjected to image processing with emphasis on contrast enhancement. Films have been exposed to similar radiation doses in order to compare the contrast levels (solid concrete:void) as a function of concrete thickness.

It has been found that the contrast is greater for imaging plates compared to film (Agfa FD8p) for all of the images, but the difference between them increases with increasing concrete thickness. It was not possible to detect the same minimum size of cavity with film compared to imaging plates. The 15 mm square void could be detected in 1200 mm thick concrete with the imaging plates.

In order to establish if there is a relationship between the thickness of the concrete and the size of the void, a void density factor was calculated for all the void sizes and for four different thicknesses of concrete. The factor is defined as:

Vdf = (Dv-Dc) x 100
              Dc

Where Dv is the SAL value of the middle of the void and Dc is the corresponding SAL value adjacent to the void. The latter is found as the average of the SAL value on both sides of the void in the same horizontal position, as the void SAL value is measured. (The SAL value is the grey-level value in the range 0 to 4095 (12 bit system)).

Fig 7: Void density factor plotted as a function of the size of the void for four different thicknesses of concrete.

Conclusion - High Energy radiography and The AGFA Strukturix DPS-system

The most significant advantage of this system is the fact that exposure times are reduced by on average 60 % compared to conventional rapid film. Exposure times are very restricting otherwise when inspecting concrete. Also, the sensitivity of the method in detecting voids has been demonstrated and can be said to be satisfactory for most purposes.

High energy radiography remains today the only non-destructive technique, which can reliably determine the condition of the pre-stressed cable ducts and cables. The sensitivity of the method in determining loss of section to the cables in congested ducts has not been established.

© NDT.net - info@ndt.net

|Top|