|NDT.net - April 2003, Vol. 8 No.4|
A new and rapid ultrasonic pulse echo technique has been used with great effect to scan the concrete lining of two railway tunnels. The instrument used was the A 1220 from MSIA Spectrum in Russia. Working side by side two engineers from FORCE Technology were able to make a full and detailed survey of the concrete thickness and homogeneity along 1300 m of tunnel, covering an area of over 10,000 square metres in less than two months. This is an example of the advances that have recently been made in non-destructive testing of concrete and demonstrates the capability of modern technology in providing the much needed post construction checks of civil engineering structures.
In the Autumn of 2001 the go-ahead was given for completion of the railway tunnel through a major ridge in the south of Sweden. The work began several years ago and has been interrupted by several set-backs. In the northern entrance to the twin tunnel some 1200 m have been completed, of which 650 m have been constructed with an inner 600 mm thick concrete lining cast in-situ. This concrete lining has been cast against a 4 mm plastic water- proof membrane, sandwiched against an inner concrete lining which in turn has been cast against the rock face. The inner concrete lining investigated is of high quality, with a 28 day compressive strength of 65-70 M Pa and maximum aggregate size 25 mm. The tunnel is designed for a ground water pressure of 15 bar from the high ridge and great care has been taken to prevent leakage from occuring in the 120 year design life-time of the tunnel. In 2000 a small leak did occur at one position due to a hidden cavity in the concrete adjacent to the water-proof membrane. This initiated a full survey of the tunnel ceiling covering an 8 m span and along the full length of the completed tunnels.
A number of alternative NDT methods were investigated at the outset, including ground penetrating radar, impact echo and ultrasonic pulse echo. The considerable size of the structures prompted consideration of the best approach to the problem. Initially it was thought best to adopt a method such as radar that would enable a rapid scan of the concrete surfaces to locate any deviations from the normal response that might indicate problem areas. This could then be followed up by more detailed examination of the suspected areas, using a technique such as ultrasonic pulse echo.
|Fig 1: Sections showing the part of the tunnel inspected|
The minimum acceptable defect size was defined as 200 mm positioned adjacent to the water-proof membrane. In addition it was stated that sharp edges, or changes in thickness at the membrane should be identified, as this type of defect could cause membrane tear. As in most civil engineering problems it is extremely difficult to predict the performance of NDT and therefore each method was tried on site before any conclusions were made about their suitability in this case.
This technique is normally used to locate reinforcing bars due to the high resolution of the method and the fact that radar waves are sensitive to steel reflectors in concrete. Due to the great difference in electro-magnetic impedance of concrete and steel the radar waves are almost completely reflected from reinforcing bars. Air voids do not cause such clear reflections, as only 50% of the radar wave energy is reflected. Also, conductive materials like concrete tend to cause severe attenuation of the waves and restrict penetration. However, in favourable circumstances this technique will indicate air vois in concrete and it was thought that the thickness of the inner liner could be determined. The greatest advantage of the technique is speed, as scanning may be done continuously and in principle without direct contact with the concrete surface. Radar was considered to be a viable alternative also due to the fact that the concrete lining is un-reinforced, with the exception of two short sections. Three different radar antennae were used in initial tests – a 1 GHz, 800 MHz and 500 MHz from Malו Geoscience, Sweden. A number of scans were made across the apex and along the length of a tunnel section. After considerable data analysis the conclusion was made that radar could have potential as an alternative technique for measurement of the concrete thickness. The method was considered to be a little uncertain, in terms of data interpretation, and we were unsure of its capability in detecting internal voids in the concrete. The reflections from the membrane could be identified best if there was some unevenness in thickness. However, it was feared that the bottom reflection could vanish in the course of processing if the concrete surface and membrane were perfectly parallel.
The best GPR profile recorded at Hallandsוs.
Fig 2: Examples of raw and processed radar data from the tunnel.
Initial tests had shown that a clear bottom echo from the interface between the concrete lining and membrane could be obtained at frequencies up to 125 kHz. It could also be used for more detailed examination to define internal voids or other reflectors if necessary. The A 1220 Ultrasonic Pulse Echo unit employs an array of dry-contact transducers which are spring-loaded or cushioned, which enables good contact with concrete surfaces without the need for surface preparation.The mechanical waves are shear horizontal and the frequency may be varied between 33 kHz and 250 kHz.
|Fig 3: The A 1220 Ultrasonic unit.|
The measurements were made by two operators, each with their own unit. This required access at arms length to the concrete surfaces. Normally, if good access is available then approximately 300 measurements can be made in 1 hour by one operator. Clearly the success of the method in this case was dependant on good access, mobility and positioning. It was estimated that if the practical difficulties could be solved then two operators could scan 2 – 3 sections per day, or around 350 square metres of concrete.
In the event, this technique was chosen as it offered the possibility of both reasonably rapid scanning of the entire tunnel and simultaneous reliable and detailed results for the defined problem case. It was estimated that each tunnel would require two operators over a period of 4 weeks in the field.
Each tunnel has been fitted with a rolling scaffold for construction of the concrete and membrane fitting. Part of this scaffold train was modified for testing work and a grid of steel wires were stretched between the two ends as a guide for positioning during test operations. Global positioning was possible as the tunnels had been cast in 16 m long sections with clearly visible casting joints. To assist the operators the ultrasonic units were mounted in a specially constructed ‘arm’ as shown in the photograph below.
Measurements were made at a spacing of 400 mm in x- and y-directions over the length of each section. The frequency of testing was increased in a 2 m span across each casting joint, as it was feared that the likelihood of defects would be greatest in this region. In this region a test spacing of 200 mm was used.
Scans were made along the length of each section and along 20 parallel lines. An example of such a B-scan is shown in Fig. 4 below.
|Fig 4: Typical B-scan showing the bottom echo from the membrane along 16 m of tunnel.||Fig 5: Typical C-scan showing concrete thickness over an area of 128 square metres|
For each section the data has been digitized and presented as a map of concrete thickness in the form of a C-scan presentation, as shown in Fig. 5 below.
Unlike many other materials traditionally tested by NDT, the velocity of ultrasonic wave propagation in concrete cannot be predicted accurately. Great differences in wave velocity can be expected for different structures, and some variations can also be expected in one and the same structure, particularily one of this size. The pulse-echo methods used measure the time of propagation of waves. To determine the thickness of the concrete the characteristic wave velocities must be determined and adjusted where necessary, for example due to local variations in strength.
It should be pointed out that no coring or physical intrusion into the concrete to determine exact thickness for calibration was possible.
Three methods have been used to determine wave velocities – a novel technique using two A 1220 antennae which provides both surface and average (with depth) shear wave velocities; Rayleigh wave velocity measurements by the Spectral Analysis of Surface Waves technique; and P-wave velocities by Impact Echo and traditional in-direct measurement using the PUNDIT. The methods of greatest interest and those adopted in this exercise are the former, i.e. shear and Rayleigh wave velocity measurement.
Shear wave velocity
Using two antennae placed at increasing separation on the concrete surface the direct or surface wave is registered as well as the waves reflected from the membrane. The success of this technique was somewhat surprising and it provided valuable information. An example of such a measurement is shown below.
|Fig 6: Shear wave velocities determined using the A 1220 in Common Mid-Point mode. The upper and lower curves are a result of wave transmission near the concrete surface and from membrane reflections respectively.|
Each alternate section was tested using the SASW-technique, with equipment from Olsen Engineering, USA. Considering that the casting work was conducted over a period of 1 year it was reasonable to expect that there may be both local and global variations in wave velocities. Some 450 Rayleigh wave measurements were conducted in all.
The SASW-technique is one which is used quite often for geophysical surveys. The velocity of the Rayleigh wave is related to the shear modulus (stiffness) and density of the material. The reults shown below are in the form of a dispersion curve, so-called as the principle will demonstrate variations in R-wave velocity with wavelength. Since waves of greater length are influenced by material at greater depth than short wavelengths then it is possible to create a display of the variation in stiffness of the material with depth.
In this investigation the average R-wave velocity was found to be 2530 m/s with little variation. This agreed well with the measured shear wave velocity of 2820 m/s.
|Fig 7: Typical SASW dispersion curves showing the variation of Rayleigh wave velocity with depth from the concrete surface.|
It was found that this technique also provided other useful information, as demonstrated above in Fig. 7. At wavelengths corresponding to the actual thickness of the concrete the dispersion curves can be seen to deviate sharply. The thicknesses indicated using this method cannot be expected to be as exact as those indicated by ultrasonic pulse echo. This is in part due to the poorer resolution of the technique but also due to the ‘global’ nature of the method – much larger volumes of concrete are encompassed by a single test than UPE. The correlation between SASW and UPE was however found generally to be surprisingly good.
|© NDT.net - firstname.lastname@example.org|||Top||