·Table of Contents
SQUID System for Magnetic Inspection of Prestressed Tendons on Concrete Bridges
H.-J. Krause, W. Wolf, W. Glaas, E. Zimmermann, M.I. Faley
Forschungszentrum Jülich, D-52425 Jülich, Germany
Forschungs- und Materialprüfanstalt Baden-Württemberg, D-70569 Stuttgart, Germany
G. Neudert , U. Gampe
Siempelkamp Prüf- und Gutachtergesellschaft mbH, D-01099 Dresden, Germany
Bundesanstalt für StraBenwesen, D-51427 Bergisch-Gladbach, Germany
Superconducting Quantum Interference Devices (SQUID) are magnetic field sensors with the highest sensitivity and largest dynamic range known to date. In a joint German R+D project, we are developing a multi-channel SQUID system for nondestructive detection of tendon ruptures in prestressed members of bridges. The system uses the magnetic stray field measuring method. The tendons are magnetized by means of a yoke magnet scanning along the member. The magnetic stray field during magnetization or the remanent field subsequent to magnetization is measured using four SQUID sensors mounted in the yoke. Signals from stirrups of the mild steel reinforcement are suppressed with two types of techniques: either a best fit of typical stirrup signals to the stray field signal and their subtraction, or the comparison of remanent field signals after changing the magnetization direction of the stirrups. Subsequent correlation analysis with the dipolar signal of a typical void yields a rupture probability distribution and rupture signal amplitudes. Results of measurements on a prestressed concrete highway bridge are presented. At three locations, signal amplitudes above the threshold values were found. Two of them were verified by opening the bridge deck. In one case, one out of eight wires in the strand was broken, in the other case two. Thus, it was shown that the unsurpassed sensitivity and dynamic range of SQUID sensors can be applied to localize ruptures in prestressed members of bridges.
The Federal road network of Germany contains about 13,000 prestressed concrete bridges. Due to constantly growing volumes of traffic and higher loads per axle, these structures are subject to increasing loads. Regular inspections of all engineering structures consist mainly of a visual inspection, thus deterioration and damages are typically identified rather late. Non-destructive test methods (NDT) may provide a relatively quick and inexpensive means to establish whether a bridge is still in a serviceable condition. Particularly important is the non-destructive identification of the position of defects in prestressed concrete bridges, as the stability of the entire structure may be affected. The magnetic stray field measuring method (Section 2) is well suited for this inspection because it allows to detect cracks in reinforcement bars [1,2,3,4,5,6,7]. With the technique and subsequent signal analysis, it has been shown that a single cracked rebar can be found in post-tensioned members, even though the magnetic signature of the crack is damped significantly due to the shielding effect of the surrounding flawless rebars and the jacket tube around the bars. Typically, the flaw signal is hidden among signals of mild steel reinforcements (stirrups) located close to the concrete surface. Provided that the magnetic field sensors have sufficient sensitivity and linearity, flaw signals may be separated from structural disturbance signals by means of a specially developed signal analysis (Section 3). Superconducting Quantum Interference Devices (SQUIDs, Section 4) are the magnetic field sensors with the best field sensitivity and the largest dynamic range known to date. In Section 5, we present a newly developed system incorporating four SQUID sensors for the inspection of rebars in bridges. The new measuring equipment has already been successfully used in field measurements, as shown in Section 6.
2. Magnetic stray field technique
The principle of the magnetic stray field measurement technique is as follows: the tendon hidden in the concrete is magnetized using an exciting magnetic field Ho applied from the outside of the concrete by means of a yoke magnet. This exciting field generates a magnetization in the reinforcement bars. Local disturbances of the distribution of this magnetization due to ruptures or reductions of the cross section cause the emanation of a magnetic leakage flux (stray field) from the member. For the generation and the measurement of the stray field, a probe containing the magnetization device (yoke magnet) and the sensors is moved along the direction of the prestressed tendon outside the concrete surface (see Figure 1).
Fig 1: The principle of magnetic leakage flux measurement.|
The stray field measurement is either conducted during magnetization by the exciting field (active field) or as a residual field measurement. In the latter case, the stray field is caused by the remanent magnetization of the steel after switching off the magnetization device. In the active field measurement, ruptures of the longitudinal rebars appear as a local maximum.
3. Signal analysis
The stray field signals are affected not only by ruptures of the tendon under consideration but also by the stirrups close to the probe. Signal analysis methods for the suppression of signals from the stirrups were developed. The magnetic signature of a stirrup shows a considerably different shape in the active field, in comparison to the residual measurement. Due to the hysteresis behavior of the ferromagnetic material, the active field signal of a stirrup has an asymmetrical shape. The polarity of the (antisymmetrical) residual field measurement signal of the stirrup depends on the location where the yoke magnet has been switched off during the preceding magnetization of the member. A typical stray field and two residual field signals of a single stirrup are shown in Figure 2. The residual field measurement R1 is performed after the yoke magnet is switched off at the end of the measurement length. In contrast, the measurement R2 denotes the case where the exciting field is switched off at the starting position after a full magnetization scan. The residual field signature R2 of the stirrup is inverted compared to R1.
Fig 2: Stray field and remanent field signals of a stirrup . The remanent field R1 was measured after switching off the yoke magnet at x=500 cm, the residual field R2 was recorded after turning off the magnet at x=0.|
This feature directly leads to the first method of stirrup signal suppression: Addition of the two residual field measurements R1 and R2 suppresses the residual field signals of the stirrups, whereas rupture signals are unaffected. Figure 3 (left) shows an example for this approach. Nevertheless, the polarity of the residual field signature of a rupture is not well defined. It depends on the magnitude of the preceding magnetization, the cross section of the flaw, and the distance probe-tendon. Due to this non-uniformity, the exclusive use of the residual field measurement for the evaluation of the integrity of tendons seems not to be sufficient . The position of the stirrups can be obtained from the local maxima (R2) or minima (R1) of the derivative of the axial residual field component.
Fig 3: Examples of the elimination of stirrup signals. Left: adding two remanent field measurements with inverse magnetization polarity of the stirrups, right: Best-fit correction of stray field measurement.
In order to separate the signals from the reinforcement close to the surface, several active field measurements are conducted at successively increasing magnitude of the exciting field. Because of the strong decrease of the exciting field with distance, the measurement at low excitation fields gives dominating signals from the mild steel reinforcement close to the surface. Due to magnetic saturation, increasing excitation field only slightly alters the magnetization of the steel close to the surface. The signal increase at higher values of the exciting field Ho is mainly caused by the steel structures (rebars) deeper in the concrete.
The second technique for the elimination of the signals from the stirrups is direct subtraction of idealized signals. First, the exact position of the stirrups is calculated from a residual field signal. Then, the signal portion of the stirrups is determined by means of the best fit method with regard to the signal of a single stirrup (Figure 2). Finally, the signal portion of the stirrups is subtracted from the measured signal. The essential portion of the signals of the tendon remains (Figure 3, right).
4. SQUID magnetometer sensor
SQUIDs are extremely sensitive magnetic field sensors. The are fabricated using High Temperature Superconducting ceramic thin films and have to be cooled down to liquid nitrogen temperature (-193°C) in order to operate. They can be regarded as magnetic flux to voltage converters with a periodic transfer function, with a periodicity equal to the magnetic flux quantum F0=2×10-15Tm2. For our application, SQUIDs which can be opeated in strong magnetic fields were developed , see Figure 4. When exposed to our maximum excitation fields (up to 15 mT), no degradation of performance was observed, the noise increase is due to 1/f noise of the current source driving the excitation magnet. For cooling the SQUID sensor array, a mobile liquid nitrogen cryostat was developed. Since the SQUID is mounted on a sapphire finger in the vacuum space of the cryostat, it can be operated orientation-independent. One refill of liquid nitrogen lasts for a working day.
Fig 4: Packaging, Layout and magnetic field resolution of the SQUID sensors used for rebar inspection.
5. Setup of the system
The system consists of 4 SQUID magnetometers for high fields, in the cryostat, mounted in a yoke magnet on the translation stage. In order to check consistency with the previous measurement technique, 4 Hall probes are also integrated with the system. A multi-channel readout and control electronics using digital signal processors for SQUID control and signal preprocessing  is used to acquire the data from the magnetic sensors and the position encoder. Due to the implemented flux quanta counting, the digital SQUID electronics achieves a dynamic range of 195 dB/ÖHz with the high-field magnetometers. It is connected to a computer via an optical link. Figure 5 shows schematically the setup of the system.
Fig 5: Schematic setup sketch of the SQUID system for magnetic inspection of rebars.
6. Measurements on a German highway bridge
As an example for a field test of the system, we present results of SQUID measurement on the valley bridge Michelsrombach (A7 near Fulda) . At selected measurement positions, tendons were localized using the Ground Penetrating Radar (GPR) SPR-Scan from ERA Technology with a 1 GHz antenna. Subsequently, magnetic measurement scans were performed along the marked tendons, see Figure 6.
Fig 6: Schematic of the measurements conducted on the valley bridge Michelsrombach. First, the tendons were localized using Ground Penetrating Radar . Subsequently, the prestressed steel rebars were inspected by scanning the yoke with the SQUID system along the member .|
After Radar localization, magnetic measurement scans were performed along the tendons. The investigation was carried out in scan segments of 260 cm length. At every location, nine scans were conducted. After a recording of the pre-existing field, scans with increasing exciting field were performed, followed by scans with decreasing field and subsequent remanent field scans after different premagnetization states. Before each measurement scan, the SQUID sensors were heated just above the critical temperature of the superconducting film in order to eliminate trapped magnetic flux. The scan velocity was 0.1 m/s. Due to the SQUID heating, the total measurement time at one location was 20 min.
Figure 7 shows the correlation coefficient and the crack signal amplitude of the evaluated stray field and remanent field scans at a selected location of the bridge.
Fig 7: Correlation coefficients and Rupture signal amplitudes.|
Indication A at x = 45 cm yielded a correlation coefficient and a rupture signal amplitude below threshold. Opening of the concrete and the jacket tube protecting the rebars showed that the indication emanated from a loose end of a rebar, see Figure 8 A. Indication B at x = 82 ± 5 cm, however, gave an correlation coefficient and a rupture signal amplitude well above our threshold. The opening of the bridge deck (Figure 8 B) confirmed that two of the eight rebar wires were cracked at this location.
Fig 8: By opening the bridge deck, the indications A and B were confirmed as a loose end and a rupture.|
6. Conclusions and outlook
Ruptures of steel in prestressed tendons can be detected by measurement of the magnetic stray field. A system utilizing SQUID sensors with unsurpassed sensitivity was developed. Stray field measurements with different magnitudes of the exciting field were taken as well as remanent field scans after different premagnetization states. Signals of the mild steel reinforcements (stirrups) were suppressed efficiently using a sophisticated signal analysis including best fit estimation of the stirrup signals, comparing residual field measurements after stirrup magnetization inversion, and crack signal correlation analysis. The applicability under field conditions was demonstrated during bridge measurements. Magnetic indications of rebar cracks were verified by opening the bridge deck.
Due to the unrivalled sensitivity of the SQUID sensors, the system is particularly well suited for periodic observation of tendons in order to record small changes. Periodic monitoring of prestressed members over a long period of time becomes feasible.
Support by German BMBF under contracts No. 13N7249/2, 13N7250, 13N7251/3 and by German BMV under contract No. 98 245/B4 is gratefully acknowledged.
- F. Kusenberger, N. Barton, Detection of flaws in reinforcement steels in prestressed concrete bridges, Final Report FH-WA/RD-81/087, Federal Highway Administration, Washington, DC (1981).
- B. Hillemeyer, C. Flohrer, Zerstörungsfreie Ortung von Spannstahlbrüchen in Spannbeton-Deckenträgern, DGZfP, Berlin (1991).
- G. Sawade et al., Signal analysis methods for remote magnetic examination of prestressed elements, Proc. Intl. Sympos. on NDT in Civil Engineering (NDT-CE), Vol.II, DGZfP, Berlin, pp.1077-1084 (1995).
- H. Scheel, Spannstahlbruchortung an Spannbetonbauteilen mit nachträglichem Verbund unter Ausnutzung des Remanenzmagnetismus, Ph.D. Thesis, D 83, Technical University Berlin, Berlin (1997).
- G. Sawade, U. Gampe, H.-J. Krause, Non Destructive Examination of Prestressed Tendons by the Magnetic Stray Field Method, in: Proc. 4th Conf. on Engineering Structural Integrity Assessment, Cambridge, U.K., pp. 353-363 (1998).
- A. Ghorbanpoor, Magnetic-based NDE of steel in prestressed and post-tensioned concrete bridges, Proc. Structural Materials Technology III, San Antonio, Texas, pp. 343-349 (1998).
- J. Krieger, H.-J. Krause, U. Gampe, G. Sawade, Magnetic field measurements on bridges and develop-ment of a mobile SQUID system, Proc. Intl. Conf. on NDE Techniques for Aging Infrastructure & Manufacturing, Newport Beach, California, USA, pp. 229-239 (1999).
- M.I. Faley et al., Operation of the HTS dc-SQUID Sensors in High Magnetic Fields, IEEE Trans. Appl. Supercond. 9, 3386-3391 (1999).
- E. Zimmermann et al., HTS-SQUID Magnetometer with Digital Feedback Control for NDE Applications, in: Rev. Prog. QNDE, Vol. 16B, Eds.: D.O. Thompson, D.E. Chimenti, Plenum, NY, pp. 2129-35 (1997).
- J. Krieger, E. Rath, H.-J. Krause, Anwendung zerstörungsfreier Prüfverfahren bei Betonbrücken: Messungen an der Talbrücke Michelsrombach, Berichte der BASt, Bergisch Gladbach, in press (2000).