Bundesanstalt für Materialforschung und -prüfung

International Symposium (NDT-CE 2003)

Non-Destructive Testing in Civil Engineering 2003
Start > Contributions >Lectures > Tendon Ducts: Print

Fast Location of Prestressing Steel Fractures in Bridge Decks and Parking Lots

Horst Scheel, Technische Universität Berlin, Institut für Bauingenieurwesen, Fachgebiet Baustoffe und Baustoffprüfung, Germany
Bernd Hillemeier, Technische Universität Berlin, Institut für Bauingenieurwesen, Fachgebiet Baustoffe und Baustoffprüfung, Germany


The measuring speed of the Remanent Magnetism Method (RM-Method) can be enhanced significantly by replacing the time-consuming multi-step-magnetization by a single-step-magnetization.

The RM-Method [SCHE1], [SCHE2], [HIL2] allows identification of potentially unsafe conditions in pre-tensioned and post-tensioned concrete structures by locating fractures of single wires, even if they are bundled with intact wires. Once they have been pre-magnetized with an electromagnet, the magnetic field of tendons is measured at the concrete surface. Fractures produce characteristic magnetic leakage fields, which can be measured with appropriate sensors at the concrete surface.

Large yoke-shaped magnets have been constructed to magnetize transverse tendons in bridge decks of lengths up to 4m in a single process. Measuring the magnetic flux density of an entire bridge deck simplifies the comparison of data from measurements at different times, which would be helpful for the monitoring of the structure.

magnetic method, non-destructive testing, non-destructive evaluation, leakage field, prestressed concrete, prestressing steel, steel fracture, bridges, parking decks, corrosion


There are various reasons for fractures in prestressing steel tendons, some of which have led to structural collapses [HIL1]. In bridge decks and parking structures de-icing salts are the reason for the fractures in most cases. All types of prestressing steel are at risk of breaking if they come in contact with corrosive media. In bonded prestressing structures this happens if the top surface of bridge or parking decks are not adequately waterproofed and chlorides are allowed to reach and depassivate the prestressing steel wires. In unbonded prestressing or poorly grouted bonded prestressing, only water is needed to cause corrosion [NüR].

Some types of heat-treated prestressing steel are so sensitive to stress-induced corrosion that fractures have been found even if the wires were embedded in alkaline mortar in high-quality grouted, intact ducts.

Failure of prestressed concrete members often happens abruptly.


The magnetic field resulting from a magnetized tendon or a magnetized steel wire is comparable to the magnetic field of a bar magnet. In the vicinity of a fracture a magnetic dipole-distribution is formed and, accordingly, a magnetic leakage field in the surrounding region. The transverse component of the magnetic flux density, measured at the concrete surface, is shown in fig.1. The characteristic leakage field allows us to detect the fracture of a single prestressing steel wire (fig.1).

Fig 1: A magnetized prestressing steel wire has a magnetic field comparable with the field of a bar magnet. At a fracture additional poles are forming that gen-erate a characteristic leakage field (left). The transverse component of the leakage field is shown on the right. It allows to draw conclusions regarding fractures of single prestressing steel wires inside the duct.



Location of prestressing steel fractures in bridge decks and parking structures involves the following four processes:

  • Magnetization of the prestressing steel
  • Scanning of the magnetic flux density
  • Data acquisition and processing
  • Data interpretation

Each of these steps has to be optimized in order to speed-up the testing of these structures.


Necessity of Magnetization
In order to draw unequivocal conclusions from the magnetic flux density measured at the concrete surface to potential fractures, the external magnetization of the prestressing steel has to generate a magnetic state where all irreversible magnetization processes in the prestressing steel are completed. This is necessary in order to erase the unknown magnetic history of the steel. The magnetization of the tendons is performed from the concrete surface using an electromagnet. A remanent magnetization of the tendons is achieved up to a concrete cover of 30cm [SCHE2], when yoke-shaped electromagnets with a length of approximately 40cm and a mass between 35kg and 65kg are used. The disadvantage of using these comparably small magnets is the time consuming magnetization procedure, because the magnets have to be lead several times along the tendons until the sufficient magnetization is achieved [SCHE2].

The position of the tendon has to be known along the entire magnetized length within a tolerance of about 10cm. Consequently it is often necessary to locate the tendons using ground penetration radar before the magnetization may begin.

The New Single-Step-Magnetization
In cases where measurements can be performed from the top surface of the concrete component the use of larger magnets is feasible. The first idea was to construct a separable electromagnetic yoke that could reach from one end of a transverse tendon in a bridge deck to the opposite end. For German standard conditions this translates into a length of the yoke of up to 18 m. Numerical simulations result in the possibility to saturate tendons of this length with an electromagnetic yoke of the same length magnetically. Laboratory tests showed that a magnetization of standard mild steel was no problem on a length of 10 m, however it was not possible to succeed in this with prestressing steel. The cause is the higher magnetic coercivity of prestressing steel compared to mild steel. Prestressing steel is not only mechanically but also magnetically harder than mild steel.

As a consequence of this result and for a better handling we constructed an electromagnet with a length of "only" 4m (fig. 2). The legs of the electromagnetic yoke have a height of about 1 m and consist of iron cores and coils with an inner diameter of approximately 35 cm. The magnetic yoke with a mass of approximately 1,5 t is operated with a controllable DC power supply with a maximum power output of approximately 16 kW. With this magnet, whose length corresponds to the width of one traffic lane (3,75 m), it is possible to magnetize a 3,50m long section of a tendon in one step.

Fig 2: The 4 m long electromagnetic yoke.

Magnetization of the tendons can be performed by driving the magnet, orientated transversely to the bridge axis, along the bridge deck over the selected section, which will typically be the edge section, where we have statistically the highest corrosion rates.

Position of the tendon is of little significance for the magnetization process. However both ends of the magnetized length of each tendon must be sufficiently close to the legs of the electromagnetic yoke in order for the magnetic flux to form a loop through the magnetized length of the tendon. With the diameter of the iron cores of the coils of the electromagnetic of 35 cm, a deviation of 35 cm on a length of 4m from a given orientation is allowed. Even in bridge decks a larger deviation from the planned orientation is highly unlikely. Therefore it is not necessary to locate the tendons with ground penetration radar before the magnetization process is performed.


To date, the magnetic flux density is measured along a single tendon, but it is planned to replace the current measuring vehicle with 10 Hall probes by an array of 256 or 512 magnetic probes arranged equidistantly in one line. This array will be aligned along the direction of the transverse tendons at a distance of 2 m behind the electromagnet. The vertical component of the magnetic flux density will be measured at the concrete surface at a resolution of 1cm along the bridge axis.

The type of sensors that will be used for the array has not been selected yet. The sensors have to cover a range from 1mT to 300 mT to measure the residual field of the magnetized tendons. For an additional measurement of magnetic flux density along the magnetized tendon section in the active field, the upper limit of the measuring range has to be extended.

The component of the magnetic flux density orthogonal to the surface of the selected bridge deck or parking structure will be scanned in a raster of approximately 1cm x 1cm during the same process as the magnetization. Presently, the measuring speed is aimed at 1 m/s. An entire bridge deck with a width of 18 m and a length of 500 m could therefore be scanned in 6 x 500 s. One section is scanned in 500 s and the bridge deck has to be split up into 6 sections with a width of 3 m each. Under ideal conditions the measuring work to locate prestressing steel fractures in this bridge would be completed within an hour.


A 12-bit resolution is sufficient for our purpose. The maximum data acquisition rate (sensor-array with 512 sensors) will be 7,5 kB/s. Data acquisition does not limit the measuring speed. More sensors or a higher resolution along the measuring path are possible if desired. The amount of data for a 500 m long and 18 m wide bridge deck is approximately 220 MB. By today's standards, this is comparably small.

Beside different filter techniques data processing consists of a routine that locates the magnetic poles of the tendon sections and relates them to find the position of these sections. The two-dimensional array of the vertical component of the magnetic flux density is then projected onto each single tendon. The results are one-dimensional arrays of the vertical component of the magnetic flux density along the single tendon sections.


Data interpretation does generally not vary from the interpretation of data gathered with the conventional RM-Method. The influence of interfering signals seems to decrease with the new method of magnetization, because the strong magnetic poles of the magnetic yoke do not interact with the mild reinforcement along the entire magnetized tendon section but only in areas close to the poles of the 4 m long electromagnetic yoke. Whether this will be an advantage or additional interfering signals will occur, must be determined from future field testing.


After the practicability of the magnetization with the large electromagnetic yoke has been demonstrated in the lab and after it has been shown that the data acquisition can be performed using commercial components, we are now starting to build a vehicle to drive the measuring device, consisting of the electromagnetic yoke with power supply, the sensor-array, the control units and a computer. Special demands on this vehicle are:

  • a sufficiently slow and constant speed,
  • the capability to carry a load of approximately 2 t,
  • a sufficiently fast-controlled hydraulic system to keep the device close to the rough surface of the component to be tested while preventing contact between device and surface,
  • no ferromagnetic parts are allowed to prevent magnetic short circuit.

Completion of a functional vehicle is scheduled for April 2004.


Upon completion of the vehicle carrying the measurement devices as well as the works on the sensor-array in 2004, field tests will start in order to optimize the measuring speed of the system and to develop methods for the physical and numerical suppression of interfering signals.


Especially in prestressed bridge decks and parking structures, fractures of prestressing steel wires are common. In most cases, the reason for the fracture is chloride-induced corrosion. An actually required application of the RM-Method to detect potential prestressing steel fracture can often not be performed due to the shut-down times involved. Therefore it is planned to increase the measurement speed by magnetizing the tendons in one step using a larger mobile magnet and by scanning the magnetic flux density on the entire surface of the bridge deck.

The feasibility of rapid magnetization of tendons using an appropriate electromagnetic yoke was shown in laboratory tests. Furthermore, data acquisition and interpretation are of no general difficulty. At present, a vehicle to drive the measurement devices is being developed. This work is to be completed by spring 2004. With this a fast procedure for locating prestressing steel fractures in transverse tendons of bridge decks or tendons other plane structures, such as parking decks, would become available. The measuring speed would be increased by several magnitudes. As a result, structural testing of the prestressing steel, required to maintain structural safety, could be performed without severe disruption of traffic.


The reported research project is funded by the building authorities of Berlin. The support of Mr. Kollotschek and Mr. Weyer is greatly acknowledged by the authors.


  1. [HIL1]-Hillemeier, B., (1993), Das Erkennen von Spanndrahtbrüchen an einbetonierten Spannstählen, Vorträge Betontag 1993, Deutscher Beton-Verein E. V, Wiesbaden
  2. [HIL2]-Hillemeier, B., (1995), Assessment of Structural Stability of Prestressed Concrete by Non-Destructive Detection of Steel Fractures in the Proceedings Vol. 1 of the International Symposium Non-Destructive Testing in Civil Engineering (NDT-CE). Ed. by G. Schickert, H. Wiggenhauser, Deutsche Gesellschaft für zerstörungsfreie Prüfung e. V., p. 23 - 29.
  3. [NüR]-Nürnberger, U. (1989), Mehr Sicherheit im Spannbetonbau; Chloridgehalte und Feuchte beachten; Korrosion behindern. Maschinenmarkt 95, 30-40, 68-72 and 128-134
  4. [SCHE1]-Scheel, H.; Hillemeier B. (1997), Capacity of the remanent magnetism method to detect fractures of steel in tendons embedded in prestressed concrete, NDT & E International, Vol. 30, No. 4, pp. 211-216, 1997, Elsevier Science Ltd.
  5. [SCHE2]-Scheel, H., (1997), Spannstahlbruchortung an Spannbetonbauteilen mit nachträglichem Verbund unter Ausnutzung des Remanenzmagnetismus, Dissertation, Technische Universität Berlin.
STARTPublisher: DGfZPPrograming: NDT.net