NO339677B1

NO339677B1 - Method for detection of degradiation in metal structures

Info

Publication number: NO339677B1
Application number: NO20150286A
Authority: NO
Inventors: Harald Horn
Original assignee: Ferrx As
Priority date: 2015-03-03
Filing date: 2015-03-03
Publication date: 2017-01-23
Also published as: NO20150286A1; RU2017130168A; WO2016140579A1

Description

The present invention is related to a method for detection of degradation in metal structures, in particular detection of stress, fatigue, metal loss, cracks and similar in ferromagnetic steel.

Background

One method for measuring the condition of steel structures involves analysis the voltage drop between contact points on the steel structure. An excitation current is applied to first contact points in the steel structure and the voltage drop response is measured between said second contact points in the steel structure. Initially, the voltage response exhibits a high voltage where electrons are moving along the surface of the metal structure followed by a transient phase with a sharp decrease. This is also called the skin effect. In the transient phase, the electrons start to travel down into the internal of the metal structure, whereupon the voltage attains a stationary phase with no voltage change. A comparison of this voltage response with calibration data from an ideal undamaged metal structure may reveal for example cracks or corrosion pits. Deviations in the transient part of the voltage course may indicate defects in the surface, such as micro cracks. Deviations further into the transient phase and into the stationary phase may indicate defects in the internal of the metal or at the inside of for example a pipe wall (provided that the current is applied at the outside of the pipe).

A method of this type is described in NO Patent 150 136. In this publication, the transient part of the voltage course is considered to be noise, and only the stationary phase is analyzed.

A similar method for measuring the condition in steel structures is described in NO Patent 323 931. Here, both the transient and the stationary part of the voltage response are measured and compared to calibration values. However, this patent analyzes both the transient and the stationary part of the voltage drop. The latter method provides a more accurate determination of the condition of a metal structure and enables an earlier prediction of failure.

WO 92/18839 Al describes a method to measure mechanical stress and fatigue in steel. This publication mainly describes a method based on measuring the transient voltage response from an excitation current. Two such transient voltage drops are measured between a number of contact points for different conditions of the structure. The measured transient voltages are compared and make the basis for computation of the mechanical stress and degree of fatigue. However, this publication does not describe how the method responds to fatigue. Furthermore, these measurements do not consider the information inherent in the stationary part of the voltage response, which includes significant information about the fatigue (page 4 line 35, page 5 lines 8-10 and 6-30, page 8 lines 19-20, claim 1 and 3).

However, there is still a need for a more accurate detection of cracks and metal loss to provide a more accurate and earlier prediction of type and size of failure.

Object

An object of the present invention is to provide a method for detection of cracks and/or metal loss in metal structures that enables determination of location, geometry, and size of cracks and metal loss, and into which it is discriminated between cracks and metal loss which is not possible with the method described in NO Patent 150136, and has not been utilized by the method described in NO

Patent 323 931. Another object is to enable characterization of the type of change in metal properties, such as even corrosion or localized corrosion or cracks, or micro cracks

The invention

These objects are achieved by a method in accordance with the characterizing part of claim 1. Additional advantageous features appear from the dependent claims.

The present invention discloses a method for detection of degradation in metal structures, in particular detection of stress, fatigue, metal loss, cracks and similar in ferromagnetic steel. Nonlimiting examples of structures to monitor include railway rails, drill pipes, subsea risers, bridges, pipes, and similar structures subjectto stress and degradation.

At least one first set of electrically conductive contact points is arranged at a surface of the metal. Each set of contact points comprises a pair of contact points arranged at a mutual distance. At least one second set of electrically conductive contact points is arranged to the same surface of the metal, wherein each second set of contact points comprises a pair of mutually displaced contacts. The second set of contact points is arranged at a location between the respective contact points in the first set of contact points.

The steps below recite the method in accordance with the present invention:

a) Applying an excitation current to the first set of contact points.

b) Measuring the voltage response between the contact points in the second set of contact points.

c) Analyzing both the transient part and the stationary part of the voltage response.

d) Comparing the measured voltage response with calibration data or with measured value from known measurement conditions, and calculate deviation values to quantify deviation of the voltage

response from the calibration data as a function of time after current onset.

e) Repeating steps a) to d) above for a period of time to provide a set of separate voltage measurements for the selected location in the metal structure.

The improvement comprises the further steps of:

f) Providing normalized deviation values from the deviation values as a function of time after current onset. g) Choosing a value from the normalized deviation values at a point where differences between respective normalized deviation values gives the best sensitivity and resolution, which corresponds to the largest difference in time after current onset between the normalized deviation values for the actual application. h) At the selected normalized deviation value, selecting one or more values for time after current onset from the respective measurements.

i) Comparing said at least one selected value for time after current onset in step h) with another value for time after current onset from the same location in the metal related to different measurements, or with a measurement tåken at a location håving damage with known dimensions, thus determining the geometry of the damage.

The absolute values of the voltage, occurrence of voltage changes at certain points in time including comparison data from numerous measurements over a period of time, enable directly determination of actual depth of localized corrosion or cracks independent of their length and width which up to now has been unattainable. As a result, an early prediction and characterization of failure may be obtained, preventing accidents and planning repair or replacement of metal components subjected to such damage.

Pipes, for example, are often provided with a coating on the outside, and corrosion often occurs at the inside of the pipe. It is therefore of interest to detect material changes, such as metal loss and cracks, which is not available to visual inspection. By inspecting the initial part of the deviation between measured voltage and a reference voltage obtained at known conditions, and the whole transient response and stable response voltage, it is possible to determine the location of metal change whether it is on outer surface, inside the metal, or on inside surface of a pipe. A sharp rise followed by a decrease of voltage deviation in the very beginning of the transient phase, indicates metal change at the same side as the contact points on the pipe being monitored. To the contrary, a zero voltage deviation in the early part of the signal followed by a slight rise in voltage deviation indicates metal change at the opposite side of the structure being monitored.

In some embodiments, it may be of importance to arrange a row of contact points in a specific direction, because damages are expected to develop in a specific direction along the structure. However, in yet other embodiments, directional development of damage may be unpredictable. A crack which is developing in a direction along the row of contact points may be difficult to detect. Subsea raisers for example, are subjected to both bending and tension. Here cracks may develop in several directions. In circumstances like this, it is desirable to measure in two or more directions by arranging two or more adjacent sets of measurement devices (each device comprising first and second contact points, DC source, and measurement circuit), arranged at a mutual angle between 0 and 90 degrees. In this way, the method of the present invention may become less sensitive to varying directional development of damages in the metal.

Moreover, ferromagnetic structures which have been subjected to stress will typically exhibit changed magnetic properties. In order to remove unwanted influence from any changes in magnetic properties upon the measurements tåken in the method of the present invention, a measurement is advantageously split up into numerous consecutive discrete measurements with a short time there between, typically four to five consecutive discrete measurements. The time between each consecutive measurement is typically a few milliseconds. This step will demagnetize the ferromagnetic metal and exclude any possible unwanted influence from changes in magnetic properties caused by previous stress or load. It should be noted that a comparison of a similar measurement tåken at a former point in time may determine maximum stress the structure was subjected to between the two points in time, e.g. from a measurement tåken at point in time before the structure was tåken into use, or from the former measurement. This is discussed in further detail below.

In accordance with the invention the rise time, level and duration of the excitation current is advantageously controlled in order to optimize the same for the application in question. Moreover, the results obtained from the method in accordance with the present invention, may advantageously be used to calibrate data models used for calculation of fatigue of the whole structure, e.g. subsea risers and wind mill towers, to improve the precision of such models.

An example of use of the method in accordance with the invention is monitoring of a subsea rise at an offshore platform. An initial measurement is tåken, e.g. at the platform deck, before the riser is tåken in use. At one or more points in time after the riser has been used, measurements can be tåken, e.g. at the platform deck, thus detecting changes in the steel as a consequence of stress applied during its use.

Definitions

The term "metal change" or "material change" as used herein, refers to a physical change in a metal structure which is affecting resistivity/resistance and/or permeability. Resistance may be affected by e.g. dislocations, micro cracks, (larger) cracks, and all types of metal loss. Accordingly, the term "metal change" is intended to include dislocations, micro cracks, cracks, all types of corrosion and erosion. The terms may have been used interchangeably, but a specific physical change in the metal is not intended to exclude other physical changes in terms of measurement of changes in the metal in general.

The terms "reference value", "reference data", or "calibration data" as used herein, refers to a voltage measured as described above, of an undamaged ideal metal structure or to a voltage measured at a location with known conditions.

The term "metal" as used herein, refers to any metal, metal alloys, including structures comprised by the same, such as pipes in process plants, subsea risers and drilling pipes at offshore oil facilities, and ra i I way ra i Is.

The term "transient voltage" or "voltage response" as used herein, is meant to include the voltage measured at a pair of the second set of contact pins.

The term "contact point" has been used herein to include "contact pins", as used in the art and in former patent publications mentioned above.

The term "current onset" as used herein, is meant to define the point in time when direct current is applied to the contact points at the metal structure, i.e. at t=0. Accordingly, "current onset" refers to the start of the pulse shaped direct current.

The term "measurement device" as used herein, is meant to include the components necessary to perform a measurement at a single location. Accordingly, each measurement device comprises the first and second contact points, the DC source, and the measurement circuit as mentioned above and described below in relation to Fig. 1.

Drawings

The invention is described in further detail below with reference to drawings, where:

Fig. 1 illustrates a schematic cross-section through a metal structure with a device for taking measurements, Figs. 2A and 2B shows, in a schematic representation, the course of the excitation direct current pulse as a function of time, and the course of the voltage response as a function of time, respectively, Fig. 3A is a schematic illustration of deviation curves, showing deviation of measured voltage responses as a function of time after current onset, compared to a reference measurement at a new steel plate. Each curve represents different degree of corrosion in the opposite side of the plate.

Fig. 3B illustrates a normalized deviation diagram from the deviation curves in Fig. 3A,

Fig. 4A is a normalized deviation diagram similar to Fig. 3B, but with a value for time after current onset established at a selected value for normalized deviation, Fig. 4B is a diagram showing the depths of material losses in mm for respective values for time after current onset from the normalized deviation midpoints indicated in Fig. 4A, Fig. 5 is a drawing showing a metal piece with two corrosion pits håving different lengths but identical depth and width, one deviation diagram similar to Fig. 3A, and one normalized deviation diagram, Fig. 6 is drawing similar to Fig. 5, but with metal loss at two different depths, and same length and width, Figs. 7A-7C are schematic voltage deviation diagrams illustrating location-dependent and size-dependent responses in the initial phase of the measurement, immediately after current onset and also the rest of the response. Fig 7A and 7B show typical response to crack initiation and growth in the same surface as the contact pins, while Fig 7C shows typical response to crack or corrosion in the surface at the opposite side.

Detailed description

Fig. 1 provides a schematic illustration of a metal piece 10 in a cross-section through the same. A first set of electrically conductive contact points 12, 13 are arranged mutually distanced at the surface of the metal piece 10. The respective contact points are connected to a source of direct current (DC) included in unit 11, to apply excitation current to the contact points. A second set of contact points 14, 15 comprising a pair of mutually distanced contact points are arranged between the pair of contact points 12,13 in the first set of contact points. The second set of contact points are connected to a voltage measurement circuit in unit 11 to amplify voltage measured at the second set of contact points 14, 15 and to digitize measurement values and additional equipment 16 to perform the control of measurements, data acquisition and data storage and upon request transfer the data to a PC or similar 17 that computes numerous figures from the voltage course over time.

An apparatus described above may include numerous pairs of contact points at numerous locations at the structure being monitored. The apparatus is known from the prior art, e.g. from NO 323931 Bl, and is not described in further detail here.

Fig. 2A shows the course of the excitation current as a function of time, in amperes as a function of time. As can be seen from the diagram, the current applied is constant. Fig. 2B shows the course of the voltage change in u.V as a function of time. The curve indicated at a(t) represents measured values, whereas the curve indicated at i(t) represents reference values provided from an ideal condition or from measurements tåken at known conditions. After onset of the current, the voltage drops rapidly in a substantial non-linear manner. This is called the transient phase of voltage drop. After a while, the voltage drop flåttens out and the voltage becomes constant. This is called the stationary phase of the voltage drop. The upper curve indicated at a(t) exhibits higher voltage values during the last part of transient phase, indicating higher resistance in the metal caused by metal change from, e.g., cracks or pits. Further analysis of this information enables characterization of the material change with regard to size, geometry and depth. This is described in further details below. The course of voltage change in the very beginning of the DC onset may also be used to determine the location of the metal change. This part of the curve is also illustrated schematically in Figs. 7A-7C described below.

The diagrams shown in Figs. 3-6 are provided from simulated values from a computer model, with input from observations made during physical measurements. These values have been verified by real tests by stepwise increase of metal loss depth or crack depth in metal samples made by gradually increased cut-outs in the metal samples.

Fig. 3A shows a deviation diagram with eight curves showing voltage deviation from reference measurement in parts per thousand as a function of time (milliseconds) after current onset, from eight separate consecutive measurements over time at the same location of a metal structure. The

Iowermost curve represents the first measurement with a (low) metal loss of 10%. The remaining curves show increasing metal loss, where the uppermost curve represents 80 % metal loss. The arrow indicates the direction of increased depth of metal loss in relation to the absolute deviation values.

The deviation values during the whole (transient and stationary phase) course of voltage change in a measurement, constituting the curve in Fig. 3A for example, may be calculated as follows:

D2(t) = (v2(t)M(t) -1),

Where index 2 is measurement number 2, tåken at a specific location in the material being monitored, V^t) is a reference voltage measured at known conditions (reference), V2(t) is a voltage measured at the same point at a later time. There may be any number of measurement values n within a series of measurements, such as 10, 50, 100, 200, 500 or 1000, tåken at any desirable interval, such as daily, weekly, one time every month, every second month. Accordingly, when the measured voltage is equal to the initial voltage at a selected point in time, the deviation is zero. Håving increased resistance in the metal, e.g., caused by micro cracks, Vn(t) may have become larger than the reference voltage at the selected point in time. The deviation may be calculated in numerous ways. A value calculated according to the above may for example be multiplied by 1000 to increase the numeric resolution.

Whereas the absolute value of the voltage deviation as discussed above may provide information about the metal loss, the diagram itself does not tell whether the metal loss is caused by a crack or by a pit, or the actual depth of the localized attack. Accordingly, the measured values must be subjected to further processing and analysis to provide information about location, geometry and size.

Fig. 3B illustrates a diagram with normalized voltage deviation values as a function of time (ms) from the values plotted in Fig. 3A. The arrow indicates the direction of increased depth of metal loss. The normalized voltage deviation may be calculated as described below. A voltage value from the stationary phase (DC) of a selected measurement and the voltage value from the reference at a selected point in time, are used to calculate a deviation coefficient Dn(DC):

Dn(DC)=Vn(DC)/Va(DC)-l,

where Vn(DC) is voltage measurement number n tåken at a specific point in time in the stationary phase of a selected measurement, whereas Vi(DC) the reference voltage, preferably tåken at the

same point in time on the DC part (stationary phase) of the voltage response. This coefficient Dn(DC) is then used to calculate normalized deviation values NDn(t) from the remaining deviation values Dn(t) within one and same measurement:

where n represents measurement number/?, such as the second measurement at a selected location, D„(t) is voltage deviation from reference at a specific point in time t of the response voltage, and Dn(DC) is the deviation coefficient calculated above. At t=0, or in other words at current onset, the relation between the absolute deviation at t=0 is small compared to the deviation constant in the stationary phase for crack or corrosion development on the other side. In the stationary phase, Dn(t) and Dn(DC) approaches each other, and the normalized value NDn(t) approaches 1. Accordingly, ND„(t)=[0..1]

Characterization of metal loss

Fig. 4A corresponds to Fig. 3A, where the first measurement is indicated by the square and dotted lines. The arrow indicates measurements with increasing metal loss from the remaining measurements. The leftmost curve corresponds to the uppermost curve in Fig. 3A, representing the highest metal loss in this measurement series. The normalized deviation diagram will be used to provide the depth of the metal loss. From the normalized deviation diagram, it is also possible to determine whether an increasing metal loss occurs along the surface of the metal structure or grows into the metal structure. Furthermore, by combining information in 3B and 3A it can be determined if a localized arrack is a crack or corrosion pit. This is explained in further detail below.

In order to find the depth of a metal loss in the opposite side, a value is selected from the normalized deviation values in a selected deviation curve. In Fig. 4A, the value for time after current onset is e.g. selected from a part of the diagram that shows highest resolution, e.g., at a part where mutual difference between consecutive measurements are highest. In this diagram, the desired region is the mid-point of the selected curve in the normalized deviation diagram at value 0.5. The value for time after current onset is directly related to the depth of the metal loss and not influenced by the volume of metal loss or cross section of cracks. In Fig 4B, respective values for time after current onset at the mid-point of the normalized deviation curve, i.e. at about 0.5, are plotted as a function of depth of material loss in millimeters. This diagram shows that there is a direct connection between time after current onset, e.g. selected at the mid-point of a normalized deviation diagram, and the depth of the material loss.

Fig. 5 is a composite drawing showing a 10 mm thick metal sample with two different metal loss sizes. In a first state, the metal sample has a metal loss of 2 mm depth, 5 mm length and 10 mm width. In the second state, the metal loss has increased to a length of 10 mm. The depth and length remain the same. The absolute deviation in voltage change for the two states is shown in the deviation diagram down to the left. The relatively low value for deviation in parts per thousand (ppt), i.e. some tens ppt, tells us that the metal loss is relatively small. In the diagram to the right, the deviation values has been normalized as outlined above, and plotted in a normalized deviation diagram. Here we can see that the curves for the two states are substantially coinciding, which tells us that the metal loss is developing along the surface of the metal in two dimensions, and not in the depth of the material. Taking factors such as material dimensions into consideration, the absolute value of the voltage deviation, together with the normalized deviation curves, may provide a direct measure of the size of the metal loss and whether it is developing in two or three dimensions. This information may also be combined with location information (metal loss at the same side as the contact points or at an opposite surface) from the very early course of deviation of voltage change, immediately after current onset and seen in relation to the later part of the deviation curves. Fig. 6 is a composite drawing similar to Fig. 5. Also this drawing illustrates a sample at two stages during development of metal loss, but here the metal loss is developing into the 10 mm thick metal sample, whereas the width = lOmm and length = 5mm remain constant. At the first measurement, the metal sample has a pit depth of 2 mm and at the second measurement pit depth is 5 mm, whereas the width and length remain constant. At the first measurement, the absolute deviation value is comparable to the first measurement of Fig. 5 (both 100 mm<3>). The volume of the second metal loss in Fig. 5 is 10 x 10 x 2 mm = 200 mm<3>, whereas the volume of the second metal loss in Fig. 6 is 5 x 5 x 10 = 250 mm<3>. However, the absolute value of the voltage deviation has about doubled from about 35 ppt to about 78 ppt. The diagram to the right in Fig. 6 illustrates the normalized deviation from the diagram to the left. Here it may be observed that there, for a selected value of normalized deviation (around 0.5), is a difference in time after current onset. This information indicates increased depth of the pit which is developing into the metal.

It can be mentioned that similar deviation values for a pit volume of 360 mm<3>with a width W, length L and depth D of respectively 6 x 10 x 6 mm in a 10 mm thick metal sample, exhibits an absolute value for voltage deviation of more than 900 ppt at the stable part of the deviation curve. On the other hand, a crack extending 6 mm into a 10 mm thick metal sample and with a width of 1 mm and length of 50 mm (300 mm<3>), exhibits an absolute value of about 370 ppt. The latter absolute deviation value of 370 ppt may be compared to the 250 mm<3>large pit from Fig. 6 which exhibits a deviation value of about 78 ppt. All the 6 mm deep crack and corrosion have similar normalized curves representing the actual depth. Accordingly, the absolute deviation value accompanied by the normalized deviation discussed above can discriminate between cracks and corrosion and can provide geometry and size of metal loss and crack into which direction the defect is developing. Analyzing the rate of the defects development over time may reveal increased speed of defect development. Figs. 7A, 7B and 7C are schematic diagrams illustrating course of voltage deviation related to size and location of defect. Fig. 7A is a typical curve for micro cracks in the surface on the same side as the contact points (pins), e.g. on the external surface of a pipe. Fig. 7B illustrates a situation similar to Fig. 7A, but where the micro cracks have merged and formed larger cracks. Both curves in Figs. 7A and 7B have a sharp rise in the very beginning of the phase immediately after current onset, whereupon the curves are falling downwards. In Fig. 7C, the sharp rise in initial voltage deviation is absent, whereupon the deviation has a slight increase. The latter situation is a typical curve for crack or metal loss on the internal surface of a pipe where the contact points (pins) are arranged at the external surface of the pipe.

As mentioned above, ferromagnetic steel structures which have been subjected to stress, e.g. risers which have been suspended from platforms and subjected to different kinds of forces, will typically exhibit changed magnetic properties, which again may affect the result from a measurement described above. In order to minimize unwanted influences of this remanent magnetization, the metal is advantageously subjected to numerous consecutive measurements, e.g. 4 or 5 measurements, each measurement including one continuous DC pulse. The step above is also denoted as "discrete measurement series". The time between each separate measurement in a discrete measurement series is typically a few milliseconds. Numerous consecutive measurements will demagnetize the metal and remove the influence from previous stress. However, the consecutive measurements discussed above are interpreted as one single measurement in the method disclosed herein.

The demagnetization step above may also be used to determine maximum stress the structure has been subjected to since the preceding measurement. A comparison of the first, and for example the fifth measurement in a discrete measurement series discussed immediately above, that reveals a change in voltage deviation between the same, indicates that the structure has been subjected to stress since the preceding measurement, and the value of the deviation is proportional with the maximum stress since the preceding measurement. By comparing this difference with a difference from such a measurement series tåken at a former point in time with known maximum stress, it is possible to determine the maximum stress the structure has been subjected to between two measurements. We have found that there is a linear proportional relation between deviation described above and maximum mechanical stress.

Exact absolute values, ratios and other figures when monitoring stress may vary with material composition, material dimensions and other properties of the metal structure. However, monitoring of crack and metal loss, the results are mainly influenced by the geometry. This information will be within the reach of a person skilled in the art, by including the present disclosure and knowledge from the literature within this field of technology. It should also be mentioned that the diagrams disclosed herein only serve as an illustration to simplify interpretation of the present method, and are not intended to limit the scope of protection. Accordingly, a tabular or array representation of the values in question would serve the intention as well.

Claims

1. A method for detection of material changes, including micro cracks, cracks, corrosion and erosion, in metal structures (10), such as railway rails, drill pipes, pipelines, risers, and bridges, the method comprising the steps of: arranging at least one first set of electrically conductive contact points (12, 13), each set comprising a pair of contact points arranged at a mutual distance, arranging at least one second set of electrically conductive contact points (14, 15), each set comprising a pair of contact points arranged at a mutual distance, wherein said second set of contact points (14,15) is arranged between the respective contact points in the first set of contact points (12, 13), the method further comprising the steps of: a) applying an excitation current to said at least one first set of contact points (12,13), b) measuring the voltage response between the contact points in the second set of contact points (14,15), c) analyzing both the transient part and the stationary part of the voltage response, d) comparing the measured voltage response with reference data and calculating deviation values to quantify deviation of voltage response from the reference data as a function of time after current onset, and e) repeating steps a) to d) above for a period of time to provide a set of separate voltage measurements for the selected location in the metal structure, characterized in: f) providing normalized deviation values from the deviation values as a function of time after current onset, g) choosing a value from the normalized deviation values at a point where differences between respective normalized deviation values gives the best sensitivity and resolution, which corresponds to the largest difference in time after current onset between the normalized deviation values for the actual application, and h) at the selected normalized deviation value, selecting one or more values for time after current onset from the respective measurements, i) comparing said at least one selected value for time after current onset in step h) with another value for time after current onset from the same location in the metal related to different measurements, or with a measurement tåken at a location håving a damage with known dimensions, thus determining the geometry of the damage.

2. The method of claim 1,characterized incalculating the depth of metal loss or crack from the selected value for time after current onset in step h).

3. The method of claim 1 or 2,characterized inanalyzing the deviation values from step d) of the initial part and the late part of the transient phase, to determine whether the metal loss is located at the same side of the metal structure or at opposite side of the contact points in the metal structure, including at an external surface or at an inner wall of a pipe.

4. The method of any one of claim 1 to 3,characterized inanalyzing the deviation value at the stationary part of the deviation curve together with the time after onset on the normalized deviation curve to determine whether the defect located at the other side of the metal structure is corrosion or crack.

5. The method any one of claim 1 to 4,characterized inperforming numerous consecutive measurements with a short time there between, to demagnetize the metal caused by any changes in magnetic properties caused by previous mechanical stress.

6. The method of claim 5,characterized inperforming fourto five consecutive measurements.

7. The method of claim 5 or 6,characterized in thatthe time between each consecutive measurement is a few milliseconds.

8. The method of any one of claim 5 to 7,characterized incomparing the measurement with a similar measurement from an earlier point in time to determine the maximum stress the structure has been subjected to since the previous measurement.

9. The method of any one of claim 5 to 8,characterized intaking a measurement that comprises several consecutive discrete measurements with short time therebetween, and calculating the deviation between the first and last measurement to determine the maximum stress the structure has been subjected to since previous such measurement.

10. The method of any one of claim 5 to 9,characterized ini) providing said first and second sets of contact points to provide a first set of contact point sets, ii) providing at least one further set of contact point sets, Ni) arranging said first set of contact point sets and further contact point sets at a mutual angle.

11. The method of claim 10,characterized inselecting the angle in the range of 0 to 90 degrees.

12. The method of any one of claim 1 to 11,characterized incontrolling the rise time, level and duration of the current excitation to optimize the same for the application in question.

13. The method of any one of claim 1 to 12,characterized inusing the measured results to calibrate data models used for calculation of fatigue of the whole structure to improve the precision of such models.

14. The method of any one of claim 1 to 13,characterized inapplying the method on a riser at an offshore platform, wherein an initial measurement is tåken before the riser is tåken in use, followed by one or more measurements after the riser has been tåken in use, thus detecting changes in the steel as a consequence of stress applied during use.