FIELD OF THE INVENTION

The present invention is in the field of a non-contact system for monitoring a metallic magnetic structure under dy namic load for detecting an impact induced propagating stress wave, and a method of determining strain in a metallic mag netic structure under dynamic load, such as a tube-like struc ture, such as a monopile for a wind turbine.

BACKGROUND OF THE INVENTION

Increasing demand for energy from renewable sources has resulted in a spectacular growth in a number of offshore wind farms on sea, such as in the North Sea. Typically wind tur bines in these farms are mounted on large top steel monopiles with diameters ranging up to eight meters and height up to more than 200 m. Typically these thin-walled cylindrical piles are driven into the seabed by hydraulic impact hammers. Due to the large forces exerted at the pile head during pile driving, the structure is prone to plastic deformations. Contact strain measurement techniques are difficult to use at the pile head during pile driving, especially on sea.

Current models of pile driving assume linear elastic mate rial behaviour. However, from practice it is known that plas tic deformations occur close to the pile top due to the high stresses in this region. These regions of plastic deformation can have a negative influence on the expected lifetime of the support structure of the offshore wind turbine. Due to the in accessibility of the pile top during installation, a method to detect the presence of this plastic region from a non-collo- cated measurement is needed.

When driving piles by hydraulic impact hammers also dy namic strain is introduced into the piles. Real-time measure ments of dynamic strain, especially at difficult accessible places, such as at sea, is often not possible. As a conse quence, information on the dynamic strain is absent.

EP2725253 A1 recites a device for metallic structure maintenance. The device uses a magneto-graphic/Magnetic Tomography technique to identify stress-related defects. The device is specifically optimized to be used for extended, non- accessible underground and underwater metallic structures in providing quality control, emergency alarms as well as time line planning for structural repairs and maintenance work. Ex amples of the use of the device include pipes for oil and gas industry. It is especially important for loaded constructions, such as pressured pipes, infrastructure maintenance, nuclear power plant monitoring, bridges, corrosion prevention and en vironment protection. The document is considered background art, wherein a magnetic field is used to detect imperfections in magnetic materials.

The present invention therefore relates to a non-contact system and a method of measuring, which solves one or more of the above problems and drawbacks of the prior art, providing reliable results, without jeopardizing functionality and ad vantages .

SUMMARY OF THE INVENTION

It is an object of the invention to overcome one or more limitations of the devices of the prior art and at the very least to provide an alternative thereto. The invention is also subject of a to be published PhD-thesis by P. Meijers of the Technical University of Delft, which publication and its con tents are incorporated by reference thereto.

Here the possibility of using the magnetic field generated by the magnetisation of the ferromagnetic pile to assess the stress state is detailed. The invention relates to a non-con- tact system for monitoring a metallic magnetic structure under dynamic load for detecting an impact induced propagating stress wave. Experiments have shown that the magnetisation of a ferromagnetic material changes with the applied stress level, even in weak constant magnetic fields, like the Earth magnetic field. Recently, experiments on the magnetic response of a large-scale ferromagnetic thin-walled cylinder under stress were reported by Viana et al . The loading therein, how ever, was quasi-static, whereas for impact loading the time scales involved are in the order of milliseconds. A numerical model is developed which couples the magneto-mechanical performance to the propagation of mechanical stress waves in the pile. The resulting magnetic field in the air region sur rounding the pile is compared to the stress history to show the applicability of the model to assess the stress state due to impact loading. Contrary to prior art considerations it has been found that a magnetic field sensor per se does not pro vide reliable results, if any. It is required to use an aniso- tropic magnetic field sensor. As an alternative a giant mag neto resistance (GMR) sensor can be used. These are passive sensors, which makes their application beneficial, particu larly at high sea. In addition, it is required to use an array of sensors, wherein sensors each individually measure magnetic field resistance. Further, it is required to use a relatively high sampling rate of >10kHz. The sampling rate is preferably higher, such as >40kHz, in order to obtain sufficient infor mation. High sampling rates typically imply also sufficient calculating capacity and strong algorithms. At least one array of anisotropic magnetoresistance (AMR) sensors, which sensors are substantially at a same first height, typically in the same plane, is used, wherein sensors are operated at said sam pling rate. Details of such sensors can be found in Ripka (ISBN 1-58053-057-5), which book and its contents are incorpo rated by reference herein. In principle only one sensor would be sufficient, however then only a vertical magnetic field component (B _{z )} and one horizontal magnetic field component (B _r) can be detected. With two sensors, typically spaced opposite of one and another, the horizontal magnetic field component is observed in two parallel horizontal directions, and hence can be determined more accurately. It is however preferred to use at least a third sensor, as then also the second horizontal direction of the magnetic field component can be determined. Each sensor may independently be tilted with respect to the vertical axis, and rotated in the horizontal plane. In order to compensate for e.g. (local) movement of a pile further sen sors may be provided. The array of sensors is spaced around the structure and the measurements can take place.

In a second aspect the present invention relates to a method of non-contact monitoring of a metallic magnetic struc ture under dynamic load, comprising providing the present system, determining the magnetic stray field (typically meas ured field minus background field) , and calculating at least one of plastic strain (also referred to as plastic defor mation) , and rigid body motion. It has been found that the ef fect of a dynamic load can be established accurately, as least as accurate as with comparable contact measurements. As the installation of contact sensors is cumbersome, full contact is difficult, if not impossible, such as at sea, the non-contact measurement provides advantages.

The present invention provides a solution to one or more of the above-mentioned problems and overcomes drawbacks of the prior art.

Advantages of the present description are detailed

throughout the description.

DETAILED DESCRIPTION OF THE INVENTION

In an exemplary embodiment of the present system the sen sors may be evenly spaced around the structure. For instance, in circular mode n sensors are evenly distributed over said circle, and at an angular distance of 360/n°.

In an exemplary embodiment of the present system at least one sensor may be an analogue sensor. It is preferred to use analogue sensors, as these provide high speed. Digital sensors could likewise be envisaged, but these are at the time of writing to the knowledge of the inventors not available with the required characteristics.

In an exemplary embodiment of the present system the array may be spaced in a circular manner.

In an exemplary embodiment of the present system sensors in the array may be synchronized in time and may be in commu nication with a high-speed data acquisition unit. In order to obtain very accurate information on the stress wave sensors in the array are preferably fully time-synchronized, i.e. operat ing within exact the same time frame, wherein the time frame may be determined by one single clock, such as a clock of a controller .

In an exemplary embodiment of the present system each sen sor may be located at a distance of 1-100 cm from the struc ture, preferably 5-50 cm, such as 20-30 cm. On sea, said distance is preferably not too small, as compensation for waves (of a supporting system or ship) is difficult. The dis tance is preferably not too large, as signal strength de creases with distance.

In an exemplary embodiment of the present system at least one further anisotropic magneto resistive sensor, preferably at least one second array of anisotropic magneto resistive sensors, may be provided at a second height, which second height is preferably 10-100 cm above or below the first height. The further sensor and/or array may be similar to the first array.

In an exemplary embodiment of the present system each ar ray may comprise each individually more than 2 sensors, pref erably more than 3 sensors, such as 4-10 sensors.

In an exemplary embodiment of the present system each ar ray may comprise a support, such as a ring, on which sensors are attached.

In an exemplary embodiment the present system may comprise a feedback loop, wherein the feedback loop is adapted to in crease or decrease a subsequent dynamic load, and/or is adapted to increase or decrease a frequency of subsequent dy namic loads. Such may relate to adapting a number of loads per time (#/min), and/or by adapting an interval between loads.

In an exemplary embodiment of the present method the me tallic structure may be selected from a tube-like structure, such as a monopile, such as a monopile for a wind-turbine, a tube for oil or gas production, for off-shore application, for on-shore application, a steel bridge, and combinations

thereof .

In an exemplary embodiment of the present method may fur ther comprise determining the geometry of the structure. For accuracy having information on the geometry may be beneficial.

In an exemplary embodiment of the present method may fur ther comprise establishing a magnetic equilibrium status of the structure.

In an exemplary embodiment the present method may comprise providing a calibration. Said calibration may be used to accu rately determine the effects of the dynamic load(s) . In an exemplary embodiment of the present method the me tallic magnetic structure may comprise a material selected from ferromagnetic material, anti-ferromagnetic material, ferri-magnetic material, and combinations thereof.

In an exemplary embodiment of the present method the me tallic magnetic structure may be a tube-like structure, such as a monopile, such as a monopile for a wind turbine, a tube for oil or gas production, for off-shore application, for on shore application, and combinations thereof.

In an exemplary embodiment of the present method a down ward moving stress wave may be measured, and a reflected stress wave may be measured.

In an exemplary embodiment of the present method an axial displacement of the structure may be measured.

In an exemplary embodiment of the present method a verti cal tangential and/or axial deformation may be measured.

In an exemplary embodiment of the present method a sam pling rate may be 10-250 kHz, preferably 20-200 kHz, more preferably 40-150 kHz, such as 50-100 kHz.

In an exemplary embodiment of the present method the feed back loop may increase or decrease a subsequent dynamic load, and/or may increase or decrease a frequency of subsequent dy namic loads, and/or may maintain dynamic load and frequency, as described above.

The invention will hereafter be further elucidated through the following examples which are exemplary and explan atory of nature and are not intended to be considered limiting of the invention. To the person skilled in the art it may be clear that many variants, being obvious or not, may be con ceivable falling within the scope of protection, defined by the present claims.

SUMMARY OF THE FIGURES

Figs. 1, 2a-c, and 3a-e show some experimental details.

DETAILED DESCRIPTION OF FIGURES

Fig. 1 shows schematics of the measurements. An array of a number of AMR magnetometers surrounding a pile provided meas urement input. Using a high-speed data acquisition system, and having information on the geometry of the pile, which can be obtained or determined in advance, processing software pro vides information of strain, rigid body motion, and plastic deformation. For calculation a one-dimensional wave propaga tion model with Rayleigh-Love correction can be used. For the magneto mechanical model Jiles's law of approach can be used.

Fig. 2a shows and experimental set-up with the AMR indi cated with a solid arrow, and the contact strain sensor with a dashed arrow. The contact monitor is attached to the pile, and the AMR at a distance of about 40 cm. Fig. 2b shows the same set-up and gives an indication of actual sizes. Fig. 2c shows an image, obtained with a camera, of plastic deformation in the pile upon applying a load, indicated with the arrow.

In an example inventors studied the results of dynamic loads. In fig. 3a and 3b a series of loads, starting at about 40 seconds, and ending at about 70 seconds was applied and the magnetic field B _r [mT] and B _z [mT] were measured. Figs. 3c and 3d show a blow-up part of the measurements. Further in these figures it can be seen that the results of the non-contact de termination and contact-monitoring overlap well. A distance of about 20 cm of the AMR sensor was found appropriate. The left- hand column of fig. 3 graphs are measured given the axial strain e _z) , measured in the prior-art way, so with a glued strain gauge. The right column focuses on the axial component of the magnetic field (B _z) measured at 20 cm from the pole.

The first row (fig. 3a) is the full signal; the second row (fig. 3b) shows an enlargement of every hammer blow; row three (fig. 3c) shows the deviation of the signal on top of the spot field, so now both signals start at about 0; In row four (fig. 3d) , both signals are normalized by dividing each signal by the peak value. The bottom row (fig. 3e) combines both normal ized signals to show that with the correct scaling (ratio max (e _z) /max {Bz) ) the signals correspond, and therefore that the elongation can be measured by magnetism.

Experiments have been performed which support the figures and advantageous effects mentioned in the description.

The research on which this patent application is based on research that has been made possible by a grant from NWO in the EUROS (Excellence in Uncertainty Reduction of Offshore wind Systems) program from NWO (#2014/13216/STW) .