WO2005036163A1

WO2005036163A1 - Method for determining corrosion on a ship

Info

Publication number: WO2005036163A1
Application number: PCT/AT2004/000347
Authority: WO
Inventors: Peter Tscheliesnig; Gerald Lackner; Hartmut Vallen; Henryk Buglacki; Jan Patkowski; Jerzy Schmidt; Szczepan Malinowski
Original assignee: Tüv Österreich (Technischer Überwachungs-Verein Österreich); Vallen Systeme Gmbh; Technical University Of Gdansk; Polski Rejestr Statkow S.A.; Foundry Research Institute; Maritime Institute Gdansk
Priority date: 2003-10-13
Filing date: 2004-10-12
Publication date: 2005-04-21
Also published as: AT412588B; ATA16112003A

Abstract

The invention relates to a method for determining corrosion on a ship, especially on a tanker, which can be carried out in a simple and economical manner. Said method is characterised by the following characteristics: sound emissions are detected and captured by means of a sensor (4, 5); characteristic variables of the sound emissions are determined; the characteristic variables of the sound emissions are evaluated; the frequency spectrum of the sound emissions is determined; the frequency spectrum is evaluated; the signals are separated according to the characteristic variables and to the frequency spectrum characteristic, etc. in a group of signals produced by corrosion, and in group produced by other sources of noise; consulting the values for evaluating the state of the boat.

Description

Method for detecting corrosion on a ship

The invention relates to a method for determining corrosion on a ship, in particular on a tanker, and a device for carrying out the method.

Due to some damaged tankers that have caused environmental disasters, tankers have to be replaced by insurance or registration organizations, such as Lloyds Register or Veritas office, to be checked for the integrity of the ship structure or the tank room. In particular, it is about determining impermissible corrosion.

Various types of tanker ships are currently operating on the oceans, namely those in which the tank wall is identical to the hull, those in which the bottom of the tank is separated from the hull, and still others in which the entire tank is separated from the hull and is connected to it, for example via spacers.

Depending on the type, a check for corrosion of all these types is more or less time-consuming, especially since the correct location of rust spots currently requires an optical inspection, which in turn means that the tank compartment is empty. This not only causes a longer stay in a port, but also an inspection of the tank room, which cannot be done without dangers or health risks for the test team. It is particularly difficult to check tankers for corrosion in which the tank wall or the tank bottom is separated from the hull.

The object of the invention is to avoid the disadvantages described above and to provide a method and a device for carrying out the method with which the inspection of a tanker is possible in a simple manner and without an additional stay of the tanker in a port caused thereby. In particular, it should also be possible to constantly check a tanker, that is, even while sailing on the open sea, so that immediate action can be taken if an inadmissible corrosion occurs.

In a method of the type described in the introduction, this object is achieved according to the invention by the following features: recording and recording sound emissions by means of a sensor, determining parameters of the sound emissions, • Evaluation of the parameters of the sound emissions, • Determination of the frequency spectrum of the sound emissions, • Evaluation of the frequency spectrum, • Separation of the signals according to the parameters and according to the characteristics of the frequency spectrum, etc. into a group of signals caused by corrosion and into a group caused by other noise sources, • using the values for an assessment of the condition of the ship.

It is already known to detect corrosion processes in different metals with the aid of a sound emission test. Both signals that are emitted by the corrosion process as well as secondary signals that are only indirectly caused by the corrosion or by the corrosion products, e.g. Rust, caused, recorded by means of piezoelectric sensors. The sound path goes from the source location, i.e. from the point of origin, via the metal or preferably via a liquid to the sensor. Based on empirical values, the signals that occur can be assigned to different types of corrosion and different damage criteria.

This method alone cannot be applied to ships, in particular tankers, without any problems, especially since ships are exposed to extraordinary ambient noise both while sailing on the high seas and in a harbor basin. These ambient noises are partly caused by mechanical effects, e.g. Waves, engine noises etc., caused and further by the ship internals themselves, such as Spacers, stiffening ribs, baffles so that frequency filtering of the signals picked up by sensors does not lead to the goal.

According to the invention, the signals caused by corrosion are preferably assigned to different types of corrosion and levels of corrosion.

The individual signals are expediently assigned to certain areas, e.g. spatial or flat areas, assigned to the ship, for which purpose to form a multi-sensor, three or more sensors are advantageously arranged at a predetermined distance and with the same sensitivity and time differences of the arrival of a signal are measured for each sensor and the direction based on these time differences, from which the signal comes is determined.

To determine a spatial direction, four sensors are preferably arranged at predetermined distances from one another and with their sensitivity rectified. According to a preferred variant, the cross-correlation function is used to determine the time differences.

In order to achieve a high accuracy of the measurements, at least two multi-sensors are provided for each room to be evaluated and their received signals are evaluated together.

The data determined by the sensor are preferably subjected to a first pattern recognition for the secretion of data, based on ambient noise, expediently after the first secretion of data, originating from ambient noise, one or more further pattern recognizations with respective secretions of data that do not originate from corrosion , be performed.

According to a preferred method variant, a time shift of the arrival times of a signal at the sensors of a multisensor is used to determine the location of the source of the signal, a function value being calculated as a function of the time shift and a maximum of the function value being determined and the time shift associated therewith as the arrival time difference of a signal is determined.

For constant monitoring of a ship, a sensor, in particular a multisensor, is expediently kept in operation while the ship is moving, preferably permanently, and the data determined by it are sent to a stationary evaluation device.

A device for carrying out the method according to the invention is characterized by at least one sensor for recording and recording a sound emission and an evaluation device for the signals received by the sensor (s) and forwarded to the evaluation device, the sensor either permanently in a ship, in particular a tanker, is arranged or can be removed and installed and fixed in the ship at a predetermined location. The latter in order to carry out periodic monitoring.

To form a multisensor, three or more sensors are preferably fixed to one another at predetermined distances from one another. At least two multisensors are expediently either fixed in the ship or fixable in the ship, the sensor or the sensors or the multi-sensor or sensors being advantageously provided within a tank space of a tanker.

For tankers, the sensor or sensors are expediently formed intrinsically safe and made of piezoelectric material, and the waves thus converted into electrical signals can be evaluated in a computer of the evaluation device, the computer advantageously having a program for determining the frequency response and a classification program based on reference data on different ones Sound sources is equipped.

The invention is explained in more detail below with reference to the drawing, with FIG. 1 schematically illustrating the basic structure of a tanker. Fig. 2 shows a tank unit of a tanker in an oblique view and Fig. 3 in longitudinal section. 4 illustrates several sensors combined to form a multisensor. 5 shows the signals received by sensors. 6 shows the assignment of signals to a source mechanism. 7a and 7b each show flow diagrams for data processing of the signals picked up by the sensors and FIGS. 8a and 8b the assignment of the signals to different source mechanisms, namely once for the raw data (FIG. 8a) and once after the location (FIG 8b).

The tanker shown in Fig. 1 has a plurality of self-contained tank units 1, which either have their own tank walls, regardless of the hull, or the side walls and / or bottom of which forms both the tank wall and the hull. The stiffening ribs and / or baffles still installed in the individual tank units 1 are not shown in detail.

Tankers of this type (e.g. for 125,000t) usually have lengths of 250m, a height from keel to deck of around 25m and a width of around 50m.

As can be seen from FIGS. 2 and 3, each tank unit 1 has at least two openings 2 directed upwards, preferably at sockets, which are closed with a cover plate 3. Sensors 4, which are preferably combined to form multisensors 5 shown in FIG. 4, extend through these openings 2 into the tank unit 1.

Each sensor 4 is made of piezoelectric material and is therefore suitable for converting a mechanical wave into an electrical signal. Each of the multisensors is arranged on a rod 6, which is fixed relative to the tank unit 1, for example on the cover plate 3, so that the position of the multisensors 5 within the tank unit 1 is precisely defined. It is essential here that the multisensors 5 are also secured against rotation about the axis of the rod 6. A multisensor 5 would be sufficient for locating sound sources, but the information from two or more multisensors 5 increases the reliability of the locating results.

Each multisensor 5 is as central as possible, but at a distance A (often predetermined by the manholes (openings 2)) from one another either permanently provided in the tank unit 1 or they are used in the tank unit 1 with periodic checking of a tanker when the ship is in port located. As can be seen from FIG. 3, the tank unit is filled up to the deck 7 with a liquid product 8.

A corrosion point 10 is indicated on the bottom 9 of the tank unit, and the sound wave field 11 emanating from this corrosion is shown in FIG. 3. It can be seen that the sound waves - these are spherical waves - reach the two multisensors 5 at different times due to the different distance from the corrosion site 10, i.e. there are time shifts, so-called arrival time differences 12 (FIG. 6) of the sound waves, not only at the multisensors 5 themselves, but also at the sensors 4 forming a multisensor 5.

These sensors 4 (hereinafter referred to as XD1 to XD4 for distinction) are combined at a precisely predetermined distance B and with their sensitivity rectified to form a multisensor 5. Four sensors per multisensor 5 are provided for locating corrosion in any direction in the room. Three sensors 4 combined to form a multisensor 5 would suffice to locate corrosion 10 within a surface, that is to say to check an individual wall of a tank unit 1.

5 shows the electrical signals converted by the individual sensors 4 (XD1 to XD4) of a multi-sensor 5 due to mechanical excitation as a voltage-time function, the abscissa representing the time and the ordinate showing the signal amplitudes of the voltage. It can be seen that the individual sensors XD1 to XD4 of the multisensor 5 receive one and the same signal at different times, and the direction can be determined from the individual arrival time differences 12 due to the position of the individual sensors XD1 to XD4 of the multisensor 5 in space. from which this signal comes can be determined. To determine the signal arrival time difference 12, each sensor XD1 to XD4 is connected to the input channel of a sound emission measuring device, which is able to measure the time and the applied electrical voltage and to store the measured values. The signal arrival time is determined by the first time that a set measurement threshold is exceeded. The signal is recorded when the measuring threshold is exceeded.

The arrival time difference 12 can be determined via a direct measurement of the arrival times using a fixed measurement threshold or with the help of the cross-correlation function.

The cross correlation function allows the arrival time difference 12 to be determined from the continuously recorded signal profiles. The signal x (t) is measured at the sensor XD1, the signal y (t) is measured at the sensor XD2. The signal y (t) is shifted along the time axis with the aid of the parameter τ until it overlaps with the signal x (t) in the best possible way. This condition is implemented mathematically by calculating the function value (numerical evaluation of the folding integral) as a function of the time shift. If the function value reaches a maximum, the time shift corresponds to the arrival time difference.

ψ _xy (τ) ... function value depending on the time shift τ If τ corresponds to the arrival time difference, ψ _xy (t) is maximum

2T ... considered time window, integration limits x (t) ... signal to sensor XDl y (t + τ) ... signal to sensor XD2, time-shifted by τ

6 shows the basic sequence of the pattern recognition applied in frequency form. The signal is represented by a voltage curve over time (voltage versus time). The calculated frequency response shows which frequencies occur with which strength in the underlying signal. The pattern recognition determines the affiliation to a considered source mechanism by comparison with reference data. On the left side of FIG. 6, two signals and their frequency response are arranged. Pattern recognition assigns a source mechanism to each signal. The right diagram shows two source mechanisms, dark symbols correspond to the reference data for corrosion and light symbols represent reference data for ambient noise. Mapping the signal in this representation (indicated by the curved arrows emanating from the frequency response) shows which source mechanism the signal under investigation is to be assigned to.

Another feature of the pattern recognition used is that characteristic values of the sound emission, such as a maximum amplitude, a signal rise time, signal duration etc., are compared and evaluated with reference data of the source mechanisms under consideration, whereupon signals which are not relevant for corrosion are eliminated and the remaining signals for evaluation the condition of the ship can be evaluated.

7a and 7b show the flowchart for data processing from the raw data to the condition evaluation of the measured tank unit 1. The raw data include all signal data recorded during a measurement. With the help of the signal parameters determined by the evaluation device, the raw data are freed from the portion by processing which makes the subsequent localization calculation more difficult or disadvantageous. The location calculation shows the noise emission sources located on the tank wall. After each step in the process, pattern recognition is applied to the remaining signals. There are different signal allocation distributions. The pattern recognition itself starts from the recorded waveform of the signal, leads to the FFT ("fast Fourier transform" to determine the frequency response) and ends with the assignment to a source mechanism under consideration with the help of a classifier (classification program, based on reference data to the individual source mechanisms).

Figures 8a and 8b illustrate two different signal allocation distributions. The x-axis shows the source mechanisms under consideration, the number of assigned signals is plotted on the y-axis, the raw data (FIG. 8a) contain, for example, 10,000 signals, of which approximately 2,500 ambient noises (25%) were assigned. According to the location calculation (Fig. 8b), the amount of data still contains, for example, 800 signals, the amount of ambient noise being significantly lower (40 pieces or 5%).

In order to determine corrosion on a wall of a ship, it is sufficient to arrange a sensor 4 on this wall. However, in order to determine the exact location of the corrosion, at least three sensors 4 are to be arranged on the wall, as is illustrated, for example, in FIG. 2 for the ship deck 7. The signals are evaluated as described above. The three sensors 4 can be combined to form a multisensor 5 ′, the three sensors 4 lying in the plane of the ship's deck 7. However, it is better to provide larger distances between the sensors 4. This embodiment variant is particularly useful for the ship deck 7 because of the sound-insulating gas phase 13 above the liquid product 8.

Claims

Claims:

2. A method for determining corrosion on a ship, in particular on a tanker, characterized by the following features: • Recording and recording of noise emissions by means of a sensor (4, 5), • Determining parameters of the noise emissions, • Evaluating the parameters of the noise emissions, • determining the frequency spectrum of the sound emissions, • evaluating the frequency spectrum, • separating the signals according to the parameters and the characteristics of the frequency spectrum, etc. into a group of signals caused by corrosion and into a group caused by other noise sources, • using the values for an assessment of the condition of the ship.

2. The method according to claim 1, characterized in that the signals caused by corrosion are assigned to different types of corrosion and levels of corrosion.

3. The method according to claim 1 or 2, characterized in that the individual signals certain areas, etc. spatial or flat areas of the ship can be assigned.

4. The method according to claim 3, characterized in that to form a multi-sensor (5) three or more sensors (4) are arranged at a predetermined distance and with their sensitivity rectified and time differences of the arrival of a signal to each sensor (4) measured and the direction from which the signal comes is determined on the basis of these time differences.

5. The method according to claim 4, characterized in that four sensors (4) are arranged at predetermined intervals (B) to each other and with their sensitivity rectified to determine a spatial direction.

6. The method according to claim 4 or 5, characterized in that the cross-correlation function is used to determine the time differences.

7. The method according to claim 4, 5 or 6, characterized in that at least two multisensors (5) are provided for each room to be evaluated (1) and their received signals are evaluated together.

8. The method according to one or more of claims 1 to 7, characterized in that the data determined by the sensor (4, 5) are subjected to a first pattern recognition for the separation of data caused by ambient noise.

9. The method according to claim 8, characterized in that after the first segregation of data originating from ambient noise, one or more further pattern recognizations are carried out with respective segregation of data that do not originate from corrosion.

10. The method according to one or more of claims 4 to 9, characterized in that a time shift (12) of the arrival times of a signal at the sensors (4) of a multi-sensor (5) is used to determine the location of the source of the signal, wherein a Function value is calculated depending on the time shift and a maximum of the function value is determined and the associated time shift is determined as the arrival time difference (12) of a signal.

11. The method according to one or more of claims 1 to 10, characterized in that a sensor (4), in particular a multisensor (5), is kept in operation, preferably permanently, while the ship is traveling and the data determined by it a stationary evaluation device can be sent.

12. Device for carrying out according to one or more of claims 1 to 11, characterized by at least one sensor (4) for recording and recording a sound emission and an evaluation device for the signals received by the sensor (s) and forwarded to the evaluation device, wherein the sensor (4) is either arranged permanently in a ship, in particular a tanker, or can be removed and installed and fixed in the ship at a predetermined location.

13. Device according to claim 11 or 12, characterized in that to form a multi-sensor (5) three or more sensors (4) are fixed to each other with predetermined distances.

14. Device according to claim 13, characterized in that at least two multisensors (5) in the ship are either permanently attached or can be fixed in the ship.

15. Device according to one or more of claims 12 to 14, characterized in that the sensor (4) or the sensors (4) or the one or more sensors (5) are provided within a tank space (1) of a tanker.

16. Device according to one or more of claims 12 to 15, characterized in that the sensor (4) or the sensors (4) are intrinsically safe and made of piezoelectric material and the waves thus converted into electrical signals can be evaluated in a computer of the evaluation device are.

17. The device according to claim 16, characterized in that the computer is equipped with a program for determining the frequency response and a classification program based on reference data for different sound sources.