NORLI PETTER (NO)
LIANG ZENG ET AL: "Excitation of Lamb waves over a large frequency-thickness product range for corrosion detection", SMART MATERIALS AND STRUCTURES, IOP PUBLISHING LTD., BRISTOL, GB, vol. 26, no. 9, 9 August 2017 (2017-08-09), pages 95012, XP020319183, ISSN: 0964-1726, [retrieved on 20170809], DOI: 10.1088/1361-665X/AA7774
SEHER MATTHIAS ET AL: "Experimental Studies of the Inspection of Areas With Restricted Access Using A0 Lamb Wave Tomography", IEEE TRANSACTIONS ON ULTRASONICS, FERROELECTRICS AND FREQUENCY CONTROL, IEEE, US, vol. 63, no. 9, 1 September 2016 (2016-09-01), pages 1455 - 1467, XP011622801, ISSN: 0885-3010, [retrieved on 20160914], DOI: 10.1109/TUFFC.2016.2583410
| 1. A method for detecting faults in plates, including the steps of: transmitting an acoustic signal towards the plate (2) from a transmitting transducer (1), receiving the acoustical signal from the plate (2) in a receiving transducer (3), wherein the receiving transducer (3) is mounted at a distance from the transmitting transducer (1), repeating said steps of transmitting and receiving in a number of test points covering at least a part of the plate, c h a r a c t e r i z e d i n the additional steps of: identifying zones of the plate wherein energy levels of the received signals are attenuated compared to other zones of the plate, comparing the energy levels of the A2 and S3 guided Lamb modes in the re ceived signals in said identified zones. 2. A method according to claim 1, where the transmitted signal is a swept pulsetrain, the received signal is filtered into two separate frequency bands represent ing the A2 and S3 guided Lamb modes, respectively, a time window is applied to said filtered signals, where the time window is located at a predetermined time offset after a peak in the signal energy of the unfiltered received signal, and the energy difference between the mean of the A2 and S3 mode signals is determined within said window. 3. A method according to claim 2 where the frequency ranges used to filter out the A2 and S3 modes are scaled to various wall thicknesses by keeping the product fd constant where f is any mentioned frequency and d is the plate thickness. 4. A method according to any of the preceding claims, further including averag ing of received signals' energy within a predefined zone. 5. Use of a method according to the preceding claims for the inspection of oil and gas pipelines, or other pipelines carrying lightweight hydrocarbon prod ucts such as diesel, condensate, or liquefied natural gas. |
Field of the Invention
The present invention relates to a method for detecting faults in plates, such as the walls of pipelines for conducting oil and gas. Background
In the oil and gas industry, there is a need for efficient testing of pipelines. The pipelines are subject to wear from corrosive fluids and sand, and deformation from movements on the seabed or in the ground. Said pipelines are also prone to developing cracks, in particular near the welding seams. Welding seams are inherent weak points due to the changes of the steel structure caused by the welding process. Cracks may develop due to stress caused by temperature or pressure cycling, movements in the ground, coating disbondment, and consequent intrusion of mineral water to the areas under stress. The structural integrity of pipelines may be tested using inspection pigs which travel inside the pipelines measuring the condition of the pipe wall. Acoustical transducers mounted on the pig are used for ultrasonic detecting of corrosion and cracks in the pipeline walls. A problem with present ultrasonic testing methods is that it is difficult to differentiate between indications of corrosion (thinning of the wall) and cracks. Even though corrosion must be considered an important factor affecting the health of the pipeline, cracks are considered to be more crucial for the pipeline, as a crack may develop and eventually cause a fatal breakdown of the pipe wall. Corrosion is handled in special ways, but cracks must be handled immediately. A scan of the pipeline wall may give an indication of the presence of a fault, but the pipeline must be physically engaged in order to determine the nature of the fault. This involves the replacement of a section of the pipeline. Often one then finds that the indicated fault is due to corrosion, which means that it was unnecessary to replace the pipeline section, as the corrosion could have been remedied in a less costly way.
Summary of the I nvention
It is an object of the present invention to devise a method for testing pipelines that may differentiate between faults caused by corrosion and cracks. This is achieved in a method as claimed in the appended claims.
Brief Description of the Drawings
The invention will now be described in detail in reference to the appended drawings, in which:
Fig. 1 is a view of the tool used for obtaining the measurements,
Fig.2 is a diagram showing the energy of the received signal versus time for a clean wall without any fault,
Fig.3 is a corresponding diagram wherein the wall is pitted (corrosion), and Fig.4 is a corresponding diagram in the case of a wall with a large crack.
Detailed Description
Fig. 1 shows the setup used in the present invention. The setup includes a cylindrical tool adapted for translatory movement inside a pipeline with wall 2. A number of transducers are mounted around the body of the tool. The transducers operate in pairs with an acoustical transmitting transducer 1 adapted to emit a pulsed signal towards the wall/plate 2 to be investigated. The signal from the transducer 1 will hit the wall at incident angles close to normal incidence and excite an acoustical signal in the wall. This signal will create waves that will be guided by the wall and that will propagate along the wall (Lamb waves). A part of this signal will leave the wall to be collected by a receiving transducer 3.
The signal emitted by the transmitting transducer 1 is a chirp (swept pulsetrain) covering a frequency range from 400 kHz to 1200 kHz.
This mean signal received from the wall is filtered into specific frequency bands each corresponding to a guided Lamb-mode. The specific modes of interest are the A 2 and S 3 modes, which correspond to the frequency ranges of 425 - 525 kHz and 650 - 750 kHz, respectively, in the present case where the wall thickness is 12.7 mm. It is important to note that the presented numbers for the frequency ranges are illustrative for the implementation of the method in the chosen case of a 12.7- mm thick steel wall. When applying the method to walls of different thicknesses the frequency numbers should be scaled so that the frequency-thickness product is kept constant.
The energy of the signal received along the wall is estimated in each chosen frequency band and low energy zones are identified. Within these low energy zones characteristic of a fault, an area is defined in which the mean signal energy is computed as a function of time.
The analysis of the resulting signals involves first the identification of low energy zones, then the comparison of the mean energies for the A 2 and S 3 modes within a time window to indicate the state of the wall. The time window is located at a fixed time offset after the signal energy in the (total) received signal reaches its peak. The location of this window is chosen to maximize the absolute energy difference between the A 2 and S 3 modes in presence of a fault within a 50 ps time interval.
Fig.2 shows the resulting diagram from an analysis of a zone of the wall without any faults. The diagram shows the received energy of the unfiltered signal and for the A 2 and S 3 modes. The unfiltered signal peaks its energy at about 30 ps, and the comparison window is located between 205 and 255 ps, i.e. the window is located 175 ps behind the energy peak and is 50 ps wide, as indicated with the stippled vertical lines. The diagram shows that the unfiltered signal (marked“All fr” in the figure) received in the comparison window is attenuated about 20 - 25 dB from the peak, and that the energy received in the A 2 band is more attenuated than in the S 3 band. The energy difference is typically about 5 dB.
Fig.3 shows a corresponding diagram from a measurement taken from an area of pitting corresponding to an area of corrosion. In the comparison window, the total signal is attenuated further 4 - 5 dB compared with the signal from the fault-free zone in Fig.2, and with the A 2 signal still about 5 - 8 dB lower than the S 3 signal.
Fig.4 shows a diagram obtained in a zone of the pipeline including a crack. The total signal is attenuated from the fault-free case in Fig.2. However, in contrast with the case of a pitting, the diagram shows that energy of the S 3 mode in this case is lower than that of the A 2 mode in the comparison window. The S 3 mode is substantially more attenuated than the A 2 mode which is the opposite behaviour from what was observed in the two previous cases. Typically, the energy difference will vary between about 3 dB to 1 4 dB, dependent on the size of the crack.
Experience has shown that the size of the energy difference is dependent on the size of the crack ( larger cracks m ean larger energy difference) , and also that clusters of cracks yield a large energy difference.
Thus, by using the inventive m ethod , one m ay identify zones of cracking and corrosion/pitting , and clearly identify the nature of the fault .
Wh ile the description only relates to testing of pipelines, the sam e techn ique m ay be adapted for testing flat plates, although then with another tool setup.