DEVICE AND METHOD FOR THE NON-DESTRUCTIVE TESTING OF A TUBE-SHAPED PRODUCT WITH A COMPLEX SHAPE

Description

[0001] The invention relates to the field of non-destructive testing of metallurgical products, in particular tubular products and more particularly tubular products with variations in internal and / or external diameters.

[0002] Long tubes are widely used in various fields of application. Examples include power generation, where so-called "boiler" tubes are used, oil and gas, where tubes are used for drilling, extraction and transport ("line pipes"), and mechanical engineering, whether in civil engineering or in the automotive and aeronautical sectors.

[0003] To better illustrate the invention, it is described in the context of generally tubular products, particularly tubes, as examples of metallurgical products. However, the invention is intended to be applied more broadly to profiles that exhibit geometric variations in their walls along their principal axis.

[0004] According to one aspect, the invention is also intended to be applied to tubular profiles which have variations in thickness on their circumference, for example hollow structural tubes with square or rectangular cross-section, called MSH profiles, or profiled tubes with hexagonal cross-section, or tubes with circular outer cross-section and hexagonal inner cross-section, or other types of cross-section which result in intentional variations in the local thickness of the tube.

[0005] Like most metallurgical products, tubes are susceptible to manufacturing defects, such as inclusions of material in the steel, cracks on an internal or external surface, or porosity. Generally speaking, any heterogeneity in the steel matrix is ​​considered an imperfection that can compromise the tube's mechanical strength in service.

[0006] This is why metal tubes are inspected immediately after their manufacture, not only to detect any defects, but also, where appropriate, to determine information useful for assessing the danger of these defects, including their size, depth, position, nature or orientation, and the compliance of these tubes with international standards.

[0007] Ultrasonic testing techniques are used. Ultrasonic waves are propagated through the tube being tested, and the resulting echoes are identified as those that cannot be attributed to the tube's geometry. Inclusions or absences of material constitute variations within the wave propagation medium and therefore generate echoes when struck by ultrasonic waves. These variations can be considered imperfections.

[0008] The intensity of the echo produced by an imperfection depends on the angle at which the wave strikes it. For a given direction of propagation of the ultrasonic wave in the tube, imperfections oriented in the corresponding direction, that is to say perpendicular to the direction of propagation, are mainly detected, with a certain tolerance, however, on the order of a few degrees, the amplitude of this tolerance generally being between 2 degrees and about ten degrees depending on the devices chosen.

[0009] A defect is defined as an imperfection that returns an echo with an amplitude greater than a threshold value. This defect is generally associated with an orientation value, which can be deduced from the inspection direction, which is the direction imparted to the ultrasonic wave by the ultrasonic sensor emitting the wave.

[0010] This predefined threshold value is set by calibration. Classically, reference defects, or standard defects, are notches of known position (depth and orientation) and dimensions, most often standardized, made in a sample tube.

[0011] The different types of defects that are most often looked for during inspections are as follows: Surface defects: o Internal or external longitudinal defects. These defects generate, in response to an ultrasonic pulse from a substantially transverse direction (i.e., a pulse located substantially in a plane perpendicular to the tube axis, or in other words, in a cross-section of the tube), an echo with an amplitude exceeding a predefined threshold. o Internal or external transverse defects, also called transverse defects. These defects generate, in response to an ultrasonic pulse generally from a longitudinal direction (i.e., a pulse located substantially in a plane containing the tube axis), an echo with an amplitude exceeding another predefined threshold. o Oblique defects. These defects generate, in response to an ultrasonic pulse generally from a direction diverging with respect to the plane containing the transducer and the tube axis, an echo with an amplitude exceeding another predefined threshold.An oblique defect generates, in response to a shot with an intermediate orientation between longitudinal and transverse, an echo with an amplitude exceeding a predefined threshold. Defects within the walls, generally called "splits," generate, in response to an ultrasonic shot with a generally radial direction, an echo with an amplitude exceeding a predefined threshold.

[0012] In practice, imperfections are not purely longitudinal or transverse, but reflect a more or less significant echo in one or the other of these directions. The orientation of an imperfection can be likened to the orientation of its largest reflective surface.

[0013] The duration of the control depends mainly on the number of shots taken, the time required for the ultrasonic waves to travel through the tube, there and back, the travel time in the possible interface coupling between the sensor and the steel and, to some extent, the time required for processing the signals captured in return.

[0014] To reconcile the requirements of production rates and safety, it has become customary to limit the number of ultrasonic shots and to search, in each tube, only for defects with certain particular orientations.

[0015] Limiting the number of shots also limits the electronic processing of data, and therefore limits the costs of the equipment needed for data processing.

[0016] There are a variety of ultrasonic sensors, generally differentiated by their complexity.

[0017] The first type of transducer is the single-element transducer (or single crystal transducer). This type of sensor has a fixed ultrasonic wave emission direction by construction. For the purposes of implementing the invention, this sensor can be motorized so as to form a directional sensor capable of emitting an ultrasonic wave with a chosen emission orientation θe, i.e. allowing the emission orientation θe to be changed.

[0018] A second type of transducer is the phased array ultrasonic transducer, also known as a linear multi-element transducer. This type of transducer comprises a plurality of electroacoustic transducing elements, in the form of piezoelectric elements, distributed across one active face of the transducer along a principal direction. For example, these piezoelectric elements can be arranged in a line with each other, forming what is sometimes called a "transducer array." Transducers with such an arrangement are said to be "one-dimensional."The transducer elements are excited simultaneously or with time offsets, sequentially, according to a time law, so as to combine the ultrasonic waves thus produced to form a deflected wave beam, possibly focused (focal point in front of the sensor), which allows a tube to be inspected for the existence of defects oriented in a corresponding direction.

[0019] An ultrasonic wave inspection device is also known, comprising a one-dimensional phased array transducer whose elementary transducers are distributed around the tube being inspected. Such a device can detect longitudinal defects and splits, but only in a small section of the tube. One-dimensional phased array sensors are the most commonly used because they are more economical to implement and allow for faster inspection.

[0020] A device for inspecting metallurgical products is also known from WO2014 / 096700. This device comprises an ultrasonic transducer having a plurality of elementary transducers (29) that can be operated independently of each other and are arranged in a two-dimensional pattern. This type of transducer makes it possible to detect defects at any angle using a single sensor, notably allowing for unrestricted orientation of the shot relative to a principal direction of the sensor.

[0021] We also know of EMAT sensors capable of generating ultrasonic waves by electromagnetic means. These sensors generally eliminate the need for a coupling device between the sensor and the element being inspected.

[0022] A non-destructive testing system for metallurgical products is also known from WO 2003 / 50527, in which a one-dimensional phased array transducer is used. Each transducer element is excited once, and then a processing circuit analyzes the overall response of the tube to this single emission, which is called a "shot" in the technique. From a shot taken along the transverse direction of the tube, it is possible to determine the presence not only of defects oriented perpendicular to this direction, but also of defects exhibiting an inclination relative to this perpendicular direction of between ±10° and ±10°.

[0023] In the remainder of this text, an ultrasonic transducer may be referred to interchangeably as a sensor, probe, or transducer, terms well known to those skilled in the art.

[0024] In practice, on a tubular product test bench, three transducers are often used: two transducers dedicated to detecting longitudinally oriented defects, allowing inspection in both directions of travel, or defects with an inclination relative to this longitudinal direction of ±20°, and a third sensor to detect defects oriented transversely to the tubular product. A fourth sensor is commonly used to check for the presence of splits and to measure the wall thickness of the tubular product. It is possible to have a fifth sensor dedicated to detecting transverse defects, in addition to the aforementioned third sensor, to perform detection in both directions of the longitudinal travel of a tubular product.

[0025] FR 2 917 832 A1 is known a system for the non-destructive testing of an axle shaft with variations in internal and external diameter comprising several ultrasonic transducers which can be moved longitudinally and transversely over different portions of the axle in order to perform ultrasonic detection of defects in these inspected portions.

[0026] FR 3 000 212 also discloses an ultrasonic testing device capable of inspecting a metallurgical product and detecting defects in any orientation. This device uses a single sensor, excited a limited number of times, which allows for a high testing rate.

[0027] According to some known embodiments, the sensors are fixed and the tube is animated by a helical movement.

[0028] According to other known embodiments, the ultrasonic sensors or probes are driven in rotation at a speed of a few thousand revolutions per minute, around a tube moving at a linear speed that can reach up to about 1 meter per second.

[0029] In other embodiments known, for example, in FR 2 796 153, a sensor is used consisting of multiple ultrasonic transducer elements surrounding the tube. The electronics allow the origin of the ultrasonic beam to be rotated around the tube by switching the groups of excited elements, and consequently replace the mechanical rotation of the sensors, described above, with an electronic scanning.

[0030] These three types of installations, all well known to those skilled in the art, are respectively called: rotating head installations, rotating tube installations, and multi-element enclosed sensor installations. When using sensors operating in electronic scanning mode, the relative rotation between the tube and the sensors is virtual. As used here, the expression "relative rotational / translational motion between the tube and the transducer arrangement" covers the case where the relative rotation is virtual.

[0031] All these techniques are used today on tubular products with a so-called constant cross-section. A tubular product with a constant cross-section is defined as a tubular product whose wall thickness is constant, or at least whose wall thickness has a constant nominal value and allows for slight dimensional variation inherent to the manufacturing processes of these tubes—that is, unintentional variations that occur within tolerance values ​​defined by standards. For example, the dimensional tolerances of API tubes are on the order of approximately -12.5% ​​to +12.5% ​​of the nominal wall thickness for the most common nominal diameters and wall thicknesses.

[0032] However, the techniques for obtaining tubular products have recently evolved and now make it possible to obtain steel tubes, possibly of great lengths and large diameters, with complex shapes, i.e. with intentional variations in thickness and / or intentional variations in their internal and / or external diameters, leading in particular to variations greater than tolerances such as those previously given of -12.5% ​​to +10% of the nominal thickness according to the API standard.

[0033] However, ultrasonic testing devices for detecting defects in tubular products are not suitable for detecting defects in tubular products with such complex shapes. In particular, automated ultrasonic testing devices designed to inspect tubular products immediately after their manufacture at an industrial rate are especially unsuitable for inspecting tubular products with complex shapes.

[0034] Therefore, there is a need to better detect potential defects within these complex-shaped tubular products.

[0035] The applicant has developed non-destructive testing techniques for tubular steel products, potentially of considerable length, generally up to 20 meters, and large diameters, up to 30 inches (approximately 77 cm), with complex shapes, i.e., variations in outside and / or inside diameters. For example, dimensional variations can be achieved by regrooving the tube. These length and diameter values ​​are not limiting to the applications of the present invention.

[0036] These tubes can have different types of sections: a section with constant thickness and constant external and internal diameters, a section with constant internal diameter and longitudinally varying external diameter, resulting in an increase or reduction in tube wall thickness, a section with constant external diameter and longitudinally varying internal diameter, resulting in an increase or reduction in tube wall thickness, a section with external and internal diameters each varying, with or without longitudinal variation in tube thickness.

[0037] Current automated inspection benches are suitable for detecting defects in tubes with constant external and internal diameters, i.e., tubes with only one section of constant wall thickness and constant external and internal diameters. However, the applicant has observed that these known benches are not suitable for the complete inspection of tubular products with sections of different types, whose external and / or internal diameters vary along their axis. Indeed, only a main section with constant external and internal diameters can be inspected with a state-of-the-art automated inspection bench.

[0038] Indeed, the applicant observed that defect detection using the principle of emitting and receiving an ultrasonic beam is based on a representative measurement of the ultrasonic beam's path, specifically the portion of the incident ultrasonic beam that arrives (or returns) at the sensor after being reflected by a defect. This ultrasonic beam path has the following important characteristics: the path distance and trajectory orientation, and the beam amplitude. These characteristics are fixed for each type of defect and each model of tubular product with a so-called constant cross-section. A tube model generally refers to a set of data such as the nominal external diameter, the nominal internal diameter or thickness, the type of steel used, etc.

[0039] The applicant has observed that as soon as a tubular product has a non-constant cross-section, i.e. with sections of different types such as those described above, detection by known devices is inoperative: the echo of the ultrasonic shot measured on the transducer may have an amplitude too low to be detected, or may not be detected at all.

[0040] The applicant observed that variations in the cross-section of a tubular product introduce uncertainties and deviations in the path of an ultrasonic beam. The present invention compensates for the influence of these cross-sectional variations in a tubular product, enabling optimized industrial inspection and improving defect detection on tubular products with non-constant cross-sections.

[0041] The present invention therefore makes it possible to improve the detectability of a defect despite a variation in the dimensions and / or shape of the walls of a tubular product. It is then referred to as a "complex tubular product".

[0042] According to one aspect of the invention, the device and method according to the invention provide a solution by adapting the parameters of the emission setpoint of an ultrasonic wave from a transducer according to the location of the transducer, in particular the longitudinal position of the sensor, by varying an emission orientation of the transducer, or by varying a gain at the emission of the transducer.

[0043] According to another aspect of the invention, the device and method according to the invention provide a solution by adapting the reception parameters of an ultrasonic signal according to the longitudinal location of the transducer, for example the gain in reception or the position of a time window for acquiring the echo signal.

[0044] According to yet another aspect of the invention, the device and method according to the invention provide a solution by adapting the emission setpoint parameters of an ultrasonic wave from a transducer and / or the reception parameters of an ultrasonic signal according to the circumferential location of the transducer, such as the emission orientation, the emission or reception gain or the position of a time window for acquiring the echo signal. THE figures 1 et 2 They show a classic metallurgical product testing installation, viewed from the front and side respectively; the figure 3 illustrates a principle of insonification of a tube wall by an ultrasonic sensor; the figure 4 illustrates a schematic cross-sectional view of the wall of a complex-shaped tube with varying external and internal diameters and illustrates the firing of an ultrasonic transducer at a transverse defect; the figure 5 illustrates an A-scan of an ultrasonic shot and its echo over time; the figures 6a-c show a cross-sectional detail of a sample tube with a variant cross-section containing longitudinal defects and two comparative graphs of the ultrasonic firing responses of a state-of-the-art device and an embodiment of the invention, according to aspects of the signal-to-noise ratio of the ultrasonic echoes and the amplitude of the ultrasonic echoes; figures 7a-e show a cross-sectional detail of a sample tube with a variant cross-section containing transverse defects and comparative graphs of the quality of the ultrasonic shot responses to transverse defects between a state-of-the-art device and a device according to an embodiment of the invention, according to aspects of the signal-to-noise ratio of the ultrasonic echoes and the amplitude of the ultrasonic echoes, and this according to two directions of inspection; the figures 8a-c show a cross-sectional detail of a sample tube with a variant cross-section containing flat-bottomed hole-type defects; as well as comparative graphs of the quality of the ultrasonic shot responses to flat-bottomed defects between a state-of-the-art device and a device according to an embodiment of the invention, according to aspects of the signal-to-noise ratio of the ultrasonic echoes, and then of the amplitude of the ultrasonic echoes; the figure 9 represents an example of implementing an embodiment of the invention on a sample with internal longitudinal notch-type defects and the corresponding C-scan obtained; the figure 10 represents in a manner analogous to the figure 9 another example of the implementation of an embodiment of the invention and the corresponding results obtained on the detection of internal transverse defects in a detection direction; the figure 11 represents in a manner analogous to the figure 9 another example of the implementation of an embodiment of the invention and the corresponding results obtained on the detection of internal transverse defects in a detection direction; the figure 12 is a schematic diagram of an acquisition channel and part of the processing electronics according to one embodiment of the invention. figure 13 is a schematic diagram of an acquisition channel and part of the processing electronics according to another embodiment of the invention.

[0045] The drawings and appendices contain elements of a definite nature. They can therefore be used not only to describe the invention but also to define it, where applicable.

[0046] The invention will be described hereafter primarily in terms of a variation of inspection parameters as a function of the longitudinal position of the transducer, corresponding to a preferred embodiment of the invention. However, the invention is also intended to apply to a variation of inspection parameters as a function of the circumferential position of the transducer, whether or not in combination with a variation of these parameters as a function of the longitudinal position of the sensor.

[0047] The invention relates to an automatic non-destructive testing device for detecting defects in a complex tubular product (3), comprising: at least one ultrasonic transducer (5) having a position defined by a longitudinal position (L) and a circumferential position (A) along the complex tubular product (3) and arranged to emit an ultrasonic beam (Em) having an emission orientation θe i (L, A); control and processing electronics (6) comprising a circuit for exciting the sensor and capturing return signals and at least one amplification stage (21,31) with a gain (G i (L ; A)), a time filter module (24) configured to apply a time filter (FT i (L ; A)) to an echo signal (Dv, Ds), and the control and processing electronics (6) is configured to define at least one ultrasonic firing parameter (Vi) as a function of the longitudinal (L) and / or circumferential (A) position of the ultrasonic transducer so as to detect defects in the tube wall, said at least one parameter being chosen from the firing emission orientation (θe i (L; A)), the gain (G i (L; A)) or the position of the time filter (FT i (THERE)).

[0048] According to one aspect of the invention, the control and processing electronics (6) can be configured to define at least two ultrasonic firing parameters (Vi) as a function of the circumferential position (A) of at least one ultrasonic transducer (5) so as to capture defects in the wall of the tube, said at least one parameter being chosen from the emission orientation of the firing (θe i (L ; A)), the gain (G i (L ; A)) or the position of the time filter (FT i (L ; A)).

[0049] Alternatively, the control and processing electronics (6) can be configured to define at least two ultrasonic firing parameters (Vi) as a function of the longitudinal position (L) of at least one ultrasonic transducer (5) so as to detect defects in the tube wall, said at least one parameter being chosen from the firing emission orientation (θe i (L ;A)), the gain (G i (L ; A)) or the position of the time filter (FT i (L ; A)).

[0050] According to one aspect, the control and processing electronics (6) is configured to define the emission orientation of the shot (θe i (L ; A)), the gain (G i (L ; A)) and the position of the time filter (FT i (L ; A)) of ultrasonic shots (Vi) as a function of the longitudinal position (L) of at least one ultrasonic transducer (5). The control and processing electronics (6) can also be configured to define at least one parameter chosen from among the emission orientation of the shot (θe i (L ; A)), the gain (G i (L ; A)) and the position of the time filter (FT i (L ; A)) of ultrasonic shots (Vi) as a function of the circumferential position (A) of at least one ultrasonic transducer (5).

[0051] In addition, the device may include at least one position sensor (7a) for determining the longitudinal position (L) of at least one ultrasonic transducer 5 relative to the complex tubular product (3). Alternatively, the device may include at least one position sensor (7a) for determining the longitudinal position (L) and the circumferential position (A) of at least one ultrasonic transducer 5 relative to the complex tubular product (3).

[0052] At least one position sensor (7a) may be selected from an incremental encoder, a rack and pinion encoder, a linear encoder, a wire encoder, a laser velocimeter, an encoder wheel or an encoder wheel ratio.

[0053] Alternatively, the device may include at least one timer (7b) to determine the relative longitudinal (L) and circumferential (A) position of the ultrasonic transducer 5.

[0054] According to another aspect, at least one amplification stage (21,31) may be an emitting amplification stage (21) having an emitting gain (Ge i (L ; A)) and the control and processing electronics (6) is configured to vary said emitting gain (Ge i (L ; A)) as a function of the longitudinal position (L) of the ultrasonic transducer (5).

[0055] In one variant, at least one amplification stage (21,31) may be a receive amplification stage (31) having a receive gain (Gr i (L ; A)) and the control and processing electronics (6) is configured to vary said receive gain (Gr i (L ; A)) as a function of the longitudinal position (L) of the ultrasonic transducer (5).

[0056] In another variant the device may include a transmit amplification stage (21) having a transmit gain (Ge i (L ; A)) and a receive amplification stage (31) having a receive gain (Gr i (L ; A)) and in which the control and processing electronics (6) is configured to vary the transmit gain (Gr i (L ; A)) or the receive gain (Gr i (L ; A)) as a function of the longitudinal position (L) of the ultrasonic transducer (5).

[0057] According to another aspect, the control and processing electronics (6) may include a parametric memory module (MEMp) capable of storing data in the form of an association between at least one longitudinal position (L) of at least one ultrasonic transducer (5) and at least one data set corresponding to emission orientation parameters of the shot (θe i (L ; A)), gain (G i (L ; A)) and / or position of the time filter (FT i (L ; A)).

[0058] Also, the control and processing electronics (6) may include a parametric memory module (MEMp) capable of storing data in the form of an association between at least one circumferential position (A) of at least one ultrasonic transducer (5) and at least one data set corresponding to emission orientation parameters of the shot (θe i (L ; A)), gain (G i (L ; A)) and / or position of the time filter (FT i (L ; A)).

[0059] In addition, the control and processing electronics (6) may include a parametric memory module (MEMp) capable of storing data in the form of an association between pairs of longitudinal and circumferential position (L; A) of the ultrasonic transducer 5 and at least one set of data corresponding to emission orientation parameters of the shot (θe i (L; A)), gain (G i (L; A)) and position of the time filter (FT i (L; A)).

[0060] According to one aspect of the invention, the parametric memory module (MEMp) can include at least one dataset corresponding to gain parameters (G i (L ; A)) in the form of receive gain parameters (Ge i (L ; A)) and transmit gain parameters (Gr i (L ; A)).

[0061] According to another aspect of the invention, the control and processing electronics (6) can be configured to emit several ultrasonic shots (Vi) for a position of the ultrasonic transducer (5), the ultrasonic shots (Vi) having emission angles θe j (L) between a minimum position orientation angle θe mini (L) and a maximum position orientation angle θe max i (L).

[0062] Thus, the control and processing electronics (6) can be arranged to perform from 2 to 8 ultrasonic shots (Vi) for a position of at least one ultrasonic transducer (5).

[0063] According to one aspect of the invention, at least one ultrasonic transducer (5) can be an ultrasonic transducer bar.

[0064] According to another aspect of the invention, at least one ultrasonic transducer (5) can be a phase-array sensor.

[0065] The invention also relates to an automatic method for inspecting tubular products with varying external or internal diameters, comprising the following steps: a. at least one ultrasonic transducer (5) is positioned at a first position (P1) b. a first ultrasonic shot (Vi) is performed by emitting an ultrasonic beam (Em) having a first orientation θe i (P1), and a first amplification in emission with a first gain in emission Ge i (P1), c. an echo returned by the complex tubular product (3) is captured and the captured echo is transformed into a received signal to which a first gain in reception Gr i (P1) is applied d. A part of the signal is isolated in a first time window (FT i (P1)) e.A second ultrasonic shot is performed by repeating steps a to d at a second position (P2), with second ultrasonic shot parameters including a second orientation θe i (P2), a second transmit gain Ge i (P2), a second receive gain Gr i (P2), a second time window (FT i (P2)) and at least one of the second ultrasonic shot parameters among the second orientation θe i (P2), the second transmit gain Ge i (P2), the second receive gain Gr i (P2), the second time window (FT i (P2)) is different from the first orientation θe i (P1), the first transmit gain Ge i (P1), the first receive gain Gr i (P1) or the first time window (FT i (P1)).

[0066] According to one aspect of the process, the first position (P1) comprises a first longitudinal position (L1) and a first circumferential position (A1) and step e) is replaced by step f) in which a second ultrasonic shot is performed by repeating steps a) to d) at a second longitudinal position (L2), with second ultrasonic shot parameters comprising a second orientation θe i (L2), a second transmit gain Ge i (L2), a second receive gain Gr i (L2), a second time window (FT i (L2)) and at least one of the second ultrasonic shot parameters among the second orientation θe i (L2), the second transmit gain Ge i (L2), the second receive gain Gr i (L2), the second time window (FT i (L2)) is different from the first orientation θe i (P1), the first transmit gain Ge i (P1), the first receive gain Gr i (P1) or the first time window (FT i (P1)).

[0067] It is understood that when the electronics are configured to define at least one ultrasonic firing parameter Vi as a function of the longitudinal position L of the ultrasonic emission means in order to capture defects in the wall of the tube, said at least one parameter being chosen from the emission orientation of the firing θe i (L), the gain G i (L) or the position of the time filter FT i (L);then the electronics is configured to define at least one ultrasonic firing parameter Vi at at least one first longitudinal position L1 of the ultrasonic emission means so as to detect defects in the wall of the tube, said at least one parameter being chosen from the firing orientation θe i (L), the gain G i (L) or the position of the time filter FT i (L), and in that the electronics is also configured to define at at least one second longitudinal position L2 at least one parameter chosen from the firing orientation θe i (L), the gain G i (L) or the position of the time filter FT i (L) different from the parameter(s) chosen at said at least first longitudinal position L1.;

[0068] We are referring to figures 1 et 2 .

[0069] An ultrasonic wave control installation includes a bench 1 supporting a complex tubular product 3 of X-axis to be controlled and an ultrasonic transducer 5, positioned near the peripheral surface of the complex tubular product 3, and connected to a control and processing electronics 6 comprising an electronic circuit for exciting the sensor.

[0070] The complex tubular product 3 can be animated with a helical movement so that the control device inspects the entire complex tubular product 3.

[0071] Alternatively, the complex tubular product 3 can rotate only relative to the bench 1, and the transducer 5 slides along the longitudinal direction of the bench 1, either synchronously with the movement of the complex tubular product 3 or sequentially. The transducer 5 can be mounted on a carriage that moves relative to the bench 1. In yet another alternative, the transducer 5 can rotate around the complex tubular product 3 while the latter is translated relative to the bench 1, either synchronously or sequentially. This generally results in two types of inspection paths: a first path called a helical path, or a second path called an incremental path by unit segments. The incremental path by segments allows the sensor to scan a circumference of the tubular product before advancing one increment in the longitudinal direction to perform another circumferential scan.This type of trajectory can have the advantage of simplifying the electronics, and for example of minimizing changes in inspection parameters when these depend on the longitudinal position of the transducer.

[0072] These trajectories allow control of the entire complex tubular product 3 using a sensor with a reduced size compared to the circumference of the complex tubular product 3. Alternatively, a larger number of sensors could be used, arranged in a ring around the complex tubular product 3, and a firing sequence could be used to ensure coverage when the complex tubular product 3 slides relative to the transducer 5.

[0073] A coupling medium, or "coupling agent" in the technique, can be interposed between the transducer 5 and the peripheral surface of the complex tubular product 3, for example in the form of a gel or water. Alternatively, the installation may include a container filled with water, or any other liquid coupling medium, in which the complex tubular product 3 and the transducer 5 are immersed. In another variant, the installation may include a water jet device, the water flow then constituting the coupling medium.

[0074] The installation is designed to inspect the complex tubular product 3 to check for defects with different orientations. The direction of an inspection corresponds to the orientation of the defects being sought within the complex tubular product 3.

[0075] In order to be able to distinguish, in the responses of the complex tubular product 3, echoes resulting from defects from those resulting from imperfections, the control installation must be calibrated for each of the inspection directions.

[0076] We now refer to the figure 3 representing a tubular product with a constant cross-section of axis X equipped with a transverse defect or internal through defect Dti, and an ultrasonic transducer 5 performing an ultrasonic shot whose beam is reflected on the internal through defect Dti and returns to the transducer 5.

[0077] Current detection techniques use ultrasonic transducers 5 positioned near a tubular product with a constant cross-section. These sensors are indirectly coupled to the tube via a liquid coupling agent, usually water. The transducer 5 has a principal direction substantially orthogonal to the X-axis of the complex tubular product 3 and therefore to the outer wall of the tubular product with a constant cross-section.

[0078] In general, ultrasonic pulses propagate through the thickness of the tubular product to the inner surface of the product wall and make multiple round trips between the inner and outer surfaces of the tubular product. In the absence of defects, the beam is reflected several times within the complex tubular product, and the absorption coefficient of the metal helps to dampen the ultrasonic wave.

[0079] To perform an ultrasonic pulse, the transducer 5 is controlled to emit an ultrasonic wave into the coupling medium, for example, water with refractive index nwater, at an incident angle θi, for example, approximately 17° to the normal of the water / steel interface surface at the point where the wave encounters the interface. This wave propagates through the coupling medium to the surface of the tubular product, and a refracted wave propagates through the material of the tubular product, for example, steel with refractive index nsteel, at a refraction angle θr, for example, approximately 40°. The relationship between the incident angle θi and the refraction angle θr is expressed by Snell's law. sin θi / V eau = sin θr / V acier with Vwater the speed of the ultrasonic wave in water and Vsteel the speed of the wave in steel.

[0080] This angle of refraction, or insonification angle, of approximately 40° in the tube material is a very effective angle for detecting internal or external defects located on the inner and outer surfaces. Indeed, a defect angle θd of approximately 40° generally allows an ultrasonic wave reflected by the defect, or echo, to follow a reverse path and return to the sensor.

[0081] We understand that the case represented in figure 3 This corresponds to the classic case of inspecting a tube with a constant cross-section and detecting an internal transverse defect Dti. However, when the tube to be inspected includes sections of different dimensions, and sections where the internal and / or external diameters of the tube vary, the internal and / or external surfaces may have slopes, and it is therefore more difficult to obtain the desired defect angle θd so that the echo of the ultrasonic shot returns to the transducer 5.

[0082] This is illustrated in figure 4 Figure 5 represents a longitudinal cross-sectional view of a portion of a complex tubular product 3 with a variable cross-section. The transducer 5 is positioned longitudinally along the X-axis so as to emit an ultrasonic wave penetrating the complex tubular product 3 at a point in which the slope of the outer wall of the tube makes a non-zero angle α with respect to the longitudinal axis X of the tubular product: this is a point in which the outer diameter of the tube varies. The normal to the outer surface of the tube at the point of penetration of the ultrasonic beam into the tube wall therefore makes an additional angle α. This angle α must therefore be taken into account when obtaining an ultrasonic beam encountering an internal transverse defect Dti with a defect angle θd of 40°.

[0083] There figure 4 This presents an additional difficulty because the ultrasonic wave encounters an internal transverse defect Dti located on a section of a complex tubular product 3, in which the inner wall has a slope with a non-zero angle β relative to the longitudinal axis of the tube. In this case, the angle over the defect θd must be equal to the angle of refraction θr plus the respective slope angles of the inner and outer walls α and β. For example, it is then necessary to adjust the emission orientation θe so that the angle θd = θr + α + β is approximately 40° in order for the ultrasonic beam reflected by the transverse defect to return to the transducer 5.

[0084] In the field of ultrasonic non-destructive testing, the following terminology is often used: "Scan" refers to a series of relative tube / sensor positions; "increment" refers to the scan step (inversely proportional to the recurrence frequency or ultrasonic firing frequency); "Ascan" refers to the graph of the electrical voltage measured across the terminals of an ultrasonic transducer, with the time of flight on the x-axis and a representation of the electrical voltage, also called ultrasonic amplitude, on the y-axis; "Bscan" refers to an image relative to a given value of the increment, with the scan corresponding to the ultrasonic firing on the x-axis, the time of flight on the y-axis, and at each point the ultrasonic amplitude converted into a gradient of gray or into colors (electronic scanning for a phased array sensor, mechanical scanning for a single-element sensor); "Echodynamic" generally refers to the graph of a curve representing the maximum amplitude received as a function of the incremental position of the transducer.For example, the shot number when there is one shot per transducer position, "Cscan" designates an image with the equivalent position in a plane space of the ultrasonic wave's firing point on the x and y axes, and representing, converted into a gradient of gray or colors, the maximum ultrasonic amplitude for that shot recorded in the considered time selector of the Ascan ("image amplitude"). In the case of a tube, a point on the x-axis of Cscan corresponds to a position along the length of the tube and a point on the y-axis corresponds to a position on the circumference of the tube. In the case of a flat product, a point on the x-axis of Cscan corresponds to a position along the length of the flat product and a point on the y-axis corresponds to a position along the width of the flat product.

[0085] There figure 5 This schematically represents the temporal shape of return signals, known as A-scan. Such a signal comprises the pulses Em of the emitted beam and the pulses of the return echoes. The A-scan signal may consist of a series of pulses Em, followed by pulses Int of interface echoes between the water and the outer surface of the tubular product. Then, in the presence of defects on the inner and / or outer surface of the tube, it may include echo signals related to defects on the inner surface Dv and echo signals related to defects on the outer surface Ds. In practice, the interface echo Int is predominant over an echo Ds due to a defect on the outer surface of the tubular product and masks this echo Ds. This is why the echo Ds of a defect on the outer surface is generally detected on the beam reflected by the inner surface of the tubular product.

[0086] In the presence of a defect on the inner surface of the tubular product, the maximum intensity of an echo Dv is detected if the intensity of the echo exceeds a threshold S1 in a time window Fe1.

[0087] In the presence of a defect on the outer surface of the tubular product, the maximum intensity of an echo Ds is detected if the intensity of the echo exceeds a threshold S2 in a time window Fe2.

[0088] The echodynamic curve is therefore the representation of the maximum amplitude of the received signal within a control window over time for each shot fired. Alternatively, the echodynamic curve could represent the maximum amplitude of the received signal as a function of the longitudinal position of the tube.

[0089] There figure 5 illustrates the principle of a detection gate whose purpose is to select a portion of the echo signals in order to possibly identify an imperfection.

[0090] Each electronic channel includes a time filter FT (e.g., a sample-and-hold circuit) connected to the transducer element to isolate successive time windows, which may present an echo relative to a desired fault (e.g., an internal or external fault).

[0091] The device according to the invention may include electronics with a time filter module 24 configured to apply at least one time filter FT i (L ; A) in order to isolate in a corresponding period Tr i (L ; A) a time window Fe i (L ; A), in which echoes Dv and / or Ds representative of the presence of defects are likely to be present.

[0092] The time position and width of a Fe i (L ; A) window depend on the speed of propagation of ultrasound in the metal and the speed of propagation in the coupling height, e.g. water height, the firing period Tr, the outside diameter and the thickness of the metal tube.

[0093] According to the invention, the positions and widths of the time windows Fe i (L; A) can be made dependent on the longitudinal position (L) of a transducer, or on the circumferential position (A) of a sensor, or the combination (L; A) of the longitudinal and circumferential positions of a transducer. Indeed, since the tube varies in external and / or internal diameters, the path of an ultrasonic wave can differ depending on the longitudinal position (L) of the transducer 5. The main parameters that vary are: the distance traveled in the coupling material, the distance traveled in the steel of the tubular product, the distance of the steel / coupling material or steel / air interfaces from the transducer 5, and the orientations of the interface surfaces relative to the sensor.For example, in one section of the tube, the path length may be shorter than the path length of the wave in another section where the tube thickness is increased. It is therefore advantageous to adjust the position and width of the time windows Fei(L) according to the sensor's position, in order to avoid using wide, resource-intensive time windows or to reduce false detections. Similarly, in a tube with a variable thickness along its circumference, the ultrasonic wave path will be modified by this thickness variation as well as by the slope(s) induced on the outer and inner surfaces.

[0094] According to another aspect, the device according to the invention may include one or more ultrasonic transducers 5. The ultrasonic transducers 5 comprise transducer elements capable of emitting or receiving ultrasound.

[0095] The device according to the invention may include, for an ultrasonic transducer 5, firing electronics (6) enabling multiple firings for a given position of the transducer 5. Said firing electronics may include common components used for each firing, and / or exclusive components reserved for each firing performed at a given position. In other words, the firing electronics may include a channel common to the different firings or a channel dedicated to each firing. In the following description, we will refer to an acquisition channel Vi associated with firing number i, regardless of the structure of the electronic components used. Thus, a common channel can successively perform the n firings of channels Vi for i varying from 1 to n or there may be n dedicated lanes to carry out the nshots. For example, the electronics can be configured to perform from one to eight shots per position of transducer 5. Preferably from two to six shots.

[0096] Preferably, the ability to perform multiple ultrasonic pulses for a given position of the transducer 5 allows for the application of several emission orientations θei(L; A) to these ultrasonic pulses, ranging from a minimum emission angle θemin(L; A) to a maximum emission angle θemax(L; A). The ability to perform multiple ultrasonic pulses for the same position of the ultrasonic transducer (5) with different emission orientations makes it possible to compensate for unintentional variations in the object being inspected. This will be illustrated in the examples.

[0097] The acquisition channels V i are therefore configured to perform a series of ultrasonic shots for a given position of the transducer 5.

[0098] There figure 12 is a schematic diagram of the control and processing electronics 6, associated by an electronic circuit with an ultrasonic transducer 5 for non-destructive testing in an example of an installation capable of implementing the invention.

[0099] The purpose of this diagram is to better show certain specific features of the invention; this view is therefore simplified and is not specific to any particular type of sensor; however, a person skilled in the art will be able to adapt this diagram according to the type of sensor used in the device.

[0100] Route V of the figure 12 includes a pulse generator 20 controlling the transmitting transducer elements.

[0101] The pulse generator 20 can be connected to a transmission amplification stage 21 whose function is to amplify the pulse signal by applying a transmission gain Ge i (L). This amplification stage 21 amplifies the electrical signal generating the ultrasonic firing.

[0102] This amplification stage 21 can be configured to adapt the emission gain Ge i (L) according to the position of the sensor, in particular the longitudinal position L of the sensor 5, and for this reason, the amplification stage 21 can be connected to a parametric memory MEMp containing amplification values ​​in relation to a longitudinal position (L) of the sensor 5.

[0103] The transmitting amplification stage 21 can be connected to a directional stage 22 configured to apply a direction θe i (L) to the ultrasonic beam. Preferably, this stage applies a timing law to the excitation of the unit transducers of the ultrasonic sensor 5. Alternatively, particularly when the sensor 5 is of the single-element type, this stage controls a sensor orientation module for the sensor 5, for example in the form of a motorized sensor support platform.

[0104] The transmission amplification stage 21 and the directional stage 22 are connected to the transmission transducer elements E and are configured to allow the transmission of an ultrasonic pulse with a power corresponding to the applied transmission gain Ge i (L) and a direction corresponding to the chosen transmission orientation θe i (L). The directional stage 22 can be connected to a parametric memory MEMp containing orientation values ​​related to a longitudinal position L of the sensor 5.

[0105] Thus, the emission amplification stage 21 and the directional stage 22 can be connected to the parametric memory module MEMp containing the orientation parameters θe i (L) and emission gain Ge i (L) functions of the longitudinal position of the sensor and the type of fault to be characterized.

[0106] The positioning module 23 is configured to return the sensor position to the parametric memory module as a longitudinal position (L) and a circumferential position (A). The positioning module 23 includes positioning electronics and at least one position sensor 7a (not shown in the diagram). figure 12 ). The positioning module 23 tells the parametric memory module MEMp what the values ​​of the active parameters in channel Vi are for a shot aimed at detecting defects of a certain type as a function of the position of sensor 5.

[0107] There figure 13 Figure 6 represents the schematic diagram of the control and processing electronics of the embodiment of the invention capable of adapting the ultrasonic firing parameters according to both the longitudinal position L and the circumferential position A of the sensor 5. The transmitting amplification stages 21 and receiving amplification stages 31 can be configured to apply a transmitting gain Ge i (L; A) or a receiving gain Gr i (L; A) respectively, depending on the longitudinal position L and the circumferential position A of the sensor 5. Similarly, the directional stage 22 is configured to apply an orientation θe i (L; A) to the ultrasonic firing beam, and the time filter FT i (L, A) is configured to apply a time window depending on the longitudinal and circumferential position of the sensor 5.In this embodiment, at least one of these parameters varies according to the longitudinal position L of sensor 5, and at least one of these parameters varies according to the circumferential position A of sensor 5. The parameter memory MEMp is then adapted to contain the appropriate parameters. The positioning module 23 is then configured to send the longitudinal (L) and circumferential (A) positions of the sensor back to the parameter memory module.

[0108] The position sensor 7a can be an incremental encoder, a rack and pinion encoder, a linear encoder, a wire encoder, a laser velocimeter, an encoder wheel or an encoder wheel ratio.

[0109] Alternatively, the positioning module may include a timer 7b instead of a position sensor (7a). This alternative is possible because automated test benches are equipped with means for relative tube movement with respect to the transducers, allowing a repeatable and predetermined relative trajectory to be established over time. However, the timer 7b may be less accurate than the position sensor 7a.

[0110] As explained previously, in the first variant, the position of sensor 5 corresponds to its longitudinal position (L). In the second variant, the position of sensor 5 corresponds to both its longitudinal (L) and circumferential (A) positions. The positioning module 23 is then configured to send the longitudinal (L) and circumferential (A) positions of the sensor to the parametric memory module. In the third variant, the position of sensor 5 corresponds to its circumferential (A) position. The positioning module 23 is then configured to send the circumferential (A) position of the sensor to the parametric memory module.

[0111] The transmitting transducer E i of sensor 5 can emit a directed ultrasonic wave beam.

[0112] The Vi channel includes a receive transducer Ri, which may be the same transducer as the transmit transducer Ei or a different transducer. A receive transducer Ri can receive any echoes of the transmitted signals and convert them into a corresponding electrical signal. The receive transducer Ri can be connected to a receive amplification stage 31, whose function is to amplify the electrical signal received by the receive transducer Ri.

[0113] The receiver amplification stage 31 can be configured to apply amplification to the received signal with a receive gain of channel Vi, denoted Gri(L), which is chosen according to the longitudinal (L) and / or circumferential (A) position of the transducer 5 and the nature of the imperfection being sought. The ability to modulate the gain of this receiver amplification stage improves fault detection, similarly to the configurable amplification in the transmit mode. When this stage is configured with analog amplification, the advantage is that it amplifies the received signal while limiting, to some extent, the amplification of received noise. When this stage is configured with digital amplification, the signal is amplified in the receive mode, but the disadvantage is that it amplifies noise more than analog amplification.

[0114] The receiving transducers Ri can also be connected to a time-domain filter module 24 configured to apply one or more time-domain filters FTi(L). Each time-domain filter FTi(L) isolates a time window in which the echo from the ultrasonic shot is likely to be present. The function of a time-domain filter FTi(L) is therefore to select the portion of the received signal corresponding to the time window in which an echo Dv, Ds from an ultrasonic shot deflected by a typical defect is likely to return to the receiving transducers R, and to specifically process the signal over a chosen time period. This reduces the memory and processing power requirements of the electronics and also prevents the measurement of an echo that does not correspond to the expected echo from an ultrasonic shot performed to detect a specific type of defect, for example, a secondary echo.

[0115] The time filter module 24 can be linked to the parametric memory module MEMp containing the position parameters of the time filters as a function of the longitudinal position of the sensor (L) and the type of fault to be characterized, and possibly in a variant both of the circumferential position (A) of the sensor and its longitudinal position (L).

[0116] The MEMp memory can be configured to hold data relating to the positions and widths of the Fe i (L) windows as a function of the sensor position (L). Correspondingly, the time filter module 24 includes time filters FT i (L) configured to modify the positions and widths of the Fe i (L) time windows for each channel V i acquisition.

[0117] In one variant, the detection gates have the same duration or length for the same type of fault. The representative quantity is then the opening time of the detection gate, or its starting position, which is generally implemented as a delay relative to the shot fired.

[0118] In a second variant, the start and end positions of the detection gate are modified, thus generating detection gates of variable length. In this variant, two representative quantities are therefore used to configure the FT i (L) values.

[0119] The time filter module 24 is followed by a processing module 25, comprising a threshold detector which detects the maximum intensity of the received signal Ds, Dv. This processing module 25 is connected to a channel acquisition memory 26 to record the maximum intensity of the echo of each channel Vi.

[0120] When the channels Vi have their own electronics, a channel acquisition memory 26 is connected in the same way to the analog channels of the device, for example the memory 26 is connected to each of the channels V 1 V 2 V 3 ,...V 8 .

[0121] When there is a single electronics unit for multiple shots, module 26 is configured to store the maximum intensity of the signal received from each channel V i associated with an ultrasonic shot i.

[0122] The acquisition memory for channels 26 can be connected to a processing module 27 configured to generate echodynamic curves. The processing module 27 can also generate A-scans and B-scans from the highest intensity received signal among the i ultrasonic shots performed at a given position for a selected defect type, and this processing module 27 can also generate a C-scan of the inspected tubular product.

[0123] The channel acquisition memory 26 can be connected to a threshold comparator 28. The threshold comparator 28 compares the maximum intensity of the received echo signal and the alert trigger threshold level stored in a dedicated alert threshold memory 29. This comparator can trigger the operation of an alert module to the operator 30.

[0124] In another embodiment of the invention, the electronics 6 is configured to define at least one parameter of the ultrasonic shot as a function of the longitudinal position (L) and the circumferential position (A) of the ultrasonic emission means so as to capture defects in the wall of the tube, said at least one parameter being chosen from the emission orientation of the shot θe i (L ; A), the gain G(L ; A) or the position of the time filter FT i (L ; A).

[0125] The applicant carried out several series of tests on sample tubular products with expressly produced defects in order to determine the quality of detection of these defects.

[0126] In a first example, the sample tube (1) with axis (X) of the figure 6a It features several sections, 1a to 1e, and is developed as follows: Section 1a has constant external (Dext) and internal (Dint) diameters. Section 1b has non-constant external (Dext) and internal (Dint) diameters, with the external diameter (Dext) increasing from section 1a to section 1c and the internal diameter decreasing in the same direction. Section 1c has an external diameter increasing from section 1b to section 1d and a constant internal diameter. Section 1d has both an external and an internal diameter increasing from section 1c to section 1e. Section 1e has a constant external diameter and an internal diameter increasing from section 1d.

[0127] Longitudinal notches dl i (identified dl 1 to dl 10) of length 25 mm were made on the sample tube of the figure 6a The depth of the notches is constant, and the bottom wall of the notches is therefore substantially parallel to the surface in which these notches are produced, which is not represented in the figure 6a .

[0128] The tube of the figure 6a was subjected to inspection by a state-of-the-art automatic defect detection device and then with a device according to the invention, in order to measure the intensities of the return echoes of each longitudinal notch dl i . In this device according to the invention, only the emission angle θe(L) is varied on an acquisition channel.

[0129] The result of this test is shown in figures 6b And 6c , which represent graphs with the notch number dl i corresponding to the notches on the x-axis figure 6a and on the y-axis the signal-to-noise ratio for the figure 6b and the amplitude loss in dB relative to reference notch #5 on the figure 6c .

[0130] Graph 6b represents three curves: the threshold curve delimiting a minimum signal-to-noise ratio of a notch chosen at 12 dB, the curve of signal-to-noise ratios recorded with a state-of-the-art device (Lex), the curve of signal-to-noise ratios recorded with a device according to the invention (Linv).

[0131] Graph 6c represents two curves: the amplitude loss curve in dB relative to reference notch no. 5 recorded with a state-of-the-art device (Lex), the amplitude loss curve in dB relative to reference notch no. 5 recorded with a device according to the invention (Linv).

[0132] We see that notches 5 and 10 are raised to the same level by the prior art device and by a device according to the invention, which is normal, notches 5 and 10 are located in a section of the tube with a constant cross-section and the difference in construction between the two devices is inoperative in this section.

[0133] On the other hand, notches 2, 3, 6, 7 return very weak echoes in the state-of-the-art device, with levels below the detectability threshold, whereas the device according to the invention makes it possible to obtain high-level echoes, above 23 dB in all cases.

[0134] It can therefore be deduced that the state-of-the-art device does not detect notches 2, 3, 7 or even 6 in production inspection mode, whereas the device according to the invention makes it possible to detect these notches.

[0135] There figure 6c This shows the amplitude loss between a reference notch, here dl5, and the intensity of the echoes on the other notches. The signal level is conventionally set at 0 dB on reference notch #5 (or dl5). The lowest received intensity is 13 dB lower with the device according to the invention, whereas the difference reaches 35 dB across 3 defects with the prior art device.

[0136] The sample tube of the figure 7a is equipped with transverse notches dt i (dt 1 to dt 18 ) on a series of varied sections similar to the example described previously. It should be noted that transverse defects can be detected using shots with a chosen orientation in a longitudinal plane, as shown in figures 3 et 4 Transverse defects can be detected in both directions of the longitudinal axis. figures 7b , 7c, 7d, 7e show the signal-to-noise ratio values ​​and echo amplitudes measured in a first direction of inspection and then in the second direction of inspection, for a state-of-the-art device and the device according to the invention used in the test of figures 6a-c that is, by only varying the emission angle θe(L) on an acquisition channel.

[0137] We note that for the inspection in a first sense of the figures 7b , 7c There is a clear improvement in the return amplitude measured on certain internal or external transverse defects. More specifically, with the state-of-the-art device, defects numbered Dt 2, Dt 3, Dt 11, Dt 15 and Dt 16 are detectable at a marginal level or may not be detected at all since the signal-to-noise ratio of their received echoes is equal to or less than 12dB.

[0138] With the device according to the invention, all these defects are detected. Furthermore, the signal-to-noise ratio of the received echoes is high, exceeding 20 dB, which allows for a clearly distinct echo from background noise. The device according to the invention also provides improved homogeneity between the different intensities of echoes measured on internal transverse defects, with a difference of less than 12 dB between defect Dt 1 and defect Dt 6. This difference is less than 25 dB for external defects Dt 10 and Dt 12.

[0139] THE figures 7b-e allow us to observe that a device according to the invention, which varies the angle of orientation of a shot, provides better results than the prior art device, particularly on the tube section whose cross-section has an outer diameter that varies longitudinally. In the direction of detection of the figure 2 The homogeneity of the recorded echo intensities is improved for both internal and external transverse defects, with deviations of less than 12dB in the signal-to-noise ratio for both internal and external transverse defects.

[0140] We will see later with the examples of figures 9 à 11 that a device according to the invention can also achieve a very good level of homogeneity by varying both the emission orientation θe i (L) and the gain G i (L).

[0141] There figure 7c This shows that there is still a significant loss of echo amplitude on certain defects, especially for external defects, which are detected by the echo of an ultrasonic beam reflecting off the internal surface of the tubular product. The path is therefore particularly long, and the attenuation of the return echo is more sensitive to variations in internal and external diameters. The attenuation can still be on the order of 25 dB on defects 12 to 14. This response can therefore be further improved, as will be described below in an embodiment in which the electronics 6 are arranged to vary at least two parameters of the ultrasonic beam as a function of the longitudinal position (L) of the ultrasonic emission means, chosen from among the beam emission orientation θe i (L), the gain G i (L), or the position of the time filter FT i (L).

[0142] THE figures 8a et 8b-c respectively show a sample tube in which 4 flat-bottomed holes Tfp i have been made, used as standard defects with regard to the detection of defects in the wall called "doubles", and the comparative curves of a state-of-the-art device Lex and a device according to the invention Linv, both by a measurement of the signal-to-noise ratio ( fig. 8b ) and by measuring the signal amplitude in dB ( fig. 8c ). The device's electronics are arranged to vary the firing orientation parameter θe i (L) and the position of the time-domain filter FT i (L) along the longitudinal axis, without varying the gain, in a manner analogous to the tests carried out and presented within the framework of the figures 6 And 7 .

[0143] The flat-bottomed holes Tfp i have a diameter of 6 mm and a depth equal to half the local thickness of the part. It was decided to make the bottom parallel to the axis of the tube, and not parallel to the inner wall of the tube.

[0144] There figure 8b shows an improvement of approximately 5dB in the signal-to-noise ratio over a section in which the external diameter varies longitudinally. figure 8c This demonstrates an improvement of more than 15dB in the amplitude level of the return signal on the same section. The variation in internal diameter alone on a section does not appear to affect the measurement quality with regard to the detection of flat-bottomed holes.

[0145] These results show that a device according to the invention also improves the detectability of defects of the split type, since the intensities of the return echoes are greater for flat bottom holes Tfp 3 and Tfp 2.

[0146] No difference was observed between the flat-bottomed holes Tfp 1 and Tfp 4 because these holes are made on sections of tube with a constant outside diameter, and the slope of the inner wall has no influence on the measurement, as the ultrasonic pulse is directed at the flat bottom. The improvement stems from the choice of the emission orientation θe(L) and the position of the detection time filters FT i(L) as a function of the longitudinal position of the sensor.

[0147] Other sample tubes with different thicknesses and variation configurations were tested and showed similar results, i.e. showing a great improvement in the detectability of defects of all types with a device according to the invention compared to a state-of-the-art device.

[0148] There figure 9 is composed of figures 9a, 9b, 9c and aims to show a complete set of tests with a representation in figure 9a of a tubular product profile with varying external and internal diameters, equipped with three internal longitudinal notches designated by A, B, C.

[0149] There figure 9b represents a C-scan 95 obtained after scanning the tubular sample 93 by a device according to an embodiment of the invention, in which the three ultrasonic firing parameters Vi vary according to the longitudinal position of the ultrasonic sensor 5.

[0150] The C-scan 95 of the figure 9b reveals three zones corresponding to three sets of inspection parameters. These sets of parameters are shown by the figure 9c Figure 97 shows the minimum and maximum values ​​of the emission angle θe i (L) for each zone. It should be understood that all values ​​in 3° increments between the minimum and maximum values ​​are inspected. Thus, in zone 1, the emission angle θe i (L) varies between 0° and 6°. Three shots are taken for each inspection position through three electronic channels with θe 1 ([1])=0°; θe 2 ([1])=3°; θe 3 ([1])=6°. The notation [1] represents all axial and circumferential positions that the sensor 5 reaches in zone 1 for inspection purposes. In zone 2, two shots are taken per position at -6° and -3°. These values ​​are the same in zone 3.With regard to the detection of longitudinal notches, the person skilled in the art understands that the orientation of the ultrasonic beam emitted in the case of the inspection of longitudinal defects includes a transverse component contained substantially in a cross-section of the tube, and obtained by a mechanical orientation of the ultrasonic transducer 5, with a "mechanical angle" θm chosen at about 17° with respect to a normal to the axis of the tubular product passing through the point of intersection between the emitted beam and the outer surface of the tubular product, as well as a longitudinal component corresponding to the emission angle θe(L) obtained in the embodiment of the . figure 9 by electronic control of a phase-array transducer. In the case of this inspection, the mechanical angle θm is fixed while the emission angle θe(L) varies according to the longitudinal position L of the transducer 5.

[0151] Figure 98 shows the opening and closing values ​​of the time window FT i (L) and therefore the positioning of the detection gate for each zone, given in mm in the water. A gate in zone 3 is delayed compared to a gate in zone 2, which is itself later than a gate in zone 1. This allows for consideration of a longer ultrasonic wave path in zone 3, where the longitudinal notch C is furthest from the transducer 5, compared to zone 1, where the longitudinal notch A is closer to the transducer 5.

[0152] The gain G i (L) and its variation with respect to the longitudinal position of the transducer 5 is shown in graph 99. The gain is increased by 1 dB in zones 2 and 3 compared to the gain used in zone 1. In this example, it is the receive gain G i (L) that is modified as a function of the longitudinal position of the transducer 5.

[0153] The C-scan obtained shows that all longitudinal notches are detected satisfactorily, and that there are no spurious echoes causing a "phantom" defect. Furthermore, the signal-to-noise ratio is very consistent across the three defects, at 13 or 12 dB.

[0154] There figure 10 is composed of figures 10a 10b 10c and shows an example of the result obtained by implementing the invention for detecting internal through defects by varying only the gain of the ultrasonic beam. Here, the part is inspected in direction 1, i.e., a scan from defect A to defect I.

[0155] There figure 10a shows a partial section of a tubular product profile 103 having transverse notches 104 distributed longitudinally on the inner surface of the tubular product 103 and numbered by the letters A to I. Each notch has a length of 10 mm.

[0156] There figure 10b The C-scan 105 shows the result of inspecting the tubular product 103 with a device according to the invention. This C-scan clearly identifies each notch A to I. The device according to the invention made it possible to identify all these notches. The C-scan 105 shows the figure 10b Numbered zones 1 to 8 correspond to a virtual segmentation of the tubular product along the longitudinal axis. These zones correspond to sets of parameters including the emission orientation θje(L), the positioning of detection gates FTi(L), and the gain Gi(L). In conjunction with the C-scan 105, the average signal-to-noise ratio (SNR) 106 measured on each detected defect is displayed.

[0157] There figure 10c shows the parameter values ​​for each zone 1 to 8. In the context of the example of the figure 10 The emission orientation is chosen using the values ​​shown in Figure 107. The minimum and maximum values ​​are the same for each zone, namely between 34° and 43°. The device is configured to fire every 3° between the minimum and maximum limits. Consequently, for each firing position, the device performed a total of 4 firings at the following angles: 34°, 37°, 40°, and 43°. The firing orientations are therefore invariant regardless of the longitudinal positioning (L) of the sensor 5 used.

[0158] There figure 10c Figure 108 also shows the positioning of the detection gates. The position values ​​of these gates are given in millimeters in the water, which is the coupling medium chosen for the experiment. These values ​​are the same in all zones 1 to 8. The positions of the detection gates are therefore invariant regardless of the longitudinal positioning (L) of the sensor 5 used.

[0159] There figure 10c Figure 109 shows, for each zone 1 to 8, the additional gain values ​​in dB added to the base gain to form the gain Gi(L). The additional gain is, for example, 3 dB in zone 1, 7 dB in zone 3, and 10 dB in zone 7. The gain Gi(L) is the same for each of the four shots of different orientations made at a given longitudinal position (L). Thus, in zone 1, for a given position, a first shot is made at an emission angle in the steel of 34° with an additional gain of 3 dB, a second shot is made at 37° with an additional gain of 3 dB, a third at 40° with 3 dB, and a fourth shot at 43° with 4 dB.

[0160] The C-scan 105 of the figure 10b This shows that the single variation of the gain Gi(L) allows for the detection of all imperfections, thus yielding a satisfactory initial result. However, the resulting C-scan exhibits a non-uniform signal-to-noise ratio that varies between 18 and 25 dB depending on the area. The C-scan also reveals other defects, corresponding to notches made on the same tubular component for other experiments, for cost-saving reasons.

[0161] There figure 11 shows a test carried out in a second stage, with the same tubular product 104 equipped with notches A to I, the inspection being carried out this time in direction 2, i.e. with a sweep from defect I to defect A, with a device according to the invention in which the set of parameters is different from that used in the context of the figures 10a-c , in that it uses a variation of the three parameters along the longitudinal axis of the tubular product 103. The new set of parameters is represented in figure 11c and the corresponding results are represented in figure 11b .

[0162] The parameter set of the figure 11c is cut along the longitudinal axis of the tubular product into 4 zones numbered 1 to 4, which is half the number of zones as before. figure 10a-c .

[0163] There figure 11c Figure 119 shows the evolution of the gain Gi(L) by representing the additional gain applied in each zone, in dB, relative to a base gain value, which is chosen as the reference gain for the inspection. The additional gain is thus zero in zone 1, 2 dB in zone 2, 1 dB in zone 3, and 2 dB in zone 4. In this example, these gain variations compensate for a longer path length of the ultrasonic waves, for example, in zone 2 of the figure 11a-c , to compensate for a weaker response from the transverse notch C which is angled and which therefore could reflect the signal less weakly in the direction of firing.

[0164] There figure 11c Figure 118 shows that the positioning of the FT i (L) detection gates differs depending on the zone. This positioning is, for example, between 220 mm and 240 mm underwater in zone 2 and between 240 mm and 270 mm in zone 4. A detection gate in zone 4 is therefore not only positioned later than a gate in zone 2, but is also wider than a gate in zone 2.

[0165] There figure 11c Figure 117 shows the firing orientations for a given position, representing the maximum and minimum values ​​of the firing angle in the steel in degrees per zone from zone 1 to zone 4. Firings are made at these extreme values ​​and also in 3° increments between the two extreme values. The number of firings per zone also varies. In zone 1, there are two firings per position at angles of 37° and 40°; in zone 2, three firings are made per position at angles of 55°, 58°, and 61°; in zone 3, there are also three firings at 40°, 43°, and 46°; and finally, in zone 4, four firings are made per position at angles of 34°, 37°, 40°, and 43°. The emission angle is greater in zone 2 due to the slope of the inner wall of the tubular product in this zone.The number of shots is higher in zone 4 because there are different types of diameter variation in this zone, even if these variations are of smaller magnitude than in zone 2.

[0166] The C-scan 115 shows in figure 11b that all notches are correctly detected, despite dividing the tubular element into only 4 zones. Therefore, it does not appear necessary to create as many zones defining parameter sets as there are sections of different types along the tubular product. The feedback intensity on the defects is homogeneous (intensity is represented by color or gray level in the C-scan). Also, the noise level is almost halved on notch G in figure 11b compared with the noise level on the G notch in figure 10b Having at least two or three parameters varying according to the longitudinal position of the sensor, chosen from the gain G i (L), the emission orientation θe i (L), the time filter FT i (L) allows to improve the quality of the detection of defects in tubular products with variable cross-section.

[0167] Furthermore, the test of figures 11a-c allows the inspection to be carried out with 20% fewer shots than in the case of figure 10a-c while maintaining good defect detectability with a satisfactory signal-to-noise ratio and good homogeneity of responses.

[0168] It is generally necessary to calibrate non-destructive testing (NDT) devices to establish a threshold value for the return echo intensity for a given defect type and a given tubular product type. That is, there is typically one calibration for longitudinal notch defects, one for transverse notch defects, and one for flat-bottom hole defects, and this for each type of tubular product. For example, one type might be a tube with an external diameter of 250 mm and an internal diameter of 200 mm, while another type might be a tube with an external diameter of 315 mm and an internal diameter of 275 mm.

[0169] Notches of known position (depth and orientation) and dimensions, most often standardized, made in a sample tube, are classically used as reference defects, or standard defects.

[0170] To limit the number of shots and limit the computing power required for the device according to the invention, it is therefore preferable to calibrate said device so as to determine the values ​​of parameters chosen from the emission orientation θe(L), the gain G(L) and / or the position of a time filter FT i (L) for different longitudinal positions of the ultrasonic sensor(s).

[0171] Preferably, reference defects should be positioned at different longitudinal locations to obtain values ​​for these parameters for each type of section of the complex tubular product 3 with variable cross-section to be inspected. In other words, calibration can be performed for the different sections of the complex tubular product 3 with variable cross-section. The calibration is carried out on a sample tube similar to the complex tubular products to be inspected. Thus, the sample tube has diameter and thickness values ​​similar to a model of the tubular product to be inspected, i.e., identical sections and the same complex shapes. Furthermore, the sample tube must be made of the same material as the model, i.e., the same grade of steel, and must also have the same heat treatments and surface finishes.

[0172] The calibration process then allows, using a sample tube representative of a given tubular product and for a specific type of defect, the association of parameters such as the emission orientation θe(L), the gain G(L), and / or the position of a time-domain filter FT i(L) with each position of the ultrasonic sensor 5. These parameters can then be stored in a table such as the one shown below for parameterization based on the longitudinal position (L) of the sensor. Position longitudinale Orientation θ j e(L) Gain G(L) Portes détection FT(L) L 1 θe(L1) G(L1) FT(L1) L 2 θe(L2) G(L2) FT(L2) L 3 θe(L3) G(L3) FT(L3) ... ... ... ... L X θe(LX) G(LX) FT(LX) ... ... ... ...

[0173] We now refer to the figure 12 .

[0174] A complex tubular product with a variable cross-section, produced by industrial processes, may exhibit dimensional variations from the desired nominal values. Therefore, from the perspective of a sensor, the walls of the tubular product to be inspected may have actual wall slopes or wall slope positions that differ from the nominal slopes and positions. Consequently, it is important to remember that having a single value for parameters such as the emission orientation θe i (L), the gain G i (L), and / or the position of a time-domain filter FT i (L) does not always guarantee the best response to an ultrasonic pulse for detecting a given defect.

[0175] Thus, according to one embodiment of the invention, the device includes control and processing electronics 6 capable of performing a series of ultrasonic shots through channels V i for a longitudinal position (L) by varying for each channel V i at least one of the parameters chosen from the emission orientation θ je(L), the gain G i (L) and / or the position of a time filter FT i (L) near nominal values ​​of the parameters which are the emission orientation θe(L), the gain G(L) and / or the position of a time filter FT (L).

[0176] For example, for the inspection of a transverse defect Dt ex at a given longitudinal position P ex, the corresponding emission orientation θe(P ex) may have been determined to be 20° in the water in the longitudinal plane. Eight ultrasonic shots can be performed with emission orientations θe from 16° to 23° in the water in increments of 1°. The control and processing electronics 6 are therefore designed to perform n shot sequences, for example, from 1 to 10 shot sequences.

[0177] Each maximum amplitude of a corresponding echo is stored in the respective MemVi channel memories. These values ​​are compared by a channel comparison module, "Comp Voies," which can retain the maximum amplitude among those recorded on each of the Vi channels. Next, the maximum amplitude recorded is compared by a threshold comparison module connected to a threshold memory. The alert module issues an alert if the maximum amplitude recorded on the different channels exceeds the threshold specified for that type of fault.

[0178] In MEMp memory, the emission orientation θ je(L) can be advantageously defined by a pair of angles which can be the longitudinal component and the transverse component of the emission angle (θ N e(L) ; θ S e(L)) i which respectively represent the angle between the normal to the ultrasonic sensor and the projection of the shot in a longitudinal plane, i.e. a plane containing the axis of the tubular product, and the angle of the shot between the normal to the ultrasonic sensor and the projection of the shot in a transverse plane, i.e. a plane containing a section of the tubular product.

[0179] For example, as part of the inspection of the figure 9 The transverse component of the angle θ S e(L) is equal to the mechanical angle θm of the sensor and remains fixed throughout the inspection, typically at a value of 17°, while the longitudinal component of the angle θ N e(L) varies according to the longitudinal position (L) with values ​​ranging from -6° to 6°. For the inspection of defects such as longitudinal notches or flat-bottomed holes, the mechanical angle θm of the sensor is typically set to 0°.

[0180] In practice, this decomposition is adapted to the type of sensor used. For a linear multi-element transducer, the angle can only be modified in a single plane. Therefore, one of the two angles, longitudinal and transverse, is mechanically controlled, and the other is electronically controlled. For a multi-element matrix transducer, it is possible to electronically control both angles.

[0181] The gain Gi(L) is a gain applied to the position (L). Most often, the gain is approximately the same for all channels. It can be advantageous to vary the gain and therefore define several Gi(L) values ​​per channel for each channel, particularly to improve the detectability of split-type defects. This also improves the homogeneity of the response to oblique defects and with different emission orientations.

[0182] In a preferred embodiment, the gain G(L) is applied to the signal corresponding to the emission of the ultrasonic wave.

[0183] In another embodiment, the gain G(L) is applied to the signal corresponding to the reception of the echo from the ultrasonic shot. The receive gain G(L) is then denoted Gr(L). It is more advantageous to modify the receive gain Gr(L) than the transmit gain Ge(L) because the latter can amplify the received noise and thus be a source of false positives.

[0184] In one embodiment variant, the device includes an emission gain chosen as a function of the channel V i and denoted Ge i (L).

[0185] In another embodiment, the device includes a receive gain chosen according to the channel V i and denoted Gr i (L).

[0186] In MEMp memory, the representative data of an FT(L) detection gate can be a pair of data indicating the start of the detection time gate and the end of the detection time gate.

[0187] The MEMp memory can also be equipped with representative detection gate data for each channel Vi, denoted FTi(L), in the form of several sets of values. Indeed, for a longitudinal position L, there can be i shots with different orientations, and the position of the detection gates may need to be adapted accordingly. Furthermore, varying the positioning of the detection gates allows for compensation of dimensional variations in the wall thickness of the tubular product. In the case of a tube thicker than the desired nominal value, the actual time of flight of an ultrasonic wave is longer. It can therefore be advantageous to include a second detection gate positioned in temporal succession to the first detection gate.

[0188] Furthermore, according to one embodiment, it is possible to have two groups of channels V i presenting corresponding values ​​on the emission angles and applied gains, the two groups differing mainly by the position of the detection gates FT i (L) to allow the possible detection of internal faults with the first group of channels V i, and the detection of external faults with the second group of channels V i. Position longitudinale L1..LX Orientation θ j e(L) i = 1..n Gain émission Ge i (L) i = 1..n Gain reception Gr i (L) i = 1..n Portes détection FT i (L) i = 1..n L 1 θ 1 e(L1) = (θ N e(L1) ; θ S e(L1)) 1 Ge 1 (L1) Gr 1 (L1) FT 1 (L1)= FT START1 (L1) FT STOP1 (L1) θ i e(L1) = (θ N e(L1) ; θ S e(L1)) i Ge i (L1) Gr i (L1) FT i (L1)= FT STARTi (L1) FT STOPi (L1) θ n e(L1) = (O NE (L1) ; θ S e(L1)) n Ge n (L1) Gr n (L1) FT n (L1)= FT STARTn (L1) FT STOPn (L1) L2 θ 1 e(L2) = (θ N e(L2); Θ S e(L2)) 1 Ge 1 (L2) Gr 1 (L2) FT 1 (L2)= FT STARTI (L2) FT STOP1 (L2) θ i e(L2) = (θ N e(L2) ; θ S e(L2)) i Ge i (L2) Gr i (L2) FT i (L1)= FT STARTi (L2) FT STOPi (L2) θ n e(L2) = (θ N e(L2) ; θ S e(L2)) n Ge n (L2) Gr n (L2) FTn(L1)= FT STARTn (L2) FTSTOP n (L2) ... ... ... ... ... ... ... ... ... ... L X = (θ N e(LX); θ S e(LX)) 1 Ge 1 (LX) Gri(LX) FT 1 (LX)= FT STARTI (LX) FT STOPI (LX) θ i e(LX) = (θ N e(LX) ; θse(LX))i Ge i (LX) Gr i (LX) FT i (LX)= FT STARTI (LX) FT STOPi. (LX) θ n e(LX) = (θ N e(LX) ; θ S e(L1)) n Ge n (LX) Gr n (LX) FT n (LX)= FT STARTn (LX) FT STOPn (LX)

[0189] In an embodiment of a bench according to the invention adapted to inspect a tubular product with a profile exhibiting only intentional variations in thickness on its circumference, and therefore without intentional variation in its cross-section along the length of said tubular product, or in other words with a substantially invariant thickness along a generatrix of the tubular product, the MEMP memory can be organized in a manner analogous to that of the preceding embodiment, by replacing the variable L with the variable A.

[0190] In another embodiment of a bench according to the invention adapted to inspect a tubular product having intentional variations in thickness both along the tubular product and on its circumference of the tubular product, the MEMP memory can be organized in a manner analogous to that of the preceding detailed embodiment, by replacing the variable L with the variable pair (L; A).

[0191] The invention also relates to a non-destructive testing method for detecting defects on complex-shaped tubular products, comprising the steps of: position an ultrasonic transducer 5 relative to a complex tubular product 3 at a first position P1 relative to the complex tubular product 3; for this first position P1, perform at least one ultrasonic shot with an ultrasonic beam having at least one first emission orientation θ je(L ; A), a first emission gain Ge i (P1), and receive an echo signal and apply to this echo signal at least one first reception gain Gr i (P1) and one first time filter FT i (P1); position the ultrasonic transducer at a second position P2 relative to the complex tubular product 3; for this second position P2, perform at least one second ultrasonic shot with a second emission orientation θ je(P2), a second reception gain Gr i (P2) or a second emission gain Ge i (P2), a second time filter FT i (P2);at least one of the second transmission orientation θ je(P2), the second receive gain Gr i (P2), the second transmission gain Ge i (P2), the second time filter FT i (P2) being different from respectively the first transmission orientation θ je(P1), the first receive gain Gr i (P1) or the first transmission gain Ge i (P1), the first time filter FT i (P1). ;

[0192] In one embodiment of this process, position P1 is a first longitudinal position L1 of the ultrasonic sensor 5 and position P2 is a second longitudinal position L2 of the ultrasonic sensor 5.

[0193] This method can be applied to a set of contiguous longitudinal positions L forming an inspection section. We can then form a first inspection section in which the ultrasonic shots Vi will have the same shooting parameters forming a first set of shooting parameters, and we can form another inspection section in which the ultrasonic shots will have another set of shooting parameters, differing from the first set of shooting parameters by at least one of the parameters among the emission orientation θ je(L), the receive gain Gr i (L), the transmit gain Ge i (L), the time filter FT i (L).

[0194] It will be understood that for a longitudinal position L, ultrasonic shots can be carried out along substantially the entire circumference of the tubular product, for example over 360° with a shot distributed regularly by angular increment between 1° and 15°, carrying out shots at locations of the ultrasonic transducer 5 determined by a longitudinal position L and a circumferential position A.

[0195] In a second embodiment of this process, position P1 is a first circumferential position A1 of the ultrasonic sensor 5 and position P2 is a second circumferential position A2 of the ultrasonic sensor 5.

[0196] In a third embodiment of this process, position P1 is a first longitudinal position L1 and circumferential position A1 of the ultrasonic sensor 5 and position P2 is a second longitudinal position L2 and circumferential position A2 of the ultrasonic sensor 5.

[0197] It will be understood that, for a given position of the ultrasonic transducer 5 at a given longitudinal position L and a given circumferential position A, a series of ultrasonic shots can be performed, varying between each shot at least one parameter from among the following: the emission orientation θje(L; A), the receive gain Gri(L; A), the transmission gain Gei(L; A), and the time filter FTi(L; A). Preferably, the emission orientation θje(L) is varied between a minimum value θmine(L; A) and a maximum value θmaxe(L; A). This improves defect detection despite unintentional geometric imperfections (ovality, eccentricity) of the tubular product.

Claims

1. Automatic non-destructive testing device for detecting defects in a complex tubular product (3), said complex tubular product (3) exhibiting variations in outside diameters and / or inside diameters, said device comprising: - at least one ultrasonic transducer (5) having a position defined by a longitudinal position (L) and a circumferential position (A) along the complex tubular product (3) and arranged to transmit an ultrasonic beam (Em) having a transmission orientation (θei(L, A)); - control and processing electronics (6) comprising a circuit for exciting the ultrasonic transducer and detecting return signals; and - at least one amplification stage (21, 31) with a gain (Gi(L; A)); - a time-domain filter module (24) configured to apply a time-domain filter (FTi(L; A)) to an echo signal (Dv, Ds), characterized in that the control and processing electronics (6) are configured to define at least one ultrasonic shot parameter (Vi) depending on the longitudinal position (L) and / or circumferential position (A) of the ultrasonic transducer, which position(s) are determined and delivered by a module (23) for positioning said device, so as to detect defects in the wall of the complex tubular product, said at least one parameter comprising the orientation of transmission of the shot (θei(L, A)).

2. Device according to Claim 1, characterized in that the control and processing electronics (6) are configured to define at least two ultrasonic shot parameters (Vi) depending on the circumferential position (A) of the at least one ultrasonic transducer (5) so as to detect defects in the wall of the complex tubular product, said at least two parameters comprising the orientation of transmission of the shot (θei(L, A)) and a parameter selected from the gain (Gi(L; A)) or the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) is present.

3. Device according to Claim 1, characterized in that the control and processing electronics (6) are configured to define at least two ultrasonic shot parameters (Vi) depending on the longitudinal position (L) of the at least one ultrasonic transducer (5) so as to detect defects in the wall of the complex tubular product, said at least two parameters comprising the orientation of transmission of the shot (θei(L, A)) and a parameter selected from the gain (Gi(L; A)) or the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) is present.

4. Device according to Claim 3, characterized in that the control and processing electronics (6) are configured to define the orientation of transmission of the shot (θei(L, A)), the gain (Gi(L; A)) and the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) of ultrasonic shots (Vi) is present depending on the longitudinal position (L) of the at least one ultrasonic transducer (5).

5. Device according to either of Claims 3 and 4, characterized in that the control and processing electronics (6) are configured to also define at least one parameter selected from the orientation of transmission of the shot (θei(L, A)), the gain (Gi(L; A)) and the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) of ultrasonic shots (Vi) is present depending on the circumferential position (A) of the at least one ultrasonic transducer (5).

6. Device according to any of Claims 1 to 5, comprising at least one position sensor (7a) for determining the longitudinal position (L) of the at least one ultrasonic transducer (5) relative to the complex tubular product (3).

7. Device according to any of Claims 1 to 5, comprising at least one position sensor (7a) for determining the longitudinal position (L) and circumferential position (A) of the at least one ultrasonic transducer (5) relative to the complex tubular product (3).

8. Device according to either of Claims 6 and 7, wherein the at least one position sensor (7a) is selected from an incremental encoder, a rack-and-pinion encoder, a linear encoder, a draw-wire encoder, a laser velocimeter, a coding wheel or a coding-wheel ratio.

9. Device according to any of Claims 1 to 5, comprising at least one timer (7b) for determining the relative longitudinal position (L) and relative circumferential position (A) of the ultrasonic transducer (5).

10. Device according to any of the preceding claims, wherein the at least one amplification stage (21, 31) is a transmit amplification stage (21) having a transmit gain (Gei(L; A)) and the control and processing electronics (6) are configured to make said transmit gain (Gei(L; A)) vary depending on the longitudinal position (L) of the ultrasonic transducer (5).

11. Device according to any of the preceding claims, wherein the at least one amplification stage (21, 31) is a receive amplification stage (31) having a receive gain (Gri(L; A)) and the control and processing electronics (6) are configured to make said receive gain (Gri(L; A)) vary depending on the longitudinal position (L) of the ultrasonic transducer (5).

12. Device according to any of Claims 1 to 9, comprising a transmit amplification stage (21) having a transmit gain (Gei(L; A)) and a receive amplification stage (31) having a receive gain (Gri(L; A)) and wherein the control and processing electronics (6) are configured to make the transmit gain (Gei(L; A)) or the receive gain (Gri(L; A)) vary depending on the longitudinal position (L) of the ultrasonic transducer (5).

13. Device according to any of Claims 1 to 12, wherein the control and processing electronics (6) comprise a parametric memory module (MEMp) capable of storing data in the form of an association between at least one longitudinal position (L) of at least one ultrasonic transducer (5) and at least one data set corresponding to parameters in respect of orientation of transmission of the shot (θei(L, A)), gain (Gi(L; A)) and / or the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) is present.

14. Device according to any of Claims 1 to 12, wherein the control and processing electronics (6) comprise a parametric memory module (MEMp) capable of storing data in the form of an association between at least one circumferential position (A) of at least one ultrasonic transducer (5) and at least one data set corresponding to parameters in respect of orientation of transmission of the shot (θei(L, A)), gain (Gi(L; A)) and / or the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) is present.

15. Device according to any of Claims 1 to 12, wherein the control and processing electronics (6) comprise a parametric memory module (MEMp) capable of storing data in the form of an association between pairs of longitudinal and circumferential positions (L; A) of the ultrasonic transducer (5) and at least one data set corresponding to parameters in respect of orientation of transmission of the shot (θei(L, A)), gain (Gi(L; A)) and the position and width, in the time-domain filter (FTi(L; A)), of the time window (Fei(L; A)) in which the echo signal (Dv, Ds) is present.

16. Device according to any of Claims 13 to 15, wherein the parametric memory module (MEMp) comprises at least one data set corresponding to gain parameters (Gi(L; A)) taking the form of parameters in respect of receive gain (Gei(L; A)) and transmit gain (Gri(L; A)).

17. Device according to any of the preceding claims, characterized in that the control and processing electronics (6) are configured to transmit a plurality of ultrasonic shots (Vi) for a position of the ultrasonic transducer (5), the ultrasonic shots (Vi) having transmission angles (θej(L)) between a minimum positional angle of orientation (θemini(L)) and a maximum positional angle of orientation (θemaxi(L)).

18. Device according to the preceding claim, characterized in that the control and processing electronics (6) are arranged to fire from 2 to 8 ultrasonic shots (Vi) for a position of the at least one ultrasonic transducer (5).

19. Device according to any of the preceding claims, wherein at least one ultrasonic transducer (5) is a strip of ultrasonic transducers.

20. Device according to any of the preceding claims, wherein at least one ultrasonic transducer (5) is a phased array sensor.

21. Automatic method for testing tubular products with varying outside or inside diameters by means of a device according to any of the preceding claims, wherein: a. at least one ultrasonic transducer (5) is positioned in a first position (PI); b. a first ultrasonic shot (Vi) is fired by transmitting an ultrasonic beam (Em) having a first orientation (θei(P1)), and a first transmit amplification with a first transmit gain (Gei(P1)), the first orientation (θei(P1)) and the first transmit gain (Gei(P1)) being transmission parameters of the first ultrasonic shot (Vi); c. an echo reflected by the complex tubular product (3) is detected and the detected echo is converted into a received signal to which a first receive gain (Gri(P1)) is applied; d. a portion of the signal is isolated in a first time window (FTi(P1)), the first receive gain (Gri(P1)) and the first time window (FTi(P1)) being reception parameters of said first ultrasonic shot (Vi); e. a second ultrasonic shot is fired by repeating steps a to d in a second position (P2), with transmission parameters of the second ultrasonic shot comprising a second orientation (θei(P2)) and a second transmit gain (Gei(P2)) and reception parameters of the second ultrasonic shot comprising a second reception gain (Gri(P2)) and a second time window (FTi(P2)), characterized in that the second orientation (θei(P2)) is different from the first orientation (θei(P1)).