METHOD AND DEVICE FOR REPAIRING A DEFECT OR A DAMAGE ON A METALLIC PIECE OF MATERIAL

Abstract

A method and associated system repair a damage or a defect in a piece of metallic material. A sensor on a remote operated vehicle (ROV) measures first geometrical data of a plurality of points or areas of the defect or damage. Based on the first geometrical data, excavation data is generated of one or more zones of the defect or damage by selecting or generating one or more predefined shapes or volumes. With the ROV, the piece of metallic material is excavated according to the excavation data to form an excavated zone. With the ROV, the piece of metallic material is repaired by depositing metallic material in the excavated zone.

Claims

1-27: (canceled)

28. A method for repairing a damage or a defect in a piece of metallic material, comprising: (a) with a sensor of a remote operated vehicle (ROV), measuring first geometrical data of a plurality of points or areas of the defect or damage; (b) based on the first geometrical data, generating excavation data of one or more zones of the defect or damage by selecting or generating one or more predefined shapes or volumes from a plurality of predetermined shapes or volumes; (c) with the ROV, excavating the piece of metallic material according to the excavation data, thereby forming an excavated zone; and (d) with the ROV, repairing the piece of metallic material by depositing metallic material in the excavated zone.

29. The method according to claim 28, wherein a dimension of the predefined shape or volume is adapted based on the first geometrical data.

30. The method according to claim 28, wherein the plurality of predetermined shapes or volumes are stored in a library.

31. The method according to claim 28, wherein the sensor is a contactless sensor.

32. The method according to claim 28, further comprising, before step (d), generating or measuring second geometrical data of an undamaged part of the piece of metallic material surrounding the damage or the defect, wherein the second geometrical data is interpolated from measured data or is generated based on the first geometrical data.

33. The method according to claim 32, wherein step (b) is based on the measured first geometrical data and the second geometrical data.

34. The method according to claim 32, further comprising generating third geometrical data corresponding to the piece of metallic material without the defect or damage.

35. The method according to claim 34, further comprising: (e) measuring fourth geometrical data of the repaired piece of material and comparing the fourth geometrical data with the third geometrical data.

36. The method according to claim 28, further comprising, before step (c), estimating a geometrical data of the deepest point or area of the damage or defect relative to an undamaged portion of the piece of metallic material.

37. The method according to claim 36, wherein step (c) comprises excavating down to a level equal or below the deepest point or area of the damage or defect.

38. The method according to claim 28, further comprising: (e) removing a portion of the deposited metallic material in a finishing step.

39. The method according to claim 28, further comprising generating a paving of an entire area the damage or defect with the predefined shapes or volumes selected or generated in step (b).

40. The method according to claim 39, further comprising generating a paving of an undamaged area neighboring the damage or defect with the predefined shapes or volumes selected or generated in step (b).

41. The method according to claim 39, wherein the predetermined shapes or volumes have a constant depth.

42. The method according to claim 39, further comprising modifying or adapting a dimension of one or more of the predefined shapes or volumes.

43. The method according to claim 28, comprising scanning the piece of metallic material with the sensor along a plurality of parallel lines.

44. The method according to claim 43, comprising scanning overlapping neighboring areas with the sensor.

45. The method according to claim 28, wherein the ROV is teleoperated to reach the damage or defect and steps (c) and (d) are performed automatically by the ROV.

46. The method according to claim 28, wherein the piece of metallic material is a part of a hydro power plant.

47. A system for repairing a damage or a defect in a piece of metallic material, comprising; one or more remote operated vehicles (ROVs); a sensor configured on the one or more ROVs to measure first geometrical data of a plurality of points or areas of the defect or the damage; means for generating excavation data based on the first geometrical data, the excavating data including one or more zones of the defect or damage selected or generated from one or more predefined shapes or volumes; an excavation tool configured on the ROVs to excavate the piece of metallic material according to the zones to define an excavated zone; and the one or more ROVs further configured to deposit metallic material in the excavated zone.

48. The system according to claim 47, wherein the one or more ROVs are configured to transmit the first geometrical data.

49. The system according to claim 47, wherein the one or more ROVs are configured to receive the excavation data.

50. The system according to claim 47, comprising a library of the predetermined shapes or volumes.

51. The system according to claim 47, wherein the sensor comprises a contactless sensor.

52. The system according to claim 47, wherein the one or more ROVs comprise a welding tool to deposit the metallic material in the excavated zone.

53. The system according to claim 47, wherein the one or more ROVs comprise a first ROV and a second ROV, wherein: the first ROV comprises the sensor and is configured to transmit the first geometrical data; and the second ROV is configured to receive the excavation data and comprises a tool for excavating the piece of metallic material to define the excavation zone, the second ROV comprising means for depositing metallic material in the excavated zone.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0055] FIG. 1 shows a runner of a hydro turbine to which the invention can be applied;

[0056] FIGS. 2A and 2B show examples of cavitation damage in a runner of a hydro turbine;

[0057] FIGS. 3A and 3B show examples of cracks in a runner of a hydro turbine;

[0058] FIGS. 4A-4D show exemplary steps of a process according to the invention;

[0059] FIGS. 5A-5C show examples of defects in a turbine runner;

[0060] FIGS. 6A-6C show examples of predefined shapes which can be used in a process according to the invention;

[0061] FIG. 7 is a schematic representation of a system according to the invention, including a remote operating vehicle (ROV), a computer and a display device;

[0062] FIG. 8 illustrates an aspect of an embodiment of a process according to the invention;

[0063] FIGS. 9A-9D show another aspect of an embodiment of a process according to the invention;

[0064] FIGS. 10A-10B illustrate an aspect of a scanning step of an embodiment of a process according to the invention; and

[0065] FIG. 11 illustrates a deposition step of an embodiment of a process according to the invention.

DETAILED DESCRIPTION

[0066] Reference will now be made to embodiments of the invention, one or more examples of which are shown in the drawings. Each embodiment is provided by way of explanation of the invention, and not as a limitation of the invention. For example features illustrated or described as part of one embodiment can be combined with another embodiment to yield still another embodiment. It is intended that the present invention include these and other modifications and variations to the embodiments described herein.

[0067] The invention applies to any piece of metallic material, such as bulb, Kaplan, Francis, pump turbine runners or pump impellers or other metallic components of a hydroelectric power plant. Other industrial applications exist.

[0068] FIG. 1 represents a hydraulic machine 1, more precisely a pump turbine 2, for converting hydraulic energy into electrical energy. A shaft 3 is coupled to a runner 7 and to the rotor of a generator (not shown on the figure) which also has an alternator that converts mechanical energy into electrical energy. The machine also includes a spiral case 4 that is supported by concrete blocks 5 and 6.

[0069] A penstock (not shown on the figure) extends between a non-represented upstream reservoir and the spiral case 4. When the turbine is operated in a generating mode, this penstock forces a water flow to power the turbine. Water flows between blades 8 of the runner 7 and rotates them around an axis x-x of the shaft 3.

[0070] The machine further includes: [0071] a distributor, comprising a plurality of movable guide vanes or wicket gates 9 that are evenly distributed around the runner; each having an adjustable pitch around an axis parallel to x-x and can be swivelled to regulate the water flow rate, all guide vanes being oriented with a same angle relative to a closed position; [0072] a pre-distributor, which includes a plurality of fixed vanes (or stay vanes) 10, evenly distributed around the axis of rotation x-x, and disposed upstream of and around the distributor.

[0073] Below the runner, water is evacuated through a draft tube and tail water tunnel 11 (discharge ring and/or draft tube).

[0074] A hydraulic machine can have many other parts or components, such as a trash rack, one or more gate(s), one or more stoplog(s), a penstock, one or more valve(s), a stay ring.

[0075] During the use of the machine to which they belong, all these parts or components, usually made of a metallic material, can possibly be damaged by any one or combination of erosion by cavitation, erosion by abrasion, ballistic impacts from solids conveyed in the water flow, fatigue, corrosion, wear, a combination of those damage modes that can generate local surface deformation(s), cracks, wear, impacts, material loss(es), or coating damages.

[0076] Such a hydraulic machine therefore must be repaired.

[0077] The other kinds of turbines or pump impellers are also concerned by these problems: at least one of the above-mentioned defects or damages, must sometimes be repaired. In particular, the runners of such turbines include blades that can suffer from the above mentioned damages that must be repaired.

[0078] Cavitation is the vaporization of a volume of a liquid. When this happens adjacent to a solid, the fluid momentum associated with the collapse of the vapour volume can damage the solid material. This problem can be persistent when high velocities and forces are used to move the fluid, as in a turbine or pump. Often, once cavitation damage has occurred in a given location, the likelihood of additional damage in the same place increases.

[0079] More precisely, cavitation damage usually consists of a surface damage from which material of the hydro turbine or of the runner of said hydro turbine is removed. A cavitation damage can have many different sizes. It can for example have a surface size of about 20 cm20 cm, but there are also very small cavitation damages (for example at the beginning of the damage) and then the damage can become larger. A first example of cavitation damage 12 (a crown cavitation damage, at a blade to crown junction) is shown in FIG. 2A. A second example is shown on FIG. 2B which represents a clearance cavitation damage 14 (blade tip diameter for a Kaplan turbine).

[0080] A cavitation damage can occur anywhere on both sides of a blade 8 (FIG. 1), on the crown and on the band surface (not shown on FIG. 1), at locations that are very difficult to reach and therefore in situ reparation is very difficult to achieve. For Francis or pump turbines or pump impellers, the inter-blade distance can be limited and can be for example of about 150 mm or more generally between 500 mm or 300 mm and 50 mm or 100 mm or 200 mm. Even for the largest inter-blade distances (larger than 400 mm), it is very difficult to repair the surfaces of the blades in situ.

[0081] A crack is another example of a defect of a metallic component (for example of the runner of said hydro turbine) to be detected. Cracks are usually surface cracks and then evolve into through-cracks after some time of propagation. Cracks may have for example a size of around 5 mm (or more) length, 1 mm (or more) depth, 0.1 mm (or more) wide. A first example of a crack 16 is shown on FIG. 3A, which is a small crack (about 20 mm at band outlet) and should be repaired in a short or medium term in order to avoid a further propagation. A second example is shown on FIG. 3B, which represents an advanced crack 18 (600 mm long at blade inlet). In this case, the hydraulic machine must be stopped in a short or medium term in order to avoid extreme failure. The repair can be done on site (by excavation, welding, grinding).

[0082] Cracks can be for example located at the inlet or the outlet of the runner, at the junction between the blades and the crown or the band, which are the weakest locations of the runner where the highest stress concentrations occur. For Francis turbines, the majority of cracks occur at the runner outlet.

[0083] In an embodiment of the invention, a defect or a damage of a piece of metallic material, for example a cavitation damage or a crack, can be first identified and/or located by one or more sensors, for example an Eddy Current sensor or a contactless sensor, for example at least one optical sensor, preferably selected from among a laser, a camera, a laser scanner, a lidar and a telemeter. This step can be performed further to, or during, a visual inspection of said metallic material wherein an operator identifies a damage or defect and possibly delimites an area of interest the includes the entire defect or damage and its surrounding area.

[0084] One or more of the above mentioned sensor(s) can be added on an existing ROV (Remote Operated Vehicle), preferably rail free, or AUV (autonomous underwater vehicle) and can operate for example both in water and in air.

[0085] First data, for example first geometrical data, of the damage or of the defect can be generated from the data of the sensor inspecting the defect or the damage.

[0086] In this application, geometrical data can include a plurality of coordinates or of XYZ coordinates of several points situated on the inspected surface, in particular the surface of the defect, which is not a flat or an even or regular surface. Preferably, as explained below, the inspected surface also includes an undamaged area (which has an upper surface that is flat or even or regular) close to the defect or surrounding it. Said points may possibly be separated by a given distance (the resolution) and/or the geometrical data may be associated with a certain confidence or accuracy. Such geometrical data can be used to form a 3D image (or a 3D scan) of the inspected part or component and/or to give a representation of the surface. In combination with the above means, a device according to the invention can include a lighting device in the visible range, for example a lamp or a laser or a LED. A lighting device may allow a sensor to measure faster and/or with more precision any of said first data or first geometrical data, or any of the second to fourth geometrical data disclosed below, or data from which any of said geometrical data can be derived.

[0087] In a variant, a device according to the invention is used in combination with another ROV or AUV which carries or includes a lighting device in the visible range, for example a lamp or a laser, for the same purpose as mentioned above.

[0088] FIG. 4A is a schematic representation of a cross-section of a simple defect or damage 20 and of its undamaged surrounding area 24 in a part or component 22, for example a blade of a runner.

[0089] As explained above, first geometrical data or coordinates, for example data of points P.sub.1d, P.sub.2d, P.sub.3d, of the upper surface of said defect(s) or damage(s) are measured or generated based on data provided by one or more sensor(s) during the scanning of the area of interest, which can be performed during a visual inspection of the metallic material or at a later stage.

[0090] Based at least on said first geometrical data, it is possible to generate data of one or more zones of said defect(s) or damage(s) in order to excavate said one or more zone, in other words in order to eliminate more material from the piece 22. Thus, for example, data of a surface 26 (represented with a dashed line on FIG. 4A) located under the upper surface of the defect are generated, which will form a receiving surface for a material to be deposited in said defect(s) or damage(s).

[0091] In a particular embodiment, the deepest points of the damage or defect can be identified by comparing at least the Z coordinates of said first geometrical data or coordinates, with a theoretical undamaged surface of the piece (for example based on data of the third geometrical data or coordinates as explained below).

[0092] Excavation can then be performed. FIG. 4B shows the same part or component 22 after excavation, 20e designating the excavated damage or defect. Preferably, one or more corner(s) of the excavated zone has a radius larger than a defined mm, for example larger than 2 or 2.5 mm, so that the material can then be added into the excavated zones without gaps.

[0093] FIG. 4C shows the same part or component 22 after material 28 was deposited in said excavated defect(s). This step for example implements addition of material by welding. The added material can be for example homogeneous material, or heterogeneous material, including functional grades.

[0094] At any of the above steps, second geometrical data or coordinates, for example of points P.sub.1, P.sub.2, P.sub.3, of the upper surface of at least one undamaged part or area 24 of the piece of material, preferably surrounding the damage or the defect 20, can be measured or generated based on data provided by said one or more sensor(s). Such data can be used as reference data, the undamaged part or area 24 being considered as a reference part or area. Preferably, said second geometrical data or coordinates are measured or generated during the same step as the above first geometrical data or coordinates. In particular, based on the data generated during the above scanning, a damaged area, as well as a reference area or surface, which is the undamaged area surrounding said damaged area, can be delimited automatically and/or by an operator. To do so, an operator can define, in a control system (for example the system bearing reference 124 on FIG. 7, described below), at least one threshold or cut-off value Z.sub.0 along the Z axis (perpendicular to the surface of the area): any scanned surface that is at an elevation along the Z axis below the threshold or cut-off value Z.sub.0 can be included in the damaged area and should preferably be repaired. The reference area and damaged area (for example surface 40 on FIG. 5B, see comments below concerning this figure) are hence delimited. In a process according to the invention, an operator can also further refine the delimitation based on his observations, know-how and experience. Based on this reference area, an extended reference surface can be calculated: it can be an interpolation of the reference area or it can be generated based on previous measurements or a 3D model of the initial surface at the location of the defect or damage before the defect or damage occurred.

[0095] As can be seen on FIG. 4C the upper surface 30 of the added material may raise above the expected or theoretical level or surface 32 of said material. For this reason, the surface of the repaired area can be further finished as can be seen on FIG. 4D, thus forming an upper surface 32 which has the expected level. For this purpose it may be helpful: [0096] to generate third geometrical data or coordinates of the upper surface of the undamaged or theoretical component, for example of points P.sub.1u, P.sub.2u, P.sub.3u of the expected level 32. The third geometrical data or coordinates may result from previous measurements made on an undamaged part or component and may have been stored in a memory or may result from calculations or simulations. The third geometrical data or coordinates can be obtained by interpolation from said second geometrical data or coordinates or may be generated before starting any generation or measurement of the above mentioned first and/or second geometrical data or coordinates. The above described data of the extended reference area can form part of said third geometrical data; [0097] to measure or generate fourth geometrical data or coordinates, for example of points P.sub.1r, P.sub.2r, Par of the upper surface of the repaired defect, after the material 28 was deposited in said excavated damage or defect(s).

[0098] The third geometrical data or coordinates can be compared with the fourth geometrical data or coordinates in order to decide how the upper surface 30 should be further finished to obtain the required or expected upper surface 32, as represented on FIG. 4D.

[0099] The component of the example illustrated on FIGS. 4A-4D has a flat upper surface that is particularly adequate for the purpose of this explanation. But the invention also applies to much more complex parts, for example like: [0100] the one illustrated in the example of FIG. 5A (cross-section) and FIG. 5B (top view) which represents a concave portion 46 having a curvature between a crown or a band 42 and a blade 44 and having a defect 40 extending over these areas and in particular in the concave portion 46; FIG. 5B shows the reference area in dark grey, surrounding the defect (or damaged area) 40 in light grey; [0101] or the one illustrated in the example of FIG. 5C which represents a blade 52 having a damage or defect 60 at its end, in a convex area of the blade.

[0102] The invention can be applied to any other part of a hydraulic machine, for example to a trash rack, one or more gate(s), one or more stoplog(s), a penstock, one or more valve(s), a spiral case, a distributor, a stay ring, one or more guide vane(s), one or more stay vane(s), wicket gate(s), the draft tube, or the runner. It can also be applied to a metallic piece of any other kind, for example in oil and gas applications, in pressure vessels or in the ship building industry.

[0103] In a particular embodiment, a method of repairing a piece of material according to the invention, in particular as described above, can comprise the following steps: [0104] STEP 1: a surface shape assessment (Post damage assessment), for example by implementing the generation or the measurement of first geometrical data as explained above; [0105] STEP 2: a substrate shapes preparation by excavation of the damaged area, as explained above; [0106] STEP 3: a shape rebuilding by material addition or deposit on substrate for instance, as explained above;

[0107] It may further include: [0108] STEP 4: a surface finishing, as explained above, and/or: [0109] STEP 5: a surface shape assessment (Repaired surface control).

[0110] According to an embodiment of the invention, a process according to the invention may involve one or more of the following (in particular during above step 1): [0111] during a visual inspection of a metallic material, an operator can identify a damage or defect and delimit an area of interest that includes the entire defect or damage and its surrounding area; the scanning of the area of interest may be performed, for example during such a visual inspection, or at a later stage; [0112] based on the data generated during the scanning, a damaged area, as well as a reference area or surface, which is the undamaged area, or at least part of it, surrounding said damaged area, can be delimited automatically or by an operator. To do so, an operator can define, in a control system, at least one threshold or cut-off value. Any scanned surface that is at an elevation on the Z axis below the threshold or cut-off value is included in the damaged area and must be repaired. The reference area and damaged area (for example surface 40 on FIG. 5B) are hence delimited. In a process according to the invention, an operator can also further refine the delimitation based on his observations, know-how and experience; [0113] based on this reference area, an extended reference surface can be calculated: it can be an interpolation of the reference area or it can be generated based on previous measurements or a 3D model of the initial surface at the location of the defect or damage before the defect or damage occurred. The data of the extended reference area forms part of the third geometrical data.

[0114] According to an aspect of the invention, the data of the surface 26 (represented with a dashed line on FIG. 4A) which is the lower surface of the planned excavation located under the upper surface of the damage or defect can be generated from a plurality, or a library, of predetermined shapes which are adapted to the repair process.

[0115] FIG. 6A shows a plurality of such predetermined shapes combined to form a theoretical shape corresponding to the example of FIG. 4A.

[0116] More generally, each predetermined shape may have a constant depth with respect to the extended reference surface (namely with the bottom parallel to the upper surface, both of them being either flat or curved), as illustrated on FIG. 6B (a view in the XZ plane); the depth of each predefined shape can be, however, adapted to the depth of the excavated area by a control system (for example system 124) or an operator based on first geometrical and third geometrical data.

[0117] FIG. 6C shows a possible front view of a primitive. A primitive is a predefined volume, which for example has: [0118] a cross-section in XZ plan in the form of one of the predefined shapes of FIG. 6B, the predefined shape having an adjustable depth along the z axis as explained above; [0119] and a top view surface in the XY plan which is in one of the forms of FIG. 6C (but other forms may be possible); the dimensions of the surface in XY plan of each primitive are determined based on predefined top view surfaces stored in a library of the control system and preferably cannot exceed the working envelope (as explained below).

[0120] These primitives and shapes can be memorized in one or more memory area of the system, for example of one or more computer(s) 124 and/or in the ROV itself. Here and in this application, other systems like one or more processor(s), or one or more microprocessor(s), or one or more microcontroller(s), or one or more controller(s) can be implemented instead of a computer.

[0121] One or more ROV(s) (Remote Operated Vehicle), preferably rail free, can be used to implement the invention. Preferably, one or more of said ROV(s) has magnetic wheels, so that it/they can easily remain in contact with the surface which must be scanned and/or repaired, whatever the orientation of the ROV.

[0122] One example of a ROV which can be implemented is the VT34 FW crawler of Dekra (see https://www.dekra-visatec.com/en/detail/index/sArticle/31). It is quite compact (14055185 mm) and is able to access a large number of turbine runners, preferably magnetic. It can be equipped with a camera having one of the above-mentioned features.

[0123] Another ROV which can be used in the frame of the invention is the GNOM Baby (see https://gnomrov.com/products/gnom-baby/): it is a very compact swimmer operating as a submarine and equipped with a camera. It is compact enough (21180150 mm) to enter in the specified 200 mm interblade distance.

[0124] Another device which can be used is the Bike of Waygate Technologies, see for example https://inspection-robotics.com/bike/.

[0125] A ROV is mounted on wheels and moves along the inspected surface.

[0126] A ROV 120 (FIG. 7), in particular any of the above devices, can be equipped with one or more sensor 121 as mentioned above and possibly a lighting device, adapted to implement the invention. It is in contact with the metallic piece of material to be repaired, for example a blade of a runner.

[0127] The ROV 120 can be driven by electric power, for example with help of a battery and a motor 123 (FIG. 7). A power source can be connected via a tether 122 or can be onboard and the ROV can be controlled wireless or via tether 122. It performs one or more steps of the invention automatically (examples of steps automatically performed are given below). For example, an operator starts the process and then the different steps of the invention are performed automatically.

[0128] The ROV can also comprise one or more tools for repairing surfaces, as explained below.

[0129] A plurality of ROVs can be used at the same time to inspect different locations of a runner. A plurality of ROVs can be used, wherein at least one inspects one or more locations of a runner and at least another one repairs one or more locations of a runner.

[0130] In FIG. 7, the ROV 120 is equipped with one or more camera(s) 121 to perform imaging and thereby provide data of the damage or defect and possibly also of a reference surface which is next to or surrounds the damage or defect. It can be connected by a tether 122 (which can comprise cables or wires to exchange data with the ROV or to control the ROV or to provide the electric power to the ROV) to a computer 124 (for example a microcomputer or a portable computer) that controls the path followed by the ROV or the camera(s) or a tool and which sends instructions to perform the appropriate inspection and/or repair steps. A tether-less or wireless ROV is also possible, see for example https://www.hydromea.com/exray-wireless-underwater-drone/; that operates autonomously with wireless connection.

[0131] Data, in particular the images from a camera(s), scanner, or from any other sensor(s), can be processed (in real time or not) by a processor in the ROV or can be transmitted to computer 124 for processing, in particular to generate one or more of the above first to fourth geometrical data. A 3D representation or view of the damage or defect can also be generated or an image can be displayed on a display or screen 126 so that an operator can decide whether the hydraulic machine should be repaired. It is also possible to make measurements and generate a map of the damage or defects detected on a blade or on any portion of a metallic piece of material; such a map, giving information about the location of the damages or defects, can be generated by computer 124 and possibly displayed. The illustrated system can be adapted to several ROVs performing inspection of different parts, or repairing different parts, at the same time.

[0132] One or more ROV can perform the operations or steps of a method according to the invention, for example the operations or steps explained above in connection with FIGS. 3A-3D; accordingly, one or more ROV can perform: [0133] a 3D scan (damage characterization, assessment and recording, excavation characterization and recording, surface control (intermediate and/or final stage control)); [0134] excavation (for example) and possibly a surface finishing step; [0135] and then filling by deposition on the substrate after excavation.

[0136] In case of use of several ROVs, each one is preferably able to register its position on the surface with respect to the position of the first ROV and the position of the area of interest (respectively damaged area and excavated area) in order to ensure that the tool path (respectively excavation strategy and material addition strategy) which is defined, for example on a 3D numerical model generated from a 3D scan, and particularly the origin points, are correctly localized on the real surface.

[0137] In some cases, a ROV 120 has a working envelope covering a portion of the inspected piece which is less than the extent of a damage or defect. A working envelope corresponds to the area that can be scanned, excavated, or repaired for one position of the ROV. It may depend on several constraints: for example obstacles in the surroundings of the damage, or the size of the robot its arm(s), or the tool(s). For this reason, a set of positions of the ROV can be defined and the ROV can be moved from one of said positions to the next one. In particular, if the working envelope of the ROV is smaller than the damaged area, the operator will have to move the ROV. A position of the ROV can be registered by the operator so that once the ROV is driven to such position, then next steps can be performed automatically.

[0138] As explained before, the operator is able to delimit the damaged area 20. As illustrated in FIG. 8, based on this data, the operator is able to select a number of primitives or predefined shapes from a library in a control system (for example system 124) to create a paving over the entire area of the damage or defect 20, corresponding to areas O.sup.1-O.sup.5. Alternatively, a control system can perform this step automatically based on predefined primitives and shapes stored in a library. Preferably, the operator can then adjust the selected primitives or predefined shapes based on his own experience and know how.

[0139] The plurality of primitives defines the working area. Each O.sup.1-O.sup.5 area in this example encompasses a part of the damaged area and a part of the reference area.

[0140] In the example of FIG. 8, an operator has identified an area of interest during an inspection. In this example, the working envelope of the ROV is larger than the scanned surface 21 and larger than the working area, which is the external perimeter of all primitives. Therefore only one position of the ROV has to be registered to perform the scan and the repair process in each of the O.sup.1-O.sup.5 areas.

[0141] Each area O.sup.1-O.sup.5 is the top view of a primitive. The shape and dimensions of O.sup.1-O.sup.5 areas depends on primitives available in a library which have been predefined based on excavation or welding limitations, for instance, and prequalified.

[0142] Each primitive can have a cross-section from among the cross sections illustrated on FIG. 6B. Its depth can be set up or adjusted to be the difference between the elevation of the extended reference surface and the local deepest point of the damaged surface at this location (obtained by comparing at least part of first and third geometrical data).

[0143] Thanks to the decomposition of the damaged area into primitives having cross sections in the form of predefined shapes, for example those of FIG. 6B, some steps of the process, for example those described below, can be automatically carried out.

[0144] The terms predefined and prequalified used in the previous paragraphs mean that the excavation, material deposition, and finishing sequences using a ROV system and a robotized repair method according to the present invention may have previously been tested and validated on samples that have the same shapes as the primitives and across a range of dimensions that are available in the libraries. This is a significant advantage of the invention since the use of simple predefined shapes and primitives for which the repair method has been previously tested and qualified enables to carry out some steps automatically, to have a repeatable process, optimizing repair time and significantly limiting the risk of defects resulting from the repair sequence.

[0145] As illustrated on FIGS. 9A to 9D (on which the dimensions indicated are just examples), the entire area of the damage or defect, and preferably of at least part of the surrounding area, may be covered by one or more primitives. For example, a single primitive can be defined for a small damage or defect 20 (FIG. 9A), or a number of n primitives forming a single row, or a plurality of p.sub.r rows each comprising a number of n.sub.i primitives (1ip.sub.r), p.sub.r=1 for FIG. 9B, p.sub.r=2 for FIG. 9C, p.sub.r=3 for FIG. 9D. The operator combines one or more primitives to create a paving covering the entire area to be repaired.

[0146] Depending on the size of the defect or damage, different scanning strategies can be implemented. For an optical sensor, the width of the area that it is able to scan depends on the distance between said sensor and the part. The scanning strategy is therefore dependent on the overall space available around the area to be scanned. Moreover, for very marked topologies, several scanning angles may be necessary to correctly describe the actual facies (peaks or hollow areas).

[0147] As illustrated on FIG. 10A (on which the dimensions indicated are just examples), scanning to measure data to generate one or more of said first to fourth geometrical data can be performed along a plurality of parallel lines with a same sensor 121, which is an effective and rapid way to scan (on FIG. 10A a same sensor is represented at successive positions 121a, 121b, 121c, 121d). Preferably, as illustrated on FIG. 10B, overlapping neighbouring areas can be scanned with a same sensor 121, which has 2 benefits: there is no risk to miss any geometrical data between two neighbouring scanning lines and the overlap enables to geometrically fit the data of the second line with the first line. The pitch of the sensor, from one scanning position to the other, can be determined as a function of the width of the width W (for example 30 mm) of the beam on the surface minus the width of the overlap (for example 1.5 mm and/or 5%).

[0148] As illustrated on FIG. 11, the steps of a repair process, for example, the excavation of the damaged or material deposition in the excavated area or the finishing of the rough repaired area, can follow a predefined process strategy that can be stored in a computer program. Predefined process strategies mean that for a given primitive, the trajectory and geometric parameters of the excavation, material deposition or finishing sequence according to the invention have been pretested and prequalified on samples that have the same shapes as the primitives and across a range of dimensions that are available in the primitives library. The strategies are based on optimisation criteria and specific process requirements. For example, the strategy may have been predefined to improve soundness of deposited material or limit discontinuity. In particular, the material deposition can be optimized from the metallurgical viewpoint (for example thermally and/or to optimize compactness and/or to avoid discontinuities). For example, the tools can follow a meandering path which can be determined or optimised by a computer program. The use of pretested and pre-validated process strategies is another major benefit of the invention. It is efficient, and the excavation, material deposition, and finishing steps can be done automatically, which reduces the risk of creating a defect and is repeatable.

[0149] As explained above in connection with FIG. 7, a control system (for example a computer) 124 can be implemented to perform one or more steps of a process according to the invention.

[0150] Some steps of a process according to the invention can be controlled by an operator, for example through the control system 124; examples of such steps are: [0151] driving the ROV from any access point (for instance manhole in the spiral case or draft tube) to the area of interest [0152] the selection of one or more scanning strategy of the area of interest or part of the area of interest, for example from among predefined strategies stored in a control system memory or the adaptation of such strategy/strategies; [0153] the setting or adjustment of the boundaries of the reference surface (undamaged) and of the damaged area; alternatively, this step can be performed automatically by a control system that is programmed or adapted to automatically delimit undamaged surface(s) and damaged surface(s) and provide a first draft of delimitation to an operator who can then adjust manually; [0154] the calculation of the theoretical former surface by polynomial interpolation of the reference surface using a computer program [0155] the selection of the needed predefined shapes/primitives based on geometrical considerations of the damaged area; alternatively, this step can be performed automatically by a control system (for example system 124, 126) that is programmed or adapted to automatically select relevant primitives/predefined shapes in a library and based on geometrical data; [0156] the definition of a paving of the damaged surface using the selected primitives. The paving subdivides the damaged areas into smaller areas (the primitives having a cross-section in the shape of one of FIG. 6B) that can be repaired using prequalified process strategies. Alternatively, this step can be performed automatically by a control system (for example system 124, 126) that is programmed or adapted to automatically create a paving of the damaged area and provide a first draft of paving to an operator who can then adjust it manually; [0157] the manual generation of an excavation volume based on the selected primitives/predefined shapes, for example using the cross-sections of the damaged area to identify the deepest level for each primitive. The excavation volume may be generated automatically or by the operator by setting/adapting the depth of the predefined shape to the deepest level of the damaged area; [0158] the registration of the ROV location using a dedicated sensor or using a tool tip as a 3D contact control machine. For example, the operator can teleoperate the positioning of the vehicle, but the exact position will be determined based on information captured by a sensor. For example, a laser sensor that makes it possible to determine the position of the sensor in relation to one or more reference point(s) or area(s) on the progression surface, and therefore to derive the position of the vehicle therefrom and by extension of the arm, the tools and the working area in relation to these points, which enables to match the digital model to the physical work area; [0159] the manual selection, setting, adjustment of excavation sequence(s), or the material addition sequence (by welding for instance) for each selected primitive area, for example using predefined or pretested or prequalified strategies stored in the control system 124 memory, for example having a trajectory following zigzags or overlapping parallel lines or centripetal or centrifugal patterns or a contour path, being performed in several layers depending for example on the depth of the damaged area; [0160] the excavation sequence start order or the material addition sequence start order; [0161] the setting of the needed final surface shape; by comparing the fourth geometrical data with the third geometrical data, the operator can set a threshold upon which the finishing operation is carried out; [0162] the manual selection, setting, or adjustment of one or more surface finishing sequence(s) for one or more of the repaired area(s) using predefined, pretested, or prequalified strategies stored in the control system memory, for example having a trajectory following zigzags or overlapping parallel lines or centripetal or centrifugal patterns or a contour path, being performed in several layers depending for example on the depth of the excavated area; [0163] the surface finishing sequence start order; [0164] the repaired surface scanning strategy selection and adaptation, for example based on predefined strategies, for example scanning strategies, which can be stored in the control system memory), for example having a trajectory following zigzags or overlapping parallel lines as in FIGS. 10A and 10B or centripetal or centrifugal patterns or a contour path.

[0165] Some steps of a process according to the invention can be automatically performed (based on an algorithm which comprises the instructions for performing the required steps), for example by the control system 124; examples of such steps are: [0166] the scanning of the damaged area(s) and of the surrounding undamaged area(s) once the scanning strategy has been selected; [0167] the automatic generation of a 3D model including damaged or undamaged area(s); [0168] the automatic generation of a delimitation between reference surface and damaged surface; such delimitation may then be validated or adjusted by an operator; [0169] the automatic calculation of the working envelope in 3D space which for example is based on a combination of several individual working envelopes of the ROV, comprising for example its arm, the tool(s), or the sensor(s); [0170] the automatic generation of a paving similar to FIG. 8 of one or more damaged surfaces using primitives/predefined shapes; such paving may then be validated or adjusted by an operator; [0171] the automatic generation of the cross-sections based on first and second geometrical data that are needed to identify the location of the deepest level of each primitive/predefined shape; [0172] the excavation, the material addition, the surface finishing, or the surface scanning of the repaired area and surrounding unrepaired area; [0173] the automatic generation of a 3D model including repaired and non-repaired areas; [0174] the automatic generation of a delimitation between reference surfaces or repaired area(s), such delimitation may then be validated or adjusted by an operator; [0175] the automatic compliance evaluation based on geometric criteria and tolerances, for example fixed by the operator; [0176] the automatic generation of a preliminary control report.

[0177] Being automatically performed, any of the above steps may be launched by an operator and once launched, such step does not require human intervention.

[0178] A plurality of devices according to the invention can be used at the same time to inspect different locations of a device, for example a runner. For example, one sensor of one such device provides vision with an imaging device, for example a camera, while another one provides light or other wavelength (for example: IR) analysis with another camera.

[0179] An example of a process according to the invention can comprise one or more of the following steps: [0180] a step of scanning the damaged area and the surrounding undamaged area (the so-called reference area); [0181] a step of automatic generation of a 3D model including damaged and undamaged area; [0182] a step of a control system performing an automatic delimitation between reference surface(s) and damaged area(s); [0183] a step of manual setting and/or adjustment, by the operator, of the boundaries between the reference surface s and the damaged area; [0184] a step of calculation of the theoretical former surface (interpolated from the reference surface); [0185] a step of automatic calculation of the working envelop in 3D space by the control system; [0186] a step of an automatic generation of a paving of the damaged area using primitives; [0187] a step of manual setting and/or adjustment of the paving, by the operator, using primitives/predefined shapes; [0188] a step of automatic identification using a control system of the deepest points normal to the interpolated surface in each primitive area by comparison between theoretical former surface and z value for each (x,y) point; [0189] a step of automatic generation of cross-sections based on geometric parameters defined in the control system, for example to locate the highest and/or deepest points for each primitive; [0190] a step of manual selection and/or setting and/or adjustment of the cross sections by the operator; [0191] a step of automatic generation of a preliminary excavation volume based on primitives/predefined shapes selection (for example by, or with the help of, the control system) shown to the operator on the cross-sections (the bottom of the excavation is to be defined as a translation in depth of the extended reference surface down to the deepest level, to ensure the bottom of the excavation is parallel to the reference surface); [0192] a step of manual setting and/or adjustment of excavation shapes and primitives/predefined shapes selection by the operator; [0193] a step of automatic generation of a preliminary excavation sequence for one or more of the primitive areas by the control system; [0194] a step of manual setting and/or adjustment of excavation sequence by the operator; [0195] a step of machining sequence start and/or end of excavation sequence; [0196] a finishing step of material deposition; [0197] a finishing step.

[0198] One or more of the above steps can be repeated, in particular: [0199] the step of scanning of the damage area and the surrounding undamaged area; [0200] the step of automatic generation of a 3D model including damaged and undamaged area; [0201] the step of calculation of the theoretical former surface (interpolated from reference surface); [0202] the step of automatic calculation of the working envelop in 3D space by the control system.

[0203] The process according to the invention can further comprise: [0204] a step of automatic evaluation (by the control system) of the excavation results by comparison between excavated surfaces, for example as defined by the operator, using one or more primitive(s) from a preceding step and resulted excavated surfaces; [0205] a step of manual control by the operator, for example on the cross-sections, that the excavation volumes are compliant with the needs; [0206] a step of quality control of the excavation, based for example on the bottom or the sides of the excavated zone; for example a defect library may have been previously established and excavation performed or can be compared to one or more elements of said library.

[0207] Compared to conventional repair strategies, the invention offers the following advantages: [0208] the time of repair is reduced as it is carried out on site; [0209] the safety of repair is increased as it does not require humans to enter the runner and no handling of heavy loads (no need to dismantle the runner); [0210] the repair can be carried out in constrained areas, where welding by a human would be very difficult; and [0211] the quality of the repair can be improved thanks to the use of prequalified reproducible sequences by a ROV.

[0212] Compared to the known repair strategies implementing a robot: [0213] the robot can access all areas of the runners (this includes inter-blade areas as low as 150 mm); [0214] the robot can work on smaller runners (surfaces with minimum curvature radius of 10 cm); [0215] the robot is well anchored on curved surfaces or can handle surface transitions to navigate from the draft tube or spiral case to the defect area (using remote control); the welding can be more precise and faster as it can be at least semi-automated.