The present invention relates to a method for producing or repairing a turbomachine part by means of a laser beam. It also relates to a turbomachine part produced using this method.
In such a method, also known as Laser Metal Depositio, a nozzle sprays a metal powder towards a substrate so as to produce the part by means of successive depositions, one on top of the other, of layers or beads, in a deposition direction. The powder considered, which typically is a mixture of metal powders, is molten by the Laser beam. The thickness of each bead typically ranges from 0.05 mm to 1 mm. Low thickness should be preferred if the surface condition is to be optimized.
This method makes it possible to produce large dimension parts, specifically as regards height (Z axis in an orthonormal X, Y, Z system). However, it is difficult to obtain guaranteed final dimensions of the part (or portion of the part) thus produced.
Variations in the powder rate, the reloading speed of the spray nozzle, the laser power or the temperature of the part, may cause variations in dimensions, specifically variations in the height of the deposited layers. Locally, such variations may have little impact; but these may increase when and as the layers are deposited and even cause an unstable deposition resulting in the generation of a saw-teeth like deposition.
One aim here is to provide a solution to all or part of the drawbacks above.
For this purpose, the proposed method for producing or repairing by spraying metal powder using a laser beam is characterized in that:
Time-setting the initial production routine through one or more modification(s) in the initially defined trajectory of the nozzle will make it possible to be closer to the expected final dimensional characteristics of the part.
A better surface condition can thus be expected, without saw-teeth being as visible as before.
As it affects one or more geometrical parameter(s), this solution further avoids having to modify process parameters: laser power, reloading speed between two successive depositions of layers, powder rate, . . . .
In order to be able to guarantee that the height of the thus deposited material is the expected one, it is recommended that the regulation test implemented should:
When implementing the above, it has been found that it could be more efficient not to modify the distance between the nozzle and the top of the layer opposite thereto, and rather, preferentially,
Similarly, it is provided:
Modifying, over time, the number of steps when successively depositing layers and/or the number of layers still to be deposited will a priori be easier and safer to be implemented and to be controlled than varying the thickness of such layers, during the production of the part.
In connection therewith, it is additionally proposed:
Keeping such determined distance and/or conditions of deposition aiming at depositing stacked layers having the same thickness will enhance the stability of the solution.
One advantage of the solution also relies in a quick production. Besides, it was found that waiting for some time was necessary for the uncertainties of accuracy in measurements to be low enough.
This is the context wherein it is provided that, during the production of the part, —said real distance should be measured,
Preferentially, it is recommended during the production of the part:
Another problem to be solved was defining how to obtain the above-mentioned distance data.
A solution using a measuring autofocus camera has been preferred.
It has thus been proposed to use a camera with an autofocus system to obtain the theoretical reference distance and the real distance.
Another corollary problem to be solved was obtaining reliable measures, which are not dependent on the conditions of the creation of the molten bath on the part being produced, opposite, along the laser axis.
Specifically when the laser beam is emitted along said Z axis, measuring the theoretical reference and real distances away from said Z axis has thus been preferred, in parallel to this axis or at an angle (A) having a projection parallel to said Z axis.
It will thus be possible to prevent the camera from aiming at the molten bath.
The invention will be better understood and other details, characteristics and advantages of the invention will appear upon reading the following description given by way of a non-restrictive example while referring to the appended drawings wherein:
The substrate 5 is a conventional support in the field, and adapted to successive layers 111, 112, 113, . . . , 11i . . . of sprayed material being deposited thereon (
Hereunder, we will disclose a situation where, as illustrated, the nozzle 1 sprays, in a layer deposition direction, here (substantially) vertically, along a Z axis, the metal power 3 towards the substrate, for making the height (portion) of the expected part.
The laser beam 13 is thus emitted toward the substrate, along the Z axis, and, in this case, the theoretical reference and real distances mentioned above will be measured in parallel to this axis or at an angle A (
The method developed here could however be implemented along either one of the other two X, Y axes of the conventional orthonormal X, Y, Z system (
In the preferred case shown: the nozzle 1 comprises two concentric cones 16a, 16b coaxial with the Z axis.
Originating from a laser source 17, and if need be, using a mirror 19, the laser beam 13 is emitted vertically towards the substrate 5 at the center of the central cone 16a.
The mixture 3 of the metal powders 3a, 3b circulates in the external cone 16b, and it is sprayed out of same, downwards, toward the substrate 5, via a carrier gas 21b. Another gas 21a surrounds the beam 13 in the internal cone 16a.
The deposition of the material resulting from the mixture 3 may be uneven because of the laser beam 13. For instance, specifically if the two cones are no longer centered, more material may be spread on one side than on the other one.
Metal powders may be titanium alloys (TA6V, Ti71, 6242, . . . ), nickel and cobalt based alloys (Inco718, Hastelloy X, René77, René125, HA188) and steels (Z12CNDV12, 17-4PH).
The successive depositions of the layers 111, 112, 113 . . . 11i . . . of materials will thus be stacked on the substrate 5, until the expected part 7 is obtained.
The block diagram of
The following steps are then carried out successively in a sequence, during the production of the part:
Two options then exist:
In the mean time, a test step 39 or 41, respectively, has been carried out, with two options again:
If step 42 or 44 has been reached, it means that at least one step of deposition is still to be carried out, and that a return, i.e. a loop is provided, in both cases, on one of the lines 50, again at step 33, to repeat the steps 33 to 39 or 41 a certain number of times, and so periodically reset the manufacturing program 34 if necessary, and adapt, in real time, the trajectory of the nozzle while measuring a real distance Di at each step 33.
As regards the modification in the trajectory at step 42, acting on the steps of forming the layers, and specifically on the number of layers still to be deposited has been preferred.
Specifically, it has been understood from the foregoing that the production of the part 7 is obtained by moving, in successive steps, the nozzle 1 away from the substrate 5, and when a layer has been deposited, from the previous layer 11i of deposited material, when and as the layers are stacked.
The definition and the storing into the memory 29, in step 31, of the predetermined trajectory of the nozzle 1 will thus preferably include those of a predetermined number of such steps of depositions.
And it will be possible to modify the trajectory of the nozzle 1 by changing the predetermined number of steps.
In practice, it is recommended that the execution of one of the above layers 11 should correspond to one step of deposition.
Thus:
In this case, the trajectory of the nozzle will then be modified in the memory 29 by substituting said predetermined number of layers to be deposited with the modified number nc of layers 111, . . . 11i still to be deposited, with nc being a positive or negative integer.
In this respect, the above-mentioned threshold (D0−Di) leading to the step 39 or 41 will advantageously be equal to the thickness of a layer 11i, i.e. typically 0.1 mm.
Then, if the distance Di is shorter than D0 by more than 0.1 mm, for example 0.2 mm, the program will add two steps of deposition, i.e. two layers. But it will remove three layers if the reading and the calculation indicate a distance Di of +0.3 mm as compared to D0. And there will be no modification in the steps if the reading and the calculation indicate a distance Di shorter or longer by less than 0.1 mm as compared to D0.
As mentioned above, in the illustrated example, this is according to the height of the part 7, (in particular) along the Z axis (or substantially parallel thereto):
In practice, it is recommended that the/each “skip” mentioned above in the reloading program should be executed at the Z coordinate measured (D1 . . . Di) for X and Y values close to those where the nozzle 1 is located when its trajectory is reset. A restart routine can then also be used to manage the laser power and/or the reloading speed accordingly.
In such a situation of production controlled along the Z axis, it is also recommended that the expected completion of the part 7 should be performed by successive depositions of layers 35a, . . . 35i all having the same thickness e, on top of each other, by moving the nozzle away from the substrate, for each layer, here along the Z axis.
This will simplify the control of the correct progress in height of the part and will further avoid creating other surface irregularities (above-mentioned saw-teeth).
Preferably, this search for a relatively simple control of the compliance with the dimensional constraints of the part will also concern a limitation of the measures D1 . . . Di and/or the modifications in the trajectory of the nozzle.
Thus it is recommended that, during the production of the part, the real distance Di should be measured and the trajectory modified only after the deposition of several layers 11i, if said deviation is reached.
In this respect, it may in particular be provided to perform several successive measurements of said real distance and (if said deviation is reached) to change the trajectory of the nozzle 1 only after the deposition of several layers with respect to the previous measurement.
In
It is therefore preferably by using the autofocus that the initial distance D0 (when no layer has been deposited yet) as well as the real positions of the optical system 15a of the lens of the camera 15 (zone 1 above), will be calculated and recorded, by aiming at the surface 35i of the last deposited layer 11i. Once the image is sharp thanks to the autofocus, the position of the nozzle relative to the part can thus be deduced.
The shootings will provide measurements parallel to the laser beam 13 directed towards the substrate, i.e. along the Z axis (or substantially parallel thereto).
Referring to the above explanations, the device will then operate as indicated hereunder, in connection with such distance measurements:
initially, as shown in
It should thus be understood that such reloading distance Dc will preferably be preserved at each step of the part production, with the nozzle moving away from the thickness e of one layer, for each deposited layer 11i.
As a matter of fact, it is recommended that, if such determined distance Dc between the substrate and one end of the nozzle facing the substrate matches the theoretical reference distance D0, this distance Dc should be maintained during the production of the part, preferably when starting depositing each layer.
With the nozzle being set at the distance Dc, a clear image of the free surface 5a of the substrate 5 taken by the camera at this initial time, using the autofocus 45, adjusted accordingly, will thus define the theoretical reference distance D0. The autofocus 45 is then preferably calibrated.
Then, the real distances D1, D2, . . . Di will then, as explained above, successively measured during the production of the part, with the nozzle still being a priori so positioned as to comply with the reloading distance Dc, using the autofocus 45 adjusted accordingly and as shown in
To prevent the camera from aiming at the molten bath, along the Z axis, which might make the focusing of the camera inaccurate, specifically when the surface receiving the laser beam is melting, the Z0 axis is here shifted aside (distance e1).
In connection therewith, two operating modes are possible:
In addition to the fact that it could have been produced using the technique disclosed above, the vane 7 could also have been repaired, in case of wear. Reference 53 also refers to a section plane of such blade intended to be replaced. The free end surface of the segment 7a of the blade still in position would define the above-mentioned surface 5a of the substrate 5.
Number | Date | Country | Kind |
---|---|---|---|
1463266 | Dec 2014 | FR | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/FR2015/053658 | 12/18/2015 | WO | 00 |