Impeller manufacturing method

Abstract

The present invention relates to a method for manufacturing an impeller, in particular an impeller for a turbomachine, said impeller comprising vanes forming channels (540) therein, the method comprising the steps of: a) forming a raw impeller part (510a) by additive manufacturing; b) removing from said raw impeller part (510a) by post-processing material in a first area (551) and a second area (552) of said inner surface, at an inner surface of at least one channel, to obtain an intermediate impeller part; and c) removing from said intermediate impeller part by machining material in said first area (551).

Description

[Technical field]

[0001] The present invention relates to a method for manufacturing a closed impeller for a radial turbomachine. [Background technology]

[0002] Turbomachines such as compressors, expanders, and pumps can be used in a number of different applications. For example, cryogenic applications, i.e. applications using very low temperature process gases, such as air separation plants, often use cryogenic turbomachines such as turboexpanders and / or turbocompressors. Such turbomachines typically include an expander impeller and / or a compressor impeller fixed to a shaft.

[0003] The impellers of such turbomachines can have an open or closed design, in which the impeller vanes form channels. The advantage of a closed design is that the impeller has a higher isentropic efficiency. Under certain process conditions, i.e. often very low specific speeds and high rotational speeds, the ideal geometry of such channels, which would provide the highest efficiency, cannot be machined by conventional subtractive techniques. In these cases, the impeller must remain open (i.e. the impeller is of open design), with associated efficiency losses. It is therefore an object of the present invention to provide an improved impeller. A similar process has already been described in EP 3281728 or US 2019 / 0134779. DISCLOSURE OF THEINVENTION

[0004] This object is achieved by providing a method for manufacturing an impeller.Embodiments of the invention are the subject of the dependent claims and the following description.

[0005] The present invention relates to the manufacture of closed impellers for radial turbomachines, such as expanders, compressors or pumps. Such impellers comprise vanes forming channels (for the working medium) within the impeller.

[0006] As mentioned above, the ideal geometry of such channels that results in the highest efficiency is typically not machineable by conventional subtractive techniques. In these cases, the impeller must remain open, i.e., it is an open design. Additive manufacturing techniques make it possible to build closed impellers, either by supporting the downskin area with anchoring structures when necessary, or by using even more difficult techniques that allow printing the downskin surface without supports up to an angle of 0° with respect to the horizontal. However, after removing the supports, or even in the absence of any supports at all, the surface roughness of the inside of the channel may deviate significantly from the expected norm, and the efficiency gains resulting from closing the impeller may be lost due to the surface roughness, as has been found.

[0007] Post-processing techniques can improve the impellers obtained by additive manufacturing, but not to the level of milled components. In addition, surface roughness has been found to have a very detrimental effect on the fatigue life of additive printed components. The fatigue life of an as-printed component typically needs to be ten times smaller than that achieved by the same part machined with, for example, Ra=1.6 μm.

[0008] Furthermore, closed impellers have been found to be subject to vibration modes in their operating speed range. When these modes are excited, some areas inside the channels, mainly at the surface, are subjected to very high cyclic loading. It is therefore unacceptable to maintain a surface roughness less than that provided in the as-milled condition in areas subject to fatigue loading.

[0009] A method for manufacturing an impeller in multiple steps is then proposed. First, a raw impeller part is formed by additive manufacturing. For example, so-called powder bed fusion additive manufacturing can be used. Then, at the inner surface of at least one channel, material is removed from said raw impeller part in a first area and in a second area of ​​said inner surface by post-treatment to obtain an intermediate impeller part. Such post-treatment can comprise at least one of the following techniques: chemical surface treatment, physical surface treatment, mechanical surface treatment or electrical surface treatment. Then, material is removed from said intermediate impeller part in said first area by machining. The first area is preferably located closer to the opening of the channel than the second area. It is preferable to carry out step b) first and then step c), but they can also be carried out in a different order.

[0010] This allows closed impellers with very long, thin and twisted channels that cannot be machined by any subtractive technique alone to be manufactured without compromising isentropic efficiency and fatigue life. The first zone, which is the majority of the channel, can be machined, but the non-machinable parts (the first zone) can be provided with a sufficiently smooth surface by post-processing of the additively manufactured part. It is noted that the second zone can also be defined to include surface parts that are machineable, but may only be machined by very complex machines, etc.

[0011] Preferably, after step b), the inner surface in the first and second zones has a roughness R of 3 to 10 μm, preferably 3 μm. a 1.

[0012] After step c), the inner surface in the first zone has a roughness Ra2 which is smaller than Ra1 and preferably equal to or less than 1.6 μm.

[0013] An advantageous method of forming the raw impeller part is such that the raw impeller part includes, at the inner surface of the at least one channel, excess material that exceeds the design contour of the inner surface, the excess material extending into the channel and forming the actual contour of the inner surface (as opposed to the design contour), such that in the second zone, the thickness of the excess material is smaller than the thickness in at least a part of the first zone. The design contour is the contour or geometry that the channel or its surface will have after all manufacturing steps are completed. To achieve a suitable thickness of the excess material, a model for additive manufacturing can be adjusted to include the excess material relative to the actual desired surface geometry.

[0014] The thickness of the excess material is preferably uniform within the second region such that the desired design contour of the inner surface within the second region is achieved, typically in a post-processing step that uniformly removes material from the surface.

[0015] Preferably, in an intermediate region, which is part of the first region and connects to the second region, the thickness of the excess material changes from the thickness in the second region to the thickness in the first region (outside the intermediate region). In particular, the transition of the thickness of the excess material at the boundary between the second region and the intermediate region is tangential to the design contour of the inner surface. Similarly, the transition of the thickness of the excess material at the boundary between the intermediate region and the remaining part of the first region may be tangential to the design contour of the inner surface. In other words, the actual contour formed by the excess material at the boundary has a distinguishable or differentiable form. This avoids abrupt connections between the regions after machining. At least a part of the excess material remaining in the intermediate region may also be removed by machining.

[0016] The proposed method is particularly applicable to materials that, when printed (i.e., additively manufactured), have many boundary voids and poor surface finishes. Boundary voids are highly detrimental to fatigue life. Poor surface finishes are detrimental to the fatigue life and isentropic efficiency of the impeller.

[0017] Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings.

[0018] It should be noted that the features mentioned above and those further described below can be used not only in the respective indicated combinations but also in further combinations or alone, without departing from the scope of the present invention. [Brief description of the drawings]

[0019] [Figure 1] 1 illustrates a turbomachine having an impeller according to a preferred embodiment of the present invention; [Diagram 2] 3 shows an impeller according to a further preferred embodiment of the present invention. [Diagram 3] 1 shows an impeller channel having multiple different sections. [Figure 4] A method according to a preferred embodiment of the present invention is illustrated in a flow diagram. [Figure 5a] 1 shows an impeller channel having multiple different sections. [Figure 5b] 1 shows an impeller channel having multiple different sections. [Figure 5c] 1 shows an impeller channel having multiple different sections. [Figure 5d] 1 shows an impeller channel having multiple different sections. [Figure 5e] 1 shows an impeller channel having multiple different sections. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Fig. 1 shows a schematic representation of a turbomachine 100 according to a preferred embodiment of the invention. The turbomachine 100, for example a cryogenic turbomachine, is configured as a compressor and an expander, i.e. both are combined in one turbomachine. The turbomachine 100 thus comprises two impellers, impeller 110 and impeller 120, both mounted on a shaft 130. The turbomachine 100 comprises channels 112 and 114 on the side of the impeller 110, which are used as inlet and outlet channels, respectively, for a working medium or working fluid (see arrow 113) that is compressed and then withdrawn (see arrow 115). The turbomachine 100 further comprises channels 122 and 124 on the side of the impeller 120, which are used as inlet and outlet channels, respectively, for a working fluid that is expanded. The impeller 110 is thus a compressor impeller and the impeller 120 is an expander impeller.

[0021] Fig. 2 shows diagrammatically an impeller 210 according to a further preferred embodiment of the invention. Whereas the impellers 110, 120 of Fig. 1 are shown diagrammatically, the impeller 210 is shown as a closed impeller, i.e. an impeller of closed design. The impeller 210 comprises a number of vanes 234 surrounded by two shrouds or plates 230 and 232. In this way, a channel 240 is formed between two of such vanes and the plate. As can be gathered from Fig. 2, the machining of such a channel 240 may be complicated or even impossible altogether, particularly in the inner area.

[0022] 3 shows a schematic representation of a channel 340, which may be one of or similar to the channels 240 shown for the impeller 210 of FIG. 2. The channel 340 is shown in cross section and top view of the channel. An inner surface of the channel 340 is indicated by reference numeral 341. As previously mentioned, some parts or areas of such an inner surface may not be machineable due to the particular geometry of the channel 340. Such areas that are not machineable are indicated (encircled) at 352. However, other areas outside of the area 352, such as area(s) 351, are machineable.

[0023] Additionally, the impeller can be analyzed to determine vibration modes and locate areas subject to fatigue loading when one of the critical frequencies is excited. The area that may be subject to high fatigue loads is indicated (encircled) by reference number 353 in Figure 3. If area 353 overlaps area 352, some fatigue problems may occur when the impeller is in operation. If areas 352 and 352 are completely separated as in the schematic diagram of Figure 3, this ensures that no fatigue problems will occur.

[0024] Figure 4 shows by means of a flow diagram a method according to a preferred embodiment of the invention. The method comprises several different steps. In step 400, a raw impeller part is formed by additive manufacturing. In step 402, material in a first area and a second area on the inner surface of at least one channel is removed by post-processing from said raw impeller part to obtain an intermediate impeller part. In step 404, material in said first area is removed by machining from the intermediate impeller part.

[0025] These steps are explained in more detail below with reference to Figures 5a-5c, which show an impeller channel with several different zones. Figure 5a shows a raw impeller part 510a, which is for example additively manufactured according to a 3D raw model. Such a raw model and the resulting raw impeller part differ from the final or design impeller or the individual model, since extra material will be added to the inner surface, preferably to all areas of the inner surface. Figure 5a shows an impeller channel 540, which may for example essentially correspond to the channel 340 shown in Figure 3. The design contour of the inner surface of said channel 540 is shown by the line 541.

[0026] The raw model according to which the raw impeller part 510a is additively manufactured includes excess material 560 at the inner surface of the channel 540 beyond the design contour 541. The excess material 560 extends into the channel. Thus, the actual contour 561 of the inner surface of the raw impeller part 510a differs from the design contour 541. It should be noted that such excess material is typically the same as the material of the rest of the impeller. The design contour 541 shown in FIG. 5a is not visible or palpable in the actual raw impeller part. Rather, this design contour 541 corresponds to the final inner surface of the channel after all manufacturing steps are completed.

[0027] The thickness of said excess material 560 is generally indicated with the reference d. Depending on the area of ​​the inner surface, this thickness is different. In the second area 552 (not machineable, note that there may be several first areas), the thickness is smaller than at least a part (far from the second area) of the first area 551 (machinable). The excess material in such area 552 corresponds to the excess material having a thickness that will be removed by a post-processing step. The excess material in such first area 551 corresponds to the excess material having a thickness that will be removed by a post-processing step, plus the excess material that will be removed by a machining step. The thickness of such excess material added in said first area 551 is preferably at least 1 mm, but does not completely close the thinnest entrance of the channel. This thickness will be partially reduced during the post-processing step and will finally be completely removed by the machining step.

[0028] In the intermediate region 555, i.e. the intermediate region 555 which is part of said first region 551 (i.e. the intermediate region is machinable) and which connects to the second region 552 (there may be several such intermediate regions), the thickness of the excess material changes from a smaller thickness d2 as in the second region 552 to a larger thickness d1 in the first region 551 (far from the second region and outside the intermediate region). Such a gradual transition between the thickness of the second region 551 and the thickness of the first region 551 is shown in Figure 5a and in more detail in Figure 5b.

[0029] The extra material (or thickness layer) in said intermediate region 555 shall preferably have a particular shape. The connection between the thickness in said intermediate region 555 and the thickness in said second region 552 at the boundary between the second region and the intermediate region shall preferably be tangential. A tangential connection shall specifically mean that the actual profile of the inner surface in this region has a differentiable form. Also, the connection between the intermediate region 555 and the remaining part of the first region 551 shall preferably be tangential, as indicated by the double arrow in Fig. 5b.

[0030] The transition from the lower to the higher thickness of the excess material must begin near or at the boundary of the second region 552 (which is not machineable), otherwise there may be a risk of an abrupt connection between the first region 551 and the second region 552 after machining. The additive manufacturing is preferably carried out without supports wherever possible. Printing techniques that allow a 0° downskin angle with respect to the horizontal are preferred. This process can also include stress relief and / or hardening heat treatments and / or HIP (hot isostatic pressing) treatments that will be performed on the geometry of the raw impeller part.

[0031] After the additive manufacturing process 400, the excess material beyond the design contour 541 will be removed (as much as possible) in the next steps 402, 404. To account for that, we adopt a new expression for the offset (excess material) depending on the ablation process, rather than on the position within the channel.

[0032] This is shown in Figure 5c. The excess material in the first and second zones and in the intermediate region is now divided into excess material for post-processing 562 (shown in bold) and excess material for machining 563. Excess material 562 and excess material 563 represent, respectively, material that is removed by post-processing and material that is removed by machining, e.g., 5-axis milling or EDM or any other subtractive machining technique.

[0033] The (surface) post-treatment step 402 can be achieved by any technique of surface post-treatment, such as chemical surface treatment and / or physical surface treatment and / or mechanical surface treatment, or any combination of these techniques.

[0034] A calibration test can be carried out to set the process parameters that will guarantee the removal of the entire material 562. The final or design contour 541 must already be present in said second area 552 after the post-treatment that removes the material 562. This can be seen in FIG. 5c. The calibration can be carried out, for example, on a simplified structure that reproduces the geometry of one channel. Such a calibration can determine, for example, what type of post-treatment needs to be applied and for how long in order to remove the material 562 as accurately as possible.

[0035] It is not a problem if some material 562 remains after this post-processing step, so long as it is only in the first region 551. Conversely, it is not a problem if a small portion of material 563 is removed during the post-processing step, so long as the final contour is maintained throughout the second region 552.

[0036] At the end of this step 402, an intermediate impeller part 510b is obtained and the channels will look like as shown in Figure 5d. The surface finish of the second section 552 will ideally be less than Ra = 10 μm.

[0037] Machining, step 404, is performed on the entire first area 551 including the intermediate region 555 to remove excess material 563, for example by conventional 5-axis milling, so that the target Ra is the same as specified for the machining operation.

[0038] The connection or boundary between the first and second area is important in terms of geometric accuracy. There should be no significant step between the surface remaining after post-processing and the surface remaining after machining. If such a step exists, it means that the post-processing step was not performed properly and that the calibration of the thickness removed by the post-processing needs to be re-adjusted.

[0039] As shown in Fig. 5e, the post-processing contour 562 is somewhat rough (see upper diagram). The connection or boundary (see arrow 570) between the first and second zones needs to be tangential (differentiable) so that the drill does not remove any more material at the end of its path inside the channel. Thus, the path for machining can be equal to the design contour 541. After the machining process, the surface in the first zone becomes smooth (see lower diagram).

[0040] The entire external contour of the impeller can then be machined so that finally the only non-machined area is the second area which will remain due to the surface treatment caused by the post-treatment method.

[0041] The present invention allows the manufacture of closed impellers where the entire channel cannot be machined. The surface finish in the areas critical for fatigue loading will be the same as that of a fully machined impeller, so no shortening of life will occur. The surface finish of the internal channels will be, for the most part, the same as that of a machined impeller, except in the inaccessible areas that will exhibit the surface finish of the post-treatment method, so the isentropic efficiency of the closed impeller will be much better than that of the same open type.

[0042] The invention applies in particular to all impellers of radial turbines, compressors or pumps, where the only limitation for installing a closed impeller is that the shape of the channel is too curved and twisted to machine the central part of the channel.

Claims

1. A method of manufacturing a closed impeller (110, 120, 210) for a radial turbomachine (100), the impeller having vanes (234) forming channels (240, 340, 540) therein, the method comprising: a) forming a raw impeller part (510a) by additive manufacturing (400); b) removing (402) from said raw impeller part (510a) by post-treatment material in a first area (551) and a second area (552) of said inner surface of at least one channel to obtain an intermediate impeller part (510b), said post-treatment comprising at least one of the following techniques: chemical surface treatment, physical surface treatment, mechanical surface treatment, electrical surface treatment; c) removing (404) material from the intermediate impeller part (510b) in the first area (551) by machining, wherein the second area is treated only by step b); A method comprising:

2. After step b), the inner surface in the first area and the second area has a roughness R of 3 to 10 μm, preferably 3 μm. ａ 10. The method of claim 1, wherein

3. After step c), the inner surface in the first area has a surface roughness of R ａ A roughness R less than 1, preferably 1.6 μm or less ａ 2. The method of claim 1 , wherein

4. The method of claim 1, wherein the first area (551) is located closer to the opening of the channel than the second area (552).

5. 2. The method of claim 1, wherein in step a), the raw impeller part (510a) comprises, on the inner surface of the at least one channel, excess material (560) that exceeds the design contour (541) of the inner surface, the excess material extending into the channel and forming the actual contour (561) of the inner surface, and wherein in the second region (552) a thickness (d) of the excess material is smaller than the thickness in at least a portion of the first region (551).

6. The method of claim 5, wherein the thickness (d2) of the excess material is uniform within the second region (552).

7. 7. The method according to claim 5 or 6, wherein in an intermediate region (555) that is part of the first region (551) and connects to the second region (552), the thickness of the excess material varies from a thickness in the second region (552) to a thickness in the first second region (551) outside the intermediate region (555).

8. 8. The method of claim 7, wherein the thickness of the excess material is uniform within the first area (551) outside the intermediate region (555).

9. 8. The method of claim 7, wherein the transition in thickness of the excess material at the boundary between the second section (552) and the intermediate region (555) is tangential to the design contour (541) of the inner surface, and / or the transition in thickness of the excess material at the boundary between the intermediate region (555) and the remainder of the first section (551) is tangential to the design contour (541) of the inner surface.

10. 9. The method of claim 8, wherein in step c), at least a portion of the remaining excess material in the intermediate region (555) is removed by machining.

Images