Method for regulating and controlling forming quality of LPBF nickel alloy inclined thin-wall structure

Abstract

The invention discloses a method for regulating and controlling the forming quality of an LPBF nickel alloy inclined thin-wall structure, and the method comprises the following steps: S1, selecting different laser powers to carry out first LPBF printing, and printing thin-wall structures with different inclination angles; s2, measuring the lower surface roughness and density of all the samples, and recording the corresponding laser power with the lowest lower surface roughness and the highest density; s3, a prediction model is constructed, optimal process parameters of other inclination angles are obtained through machine learning, and an inclined thin wall forming quality database is formed; and S4, the optimal laser power is obtained according to the prediction model, and second-time LPBF printing is conducted according to the optimal laser power. The problem that the quality of the lower surface of the nickel alloy inclined thin-wall structure prepared through LPBF is poor can be solved, cost is low, efficiency is high, and practicability is high.

Description

Technical Field

[0001] The invention relates to the technical field of metal laser powder bed melting, and in particular to a method for regulating the forming quality of an LPBF nickel alloy inclined thin-wall structure. Background Art

[0002] Thin-walled structures are a typical lightweight structure used in aerospace applications. They have the characteristics of high material utilization and meet the needs of lightweight aircraft engines. The quality and strength of key thin-walled components such as the outer shell and cabin of a spacecraft directly affect the performance and safety of the spacecraft. Traditional methods for processing thin-walled structures include milling and plastic forming, but the tool wear of formed thin-walled components is severe, the efficiency is low, and the cost is high. Laser powder bed fusion technology (LPBF) is a laser additive manufacturing technology that uses high-energy lasers to melt metal powders. The thin-walled structures formed by LPBF have high surface integrity, high efficiency, and low cost.

[0003] IN718 is the most widely used nickel-based high-temperature alloy in aerospace applications. It is widely used in thin-walled structural components such as casings and combustion chambers. The weldability of IN718 makes it suitable for LPBF forming. In the process of LPBF forming IN718 thin-walled structures, splashing and Marangoni effect will inevitably occur, and due to the combined action of gravity and capillary force, a "step effect" will be generated, which makes the roughness of IN718 thin walls formed by LPBF high, and there are problems such as powder adhesion and slag on the surface, which affects the use of IN718 thin-walled components. In order to solve the problem of high surface roughness under thin-walled structures, the forming angle of thin-walled structures can be improved by optimizing the support structure to improve the forming quality. However, removing the support is a complex and tedious process, which will cause some slag to appear on the thin-walled structure of the body close to the support. At present, domestic companies have realized the laser powder bed fusion support-free printing technology of thin-walled components, solving the adverse effects of the support structure. In addition, there is post-processing. Post-processing includes heat treatment, surface modification, advanced finishing and laser polishing, but the post-processing process is more complicated, takes a long time and has lower efficiency. Compared with other inventions, the present invention can optimize the forming quality of thin-walled structures with different inclination angles. The optimized laser power used in the present invention combined with machine learning to obtain a database of inclined thin-walled structure forming quality is an efficient method for improving the surface quality of thin-walled structures with different inclination angles. By exploring the thin-walled structure forming quality process in the early stage and expanding it through machine learning to obtain the optimized laser power for other inclination angles and forming a database, the surface roughness of inclined thin-walled components can be effectively reduced, the laser power optimization at different inclination angles can be achieved, the forming quality of the lower surface can be improved, and the density can be improved and the surface powder adhesion and slag can be reduced while improving time efficiency. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method for regulating the surface forming quality of LPBF formed nickel alloy inclined thin-walled structures, so as to improve the surface forming quality of nickel alloy inclined thin-walled components.

[0005] Technical solution: A method for regulating the forming quality of LPBF nickel alloy inclined thin-walled structures, comprising the following steps:

[0006] S1, select different laser powers for the first LPBF printing to print thin-walled structures with different tilt angles;

[0007] S2, measure the lower surface roughness and density of all samples, and record the laser power with the lowest lower surface roughness and the highest density;

[0008] S3, construct a prediction model, obtain the optimal process parameters for the remaining tilt angles through machine learning, and form a tilt thin-wall forming quality database;

[0009] S4, obtaining the optimal laser power according to the prediction model, and performing the second LPBF printing according to the optimal laser power.

[0010] Furthermore, the laser power ranges from 260W to 340W.

[0011] Furthermore, the selected tilt angles are 0°, 15°, 30°, 45° and 60°, respectively, and for each tilt angle, laser powers of 260 W, 280 W, 300 W, 320 W and 340 W are applied, respectively.

[0012] Furthermore, all samples with measured surface roughness were polished on the side with 180#, 400#, 800#, 1200#, 1500# and 2000# sandpaper in turn until the center of the sample was polished with diamond polishing liquid, and their density was calculated using ImageJ combined with binary image processing.

[0013] Furthermore, a fourth-order polynomial regression equation is used to obtain the laser power prediction set at other angles through interpolation.

[0014] Compared with the prior art, the present invention has the following significant effects:

[0015] 1. Improve time efficiency, no need for post-processing, simple and efficient, reduce surface roughness while reducing time cost, and the density of the formed thin-walled components is improved, and it can be used for almost all material inclined thin-walled components. Although the roughness of the optimized thin-walled structure is still somewhat different from industrial demand, the present invention proposes a method for improving the forming quality of thin-walled samples for reference.

[0016] 2. No post-processing such as surface processing, abrasive flow processing and other additional processes and additional tools are required, which reduces complex process flow and improves forming efficiency.

[0017] 3. The forming quality database of inclined thin-walled structures is obtained by machine learning, which can optimize the forming quality of thin-walled structures with different inclination angles. The optimized laser power at each inclination angle is predicted by reducing the amount of experiments, and the model predicts that the roughness of the thin-walled structure under the optimal laser rate parameters is lower than that of the embodiment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the overall flow chart of the present invention;

[0019] Figure 2 Schematic diagram of the lower surface roughness with a tilt angle of 0° before the process control;

[0020] Figure 3 Schematic diagram of the lower surface roughness with a tilt angle of 15° before the process is controlled;

[0021] Figure 4 Schematic diagram of the lower surface roughness with a tilt angle of 30° before the process is controlled;

[0022] Figure 5 Schematic diagram of the lower surface roughness with a tilt angle of 45° before the process is controlled;

[0023] Figure 6 Schematic diagram of the lower surface roughness with a tilt angle of 60° before the process is controlled;

[0024] Figure 7 Schematic diagram of the lower surface roughness with a tilt angle of 0° after the control process;

[0025] Figure 8 Schematic diagram of the lower surface roughness with a tilt angle of 15° after the control process;

[0026] Fig. 9 Schematic diagram of the lower surface roughness with a tilt angle of 30° after the control process;

[0027] Fig.10 Schematic diagram of the lower surface roughness with a tilt angle of 45° after the control process;

[0028] Fig.11 Schematic diagram of the lower surface roughness with a tilt angle of 60° after the control process;

[0029] Fig.12 This is a light microscope image with a tilt angle of 0° before the control process;

[0030] Fig.13This is a light microscope image with a tilt angle of 15° before the control process;

[0031] Fig.14 This is a light microscope image with a tilt angle of 30° before the control process;

[0032] Fig.15 This is a light microscope image with a tilt angle of 45° before the process is controlled;

[0033] Fig.16 This is a light microscope image with a tilt angle of 60° before the control process;

[0034] Fig.17 This is the optical microscopy image with a tilt angle of 0° after the control process;

[0035] Fig.18 This is the optical microscopy image with a tilt angle of 15° after the control process;

[0036] Fig.19 This is the optical microscopy image with a tilt angle of 30° after the control process;

[0037] Fig. 20 This is the optical microscopy image with a tilt angle of 45° after the control process;

[0038] Fig.21 This is the optical microscopy image with a tilt angle of 60° after the control process;

[0039] Fig. 22 This is a light microscope image with a tilt angle of 20° before the control process;

[0040] Fig.23 This is a light microscope image with a tilt angle of 20° after the control process. DETAILED DESCRIPTION

[0041] The present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0042] The present invention proposes a method for regulating the surface forming quality of a nickel alloy inclined thin-walled structure based on an optimized scanning strategy method, which can effectively improve the surface forming quality of a nickel alloy thin-walled structure. First, a certain range of laser powers are used for the first printing to print several thin-walled structures with common inclined angles, in order to explore the laser power with better forming quality of the lower surface at a certain inclined angle, grind and polish the inside of the thin-walled structure, and measure its density using an optical microscope and a binarization method. Secondly, the optimal process parameters for the remaining inclined angles are obtained through machine learning, forming an inclined thin-wall forming quality database, and obtaining the optimal process parameters for thin-walled structures with different inclined angles under the condition of the highest density. The flow chart of the present invention is as follows: Figure 1 , the main steps are as follows:

[0043] Step A1, the first LPBF printing, explores the parameter window of the lower surface forming quality of the thin-walled structure of nickel alloy with several typical tilt angles, and obtains the laser power parameter with good lower surface forming quality while ensuring good internal density;

[0044] Step A2, using the parameters and roughness values ​​obtained in step A1, interpolation processing is performed, that is, the optimal process parameters with better forming quality of thin-walled structures with other inclination angles are obtained by machine learning, so as to form a thin-walled structure forming quality database;

[0045] Step A3, for the database obtained in step A2, select an angle between 0° and 90° to perform comparative verification of the surface forming quality, i.e., roughness, to verify the accuracy of the method.

[0046] A method for regulating LPBF process to optimize the surface forming quality of inclined nickel alloy thin-walled structures, specifically comprising the following steps:

[0047] Step 1: Dry the IN 718 powder in a vacuum drying oven at 100°C for 8 hours to remove moisture and improve powder fluidity; and place the dried powder into a forming cavity of a laser powder bed melting device and introduce argon as a protective gas.

[0048] Step 2: Use Solid Works modeling software to build a printing model, import the model into Materials Magics slicing software for laser process parameter setting and slicing, set the laser power, laser scanning speed, layer thickness and scanning spacing during the processing, and determine the scanning strategy. Set the slicing layer thickness to 20-30μm and the scanning spacing to 30-50μm, and import the job file processed by Materials Magics slicing software into the laser powder bed melting equipment.

[0049] In this embodiment, the laser power is 260W to 340W. This laser power range is the process range in which the LPBF printing equipment used in this experiment can form IN 718 blocks without warping, deformation and burning. The scanning speed is 1000mm / s, the layer thickness is 30μm, the scanning strategy is strip scanning, the strip width is 10mm, the inter-layer rotation is 67°, the scanning spacing is 50μm, and the other parameters are unified.

[0050] Because the existing technology has achieved support-free printing, the lower surfaces of all thin-walled samples in this embodiment are not supported. Except for the area of ​​the inclined part connected to the substrate, which can be ignored due to its small area, the rest of the design length is the same.

[0051] Nickel alloy thin-walled structures with five different tilt angles of 0°, 15°, 30°, 45° and 60° are designed. These tilt angles are commonly used in the industry to optimize thin-walled structures. The thin-wall size is 8×22×1mm. The model of the thin-walled structure with each tilt angle is imported into the slicing software for slicing. The laser power, scanning speed, layer thickness and scanning spacing of each thin-walled structure sample during the processing are set, and the scanning strategy is determined. Five parallel samples are printed for each thin-wall with each tilt angle. These five tilt angles are all applied with laser powers of 260W, 280W, 300W, 320W and 340W, totaling 25 samples, and the laser power interval is very small. Based on the previous research of the research group on IN 718 nickel alloy, the laser power of the experimental equipment used in the present invention is within the range of 260W-340W to form IN 718 nickel alloy blocks without warping deformation and the like. The laser power for forming thin-walled structures without severe warping deformation is also within this range.

[0052] Step 3: Turn on the laser, use the laser powder bed melting equipment to tilt and form the nickel alloy thin-walled structure under the protection of argon, print the nickel alloy thin-walled sample, import the job file generated after slicing into the laser powder bed melting equipment, turn on the laser, and form thin-walled structures with different tilt angles according to the path planning specified by the slicing. The whole process is carried out under the protection of argon.

[0053] Step 4, after forming, cutting it from the substrate and removing the surface oil stains, to obtain an inclined nickel alloy thin-walled structure without cracks and obvious warping deformation.

[0054] Step 5, measure the lower surface roughness of the inclined thin-walled specimen of the nickel alloy by using a laser confocal instrument. The tested area of ​​each specimen is of the same size and position. Each specimen is tested in the upper, middle and lower areas to obtain surface morphology and surface roughness (Sa) data. The upper surface roughness of all specimens is low and does not need to be considered. Only the roughness of the lower surface needs to be considered.

[0055] The surface roughness of the lower surface of the thin-walled sample was tested using a laser confocal device. Three areas with a size of 1 mm × 1 mm were taken from the lower surface of each tested sample. Finally, the surface roughness result of each surface was taken as the average value of the three areas.

[0056] Step 6, grind and polish the thin-walled sample until the middle of the sample, take a light microscope photo, and use ImageJ binarization to adjust the light microscope image threshold of the thin-walled sample to calculate its density.

[0057] All samples with measured surface roughness must be ground and polished. Use 180#, 400#, 800#, 1200#, 1500# and 2000# sandpaper to grind the sides of the nickel alloy thin-walled samples, roughly grind to the center of the sample, polish with diamond polishing liquid, observe the pores after polishing, and calculate the density using ImageJ combined with binary image processing methods.

[0058] Step 7, analyze the experimental data of steps 5 and 6, and record the laser power corresponding to the minimum lower surface roughness in the range of 260W-340W and the laser power corresponding to the maximum internal density at different tilt angles as the basic data for machine learning of surface forming quality at different tilt angles.

[0059] The process parameters corresponding to the lowest roughness of the lower surface at five tilt angles of 0°, 15°, 30°, 45° and 60° were recorded, and the internal laser power with the highest density was also recorded. Since the density obtained in this experiment was relatively high, the effect of laser power on density was not considered.

[0060] Step 8: Using machine learning methods, according to the above experiments (including the roughness data before calculating the mean), a data set of LPBF forming quality of thin-walled structures is obtained, including thin-wall inclination angle and laser power parameters, that is, the training set is five inclination angle values ​​and laser power values. The fourth-order polynomial regression equation is used to obtain the laser power prediction set of other angles through interpolation, that is:

[0061] y=b0+b1cosx+b2(cosx) 2 +b3(cosx) 3 +b4(cosx) 4

[0062] Where y is the laser power, x is the tilt angle; b0, b1, b2, b3, b4 are the coefficients of the polynomial respectively;

[0063] That is, a fourth-order polynomial fitting formula is obtained, and the optimal laser power among the five tilt angles is obtained by using the laser power corresponding to the point on the curve.

[0064] Step 9, verify the prediction model, print a thin-walled structure with a tilt angle of 70° at a laser power of 300W as Example 2, obtain a better laser power with a tilt angle of 70° according to the prediction model, and print this laser power with a tilt angle of 70° as Comparative Example 2 to verify the accuracy of the prediction model.

[0065] In the following Examples 1 and 2 and Comparative Examples 1 and 2, the IN718 powder used includes the following substances in percentage by mass: Ni 52.8-54wt%, Cr 18.85wt%, Nb 4.95wt%, Mo 3.04wt%, Ti 0.94wt%, Al 0.47wt%, C 0.045wt%, and the rest is Fe.

[0066] Example 1

[0067] LPBF is used to form nickel alloy thin-walled structures with five different inclination angles, namely, nickel alloy thin-walled structures with inclination angles of 0°, 15°, 30°, 45° and 60°.

[0068] The IN718 powder was placed in a vacuum drying oven and dried at 100°C for 8 hours to remove moisture and improve powder fluidity.

[0069] Solid Works was used to build and print models with five tilt angles, and the models were imported into Materials Magics slicing software for slicing. The laser power of the thin-walled structures with different tilt angles was 300W, the scanning speed was 1000mm / s, the layer thickness was 30μm, the scanning strategy was strip scanning, the strip width was 10mm, the interlayer rotation was 67°, the scanning spacing was 50μm, and after digital model processing, it was imported into the laser powder bed melting equipment.

[0070] The laser is turned on, and a nickel alloy thin-wall structure is tilted using a laser powder bed melting device under the protection of argon.

[0071] The lower surface of the thin-walled sample was tested for surface roughness using a laser confocal device. Three areas were taken from each lower surface, and the area size was 1mm×1mm. Finally, the surface roughness result of each surface was taken as the average value of the three areas. The results of the tilt angles of 0°, 15°, 30°, 45° and 60° were as follows: Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 As shown, Figure 2 to Figure 6 The medium roughness area is 1 mm, and the numbers on the right represent the relative maximum and minimum values ​​of the lower surface compared with the average value. The difference between the maximum and minimum values ​​can reflect the surface forming quality of the thin-walled component.

[0072] All the samples with measured surface roughness were ground and polished. The sides of the nickel alloy thin-walled samples were polished with 180#, 400#, 800#, 1200#, 1500# and 2000# sandpapers in turn. The samples were polished to the center of the samples and polished with diamond polishing liquid. The pores after polishing were observed. The density was calculated by using the method of binarizing the light microscope images with ImageJ. The results of the tilt angles of 0°, 15°, 30°, 45° and 60° were shown as follows: Fig.12 , Fig.13 , Fig.14 , Fig.15 , Fig.16 shown.

[0073] Comparative Example 1

[0074] The operations of Example 1 were repeated, except that the laser powers of the thin-walled samples at various tilt angles were 260 W, 280 W, 300 W, 320 W, and 340 W, the scanning speed was 1000 mm / s, the layer thickness was 30 μm, the scanning strategy was strip scanning, the strip width was 10 mm, the inter-layer rotation was 67°, the scanning spacing was 50 μm, and the thin wall at each tilt angle was printed, that is, thin-wall structures with five laser powers were printed at each tilt angle.

[0075] Similarly, the surface roughness of the lower surface of each sample was measured. The roughness of the lower surface of each sample with an inclination angle of 0°, 15°, 30°, 45° and 60° was as follows: Figure 7 , Figure 8 , Fig. 9 , Fig.10 , Fig.11 As shown, all the sides of the samples were ground and polished, and the pores after polishing were observed. The density was calculated using the image binarization method, and the density at inclination angles of 0°, 15°, 30°, 45°, and 60° was obtained, as shown in Fig.17 , Fig.18 , Fig.19 , Fig. 20 , Fig.21 shown.

[0076] Record the roughness and density at the corresponding tilt angle, and the lowest corresponding laser power. The laser power with the highest density of all thin-walled samples is 300W. The density of all thin-walled structures is relatively high, and the effect of laser power on density can be ignored. The laser power values ​​with the lowest surface roughness are 260W, 300W, 300W, 260W, and 280W, respectively.

[0077] By using machine learning and fourth-order polynomial regression, the laser power prediction set at other angles is obtained by interpolation, that is, y = b0 + b1cosx + b2(cosx) 2+b3(cosx) 3 +b4(cosx) 4 The fourth-order polynomial fitting formula is obtained, and the optimal laser power of the tilt angle among the five tilt angles is obtained by using the laser power of the points on the curve, where b0 = -11355.61826, b1 = 67428.19997, b2 = -143274.73979, b3 = 132381.09346, b4 = -44918.93538, R 2 =0.98, confidence interval 95%.

[0078] By comparing Comparative Example 1 and Example 1, it can be seen that the density of the thin-walled samples at various inclination angles after optimization is improved, the lower surface roughness Sa is reduced, and the roughness of the thin-walled components is reduced without adding support, thereby improving the surface forming quality of the thin-walled components.

[0079] By comparing Comparative Example 1 and Example 1, it can be seen that the density of the thin-walled samples with five kinds of inclination angles after optimization is improved, and the lower surface roughness Sa is reduced, with the lowest decrease of 2. The roughness of the thin-walled components is reduced without adding support, thereby improving the surface forming quality of the thin-walled components.

[0080] Example 2

[0081] LPBF is used to form a nickel alloy thin-walled structure with a nickel alloy tilt angle of 20°.

[0082] The IN718 powder was placed in a vacuum drying oven and dried at 100°C for 8 hours to remove moisture and improve powder fluidity.

[0083] Solid Works was used to build and print models with five tilt angles, and the models were imported into Materials Magics slicing software for slicing. The laser power of the thin-walled structures with different tilt angles was 300W, the scanning speed was 1000mm / s, the layer thickness was 30μm, the scanning strategy was strip scanning, the strip width was 10mm, the interlayer rotation was 67°, the scanning spacing was 50μm, and after digital model processing, it was imported into the laser powder bed melting equipment.

[0084] The laser is turned on, and a nickel alloy thin-wall structure is tilted using a laser powder bed melting device under the protection of argon.

[0085] The lower surface of the thin-walled sample was tested for surface roughness using a laser confocal device. Three areas were taken from each lower surface, and the area size was 1mm×1mm. Finally, the surface roughness result of each surface was taken as the average value of the three areas, such as Fig. 22The roughness area in the figure is 1mm, and the numbers on the right represent the relative maximum and minimum values ​​of the lower surface compared with the average value. The difference between the maximum and minimum values ​​can reflect the surface forming quality of the thin-walled component.

[0086] All the samples with measured surface roughness were ground and polished. The sides of the nickel alloy thin-walled samples were polished with 180#, 400#, 800#, 1200#, 1500# and 2000# sandpapers in turn. The samples were polished to the center of the samples. Diamond polishing liquid was used to polish them. The pores after polishing were observed. The density was calculated by using the method of binarizing the light microscope images with ImageJ. Fig.23 .

[0087] Comparative Example 2

[0088] The fourth-order polynomial obtained by comparative example 2 predicts that the optimal laser power for a nickel alloy with an inclination angle of 20° is 312 W. The operation of example 1 is repeated to print a nickel alloy thin-walled structure, and the density and lower surface roughness are also measured.

[0089] By comparing Example 2 and Comparative Example 2, it can be seen that the optimal process parameters predicted by this model can improve the forming quality at different inclination angles and reduce the surface roughness.

[0090] The present invention has a good effect on optimizing the forming quality of nickel alloy thin-walled structures. The present invention provides a method and idea for optimizing the surface forming quality of nickel alloy thin-walled structures manufactured by laser additive manufacturing. The technical personnel in this field can modify or replace the technical scheme of the present invention without departing from the scope of the technology of the present invention, and all should be within the protection scope determined by the claims.

Claims

1. A method for regulating the forming quality of LPBF nickel alloy inclined thin-walled structure, characterized in that: The steps include: S1, select different laser powers for the first LPBF printing to print thin-walled structures with different tilt angles; S2, measure the lower surface roughness and density of all samples, and record the laser power with the lowest lower surface roughness and the highest density; S3, construct a prediction model, obtain the optimal process parameters for the remaining tilt angles through machine learning, and form a tilt thin-wall forming quality database; S4, obtaining the optimal laser power according to the prediction model, and performing the second LPBF printing according to the optimal laser power.

2. The method for regulating the forming quality of LPBF nickel alloy inclined thin-walled structure according to claim 1, characterized in that: The laser power ranges from 260W to 340W.

3. The method for regulating the forming quality of the LPBF nickel alloy inclined thin-walled structure according to claim 1, characterized in that: The selected tilt angles were 0°, 15°, 30°, 45° and 60°, and for each tilt angle, laser powers of 260 W, 280 W, 300 W, 320 W and 340 W were applied, respectively.

4. The method for regulating the forming quality of LPBF nickel alloy inclined thin-walled structure according to claim 1, characterized in that: All samples with measured surface roughness were polished on the side with 180#, 400#, 800#, 1200#, 1500# and 2000# sandpaper in turn until the center of the sample was polished with diamond polishing liquid, and their density was calculated using ImageJ combined with binary image processing.

5. The method for regulating the forming quality of LPBF nickel alloy inclined thin-walled structure according to claim 1, characterized in that: The fourth-order polynomial regression equation is used to obtain the laser power prediction set at other angles through interpolation.

Images