Workpiece performance optimization method and system, electronic equipment and storage medium

Abstract

The invention provides a workpiece performance optimization method and system, electronic equipment and a storage medium. The method comprises the steps that microstructure features of a workpiece in the LPBF forming process are extracted, and the mechanical property of the workpiece is determined according to the microstructure features; in response to a to-be-optimized area of which the mechanical property does not meet a preset condition on the workpiece, determining optimization process parameters of the to-be-optimized area according to a process optimization model; wherein the process optimization model represents a corresponding relationship among process parameters, microstructure characteristics and mechanical properties, and the process optimization model is constructed according to simulation data and / or experimental data; and optimizing the region to be optimized by adopting the optimization process parameters. The structure uniformity of all structural domains of the workpiece can be further improved, the problems of inconsistent microstructure distribution and mechanical property fluctuation caused by complex workpiece structure and non-uniform heat input are solved, and the defect occurrence rate of the workpiece is reduced.

Description

Technical Field

[0001] This disclosure relates to the field of metal additive manufacturing, and more particularly to a method, system, electronic device, and storage medium for optimizing the performance of a workpiece. Background Technology

[0002] Currently, most methods for determining the process parameters for manufacturing based on Laser Powder Bed Fusion (LPBF) rely on experience or preliminary experiments. Manufacturing workpieces with complex geometries and significant differences in size and thickness in different regions using this method presents problems such as uneven heat input, incompatibility of process parameters with workpiece structure, and fluctuations and inhomogeneities in the mechanical properties of the workpiece. Summary of the Invention

[0003] The technical problem to be solved by this disclosure is to overcome the above-mentioned defects in the prior art and provide a method, system, electronic device, and storage medium for optimizing the performance of a workpiece.

[0004] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0005] Firstly, a method for optimizing the performance of a workpiece is provided, including:

[0006] The microstructural features of the workpiece during the LPBF forming process are extracted, and the mechanical properties of the workpiece are determined based on the microstructural features.

[0007] In response to the existence of an area on the workpiece whose mechanical properties do not meet the preset conditions, the optimized process parameters of the area to be optimized are determined according to the process optimization model; wherein, the process optimization model characterizes the correspondence between process parameters, microstructural features and mechanical properties, and the process optimization model is constructed based on simulation data and / or experimental data;

[0008] The optimized process parameters are used to optimize the region to be optimized.

[0009] Optionally, the microstructural features of the workpiece during the LPBF forming process are extracted, including:

[0010] Acquire multimodal data of the workpiece during the LPBF forming process;

[0011] The microstructure features are extracted from the multimodal data; wherein the multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and molten pool image.

[0012] Optionally, the microstructural features of the workpiece during the LPBF forming process are extracted, and the mechanical properties of the workpiece are determined based on the microstructural features, including:

[0013] Multimodal data of the workpiece during the LPBF forming process is acquired, and the multimodal data is input into a performance mapping model. The performance mapping model extracts microstructural features from the multimodal data and determines the mechanical properties of the workpiece based on the extracted microstructural features. The performance mapping model is obtained by regression analysis on the multimodal data samples or by machine learning training on the multimodal data samples. The multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and melt pool image.

[0014] Optionally, extracting the microstructural features from the multimodal data includes:

[0015] Microstructural features of the workpiece are extracted from preprocessed multimodal data; wherein, the preprocessing includes at least one of the following: filtering, denoising, and region segmentation;

[0016] And / or, normalize and multiscale fusion processing is performed on the multimodal data, and the microstructural features of the workpiece are extracted from the spectral peak regions of the fusion result.

[0017] Optionally, the microstructural features of the workpiece during the LPBF forming process are extracted, including:

[0018] Based on the laser scanning path, geometric features and heat input distribution of the workpiece during the LPBF forming process, the workpiece is divided into several domains, and the microstructure features of each domain are extracted.

[0019] And / or, the workpiece is a blade; the blade's domain includes at least one of the following: root domain, blade body domain, and tip domain.

[0020] Optionally, the optimized process parameters for the region to be optimized are determined according to the process optimization model, including:

[0021] Construct optimization objectives;

[0022] Using the process optimization model as a constraint, the optimization objective is solved in reverse by a numerical optimization algorithm, and the process parameters at which the optimization objective is minimized are determined as the optimized process parameters of the region to be optimized.

[0023] Wherein, the optimization objective is COV H COV is the coefficient of variation of hardness distribution. D COV is the coefficient of variation for molten pool depth. Por w is the coefficient of variation for the porosity fraction. H w D w PAll are weights.

[0024] Secondly, a performance optimization system for a workpiece is provided, comprising:

[0025] An extraction module is used to extract the microstructural features of the workpiece during the LPBF forming process and determine the mechanical properties of the workpiece based on the microstructural features.

[0026] The process parameter determination module is used to determine the optimized process parameters of the workpiece based on the process optimization model in response to the existence of an area on the workpiece whose mechanical properties do not meet the preset conditions. The process optimization model characterizes the correspondence between process parameters, microstructure features and mechanical properties, and is constructed based on simulation data and / or experimental data.

[0027] An optimization module is used to optimize the region to be optimized using the optimized process parameters.

[0028] Thirdly, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and for running on the processor, wherein the processor executes the computer program to implement the performance optimization method for the workpiece described in any one of the first aspects.

[0029] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the performance optimization method for the workpiece described in any one of the first aspects.

[0030] Fifthly, a computer program product is provided, comprising a computer program that, when executed by a processor, implements a performance optimization method for a workpiece as described in any one of the first aspects.

[0031] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0032] The positive and progressive effects of this disclosure are as follows: This disclosure monitors the microstructural features that characterize the forming results during the workpiece forming process, and determines the optimized process parameters of the region to be optimized based on the microstructural features and the process optimization model, and then performs subsequent forming. That is, a feedback mechanism is adopted during the forming process to dynamically modify and optimize the process parameters, so that the forming process of subsequent layers or subsequent parts is gradually optimized, the uniformity of the microstructure of each domain of the workpiece is further improved, and the problems of inconsistent microstructure distribution and fluctuation of mechanical properties caused by complex workpiece structure and uneven heat input are solved, thereby reducing the defect rate of the workpiece. Attached Figure Description

[0033] Figure 1A flowchart of a performance optimization method for a workpiece provided as an exemplary embodiment of this disclosure;

[0034] Figure 2a An application scenario diagram illustrating a workpiece performance optimization method provided in an exemplary embodiment of this disclosure, which involves sampling analysis of the workpiece;

[0035] Figure 2b for Figure 2a The front view of the workpiece is shown;

[0036] Figure 2c for Figure 2a The rear view of the workpiece is shown;

[0037] Figure 3 This is a schematic diagram of the optical microscopy results of observing the long side, short side, and upper, middle, and lower parts of a first batch of blades with a layer thickness of 100 μm and a wall thickness of 2 mm using a metallographic optical microscope, as an exemplary embodiment of the present disclosure.

[0038] Figure 4 A schematic diagram illustrating the effect of etching the long side, short side, and upper, middle, and lower parts of a blade with a first-batch layer thickness of 100 μm and a wall thickness of 2 mm, as provided in an exemplary embodiment of this disclosure;

[0039] Figure 5a A density comparison diagram of an LPBF 316L blade and a sample block provided as an exemplary embodiment of this disclosure;

[0040] Figure 5b A comparison diagram of the melt pool depth of an LPBF 316L blade and a sample block is provided as an exemplary embodiment of this disclosure;

[0041] Figure 6a A hardness test sampling diagram of a hardness tester provided as an exemplary embodiment of this disclosure;

[0042] Figure 6b for Figure 6a A statistical chart of sampling locations and hardness values;

[0043] Figure 7 A schematic diagram of optical microscopy results for observing the long side, short side, and upper, middle, and lower parts of a second batch of blades with a layer thickness of 100 μm and a wall thickness of 2 mm, as provided in an exemplary embodiment of this disclosure;

[0044] Figure 8 A schematic diagram illustrating the effect of etching the long side, short side, and upper, middle, and lower parts of a second batch of blades with a layer thickness of 100 μm and a wall thickness of 2 mm, as provided in an exemplary embodiment of this disclosure;

[0045] Figure 9 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of the present disclosure. Detailed Implementation

[0046] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0047] The prefixes such as "first" and "second" used in this disclosure are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this disclosure does not constitute a limitation on the described objects. The description of the described objects is given in the claims or the context of the embodiments, and should not be construed as an unnecessary limitation. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0048] Currently, most methods for determining the process parameters for manufacturing based on Laser Powder Bed Fusion (LPBF) rely on experience or preliminary experiments, often employing a single set of parameters to manufacture the entire blade. Experiments have revealed that due to the significant differences in thermophysical behavior across different regions of 316L stainless steel during LPBF, these conventional uniform process parameters are insufficient to meet the processing requirements of all areas. This is particularly true for 316L alloy blades with extremely complex geometries, exhibiting significant differences in size and thickness between their long sides, short sides, bottom, and middle regions. This structural characteristic results in drastically different heating conditions and cooling rates in different areas of the blade during laser processing. Manufacturing 316L alloy blades using current processes often results in inhomogeneities and inconsistent properties (hardness), making it difficult to meet processing requirements.

[0049] Based on the above problems, this disclosure provides a method for optimizing the performance of a workpiece. The method of this embodiment is particularly suitable for optimizing the process parameters of workpieces with large layer thickness (layer thickness ≥ 100 μm) and extremely complex geometry (significant differences in size and thickness in different regions). This disclosure does not particularly limit the type of workpiece, and the workpiece may be, for example, a steam turbine blade.

[0050] See Figure 1 The method includes the following steps:

[0051] Step 101: Extract the microstructure features of the workpiece during the LPBF forming process, and determine the mechanical properties of the workpiece based on the microstructure features.

[0052] The microstructure characteristics of the workpiece during the LPBF forming process reflect the forming result.

[0053] Mechanical properties, also known as uniformity of structure and consistency of performance, can be characterized by, but are not limited to, at least one of the following parameters: yield strength, tensile strength, hardness, density, etc.

[0054] Step 102: In response to the existence of an area on the workpiece whose mechanical properties do not meet the preset conditions, determine the optimization process parameters of the area to be optimized based on the process optimization model.

[0055] The so-called mechanical properties not meeting the preset conditions means that the microstructure and properties of the workpiece deviate from the design requirements during the forming process.

[0056] Among them, the process optimization model characterizes the correspondence between process parameters, microstructure features and mechanical properties, and the process optimization model is constructed based on simulation data and / or experimental data.

[0057] The preset conditions can be set according to actual needs. The area to be optimized is the area on the workpiece whose mechanical properties do not meet the preset conditions.

[0058] Step 103: Optimize the region to be optimized using optimized process parameters.

[0059] Optimizing process parameters includes at least one of the following: laser power density, scanning speed, scanning path sequence, number of local remelting cycles, etc.

[0060] In this embodiment, the entire process of workpiece forming is monitored, and steps 101 to 103 are executed to optimize the process parameters. The process parameters can be optimized periodically; the optimization time of the process parameters can also be dynamically adjusted. For example, the interval between two adjacent optimizations of process parameters is related to the mechanical properties of the workpiece; the better the mechanical properties, the longer the interval.

[0061] In this embodiment, the microstructural features characterizing the forming result are monitored during the workpiece forming process. Based on these microstructural features and the process optimization model, the optimized process parameters for the region to be optimized are determined for subsequent forming. That is, a feedback mechanism is used during the forming process to dynamically modify and optimize the process parameters so that the forming process of subsequent layers or subsequent components is gradually optimized. The uniformity of the microstructure of each domain of the workpiece is further improved, which solves the problem of inconsistent microstructure distribution and fluctuation of mechanical properties caused by complex workpiece structure and uneven heat input, and reduces the defect rate of the workpiece.

[0062] In one embodiment, step 101, which involves extracting the microstructural features of the workpiece during the LPBF forming process, includes: dividing the workpiece into several domains based on the laser scanning path, geometric features, and heat input distribution of the workpiece during the LPBF forming process, and subsequently extracting the microstructural features of each domain. A domain is also known as a structural region.

[0063] In this embodiment, the division of each domain comprehensively considers factors such as laser scanning path, geometric features, and heat input distribution, ensuring that each domain has a relatively consistent thermal history and morphological conditions. Furthermore, through partitioning and monitoring, the differences in microstructure and the gradient variation of mechanical properties (e.g., hardness) at different locations such as the long side, short side, top, middle, and bottom of the workpiece are revealed. This provides a targeted basis for subsequent parameter optimization and effectively reduces the problem of local performance mismatch caused by fluctuations in workpiece microstructure.

[0064] Preferably, the domain division can be determined by combining finite element heat conduction simulation results or actual thermal imaging data of the forming process, grouping regions with similar temperature fields into one category. For example, the blade can be divided into root, blade body, and tip domains along the chord length direction, and further subdivided into different levels of domains along the azimuth direction. Alternatively, the blade can be divided into root, blade body, and tip along its length, into inlet edge and outlet edge along the chord length direction, and into inner and outer arc surfaces, for a total of 18 domains.

[0065] After partitioning, each domain is configured with appropriate laser power density, scanning speed, layer thickness, and scanning strategy to maintain uniformity of thermal input conditions as much as possible. Domain partitioning provides a fundamental guarantee for subsequent microstructural feature extraction and analysis, making the measurement of the microstructure properties of each domain comparable and specific. Thermal input distribution can be characterized, but is not limited to, through interlayer accumulated heat, and geometric features include geometric cross-sectional thickness.

[0066] In one embodiment, extracting the microstructural features of the workpiece during the LPBF forming process includes: acquiring multimodal data of the workpiece during the LPBF forming process, and extracting microstructural features from the multimodal data. The multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and molten pool image.

[0067] The crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, and phase structure spectrum are presented as spectra. The crystal orientation distribution spectrum and phase composition spectrum can be obtained using scanning electron microscopy combined with electron backscatter diffraction. The local hardness-displacement spectrum can be obtained by mechanical loading using a nanoindenter. The phase structure spectrum can be obtained through X-ray diffraction.

[0068] It should be noted that if the workpiece is divided into several domains, multimodal data of each domain can be obtained separately, and the microstructural features of the domain can be extracted from the multimodal data of each domain.

[0069] In this embodiment, microstructural features are extracted from multimodal data, which overcomes the information limitations of a single mode and can more comprehensively and accurately characterize the morphology of the workpiece during the LPBF forming process.

[0070] In one embodiment, mechanical properties matching the microstructure characteristics of the workpiece are determined based on the correspondence between microstructural features and mechanical properties. This correspondence is obtained through regression analysis of simulation data and / or experimental data, or through machine learning training of simulation data and / or experimental data.

[0071] In one embodiment, the correspondence between microstructural features and mechanical properties is characterized by the following formula:

[0072] ;

[0073] Where H is the local hardness, used to characterize the mechanical properties of the workpiece; D is the depth of the molten pool; W is the width; D / W is the depth-to-width ratio; and Af is the void area fraction. For the short side position factor, The bottom position factor, The position factor of the longer side; This is the error term; - is the regression coefficient.

[0074] The regression coefficients are obtained as follows: First, experimental tests are conducted on different areas of the workpiece to collect microstructural characteristic parameters such as molten pool depth D, molten pool width W, void area fraction Af, and location factor, and the corresponding local hardness H is measured to form a dataset of input parameters and output performance. Then, the dataset is input into a multivariate regression algorithm, and the regression coefficients are solved using the least squares method or other fitting methods. Furthermore, the regression coefficients can be corrected and optimized by combining finite element thermo-mechanical-fluid field simulation data. Through joint modeling of experimental and simulation data, stable and accurate regression coefficients can be obtained to reflect the correspondence between microstructural characteristics and local hardness.

[0075] The above D, W, D / W, Af, , , All parameters characterize the microstructure of the workpiece. This embodiment can establish a path for quantitatively evaluating hardness fluctuations (mechanical properties) based on the microstructure characteristics of the workpiece, which is more targeted and explanatory than traditional evaluation methods that rely on density or macroscopic mechanical properties.

[0076] It should be noted that the above correspondence can be represented not only by functions, but also by models or tables. This disclosure does not impose any particular limitation on this aspect.

[0077] In one embodiment, the microstructural features of the workpiece during the LPBF forming process are extracted, and the mechanical properties of the workpiece are determined based on these microstructural features, including:

[0078] Multimodal data of the workpiece during the LPBF forming process is acquired, and the multimodal data is input into the performance mapping model so that the performance mapping model can extract microstructural features from the multimodal data and determine the mechanical properties of the workpiece based on the extracted microstructural features.

[0079] The performance mapping model is obtained through regression analysis or machine learning on multimodal data samples. Multimodal data includes at least two of the following categories: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and melt pool image.

[0080] By inputting the multimodal data of each domain of the workpiece into the performance mapping model, the distribution of mechanical properties in different regions of the workpiece can be obtained. This distribution of mechanical properties can be characterized, but is not limited to, through curves (performance spectra), functions, or numerical values.

[0081] In one embodiment, each domain corresponds to a microstructural feature and mechanical property relationship or performance mapping model or process optimization model, thereby enabling the monitoring of inconsistent microstructure evolution in different regions due to differences in heat input, and targeted optimization of process parameters.

[0082] In one embodiment, before feature extraction, the multimodal data is preprocessed to extract the microstructural features of the workpiece from the preprocessed multimodal data, thereby extracting effective feature information. The preprocessing includes at least one of the following: filtering, denoising, and region segmentation.

[0083] In one embodiment, during feature extraction, the multimodal data undergoes normalization and multi-scale fusion processing, and the microstructural features of the workpiece are extracted from the spectral peak regions of the fusion result. Specifically, the spectral data is normalized and fused using an energy spectrum stability projection algorithm. The fusion result reflects the relationship between the microstructural features of each domain and the distribution of local mechanical responses.

[0084] In this embodiment, normalization and multi-scale fusion processing reduce measurement differences and noise interference during multimodal data acquisition. Finally, spectral peak features are extracted from the fusion results, and a complete domain spectrum is constructed. This domain spectrum accurately reflects the evolution of the microstructural features and local mechanical response distribution of each domain, and can serve as input data for subsequent performance mapping.

[0085] Spectral peak regions refer to local energy peak areas or texture peak areas formed by the interaction between the scanning trajectory and the molten pool boundary in metallographic or microstructural spectra. These regions often correspond to the locations where local structural differences are most significant. Spectral peak characteristics are quantitative descriptive parameters extracted from these regions, including but not limited to: peak intensity, peak width, peak position shift, contrast, and local texture orientation.

[0086] In this embodiment, instead of directly using the original spectrum as input, a domain spectrum is constructed based on spectral peak features. A domain spectrum refers to the mapping and aggregation of multidimensional features extracted from the typical peak regions of different domains (long side, short side, middle, bottom, etc.) to form a set of spectra that can reflect the relationship between the microstructural features of each domain and its local mechanical response.

[0087] The term "mechanical response distribution" refers to the response characteristics of a local microstructure within a domain to external loading or internal stress fields, such as local hardness variations, strain concentration trends, or residual stress distribution. It is not a concept independent of performance, but rather a precursor to performance: the mechanical response distribution reflects the microscopic load-bearing state and deformation compatibility of a material in different domains; while mechanical performance indicators (such as yield strength and elongation) are comprehensive representations on a macroscopic scale.

[0088] Introducing the mechanical response distribution into the domain spectrum aims to infer and predict regional fluctuations in macroscopic mechanical properties by quantitatively capturing the response patterns at the microscale. In other words, the mechanical response distribution serves as an intermediate layer of information connecting "microscopic structural features" and "macroscopic performance," and it is an indispensable input dimension for performance mapping models.

[0089] In this embodiment, a performance mapping model is used to quantitatively convert the microstructure characteristics of the domains into macroscopic mechanical properties, and a structural homogeneity factor for homogeneity assessment is calculated for each domain. In step 102, the structural homogeneity factor is used to determine whether to determine the optimized process parameters for the region to be optimized based on the process optimization model.

[0090] After obtaining the mechanical properties of the workpiece in step 101, non-uniformity analysis can be performed and control strategies can be formulated. First, by comparing the mechanical properties of each region of the workpiece, regions that do not meet the preset conditions for optimization are identified. For regions that are found to have low structural homogeneity factors or abnormal performance, targeted control measures can be taken. For example, process parameters such as laser power density, scanning speed, and scanning path sequence of the region can be adjusted, and the number of local remeltings can be increased to change the heat input conditions; or directional annealing or regional aging treatment can be performed on the region after forming to promote microstructure homogenization.

[0091] In one embodiment, this embodiment is further integrated with an online monitoring system. During the forming process, the online monitoring system acquires temperature field and molten pool images in real time and inputs them into the performance mapping model in real time. The temperature field reflects the distribution of heating and cooling rates in different regions, and the molten pool image characterizes the depth, width, and morphological features of the molten pool. This information corresponds to the grain structure evolution and mechanical properties of the workpiece. The performance mapping model uses the aforementioned monitoring data to extract and correct microstructural features, thereby dynamically adjusting the predicted mechanical properties and further determining optimized process parameters to achieve dynamic correction and closed-loop control. Through this feedback loop, the forming process of subsequent layers or components of the workpiece can be gradually optimized, and the uniformity of the microstructure of each domain can be further improved.

[0092] In one embodiment, the input parameters of the process optimization model include microstructural features (or spectra) and mechanical properties, and the output parameters include process parameters. In step 102, the current microstructural features (or spectra) and mechanical properties of the workpiece are input into the process optimization model to obtain the optimized process parameters for the region to be optimized.

[0093] In one embodiment, determining the optimized process parameters of the region to be optimized based on the process optimization model includes: constructing an optimization objective, using the process optimization model as a constraint, solving the optimization objective in reverse using a numerical optimization algorithm, and determining the process parameters at the minimum of the optimization objective as the optimized process parameters of the region to be optimized.

[0094] Among them, the optimization objective is COV H COV is the coefficient of variation of hardness distribution. D COV is the coefficient of variation for molten pool depth. Por w is the coefficient of variation for the porosity fraction. H w D w P All are weights.

[0095] Understandably, in this embodiment, the microstructural features characterizing the actual forming result of the workpiece are monitored during the forming process, and the optimized process parameters of the region to be optimized are determined based on the microstructural features and the process optimization model, and subsequent forming is carried out. The deviation between the optimized process parameters and the previous process parameters will not be too large. The optimized process parameters can be selected from the vicinity of the previous process parameters, thereby eliminating a large number of process parameters that will cause the workpiece to warp and crack, and using less data to achieve process parameter optimization.

[0096] The following example uses an LPBF 316L blade as the workpiece, with process parameters of 325W laser power, 937mm / s scanning speed, and 100μm layer thickness. Experiments are conducted to verify the effect of the method in this embodiment on improving the hardness uniformity of different areas of the blade.

[0097] (1) Sampling and analysis were performed at different locations on the leaf, see [reference]. Figures 2a-2c The sampling rules include: 6 points each on the long side L1~L6, the short side S1~S6, and the center C1~C6, 3 points on the inner side, and 3 points on the outer side, which should cover as many representative positions of the entire blade as possible.

[0098] (2) The long side, short side, and upper, middle, and lower parts of the blade with a layer thickness of 100 μm and a wall thickness of 2 mm were observed using a metallographic optical microscope. The results are shown in (see Figure 3 Numerous large-scale holes were found along the long side of the blade; the blade base had even more defects; no cracks were found at any of the sampling locations.

[0099] (3) After the blades were coarsely ground, finely ground, and polished with SiC sandpaper (220# to 5000#) using oxide polishing suspension (OPS), the sample surface was etched with reagent (1 vol% HF + 1.5 vol% HCl + 2.5 vol% HNO3 + 95 vol% H2O) for about 2 minutes. Then, the sample was observed using a metallographic optical microscope. The results are shown in (see Figure 4 This indicates that the microstructure is most uniform at the center of the blade; the molten pool at the bottom is deeper (the molten pool depth can reach 406.6 μm); and the molten pool morphology at the top is the most irregular.

[0100] (4) Statistical analysis of the density and molten pool depth of the blade and the sample block (printed with the same parameters as the blade) confirmed that the density of the blade was not uniform, the density of the sample block was slightly higher than that of the blade, and the wall thickness of the blade was much smaller than that of the sample block (blade wall thickness 2mm, sample block height 10mm), making it easier for heat to dissipate. As a result, the molten pool depth of the blade (the maximum molten pool depth in the middle part of the blade is 173.5μm) was much smaller than that of the sample block (268.4μm). For specific results, see [link to results]. Figure 5a and Figure 5b ;

[0101] (5) Hardness tests were conducted on the long side, short side, and upper, middle, and lower parts of the blade using a hardness tester, and statistical analysis was performed. The hardness of the short side (the lower part of the short side has the highest hardness value of 223 HV) was higher than that of the long side, and the hardness of the lower part was higher than that of the upper part. For specific results, see [link to results]. Figure 6a and Figure 6b .

[0102] Furthermore, using the method of this embodiment and optimizing the process parameters based on the first round of process parameters, new process parameters (i.e., optimized process parameters) are obtained. LPBF 316L stainless steel forming blades with a wall thickness of 2mm are printed using these new process parameters. The new process parameters for LPBF manufacturing are: laser power 337.5 W, scanning speed 932 mm / s, and layer thickness 100 μm. The specific steps are as follows:

[0103] (1) Sampling analysis was performed at different locations on the leaf, following the same sampling rules as in Example 1.

[0104] (2) The long side, short side, and upper, middle, and lower parts of the second batch of leaves were observed using a metallographic optical microscope. The results are shown in (see Figure 7 Similarly, numerous large-scale holes were found along the long side of the blades; there were even more defects at the bottom of the blades; and cracks were also found.

[0105] (3) After the blades were coarsely ground, finely ground, and polished with SiC sandpaper (220# to 5000#) (oxide polishing suspension, OPS), the sample surface was etched with Kroll reagent (1 vol% HF + 1.5 vol% HCl + 2.5 vol% HNO3 + 95 vol% H2O) for about 2 minutes, and then observed with a metallographic optical microscope. The results are shown in (see Figure 8 This indicates that the microstructure is most uniform at the center of the blade; the molten pool at the bottom is deeper; and the molten pool at the top has the most irregular morphology.

[0106] (4) Statistical analysis of the average melt pool depth of the first and second batches of blades showed that the average melt pool depth of the first batch of blades was wider (222.7 μm for the first batch and 174.9 μm for the second batch), and the variance was larger. The melt pool line spacing of samples at different positions of the second batch of blades was very close, indicating that the uniformity of the new blades was significantly improved. For specific results, see [link to results]. Figure 8 .

[0107] Test results show that after optimizing the process using the method of this embodiment, the standard deviation of blade hardness decreased from 15.2 HV in the first batch to 8.9 HV in the second batch. The hardness difference between the top and bottom regions decreased by about 30%, and the difference in molten pool morphology between regions was significantly reduced, indicating that the uniformity of the structure and the consistency of performance were improved.

[0108] Corresponding to the aforementioned embodiments of the workpiece performance optimization method, this disclosure also provides embodiments of the workpiece performance optimization system.

[0109] This disclosure provides a schematic diagram of a workpiece performance optimization system, which is used to implement the performance optimization method provided in any of the above embodiments. The system includes:

[0110] The extraction module is used to extract the microstructural features of the workpiece during the LPBF forming process and determine the mechanical properties of the workpiece based on the microstructural features.

[0111] The process parameter determination module is used to determine the optimized process parameters for the workpiece region whose mechanical properties do not meet the preset conditions in response to the existence of such regions. The process optimization model characterizes the correspondence between process parameters, microstructural features, and mechanical properties, and is constructed based on simulation data and / or experimental data. The optimized process parameters are used to optimize the region.

[0112] An optimization module is used to optimize the region to be optimized using the optimized process parameters.

[0113] Optionally, when extracting the microstructural features of the workpiece during the LPBF forming process, the extraction module is used to:

[0114] Acquire multimodal data of the workpiece during the LPBF forming process;

[0115] The microstructure features are extracted from the multimodal data; wherein the multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and molten pool image.

[0116] Optionally, the extraction module is specifically used for:

[0117] Multimodal data of the workpiece during the LPBF forming process is acquired, and the multimodal data is input into a performance mapping model. The performance mapping model extracts microstructural features from the multimodal data and determines the mechanical properties of the workpiece based on the extracted microstructural features. The performance mapping model is obtained by regression analysis on the multimodal data samples or by machine learning training on the multimodal data samples. The multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and melt pool image.

[0118] Optionally, when extracting the microstructural features from the multimodal data, the extraction module is used to:

[0119] Microstructural features of the workpiece are extracted from preprocessed multimodal data; wherein, the preprocessing includes at least one of the following: filtering, denoising, and region segmentation;

[0120] And / or, normalize and multiscale fusion processing is performed on the multimodal data, and the microstructural features of the workpiece are extracted from the spectral peak regions of the fusion result.

[0121] Optionally, when extracting the microstructural features of the workpiece during the LPBF forming process, the extraction module is used to:

[0122] Based on the laser scanning path, geometric features and heat input distribution of the workpiece during the LPBF forming process, the workpiece is divided into several domains, and the microstructure features of each domain are extracted.

[0123] And / or, the workpiece is a blade; the blade's domain includes at least one of the following: root domain, blade body domain, and tip domain.

[0124] Optionally, the process parameter determination module is used for:

[0125] Construct optimization objectives;

[0126] Using the process optimization model as a constraint, the optimization objective is solved in reverse by a numerical optimization algorithm, and the process parameters at which the optimization objective is minimized are determined as the optimized process parameters of the region to be optimized.

[0127] Wherein, the optimization objective is COV H COV is the coefficient of variation of hardness distribution. D COV is the coefficient of variation for molten pool depth. Por w is the coefficient of variation for the porosity fraction. H w D w P All are weights.

[0128] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs.

[0129] Figure 9 This is a schematic diagram of the structure of an electronic device according to an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the performance optimization method of the workpiece described in any of the above embodiments. Figure 9 The electronic device 90 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0130] like Figure 9As shown, the electronic device 90 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different system components (including memory 92 and processor 91).

[0131] Bus 93 includes a data bus, an address bus, and a control bus.

[0132] The memory 92 may include volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0133] The memory 92 may also include a program tool 925 (or utility) having a set (at least one) program module 924, such program module 924 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0134] The processor 91 executes various functional applications and data processing by running computer programs stored in the memory 92, such as the workpiece performance optimization method provided in any of the above embodiments.

[0135] Electronic device 90 can also communicate with one or more external devices 94 (e.g., keyboard, pointing device, etc.). This communication can be performed through input / output (I / O) interface 95. Furthermore, electronic device 90 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 96. As shown, network adapter 96 communicates with other modules of electronic device 90 via bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0136] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0137] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the workpiece performance optimization method provided in any of the above embodiments.

[0138] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0139] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the performance optimization method for the workpiece described in any of the preceding embodiments.

[0140] The program code for executing the computer program product of this disclosure can be written in any combination of one or more programming languages, and the program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0141] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A method for optimizing the performance of a workpiece, characterized in that, include: The microstructural features of the workpiece during the LPBF forming process are extracted, and the mechanical properties of the workpiece are determined based on the microstructural features. In response to the existence of an area on the workpiece whose mechanical properties do not meet the preset conditions, the optimized process parameters of the area to be optimized are determined according to the process optimization model; wherein, the process optimization model characterizes the correspondence between process parameters, microstructural features and mechanical properties, and the process optimization model is constructed based on simulation data and / or experimental data; The optimized process parameters are used to optimize the region to be optimized.

2. The workpiece performance optimization method according to claim 1, characterized in that, The microstructural features of the workpiece during the LPBF forming process were extracted, including: Acquire multimodal data of the workpiece during the LPBF forming process; The microstructure features are extracted from the multimodal data; wherein the multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and molten pool image.

3. The workpiece performance optimization method according to claim 1, characterized in that, Extracting the microstructural features of the workpiece during the LPBF forming process and determining the mechanical properties of the workpiece based on the microstructural features includes: Multimodal data of the workpiece during the LPBF forming process is acquired, and the multimodal data is input into a performance mapping model. The performance mapping model extracts microstructural features from the multimodal data and determines the mechanical properties of the workpiece based on the extracted microstructural features. The performance mapping model is obtained by regression analysis on the multimodal data samples or by machine learning training on the multimodal data samples. The multimodal data includes at least two of the following types: crystal orientation distribution spectrum, phase composition spectrum, local hardness-displacement spectrum, phase structure spectrum, temperature field distribution, and melt pool image.

4. The workpiece performance optimization method according to claim 2 or 3, characterized in that, Extracting the microstructural features from the multimodal data includes: Microstructural features of the workpiece are extracted from preprocessed multimodal data; wherein, the preprocessing includes at least one of the following: filtering, denoising, and region segmentation; And / or, normalize and multiscale fusion processing is performed on the multimodal data, and the microstructural features of the workpiece are extracted from the spectral peak regions of the fusion result.

5. The workpiece performance optimization method according to claim 1, characterized in that, The microstructural features of the workpiece during the LPBF forming process were extracted, including: Based on the laser scanning path, geometric features and heat input distribution of the workpiece during the LPBF forming process, the workpiece is divided into several domains, and the microstructure features of each domain are extracted. And / or, the workpiece is a blade; the blade's domain includes at least one of the following: root domain, blade body domain, and tip domain.

6. The method for optimizing the performance of a workpiece according to any one of claims 1-3 and 5, characterized in that, The optimized process parameters for the region to be optimized are determined based on the process optimization model, including: Construct optimization objectives; Using the process optimization model as a constraint, the optimization objective is solved in reverse by a numerical optimization algorithm, and the process parameters at which the optimization objective is minimized are determined as the optimized process parameters of the region to be optimized. Wherein, the optimization objective is COV H COV is the coefficient of variation of hardness distribution. D COV is the coefficient of variation for molten pool depth. Por w is the coefficient of variation for the porosity fraction. H w D w P All are weights.

7. A performance optimization system for a workpiece, characterized in that, include: An extraction module is used to extract the microstructural features of the workpiece during the LPBF forming process and determine the mechanical properties of the workpiece based on the microstructural features. The process parameter determination module is used to determine the optimized process parameters of the workpiece based on the process optimization model in response to the existence of an area on the workpiece whose mechanical properties do not meet the preset conditions. The process optimization model characterizes the correspondence between process parameters, microstructure features and mechanical properties, and is constructed based on simulation data and / or experimental data. An optimization module is used to optimize the region to be optimized using the optimized process parameters.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the performance optimization method for the workpiece as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the performance optimization method for the workpiece as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the performance optimization method for the workpiece as described in any one of claims 1 to 7.

Images