A method for preparing a low-oxygen shaped titanium alloy powder by cold isostatic pressing

Abstract

The present application relates to a kind of titanium alloy powder cold isostatic pressing forming low-oxygen special-shaped part preparation method, the method comprises the following steps: S1: obtaining metal alloy powder particle size distribution data and viscosity parameter, the viscosity parameter is used to characterize the flow resistance and obtain powder flow behavior initial model;S2: according to the powder flow behavior initial model, using numerical simulation method simulates the dynamic filling path of powder in corner area in cold isostatic pressing process;S3: the density gradient value at the filling insufficient position coordinate is judged to exceed preset threshold, obtain improved powder flow guide scheme, for reducing the gap formation in corner area;After determining powder filling path, the shrinkage of powder in three-dimensional space in isostatic pressing process is simulated again, and the size of mold is determined;S4: extract key pressure field distribution model from the improved powder flow guide scheme, determine the best pressing time and pressure curve;S5: by the determined pressing parameter combination driving actual cold isostatic pressing equipment operation, obtain the final verified forming process parameter.

Description

Technical Field

[0001] This invention relates to the field of hydrostatic pressing manufacturing technology, and in particular to a method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder. Background Technology

[0002] In the field of metal processing, titanium alloys have attracted much attention due to their excellent strength and corrosion resistance, especially in high-end industries such as aerospace and medical devices where they have irreplaceable value. However, how to prepare titanium alloy parts with complex shapes and stable performance through efficient forming processes has always been a key issue that the industry urgently needs to overcome. In particular, the cold isostatic pressing process in powder metallurgy technology can achieve near-net-shape forming of complex parts at a lower cost, and is therefore considered a highly promising direction.

[0003] Nevertheless, existing processes often face insurmountable limitations when processing semi-dehydrogenated or fully dehydrogenated titanium alloy powders. Many traditional methods are prone to insufficient strength or frequent defects in certain areas of parts with complex shapes due to uneven powder distribution. This problem is not simply due to insufficient technical means, but rather because of the lack of precise guidance on powder flow behavior during the pressing process, especially in areas where the shape of the part changes drastically, which significantly increases the complexity of process control.

[0004] Focusing on specific technical challenges, the flow characteristics of powder during pressing become the core factor affecting molding quality. Because the powder often does not fill complex shaped areas sufficiently, especially at the edges or corners of parts, voids or uneven density are easily formed. This flow characteristic is directly related to another key issue: insufficient powder enrichment in local areas, which threatens the structural integrity of the pressed parts in these areas. For example, in an irregularly shaped part with sharp corners, cracks or deformation risks often occur in the corner areas because the powder cannot be effectively deposited, which poses a significant hidden danger to the safety and reliability of subsequent use.

[0005] Therefore, how to effectively guide the flow direction of powder during cold isostatic pressing and achieve full filling of key areas of complex-shaped parts has become a key issue in improving the forming quality of semi-dehydrogenated titanium alloy parts. This challenge is not only related to the improvement of process technology, but also directly affects the performance and application range of the final product. Summary of the Invention

[0006] The purpose of this invention is to overcome the defects in the prior art and provide a method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder.

[0007] To achieve the above objectives, the technical solution of the present invention is to design a method for preparing low-oxygen irregular-shaped parts by cold isostatic pressing of titanium alloy powder, the method comprising: S1: Obtain the particle size distribution data and viscosity parameters of the metal alloy powder. The viscosity parameters are used to characterize the flow resistance. The surface morphology of the powder is analyzed by electron microscopy imaging to obtain an initial model of powder flow behavior. The initial model of powder flow behavior is based on inter-particle interaction and is used to simulate the powder distribution prediction in the pressing process of complex shaped parts. S2: Based on the initial model of powder flow behavior, the dynamic filling path of powder in the corner region during cold isostatic pressing is simulated using numerical simulation methods. S3: If the density gradient value at the coordinates of the insufficient filling location exceeds the preset threshold, an improved powder flow guidance scheme is obtained to reduce the formation of voids in the corner area; after determining the powder filling path, the mold size is determined by numerically simulating the shrinkage change of powder in three-dimensional space during isostatic pressing. S4: Extract the key pressure field distribution model from the improved powder flow guiding scheme, and determine the optimal pressing time and pressure curve; S5: Drive the actual cold isostatic pressing equipment by determining the combination of pressing parameters, monitor the real-time powder density sensor feedback data, determine whether the deviation between the sensor data and the simulation prediction is within the allowable range, and obtain the final verified molding process parameters.

[0008] A further preferred technical solution is that, in step S1, the metal alloy powder is detected by a particle size measuring device to obtain particle size and distribution data, which is stored in a pre-established database to obtain basic distribution information of the powder particles; Electron microscopy was used to image and analyze powder samples, extract surface morphology and morphological feature data, and store them as a high-resolution image dataset to obtain surface detail information of powder particles; For the extracted surface morphology and morphological feature data, a pre-established image processing algorithm is used to quantize the features, determine the contact patterns and potential interaction strength between particles, and if the feature quantization result exceeds the preset threshold, the image data is corrected a second time to obtain the preliminary evaluation results of the interaction between particles. Based on the preliminary evaluation results of interparticle interactions, combined with viscosity parameters and flow resistance characteristics, an initial model of powder flow behavior is constructed.

[0009] In a further optimized technical solution, in step S2, the dynamic performance of powder flow during cold isostatic pressing is analyzed by numerical simulation method, detailed calculations are performed on the filling path of the corner area, the specific location coordinate information of insufficient filling is obtained, and the distribution status of key areas is determined. Based on the obtained coordinate information of the incomplete filling location, combined with the dynamic analysis results, the density gradient data in the corner area is calculated to obtain detailed distribution information of density change; according to the determined density change information, the size change is simulated and analyzed to obtain the dimensional shrinkage rate of titanium alloy powder during cold isostatic pressing. The density gradient data is processed using pre-established analysis tools to determine whether the density change exceeds a preset threshold range. If it does, the relevant location coordinates are marked to determine the area that needs to be adjusted.

[0010] A further optimized technical solution is to construct an optimized adjustment scheme for the marked area that needs adjustment, based on the input data of pressure distribution, to obtain the adjusted pressure distribution parameters and obtain improved data suitable for the corner area; By improving the data, the parameters of the initial model of powder flow behavior are updated, the filling path of powder flow under the adjusted pressure distribution is simulated, and new filling state information is determined. Based on the new filling state information, analyze the trend of density gradient change and determine whether the preset uniformity condition is met. If not, adjust the pressure distribution parameters a second time to obtain the final optimization result. After obtaining the final optimization results, the problem of insufficient filling in the corner area was verified to determine the improvement effect in the cold isostatic pressing process.

[0011] A further preferred technical solution is that, in step S3, key information is extracted from the coordinate data of the insufficient filling positions, and a pre-established analysis tool is used to scan the distribution status of the corner area to obtain the range of abnormal distribution areas and determine the key positions that need to be processed. Based on the area range of the abnormal distribution, data is collected on the changes in density gradient to obtain corresponding measurement values. These values ​​are then compared with preset thresholds to determine whether there are any outliers outside the range. If there are outliers exceeding the preset threshold, the pressure gradient parameters are initially adjusted to generate temporary pressure distribution data, thus obtaining a preliminary corrected distribution state.

[0012] A further optimized technical solution involves simulating and analyzing the initially corrected distribution state, using the finite element analysis method to dynamically calculate the powder flow path in the corner region, obtaining intermediate results of flow guidance, and determining the direction of path adjustment. Based on the direction of path adjustment, the pressure distribution data is optimized a second time to construct the final pressure field distribution model and obtain a flow guidance scheme suitable for corner areas. The final flow guidance scheme is used to simulate and verify the formation of voids in the corner area, obtain the verified distribution data, and determine whether the preset uniformity conditions are met. If the verified distribution data still does not meet the preset uniformity conditions, the pressure gradient parameters are fine-tuned, the flow guidance scheme is updated, and the adjusted final result is obtained.

[0013] In a further optimized technical solution, in step S4, pressure field distribution data is extracted from the powder flow guidance scheme, and a swarm intelligence optimization algorithm is used to perform a preliminary analysis of the pressing parameter combination to obtain a set of initial pressing time and pressure curve configurations. Based on the initial pressing time and pressure curve configuration, data calibration is performed to address local differences in the pressure field distribution, and the range of parameter combinations after calibration is determined. Using the calibrated parameter combination range, a preliminary simulation of the density uniformity of the part was performed in a virtual forming environment to obtain density distribution data during the simulation process; If there are local non-uniformities in the density distribution data during the simulation process, the parameters are fine-tuned for the pressure field distribution in the non-uniform areas to obtain the adjusted combination of suppression parameters.

[0014] The further optimized technical solution, through the adjusted combination of pressing parameters, re-verifies the density uniformity in the virtual molding environment and obtains updated density distribution data; If the updated density distribution data still has deviations, the direction of the pressure curve adjustment will be optimized a second time to determine the final combination of suppression parameters.

[0015] A further optimized technical solution involves, in step S5, driving the cold isostatic pressing equipment to operate by pressing parameters, collecting sensor data in real time, and obtaining feedback data on powder density. The equipment is driven to run again by adjusting the pressing parameter set, and the powder density sensor data is monitored in real time to determine whether the deviation range has returned to the preset range, so as to obtain the final verified molding process parameter set.

[0016] A further optimized technical solution involves using a pre-established comparison model based on the obtained powder density feedback data to determine the range of deviation between the feedback data and the simulation prediction, and to ascertain whether the deviation is within a preset range. If the deviation exceeds the preset range, anomaly detection is performed on the feedback data to obtain the distribution area of ​​the abnormal data and obtain the preliminary location result of the abnormal area. Based on the preliminary location results of the abnormal regions, the support vector machine algorithm is used to classify the abnormal data and determine the main source categories of the abnormal data. Based on the main source categories of abnormal data, relevant records in the equipment operation log are obtained to determine whether the equipment operation status interferes with the powder density, and the analysis results of the status interference are obtained. If the state interference analysis results show that there is interference, then the equipment operating parameters are fine-tuned to obtain the adjusted operating parameter combination and determine the new suppression parameter set.

[0017] The advantages and beneficial effects of this invention are as follows: This invention discloses a comprehensive optimization method for addressing the problem of insufficient filling in corner areas during the cold isostatic pressing process of metal alloy powders. It aims to solve the operational challenges of uneven powder distribution, abnormal density gradients, and void formation in the pressing of complex-shaped parts. By constructing an initial model of powder flow behavior based on interparticle interactions, and combining it with numerical simulation methods to accurately predict the dynamic filling path and density gradient values ​​in corner areas, this invention iteratively adjusts the pressure gradient parameters when insufficient filling is detected, generating a corrected pressure field distribution model, and calculates the optimal combination of pressing parameters using a swarm intelligence optimization algorithm. If the density uniformity is not met, additive ratio data is incorporated to update the parameters, ultimately driving the actual equipment operation and verifying the process parameters. This invention significantly improves the filling efficiency in corner areas and the overall density uniformity of parts, providing reliable technical support for the high-quality molding of complex-shaped parts. Attached Figure Description

[0018] Figure 1 This is a flowchart of a method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings and examples. The following examples are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0020] like Figure 1 As shown in this embodiment, a method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder is disclosed. The method includes: S1: Obtain particle size distribution data and viscosity parameters of metal alloy powder. The viscosity parameters are used to characterize flow resistance. Analyze the surface morphology of the powder through electron microscopy imaging to obtain an initial model of powder flow behavior. The initial model of powder flow behavior is based on interparticle interaction and is used to simulate the powder distribution prediction in the pressing process of complex shaped parts. S2: Based on the initial model of powder flow behavior, the dynamic filling path of powder in the corner region during cold isostatic pressing is simulated using numerical simulation methods. S3: Determine that the density gradient value at the coordinates of the insufficient filling location exceeds a preset threshold, and obtain an improved powder flow guidance scheme to reduce the formation of voids in the corner area; simulate and analyze the size change according to the determined density change information to obtain the dimensional shrinkage rate of titanium alloy powder during cold isostatic pressing; S4: Extract the key pressure field distribution model from the improved powder flow guiding scheme, and determine the optimal pressing time and pressure curve; S5: Drive the actual cold isostatic pressing equipment by determining the combination of pressing parameters, monitor the real-time powder density sensor feedback data, determine whether the deviation between the sensor data and the simulation prediction is within the allowable range, and obtain the final verified molding process parameters.

[0021] In step S1, the particle size distribution data and viscosity parameters of the metal alloy powder are obtained. The particle size distribution data is obtained through particle size measurement, and the viscosity parameters are used to characterize the flow resistance. The surface morphology characteristics of the powder are analyzed by electron microscopy imaging to obtain an initial model of powder flow behavior. The initial model of powder flow behavior is based on inter-particle interaction and is used to simulate the powder distribution prediction in the pressing process of complex shaped parts.

[0022] Metal alloy powder is analyzed using particle size analyzers to obtain particle size and distribution data, which are then stored in a pre-established database to obtain basic particle distribution information. Based on the obtained particle size and distribution data, combined with rheometer test results, viscosity parameters are calculated to determine the flow resistance characteristics of the powder under different shear conditions.

[0023] Electron microscopy was used to image and analyze powder samples, extract surface morphology and morphological feature data, and store them as a high-resolution image dataset to obtain surface detail information of powder particles.

[0024] For the extracted surface morphology and morphological feature data, a pre-established image processing algorithm is used to quantize the features, determine the contact patterns and potential interaction strength between particles, and if the feature quantization results exceed the preset threshold, the image data is corrected a second time to obtain the preliminary evaluation results of the interaction between particles.

[0025] Based on the preliminary evaluation results of interparticle interactions, combined with viscosity parameters and flow resistance characteristics, an initial model of powder flow behavior was constructed to simulate the flow behavior of the powder and determine the behavioral response data under different pressing conditions.

[0026] Based on the flow behavior response data simulated by the initial powder flow behavior model, and combined with the geometric constraints of the pressing process for complex-shaped parts, the finite element method is used to predict the powder distribution during the pressing process, yielding the distribution prediction results. Based on these distribution prediction results, and combined with the process parameters in the actual pressing process, the parameter settings of the initial powder flow behavior model are adjusted through data comparison and iterative optimization to determine the final powder distribution prediction model.

[0027] In some embodiments, when testing metal alloy powder with a particle size measuring device, a laser particle size analyzer can be used to obtain particle size and distribution data. Assuming that the test results show that the powder particle size is mainly concentrated in the range of 10-50 micrometers and the distribution is normally distributed, this provides basic data for subsequent flow behavior analysis. Such data is stored in a database, which is convenient for traceability and analysis and helps to accurately grasp the basic characteristics of the powder.

[0028] When calculating viscosity parameters by combining rheometer test results, the stress response of the powder at different shear rates can be measured by the rheometer. For example, if the viscosity is 500 Pa·s at a low shear rate and drops to 200 Pa·s at a high shear rate, it indicates that the powder has shear thinning characteristics. This helps to determine the flow resistance characteristics of the powder during the pressing process and provides a basis for process optimization.

[0029] When using an electron microscope to image and analyze powder samples, high-resolution images of the particle surface can be obtained. If the images show that the particle surface has high roughness and many tiny protrusions, this may affect the contact behavior between particles. After storing these image data, an intuitive basis can be provided for subsequent morphological feature quantification, thereby improving the accuracy of the analysis.

[0030] When using image processing algorithms for feature quantization, parameters such as particle roundness and surface roughness can be extracted. If more than 30% of the particles have a roundness value below 0.8 and the preset threshold is 0.85, the image data needs to be corrected a second time to more accurately assess the strength of the interaction between particles. This can effectively reduce the risk of misjudgment and ensure the reliability of the assessment results.

[0031] When constructing an initial model to simulate powder flow behavior, viscosity parameters and interparticle interaction data can be combined. It can be assumed that the powder flows uniformly under low pressure, while local accumulation occurs under high pressure. This helps to predict the behavioral response under different pressing conditions and provides a reference for process design.

[0032] When using the finite element analysis method to predict the powder distribution state, the geometric constraints in the pressing process of complex-shaped parts can be simulated. Assuming that the part has narrow channels, the prediction results show that the powder is unevenly distributed in the channels, which provides important information for subsequent process adjustments and helps to reduce defects.

[0033] When adjusting the initial model parameters for powder flow behavior, the predicted results can be compared with process parameters such as pressure and speed during the actual pressing process. If the powder filling rate in the actual pressing process is 10% lower than the predicted value, then parameters such as the particle friction coefficient in the model need to be optimized to ultimately determine a prediction model that is closer to reality. This can significantly improve prediction accuracy and reduce trial and error costs in production.

[0034] In step S2, based on the initial model of powder flow behavior, a numerical simulation method is used to simulate the dynamic filling path of powder in the corner region during cold isostatic pressing, and to determine the coordinates of the incomplete filling location and the density gradient value. The density gradient value is used as the input basis for optimizing the pressure distribution.

[0035] The dynamic behavior of powder flow during cold isostatic pressing (COP) is analyzed using numerical simulation. Detailed calculations are performed on the filling path in corner regions to obtain the coordinates of incomplete filling locations and determine the distribution status of key areas. If the density gradient value at the incomplete filling location exceeds a preset threshold, an improved powder flow guidance scheme is derived to reduce void formation in corner regions. Dimensional changes are simulated and analyzed based on the determined density change information to obtain the dimensional shrinkage rate of titanium alloy powder during COP. A pre-established analysis tool is used to process the density gradient data to determine if the density change exceeds a preset threshold range. If it does, the relevant coordinates are marked to determine the area requiring adjustment. For the marked areas requiring adjustment, an optimized adjustment scheme is constructed based on the pressure distribution input, and the adjusted pressure distribution parameters are obtained, resulting in improved data suitable for corner regions.

[0036] The parameters of the initial model of powder flow behavior are updated by improving the data, simulating the changes in the filling path of powder flow under the adjusted pressure distribution, and determining the new filling state information. Based on the new filling state information, the changing trend of the density gradient is analyzed to determine whether the preset uniformity condition is met. If not, the pressure distribution parameters are adjusted a second time to obtain the final optimization result. After obtaining the final optimization result, the problem of insufficient filling in the corner area is verified. The finite element analysis method is used to simulate the dynamic performance of powder flow and determine the improvement effect in the cold isostatic pressing process.

[0037] In some embodiments, when analyzing the dynamic performance of powder flow during cold isostatic pressing using numerical simulation, the filling path in corner regions can be considered first. Due to the complexity of their geometry, corner regions are often areas prone to incomplete filling. Suppose that during the simulation, it is found that the powder flow velocity at the corner is significantly lower than that in the straight area, especially when the corner angle is less than 45 degrees, the flow path shows a clear sluggish trend. In this case, the motion trajectory of powder particles at the corner can be recorded using simulation software to obtain the specific coordinate information of the incomplete filling location. For example, if the coordinates of a key point are x=12.5mm, y=8.3mm, z=5.0mm, this area can be identified as the key distribution point of incomplete filling.

[0038] Because cold isostatic pressing applies water pressure to powder from all directions, the shrinkage in each direction is inconsistent for molds of different shapes and sizes. This usually requires custom molds for continuous trial and optimization, which is costly and time-consuming. Therefore, numerical simulation is used to analyze the shrinkage of powder before and after cold isostatic pressing based on density changes. The simulation measures the shrinkage in the x, y, and z directions of circular, polygonal, arc, and irregular shapes under different aspect ratios to obtain specific values. Molds can then be designed based on the simulation results, which can significantly reduce mold and time costs.

[0039] For example, suppose the desired shape is a cylindrical form with a maximum diameter of φ42 and a height of 170mm. In actual cold isostatic pressing, molds with different aspect ratios shrink differently. Verifying this experimentally would require significant manpower and resources. However, through numerical simulation, 1000 points (n1~n1000) are selected on the outer side of the pre-defined mold shape. The coordinate changes of these 1000 points are simulated as the powder density increases from 2.0g / cm³ to 3.2g / cm³. For instance, if the initial coordinates of n1 are (50, 0, 200), and after simulated cold isostatic pressing, the coordinates of n1 become (42, 0, 170), indicating a shrinkage rate of 16.2% in the x-direction, 15% in the z-direction, and no shrinkage in the y-direction. Similarly, the shrinkage rates from n2 to n1000 are obtained. If the shape after shrinkage basically meets the desired product requirements, the simulation is complete. The mold size and shape are then designed based on these simulation results. If the simulation results do not match the product requirements, a second simulation is performed until the ideal result is obtained. This method can quickly obtain mold data, saving considerable manpower and resources.

[0040] When calculating the density gradient data in the corner region, the packing state of powder particles can be analyzed based on the simulation results; assuming that the density in the corner region gradually decreases from the center to the edge, with a center density of 3.2 g / cm³. 3 The edge concentration is only 2.6-2.8 g / cm³. 3 The density gradient changes significantly; by processing these data using pre-established analytical tools, if the preset density change threshold is 0.5 g / cm³, 3 The actual change exceeded 1.0 g / cm³. 3 Then, the area exceeding the threshold is marked to determine the specific range that needs to be adjusted, for example, marking an area with a radius of 3mm centered at coordinate x=12.5mm.

[0041] When constructing an optimization and adjustment scheme for the marked area, the local pressure parameters can be adjusted based on the input of pressure distribution. Assuming the initial pressure is 100MPa, it can be adjusted to 120MPa, and the response change of powder flow in the simulation can be observed to obtain the improved pressure distribution parameters. This adjustment helps to improve the filling state of the corner area, and the simulation results may show that the filling rate increases from 75% to 85%.

[0042] By updating the initial model of powder flow behavior with improved data, the changes in the powder flow's filling path can be re-simulated; assuming that adjusting the pressure increases the flow velocity in the corner region by 10%, the new filling state shows that the density gradient has decreased to 0.3 g / cm³. 3 Within; if the preset uniformity condition is 0.4 g / cm³ 3 If the pressure is within acceptable limits, the requirement is met; otherwise, the pressure needs to be adjusted again to 125 MPa until the expected pressure is achieved.

[0043] When verifying the final optimization results, the finite element analysis method was used to simulate the dynamic performance of the cold isostatic pressing process. It was assumed that the verification results showed that the problem of insufficient filling in the corner area was reduced by 80% and the density distribution was more uniform. This verification method can intuitively reflect the feasibility of the adjustment scheme and provide a reliable basis for subsequent process optimization.

[0044] In step S3, if the density gradient value at the coordinates of the insufficient filling location exceeds a preset threshold, then the pressure gradient parameter is iteratively adjusted to generate a corrected pressure field distribution model, thereby obtaining an improved powder flow guidance scheme to reduce void formation in corner regions.

[0045] By extracting key information from the coordinate data of the insufficient filling locations, a pre-established analysis tool is used to scan the distribution status of the corner area to obtain the range of abnormal distribution areas and determine the key locations that need to be addressed. Based on the range of abnormal distribution areas, data is collected on the changes in density gradient to obtain corresponding measurement values, which are then compared with a preset threshold to determine whether there are any abnormal points exceeding the range. If there are abnormal points exceeding the preset threshold, the pressure gradient parameters are initially adjusted to generate temporary pressure distribution data, resulting in a preliminarily corrected distribution status.

[0046] By simulating and analyzing the initially corrected distribution state, the powder flow path in the corner region is dynamically calculated using the finite element method to obtain intermediate results of flow guidance and determine the direction of path adjustment. Based on the direction of path adjustment, secondary optimization is performed on the pressure distribution data to construct the final pressure field distribution model, resulting in a flow guidance scheme suitable for the corner region. The void formation in the corner region is simulated and verified using the final flow guidance scheme, and the verified distribution data is obtained to determine whether it meets the preset uniformity conditions. If the verified distribution data still does not meet the preset uniformity conditions, the pressure gradient parameters are fine-tuned, the flow guidance scheme is updated, and the adjusted final result is obtained.

[0047] Specifically, when extracting coordinate data of incomplete filling locations, we can first focus on the distribution of key points in corner areas. Suppose that during the scanning process, the coordinate points of a certain corner area are found to be concentrated in a narrow range, such as around x=10.2mm, y=7.5mm, z=4.8mm, which may mean that there is a significant filling defect in this area. Using pre-established analysis tools, we can perform cluster analysis on these coordinate points to determine the range of abnormal distribution areas. For example, a spherical area with a radius of 2.5mm centered on this coordinate is marked as a key processing location. This method helps to quickly locate the problem area.

[0048] For data acquisition regarding density gradient changes, density distribution values ​​within a corner region can be obtained using sensors or simulation software; assuming the measurement results show that the density in a certain area is 3.2 g / cm³ from the center... 3 Gradually decreasing to the edge at 2.6 g / cm 3 The variation range is relatively large; combined with the preset threshold of 0.6 g / cm³ 3 After comparison, some points were found to be outside the range and were marked as anomalies; this method of data collection and comparison can provide accurate basis for subsequent adjustments.

[0049] When making initial adjustments to the pressure gradient parameters, temporary pressure distribution data can be generated based on the location and density changes of anomalies. Assuming the initial pressure is 90 MPa, it can be locally increased to 110 MPa in the anomaly area, and the distribution after the initial correction can be observed. This adjustment method can provide basic data support for subsequent simulations.

[0050] When using the finite element method to dynamically calculate the powder flow path in the corner region, the movement trajectory of particles under different pressures can be simulated. If the intermediate results show that the flow path still has a certain degree of stagnation at the corner, the direction of path adjustment can be determined to be either increasing the local pressure or optimizing the geometric constraints. This dynamic calculation method helps to clarify the specific direction of adjustment.

[0051] When performing secondary optimization of pressure distribution data, a final pressure field distribution model can be constructed based on intermediate results. Assuming that after multiple simulations, adjusting the pressure to 115 MPa stabilizes the flow guidance scheme in the corner region, this optimization method can provide a more suitable solution for actual processes.

[0052] When simulating and verifying the formation of voids in the corner region using the final flow guidance scheme, distribution data can be obtained and it can be determined whether the uniformity condition is met. Assuming that the verification results show that the void ratio decreases from 8% to 3%, which is close to the preset condition, this verification method can provide a reliable reference for process improvement.

[0053] If the distribution data still does not meet the requirements, the pressure gradient parameters are fine-tuned, for example, by slightly increasing the pressure to 118 MPa, and the flow guidance scheme is updated. This fine-tuning method can gradually approach the ideal state, ensuring that the final result meets expectations.

[0054] In step S4, a key pressure field distribution model is extracted from the improved powder flow guidance scheme. A swarm intelligence optimization algorithm is used to calculate multiple sets of pressing parameter combinations, determine the optimal pressing time and pressure curve, and output the results to a virtual molding environment to verify the overall density uniformity of the part.

[0055] By extracting pressure field distribution data from the powder flow guidance scheme, a swarm intelligence optimization algorithm is used to conduct a preliminary analysis of the pressing parameter combination, resulting in an initial pressing time and pressure curve configuration. Based on the initial pressing time and pressure curve configuration, data calibration is performed to address local differences in the pressure field distribution, determining the range of calibrated parameter combinations. Using the calibrated parameter combination range, a preliminary simulation of the part density uniformity is conducted in a virtual molding environment to obtain density distribution data during the simulation process. If local non-uniformity exists in the density distribution data during the simulation process, the parameters are fine-tuned for the pressure field distribution in the non-uniform areas, resulting in an adjusted pressing parameter combination.

[0056] By adjusting the pressing parameter combination, density uniformity is re-verified in the virtual molding environment to obtain updated density distribution data. If the updated density distribution data still has deviations, the direction of the pressure curve adjustment is optimized a second time to determine the final pressing parameter combination. Based on the final pressing parameter combination, complete pressing process data suitable for the virtual molding environment is generated, and the adjusted powder flow guidance configuration is output.

[0057] In some embodiments, if the overall density uniformity index of the part in the virtual forming environment is lower than the target value, additional powder additive ratio data is obtained and incorporated into the calculation loop of the swarm intelligence optimization algorithm to obtain an updated pressing parameter combination, which is used to improve the filling efficiency of the corner area.

[0058] Specifically, the proportion data of powder additives is obtained from the database source, and combined with the density distribution of parts in the virtual molding environment, a swarm intelligence optimization algorithm is used to perform preliminary parameter iteration calculations to obtain a set of updated pressing parameters. Based on the obtained updated pressing parameters, the density distribution of parts is simulated and verified in the virtual molding environment, and the simulated density distribution data is obtained to determine whether it meets the preset standards.

[0059] If the simulated density distribution data does not meet the preset standard, data analysis is performed on local areas of the part's density distribution. Additional powder additive ratio data is extracted from the database to determine a new formulation. Using this new formulation, a swarm intelligence optimization algorithm is used to iteratively adjust the pressing parameters, obtaining the adjusted parameter combination data. If the simulation results in the virtual molding environment still do not meet the uniformity index, targeted processing is performed on the density optimization of local areas to obtain further powder additive ratio adjustment data. Based on the obtained powder additive ratio adjustment data, the pressing parameter combination is updated, and the density distribution simulation is repeated in the virtual molding environment to obtain the final density distribution verification result. The final density distribution verification result is used to determine whether the preset standard is met. If it is still not met, historical data is extracted from the database for comparative analysis to determine the final adjustment direction of the pressing parameters.

[0060] When obtaining powder additive ratio data from a database source, the influence of different ratio schemes on part density can be extracted from historical records. Assuming that the database stores the density distribution when the additive ratio is 5%, 10%, and 15% in previous experiments, the analysis shows that the 10% ratio can effectively improve the local density unevenness under specific pressing conditions. This data extraction method helps to provide a reference benchmark for subsequent optimization and ensures the rationality of parameter adjustment direction.

[0061] In one possible implementation, when simulating the density distribution of parts in a virtual molding environment, the density values ​​of different regions of the part can be analyzed by region. For example, the density value of the central region of the part is 4.5 g / cm3, while the density value of the edge region is only 4.1 g / cm3, which is significantly lower than the standard value of 4.43 g / cm3. By comparing the partitioned data, the problem area can be quickly located, providing a targeted basis for subsequent parameter iteration.

[0062] When using swarm intelligence optimization algorithms for preliminary parameter iteration calculations, the combined effect of multiple sets of suppression parameters can be simulated.

[0063] In one embodiment, the initial compression time is set to 10 seconds and the pressure to 50 MPa. After algorithm iteration, a new combination of compression time of 120 seconds and pressure of 180 MPa may be obtained. This iterative process finds a better solution by simulating group behavior, which can effectively improve the adaptability of parameter configuration.

[0064] In one possible implementation, when evaluating the simulated density distribution data, if local areas are found to still have densities below the standard value, the powder flow characteristics of those areas can be analyzed in more detail. For example, if the lower density in edge areas is due to uneven powder particle distribution, historical data with an additive ratio of 12% can be extracted from the database and adjusted to a new formulation based on the current environment. This method helps to accurately address localized problems.

[0065] When performing a second iteration of the compression parameters, the compression time can be increased or the pressure distribution can be adjusted to address the issue of insufficient density in local areas. For example, if the local pressure in the edge area is increased from 55MPa to 60MPa and the compression time is extended to 15 seconds, the simulation results show that the density value increases from 2.2g / cm3 to 2.4g / cm3. This fine-tuning method can gradually approach the preset standard.

[0066] In one possible implementation, if the adjusted parameter combination still does not meet the uniformity index, targeted processing can be carried out on local areas; for example, the proportion of powder additives in a specific area can be increased to 13%, and the virtual molding environment can be used to re-simulate and observe whether the density distribution tends to be uniform. This regional optimization strategy can effectively improve the overall consistency.

[0067] For example, when judging whether the standard has been met by verifying the final density distribution, if there is still a deviation, comparative data of historical compression parameters and density distribution can be extracted from the database. Assuming that the density distribution is relatively ideal when the compression time is 14 seconds and the pressure is 58 MPa in the historical data, it can be used as a reference direction for the current adjustment. This historical data comparative analysis can provide reliable support for the final parameter determination.

[0068] In step S5, the actual cold isostatic pressing equipment is driven by the updated pressing parameter combination, the real-time powder density sensor feedback data is monitored, and it is determined whether the deviation between the sensor data and the simulation prediction is within the allowable range, so as to obtain the final verified molding process parameter set.

[0069] The cold isostatic pressing equipment is driven by pressing parameters, and sensor data is collected in real time to obtain powder density feedback data. Based on the acquired powder density feedback data, a pre-established comparison model is used to determine the deviation range between the feedback data and the simulation prediction, and to determine whether the deviation is within the preset range.

[0070] If the deviation exceeds the preset range, outlier detection is performed on the feedback data to obtain the distribution area of ​​the outlier data and obtain the preliminary location result of the outlier area. Based on the preliminary location result of the outlier area, the support vector machine algorithm is used to classify the outlier data and determine the main source category of the outlier data.

[0071] Based on the main source categories of the abnormal data, relevant records are retrieved from the equipment operation log to determine whether the equipment operating status interferes with powder density, thus obtaining the analysis results of state interference. If the state interference analysis results indicate the presence of interference, the equipment operating parameters are fine-tuned to obtain the adjusted combination of operating parameters and determine a new set of pressing parameters.

[0072] The equipment is driven to run again by adjusting the pressing parameter set, and the powder density sensor data is monitored in real time to determine whether the deviation range has returned to the preset range, so as to obtain the final verified molding process parameter set.

[0073] When a cold isostatic pressing (COP) device is driven by pressing parameters, sensor data can be collected in real time to obtain feedback data on powder density. Sensor data typically includes key indicators such as pressure distribution and density values, aiming to monitor the real-time status of the molding process. For example, if in a certain run, the sensor detects a powder density of 2.3 g / cm³, while the expected value is 2.5 g / cm³, indicating a significant deviation, this real-time data acquisition provides a basis for subsequent analysis.

[0074] To determine the discrepancy between feedback data and simulated predictions, a pre-established comparative model can be used for analysis. Suppose the comparative model sets the deviation range at ±0.1 g / cm³, while the actual deviation reaches 0.2 g / cm³, exceeding the preset range; in this case, further investigation is needed to pinpoint the source of the problem. Such comparative models are typically built based on historical data and simulation results, enabling rapid identification of anomalies.

[0075] When detecting anomalies, preliminary location results of abnormal areas can be obtained through data distribution analysis. For example, if the detection finds that low density is mainly concentrated at the edges of parts, while the data in the central area is normal, this zoning analysis helps narrow down the problem area and provides direction for subsequent classification and processing.

[0076] When using the Support Vector Machine (SVM) algorithm to classify outlier data, the outlier data can be categorized into factors such as equipment factors and material factors. Assuming the classification results indicate that the outlier mainly stems from uneven pressure distribution in the equipment, the next step is to focus on checking the equipment's operating status. This classification method can effectively pinpoint the root cause of the problem.

[0077] When analyzing equipment operation logs, records related to pressure fluctuations or time control can be extracted. For example, if the logs show a brief pressure drop during operation, from 50 MPa to 45 MPa, this could potentially lead to insufficient density; such analysis helps determine whether the equipment status is interfering with density.

[0078] If a state disturbance is confirmed, fine-tuning of the equipment operating parameters is necessary. For example, adjusting the pressure from 45 MPa back to 50 MPa and appropriately extending the pressing time to 12 seconds yields a new parameter combination; this fine-tuning method can specifically address the problem area.

[0079] By driving the equipment again with the adjusted parameter set, the powder density data can be monitored in real time. Assuming that the adjusted density value increases from 2.3 g / cm3 to 2.5 g / cm3, the deviation returns to the preset range, verifying the effectiveness of the parameters. This repeated verification process ensures the reliability of the final molding process parameter set.

[0080] Throughout the process, the implementation methods for each technical topic are closely centered around powder density optimization, forming a complete closed loop from data acquisition to parameter adjustment. This systematic approach can effectively improve process stability, ensure the consistency of part quality, and save manpower and resources spent on actual verification.

[0081] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for preparing low-oxygen irregular-shaped parts by cold isostatic pressing of titanium alloy powder, characterized in that, The method includes: S1: Obtain the particle size distribution data and viscosity parameters of the metal alloy powder. The viscosity parameters are used to characterize the flow resistance. The surface morphology of the powder is analyzed by electron microscopy imaging to obtain an initial model of powder flow behavior. The initial model of powder flow behavior is based on inter-particle interaction and is used to simulate the powder distribution prediction in the pressing process of complex shaped parts. S2: Based on the initial model of powder flow behavior, the dynamic filling path of powder in the corner region during cold isostatic pressing is simulated using numerical simulation methods. S3: If the density gradient value at the coordinates of the insufficient filling location exceeds the preset threshold, an improved powder flow guidance scheme is obtained to reduce the formation of voids in the corner area; after determining the powder filling path, the mold size is determined by numerically simulating the shrinkage change of powder in three-dimensional space during isostatic pressing. S4: Extract the key pressure field distribution model from the improved powder flow guiding scheme, and determine the optimal pressing time and pressure curve; If there are local non-uniformities in the density distribution data during the simulation process, the parameters are fine-tuned for the pressure field distribution in the non-uniform region to obtain the adjusted combination of suppression parameters. S5: Drive the actual cold isostatic pressing equipment to run by the determined combination of pressing parameters, monitor the real-time powder density sensor feedback data, determine whether the deviation between the sensor data and the simulation prediction is within the allowable range, and obtain the final verified molding process parameters. In step S2, the dynamic performance of powder flow during cold isostatic pressing is analyzed by numerical simulation. Detailed calculations are performed on the filling path in the corner area to obtain the specific coordinate information of the incomplete filling location and determine the distribution status of the key area. Based on the obtained coordinate information of the incomplete filling location, combined with the dynamic analysis results, the density gradient data in the corner area is calculated to obtain detailed distribution information of density change; according to the determined density change information, the size change is simulated and analyzed to obtain the dimensional shrinkage rate of titanium alloy powder during cold isostatic pressing. The density gradient data is processed using a pre-established analysis tool to determine whether the density change exceeds a preset threshold range. If it does, the relevant location coordinates are marked to determine the area that needs to be adjusted. In step S3, key information is extracted from the coordinate data of the insufficient filling positions, and a pre-established analysis tool is used to scan the distribution status of the corner area to obtain the range of abnormal distribution areas and determine the key positions that need to be processed. Based on the area range of the abnormal distribution, data is collected on the changes in density gradient to obtain corresponding measurement values. These values ​​are then compared with preset thresholds to determine whether there are any outliers outside the range. If there are outliers exceeding the preset threshold, the pressure gradient parameters are initially adjusted to generate temporary pressure distribution data and obtain a preliminary corrected distribution state. In step S4, pressure field distribution data is extracted from the powder flow guidance scheme, and a swarm intelligence optimization algorithm is used to perform a preliminary analysis of the pressing parameter combination to obtain a set of initial pressing time and pressure curve configurations. Based on the initial pressing time and pressure curve configuration, data calibration is performed to address local differences in the pressure field distribution, and the range of parameter combinations after calibration is determined. Using the calibrated parameter combination range, a preliminary simulation of the density uniformity of the part was performed in a virtual molding environment to obtain density distribution data during the simulation process.

2. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 1, characterized in that, In step S1, the metal alloy powder is tested using a particle size measuring device to obtain particle size and distribution data, which is then stored in a pre-established database to obtain basic distribution information of the powder particles. Electron microscopy was used to image and analyze powder samples, extract surface morphology and morphological feature data, and store them as a high-resolution image dataset to obtain surface detail information of powder particles; For the extracted surface morphology and morphological feature data, a pre-established image processing algorithm is used to quantize the features, determine the contact patterns and potential interaction strength between particles, and if the feature quantization result exceeds the preset threshold, the image data is corrected a second time to obtain the preliminary evaluation results of the interaction between particles. Based on the preliminary evaluation results of interparticle interactions, combined with viscosity parameters and flow resistance characteristics, an initial model of powder flow behavior is constructed.

3. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 2, characterized in that, For the marked area that needs adjustment, an optimized adjustment scheme is constructed based on the input data of pressure distribution, and the adjusted pressure distribution parameters are obtained to obtain improved data suitable for corner areas; By improving the data, the parameters of the initial model of powder flow behavior are updated, the filling path of powder flow under the adjusted pressure distribution is simulated, and new filling state information is determined. Based on the new filling state information, analyze the trend of density gradient change and determine whether the preset uniformity condition is met. If not, adjust the pressure distribution parameters a second time to obtain the final optimization result. After obtaining the final optimization results, the problem of insufficient filling in the corner area was verified to determine the improvement effect in the cold isostatic pressing process.

4. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 1, characterized in that, By simulating and analyzing the initially corrected distribution state, the powder flow path in the corner region is dynamically calculated using the finite element analysis method to obtain intermediate results of flow guidance and determine the direction of path adjustment. Based on the direction of path adjustment, the pressure distribution data is optimized a second time to construct the final pressure field distribution model and obtain a flow guidance scheme suitable for corner areas. The final flow guidance scheme is used to simulate and verify the formation of voids in the corner area, obtain the verified distribution data, and determine whether the preset uniformity conditions are met. If the verified distribution data still does not meet the preset uniformity conditions, the pressure gradient parameters are fine-tuned, the flow guidance scheme is updated, and the adjusted final result is obtained.

5. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 1, characterized in that, By adjusting the combination of pressing parameters, density uniformity was re-verified in the virtual molding environment to obtain updated density distribution data. If the updated density distribution data still has deviations, the direction of the pressure curve adjustment will be optimized a second time to determine the final combination of suppression parameters.

6. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 1, characterized in that, In step S5, the cold isostatic pressing equipment is driven by pressing parameters to collect sensor data in real time and obtain feedback data on powder density. The equipment is driven to run again by adjusting the pressing parameter set, and the powder density sensor data is monitored in real time to determine whether the deviation range has returned to the preset range, so as to obtain the final verified molding process parameter set.

7. The method for preparing low-oxygen irregular parts by cold isostatic pressing of titanium alloy powder according to claim 6, characterized in that, Based on the obtained powder density feedback data, a pre-established comparison model is used to determine the deviation range between the feedback data and the simulation prediction, and to determine whether the deviation is within the preset range. If the deviation exceeds the preset range, anomaly detection is performed on the feedback data to obtain the distribution area of ​​the abnormal data and obtain the preliminary location result of the abnormal area. Based on the preliminary location results of the abnormal regions, the support vector machine algorithm is used to classify the abnormal data and determine the main source categories of the abnormal data. Based on the main source categories of abnormal data, relevant records in the equipment operation log are obtained to determine whether the equipment operation status interferes with the powder density, and the analysis results of the status interference are obtained. If the state interference analysis results show that there is interference, then the equipment operating parameters are fine-tuned to obtain the adjusted operating parameter combination and determine the new suppression parameter set.

Images