Metal powder recovery system for additive manufacturing of rocket engine thrust chamber

Abstract

The invention relates to the technical field of rocket engine thrust chamber additive manufacturing, and discloses a metal powder recovery system for rocket engine thrust chamber additive manufacturing. Comprising a powder collection module, a screening classification module, a purification treatment module, a particle size adjustment module and an intelligent control module. The powder collecting module captures residual metal powder and collects data to generate a data set; the screening and grading module performs particle size separation by using a joint algorithm; the purification treatment module is used for constructing a multi-stage purification model to remove an oxide layer and decompose organic matters; the particle size adjustment module designs a composite regulation strategy to adjust particle size distribution; and the intelligent control module establishes an optimization model to generate a parameter adjustment instruction. In addition, the system is further provided with a powder characteristic analysis module, a circulation monitoring module and an exception handling module. The system can efficiently recover the metal powder, improve the quality and reutilization value of the metal powder, reduce the cost and guarantee the quality of additive manufacturing products, and is suitable for metal powder recovery treatment in the field of additive manufacturing of rocket engine thrust chambers.

Description

Technical Field

[0001] The present invention relates to the technical field of additive manufacturing of rocket engine thrust chambers, and particularly to a metal powder recycling system for additive manufacturing of rocket engine thrust chambers. Background Art

[0002] In the process of additive manufacturing of rocket engine thrust chambers, metal powder is a key consumable with a huge usage amount and high cost. With the continuous development of space technology, the performance and quality requirements for rocket engine thrust chambers are increasing day by day. Additive manufacturing technology has been widely used in this field due to its unique advantages. However, this has also brought the problem of metal powder recycling and treatment.

[0003] On the one hand, after additive manufacturing is completed, a large amount of metal powder will remain in the equipment. If these residual powders are not recycled, it will not only cause serious waste of resources, but also significantly increase production costs. According to statistics, in some large space manufacturing enterprises, the annual cost increase caused by ineffective recycling of metal powder can reach millions or even tens of millions of yuan. On the other hand, the residual powder often contains impurities, moisture, and has a problem of uneven particle size distribution. If directly reused, it will seriously affect the quality of additive manufacturing products. For example, impurities and moisture may cause defects such as pores and cracks inside the product, and uneven particle size distribution will affect the forming accuracy and mechanical properties of the product, thereby affecting the reliability and stability of the rocket engine thrust chamber and posing a potential threat to the safety of space missions.

[0004] Existing metal powder recycling technologies have many deficiencies. Most traditional recycling methods use simple mechanical screening or single purification means, and cannot accurately control the particle size distribution of the powder and effectively remove impurities. In the screening link, ordinary vibrating screens are difficult to finely classify fine particle powders, resulting in poor particle size consistency of the recycled powder. In terms of purification treatment, conventional cleaning and drying methods cannot completely remove the oxide layer and organic matter on the powder surface, making the purity of the recycled powder unable to meet the high standards required for additive manufacturing of rocket engine thrust chambers.

[0005] Moreover, the current recycling system has a low degree of intelligence, lacking real-time monitoring and precise control of various parameters in the recycling process. It is unable to dynamically optimize according to the characteristics of different batches of powders and actual situations such as recycling efficiency and energy consumption, resulting in high energy consumption and low efficiency in the recycling process. For example, when adjusting the parameters of the screening equipment, it often relies on manual experience and is difficult to achieve the best screening effect, wasting time and increasing labor costs.

[0006] In addition, there are also no effective means for the quality assessment and recycling management of recycled powders. It is difficult to accurately judge whether the performance of the recycled powders meets the standards for reuse, and to determine their reasonable number of recycling times, which easily leads to the overuse of recycled powders affecting product quality, or the premature discarding of powders still having utilization value, further exacerbating resource waste. Summary of the Invention

[0007] The purpose of the present invention is to provide a metal powder recycling system for additive manufacturing of a rocket engine thrust chamber to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: A metal powder recycling system for additive manufacturing of a rocket engine thrust chamber, the system includes: A powder collection module, which is used to capture the residual metal powder in the additive manufacturing equipment through a multi-stage negative pressure adsorption device, and simultaneously collect data on powder particle size distribution, humidity, and impurity content to generate a powder state data set; A screening and grading module, which is used to perform dynamic particle size separation on the powder state data set by using a combined algorithm of vibration screening and air flow separation to generate a set of graded powders; A purification and treatment module, which is used to construct a multi-stage purification model based on high-temperature degassing and chemical cleaning to remove the oxide layer and decompose organic substances from the set of graded powders to generate a purified powder sequence; A particle size adjustment module, which is used to design a composite control strategy based on ball milling and spray drying to adjust the particle size distribution of the purified powder sequence according to a preset particle size threshold to generate a regenerated powder sample; An intelligent control module, which is used to establish a multi-objective optimization model based on dynamic programming to generate parameter adjustment instructions according to real-time powder recycling efficiency and energy consumption constraints.

[0009] Preferably, the combined algorithm of vibration screening and air flow separation includes: Construct a multi-layer screen structure of the vibration screening device, and monitor the vibration frequency and amplitude of the screen through an acceleration sensor; Design a vortex intensity adjustment equation for the air flow separator, and generate a stratified air flow velocity gradient according to the powder density difference; Adopt a fuzzy control algorithm to dynamically match the screening efficiency and air flow separation accuracy, and output the multi-stage particle size interval division result.

[0010] Preferably, the construction of the multi-stage purification model includes: Set a gradient heating curve in the high-temperature degassing stage, and monitor the powder surface temperature distribution through an infrared thermal imager; Configure an acid-base alternating reaction tank in the chemical cleaning stage, and use a conductivity sensor to detect the ion concentration of the cleaning solution in real time; Remove suspended impurities through the combined process of residue precipitation and centrifugal separation, and generate a quality assessment report for the purified powder.

[0011] Preferably, the design of the composite regulation strategy includes: Define the rotation speed and grinding time of the ball mill as the first regulation variable, and establish a correlation model between the energy consumption of ball milling and the powder crushing efficiency; Define the atomization pressure and hot air temperature of the spray drying tower as the second regulation variable, and construct an optimization function for the powder sphericity and fluidity; Solve the process parameter combination of ball milling and spray drying through a collaborative optimization algorithm, and generate a particle size distribution curve of the recycled powder.

[0012] Preferably, the establishment of the multi-objective optimization model includes: Define the state space as a combined vector of powder recovery rate, energy consumption index, and equipment wear rate; Construct a framework of non-dominated sorting genetic algorithm with constraints, and optimize the control parameters by using the elitist retention strategy and crossover mutation operation; Design the objective function as the weighted difference between the recovery cost and the powder reuse value, and output the optimal operation parameter configuration scheme.

[0013] Preferably, the system further includes: A powder property analysis module for characterizing the crystal structure and morphology of the recycled powder through an X-ray diffractometer and a scanning electron microscope; Input the detection data into the powder performance database, and trigger an abnormal particle size warning signal through difference comparison.

[0014] Preferably, the generation of the powder state data set includes: Synchronously collect the powder flow rate, humidity sensor readings of the negative pressure adsorption device, and the impurity distribution map of the image recognition system; Use the principal component analysis algorithm to reduce the dimensionality of the multi-dimensional data and construct a low-dimensional feature matrix.

[0015] Preferably, the system further includes: A powder circulation monitoring module for establishing a correlation model between the powder circulation times and performance degradation; Use the particle swarm optimization algorithm to dynamically adjust the new and old powder ratio, and generate the maximum circulation times threshold.

[0016] Preferably, the generation of the parameter adjustment instruction includes: Construct a mixed integer programming model with time-varying constraints, with the equipment life and recovery efficiency as boundary conditions; use the branch and bound method to solve the multi-stage decision-making problem and generate a feasible solution set of the screen aperture and air flow velocity; Evaluate the comprehensive priority of the solution set through the grey relational analysis method, and output the real-time regulation instruction sequence.

[0017] Preferably, the system further includes: An exception handling module for constructing a powder defect pattern library based on an autoencoder; Wavelet packet transform is used to extract transient abnormal signals during the recycling process to generate an identification result of the powder pollution type.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: In terms of the efficiency and accuracy of powder recycling, the powder collection module can comprehensively capture residual metal powder in all corners of the additive manufacturing equipment through a multi-stage negative pressure adsorption device, greatly improving the powder recovery rate. At the same time, data such as powder particle size distribution, humidity, and impurity content are collected synchronously to generate a powder state data set, providing an accurate basis for subsequent processing. The screening and grading module adopts a combined algorithm of vibration screening and air flow separation, constructs a multi-layer screen structure and monitors the screen vibration, designs an adjustment equation for the vortex strength of the air flow separator based on the powder density difference, and then uses a fuzzy control algorithm to dynamically match the screening and separation accuracy, realizing high-precision particle size separation of the powder and effectively improving the quality of the recycled powder.

[0019] In terms of purification treatment, the multi-stage purification model has obvious advantages. In the high-temperature degassing stage, a gradient heating curve is set and an infrared thermal imager is used to monitor the temperature distribution to ensure uniform heating of the powder, effectively removing adsorbed gases and volatile impurities; in the chemical cleaning stage, an acid-base alternating reaction tank is configured, and the ion concentration of the cleaning solution is detected in real time by a conductivity sensor to accurately control the cleaning process. Combining the combined process of residue precipitation and centrifugal separation can completely remove suspended impurities, enabling the recycled powder to meet the high-purity standard and meet the strict requirements for powder quality in the additive manufacturing of rocket engine thrust chambers.

[0020] The composite regulation strategy designed by the particle size adjustment module has remarkable results. By defining relevant regulation variables of the ball mill and spray drying tower, establishing a correlation model and an optimization function, and using a collaborative optimization algorithm to solve the process parameter combination, the powder particle size distribution can be accurately adjusted, enabling the particle size of the regenerated powder sample to meet the preset particle size threshold, and improving the applicability and reuse value of the powder.

[0021] The intelligent control module establishes a multi-objective optimization model based on dynamic programming, comprehensively considers factors such as powder recovery rate, energy consumption index, and equipment wear rate, uses the weighted difference between the recovery cost and the powder reuse value as the objective function, and optimizes the control parameters through a constrained non-dominated sorting genetic algorithm. It can generate accurate parameter adjustment instructions according to the real-time powder recovery efficiency and energy consumption constraints, realizing the intelligent and efficient operation of the system, reducing the production cost, and extending the service life of the equipment.

[0022] The system also includes multiple auxiliary modules, which further enhance its practicability. The powder property analysis module characterizes the recycled powder through an X-ray diffractometer and a scanning electron microscope, inputs the data into a database for differential comparison, and can promptly trigger an abnormal particle size warning signal to ensure the powder quality. The powder circulation monitoring module establishes a correlation model between the powder circulation times and performance attenuation, uses the particle swarm optimization algorithm to dynamically adjust the new and old powder ratio, determines the maximum circulation times threshold, and realizes the scientific management of powder recycling. The abnormal handling module constructs a powder defect pattern library based on an autoencoder, extracts transient abnormal signals using wavelet packet transform to identify the powder pollution type, and provides strong support for promptly handling problems in the recycling process.

[0023] Overall, the metal powder recycling system of the present invention innovatively optimizes from multiple links, effectively solves the problems of resource waste, poor quality, and low intelligence level existing in the metal powder recycling of the prior art, greatly improves the recycling rate and reuse quality of metal powder, reduces production costs, improves the quality and efficiency of additive manufacturing of rocket engine thrust chambers, and provides key technical support for the sustainable development of the aerospace manufacturing field. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is the working principle diagram of the metal powder recycling system for additive manufacturing of rocket engine thrust chambers of the present invention; Figure 2 is the working principle diagram of the combined algorithm of vibration screening and air flow separation; Figure 3 is the working principle diagram of the multi-stage purification model; Figure 4 is the working principle diagram of the powder circulation monitoring module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0026] Please refer to Figures 1-4 , the present invention provides a metal powder recycling system for additive manufacturing of rocket engine thrust chambers, and the system includes: Powder collection module: Reasonably arrange multi-stage negative pressure adsorption devices inside the additive manufacturing equipment. These negative pressure adsorption devices can efficiently capture the residual metal powder in the equipment. While capturing the powder, use a high-precision powder particle size distribution measuring instrument, a humidity sensor, and an impurity content detection device to synchronously collect data on powder particle size distribution, humidity, and impurity content. Integrate these multi-dimensional data collected to generate a powder state data set.

[0027] Screening and classification module: Receive the powder state data set generated by the powder collection module, and use a combined algorithm of vibration screening and air flow separation to dynamically separate the particle size of the powder. In this way, classify powders of different particle sizes and finally generate a set of classified powders.

[0028] Purification and treatment module: Construct a multi-stage purification model based on high-temperature degassing and chemical cleaning. For the set of classified powders, first use high-temperature degassing to remove the gas adsorbed on the powder surface and some volatile impurities, and then use chemical cleaning to remove the oxide layer and decompose organic substances, and finally generate a sequence of purified powders.

[0029] Particle size adjustment module: Design a composite control strategy based on ball milling and spray drying. According to the preset particle size threshold, adjust the particle size distribution of the purified powder sequence. Through the coordinated treatment of ball milling and spray drying, generate a regenerated powder sample that meets the usage requirements.

[0030] Intelligent control module: Establish a multi-objective optimization model based on dynamic programming. This model takes the real-time powder recovery efficiency and energy consumption constraint as input conditions. After model calculation and analysis, generate parameter adjustment instructions to control the operating parameters of each device in the entire recovery system to achieve the efficient and energy-saving operation of the system.

[0031] The following further elaborates on the present invention in conjunction with specific embodiments: Embodiment 1: This embodiment elaborates in detail on the specific implementation manner of the combined algorithm of vibration screening and air flow separation.

[0032] In actual operation, the multi-layer screen structure of the vibration screening device first constructs multi-layer screens. The number of screen layers is determined according to the required particle size separation accuracy, generally set to 3 - 5 layers. The aperture of each layer of screen decreases according to a certain gradient. For example, the aperture of the topmost screen is 500μm, and it is 300μm, 150μm, etc. in sequence downward. Install acceleration sensors at key positions of the screen, and the acceleration sensors are used to monitor the vibration frequency and amplitude . The vibration frequency directly affects the bouncing of the powder on the screen and the screening efficiency, and the amplitude determines the amplitude of the powder bouncing and the movement trajectory on the screen. When the vibration frequency When it is too low, the powder is difficult to jump sufficiently, and the screening efficiency will decrease; the amplitude Too large or too small will affect the screening effect of the powder.

[0033] For the air classifier, the vortex intensity adjustment equation is designed as , where represents the stratified air velocity gradient (m / s), is a coefficient related to the equipment structure (dimensionless), generally taking values between 0.8 and 1.2, and can be determined through experiments according to the actual situation of the equipment; is the powder density (kg / m³), and the densities of different metal powders are different. For example, the density of titanium alloy powder is about 4500 kg / m³, and the density of aluminum alloy powder is about 2700 kg / m³; is the acceleration due to gravity (m / s²), with a value of about 9.8 m / s²; is the height (m) inside the air classifier, which is determined according to the design specifications of the equipment and is usually between 1 and 3 m. Through this equation, a stratified air velocity gradient is generated according to the powder density difference, so that powders with different densities have different movement trajectories in the air flow, thereby achieving more accurate particle size separation.

[0034] The fuzzy control algorithm is used to dynamically match the screening efficiency and the air separation accuracy. The fuzzy control algorithm takes the vibration frequency and amplitude of the vibrating screening device, as well as the stratified air velocity gradient of the air classifier as input variables. These input variables are fuzzified. For example, the vibration frequency is divided into three fuzzy sets: low, medium, and high, the amplitude is divided into three fuzzy sets: small, medium, and large, and the stratified air velocity gradient is divided into three fuzzy sets: low, medium, and high. According to the preset fuzzy rules, these fuzzy rules are summarized based on a large amount of experimental data and experience. For example, when the vibration frequency is high, the amplitude is large, and the stratified air velocity gradient is high, the vibration frequency and the stratified air velocity gradient

[0035] In the actual application scenario, assume that a batch of titanium alloy powders is recycled. Initially, through measurement, it is known that the powder density is 4500 kg / m³, and the height inside the air classifier is 2 m, and the equipment structure correlation coefficient is determined to be 1.0 through experiments. According to the vortex intensity adjustment equation , the stratified air flow velocity gradient is calculated to be approximately 138.6 m / s. During operation, the acceleration sensor monitors the vibration frequency of the vibrating screening device to be 50 Hz, and the amplitude is 8 mm. After being processed by the fuzzy control algorithm, it is found that the screening efficiency is relatively low under the current parameters, and the air classification accuracy also needs to be improved. Therefore, according to the fuzzy rules, the vibration frequency is appropriately reduced to 45 Hz, and at the same time, the stratified air flow velocity gradient is adjusted to 130 m / s. After running for a period of time, it is found that the classification effect of the powder is significantly improved, and the powder is successfully and effectively separated according to the expected particle size range, generating a classified powder collection, providing high-quality raw materials for subsequent purification treatment.

[0036] Example 2: This example details the construction process of the multi-stage purification model. Specifically, it includes: In the high-temperature degassing stage, in order to ensure that the powder can be fully degassed and avoid damage to the powder properties due to excessive temperature, a gradient heating curve is set. For example, the initial temperature is set to 100 °C, the heating rate is 5 °C / min, after heating to 300 °C, it is maintained for 1 h, and then it is heated to 500 °C at a rate of 3 °C / min and maintained for 2 h. Multiple monitoring points of infrared thermal imagers are evenly arranged on the surface of the powder, and the surface temperature distribution of the powder is monitored through the infrared thermal imagers , where , represent the coordinates (m) of the monitoring points on the powder surface. By monitoring the temperature distribution, it is possible to timely detect whether there is uneven temperature on the powder surface. If there are areas with local overheating or overcooling, the power of the heating equipment is adjusted in a timely manner or the position of the powder in the heating equipment is adjusted to ensure uniform heating of the whole powder and ensure the consistency of the degassing effect.

[0037] In the chemical cleaning stage, an acid-base alternating reaction tank is configured. First, the powder is put into the acidic cleaning solution for cleaning. The acidic cleaning solution is generally dilute hydrochloric acid, and the concentration is controlled between 5% - 10%. After soaking in the acidic cleaning solution for a period of time, the powder is taken out, rinsed thoroughly with deionized water, and then put into the alkaline cleaning solution. The alkaline cleaning solution is usually sodium hydroxide solution, and the concentration is controlled between 3% - 8%. An electrical conductivity sensor is used to detect the ion concentration of the cleaning solution in real time , and the ion concentration The unit is mol / L. The ion concentration can reflect the content of impurity ions in the cleaning solution. As the cleaning process progresses, the impurity ions in the cleaning solution will gradually increase, and the ion concentration will also increase accordingly. When the ion concentration reaches a certain threshold, it indicates that the cleaning solution is saturated and the cleaning solution needs to be replaced in time to ensure the cleaning effect.

[0038] After high-temperature degassing and chemical cleaning are completed, suspended impurities are removed through a combined process of residue precipitation and centrifugal separation. First, the cleaned powder solution is left standing for a period of time to allow larger particles of residue to settle to the bottom of the container. Then, the upper solution is transferred to a centrifuge, and the rotation speed of the centrifuge is adjusted according to the characteristics of the powder, generally between 3000 - 5000 r / min. After centrifugal separation, the remaining suspended impurities are further removed. Finally, based on the mass change of the powder before and after cleaning, the detection results of impurity content, and the performance test data of the powder, a purification powder quality assessment report is generated. The assessment report details various performance indicators of the purified powder, such as purity, particle size distribution, oxygen content, etc., providing an important basis for subsequent judgment of whether the powder meets the usage requirements.

[0039] For example, when purifying a batch of aluminum alloy powder, during the high-temperature degassing stage, heating is carried out according to the set gradient heating curve. It is monitored by an infrared thermal imager that the surface temperature distribution of the powder is relatively uniform, and the temperature deviation is within ±5°C. During the chemical cleaning stage, the powder is first immersed in a 8% dilute hydrochloric acid solution for 30 min, and the ion concentration gradually increases from the initial 0.01 mol / L to 0.1 mol / L as monitored by a conductivity sensor. Then the powder is taken out, rinsed clean, and immersed in a 5% sodium hydroxide solution for 20 min. At this time, the ion concentration increases from the initial 0.005 mol / L to 0.08 mol / L again. After residue precipitation and centrifugal separation, the purified powder is tested. It is found that the purity of the powder has increased from the original 95% to 99%, the oxygen content has decreased from 0.5% to 0.1%, and the particle size distribution has become more uniform. Based on these data, a purification powder quality assessment report is generated, indicating that the quality of this batch of powder has been significantly improved after purification and meets the usage requirements for additive manufacturing of the rocket engine thrust chamber.

[0040] Example 3: This example details the implementation method of the composite regulation strategy: During the ball milling process, define the rotation speed (r / min) and the grinding time (h) of the ball mill as the first regulation variables. The rotation speed It directly affects the impact force of the grinding medium on the powder and the grinding effect, and the grinding time determines the degree to which the powder is ground. Establishing the correlation model between the ball milling energy consumption and the powder crushing efficiency is , where is the power of the ball mill (kW). Ball mills of different models have different powers, generally between 5 - 20 kW. The powder crushing efficiency can be calculated by measuring the change in the average particle size of the powder before and after grinding. The formula is: where is the average particle size of the powder before grinding (μm), is the average particle size of the powder after grinding (μm). In actual operation, by adjusting the rotation speed of the ball mill and the grinding time , the ball milling energy consumption and the powder crushing efficiency can be controlled. For example, when it is necessary to improve the powder crushing efficiency, the rotation speed of the ball mill can be appropriately increased or the grinding time can be extended, but this will also increase the ball milling energy consumption . Therefore, a trade - off needs to be made between the two to achieve the best grinding effect.

[0041] In the spray - drying stage, the atomization pressure (MPa) of the spray - drying tower and the hot - air temperature (°C) are defined as the second control variables. Constructing the powder sphericity and the fluidity optimization function as , the fluidity optimization function . The atomization pressure affects the atomization effect of the powder. A higher atomization pressure can make the powder form finer droplets, which is beneficial to improving the powder sphericity . The hot - air temperature affects the drying speed and surface quality of the powder. Appropriately increasing the hot - air temperature can accelerate the drying speed, but too high a temperature may cause surface oxidation or agglomeration of the powder, affecting the fluidity of the powder. By adjusting the atomization pressure and the hot - air temperature , the powder sphericity and fluidity can be optimized.

[0042] Solve the process parameter combinations of ball milling and spray drying through a collaborative optimization algorithm. The collaborative optimization algorithm can adopt genetic algorithms, particle swarm optimization algorithms, etc. Taking the genetic algorithm as an example, first randomly generate a set of initial ball milling speeds , grinding times , atomization pressures and hot air temperatures as a population. Then, according to objective functions such as ball milling energy consumption , powder crushing efficiency , powder sphericity and fluidity optimization functions , calculate the fitness value of each individual. Through genetic operations such as selection, crossover, and mutation, continuously evolve the population, and finally obtain the process parameter combinations of ball milling and spray drying that meet the requirements, and generate the particle size distribution curve of the recycled powder.

[0043] For example, for a batch of nickel-based alloy powders, the ball mill power is 10 kW. Initially, the average particle size of the powder is 100 μm. During the ball milling process, through multiple experiments, it is found that when the ball mill speed is 800 r / min and the grinding time is 3 h, the ball milling energy consumption , the powder crushing efficiency . At this time, the average particle size of the ground powder is 50 μm, and the powder crushing efficiency . In the spray drying stage, by adjusting the atomization pressure and the hot air temperature , when the atomization pressure is 3 MPa and the hot air temperature is 150 °C, according to the powder sphericity , the calculated powder sphericity ; according to the fluidity optimization function , the calculated fluidity optimization function . Through the collaborative optimization of the genetic algorithm, finally determine that the ball mill speed is 750 r / min, the grinding time is 2.5 h, the atomization pressure is 2.5 MPa, and the hot air temperature is 130 °C. Under this process parameter combination, the generated particle size distribution curve of the recycled powder meets the requirements of additive manufacturing for the thrust chamber of a rocket engine, and the sphericity and fluidity of the powder are also effectively optimized.

[0044] Example 4: This example details the establishment process of the multi-objective optimization model. Specifically, it includes: Define the state space as the powder recovery rate , the energy consumption index and the equipment wear rate of the combined vector, that is . The powder recovery rate represents the ratio of the mass of the recovered powder to the mass of the original powder, which is one of the important indicators to measure the performance of the recovery system, and its value ranges from 0 to 1, the higher it is, the better the recovery effect; the energy consumption index is used to evaluate the energy consumption during the recovery process. It is related to the energy consumption of each device in the recovery system and can be obtained by calculating the power and running time of the device. The smaller the value of the energy consumption index , the more energy-efficient the recovery process; the equipment wear rate reflects the wear degree of the equipment in the recovery system during operation, usually measured by the wear amount of the key components of the equipment per unit time. The lower the equipment wear rate , the longer the service life of the equipment.

[0045] Construct a constrained non-dominated sorting genetic algorithm framework. In this framework, the elitist retention strategy is adopted. The elitist retention strategy means that in each generation of evolution, a part of the individuals with the best fitness values are directly retained to the next generation, which can ensure that excellent individuals will not be lost due to genetic operations and accelerate the convergence speed of the algorithm. At the same time, crossover and mutation operations are carried out. The crossover operation refers to randomly selecting two individuals and exchanging some of their genes to generate new individuals; the mutation operation refers to randomly changing the genes of individuals with a certain probability to increase the diversity of the population and avoid the algorithm falling into local optimal solutions. By continuously performing elitist retention, crossover and mutation operations, the control parameters are optimized.

[0046] Design the objective function as the weighted difference between the recovery cost and the powder reuse value. The recovery cost includes equipment operation cost, raw material cost, labor cost, etc.; the powder reuse value is related to factors such as the quality and market price of the powder. By optimizing the objective function, the optimal operating parameter configuration plan is output. These operating parameters include the suction force of the negative pressure adsorption device, the vibration frequency and amplitude of the vibrating screening device, the air flow velocity of the air classifier, etc., so as to reduce energy consumption and equipment wear, improve the powder reuse value, and reduce the recovery cost on the premise of meeting a certain recovery efficiency.

[0047] Example 5: This example mainly elaborates on the specific implementation methods of the powder characteristic analysis module and the powder circulation monitoring module.

[0048] In the powder property analysis module, an X-ray diffractometer is used to characterize the crystal structure of the recycled powder. The X-ray diffractometer irradiates the recycled powder sample with X-rays of a specific wavelength. When the X-rays interact with the crystals in the powder, diffraction occurs. Based on the diffraction pattern, technicians can precisely analyze the crystal structure of the powder, determine whether there are lattice distortions, crystal defects, etc., as well as key parameters such as the type of crystal and lattice constants.

[0049] The morphology of the recycled powder is characterized using a scanning electron microscope, which can directly observe features such as the shape, size, and surface roughness of the powder particles. Under high-resolution scanning electron microscope images, it is possible to clearly see whether the powder particles are spherical, irregular, or there is agglomeration. For example, if the powder particles are observed to be spherical and have a smooth surface, it indicates good fluidity, which is beneficial for subsequent additive manufacturing processes; if agglomeration is found, the reasons for agglomeration need to be further analyzed and corresponding measures taken for treatment.

[0050] The test data is input into the powder property database, which pre-stores a large amount of performance data of metal powders from different batches and different quality grades. By comparing the test data of the recycled powder with the standard data in the database, a difference analysis is carried out. If the particle size of the recycled powder exceeds the preset normal range, the system will automatically trigger an abnormal particle size warning signal. For example, for a certain type of nickel-based alloy powder, the standard particle size range is 50 - 150 μm. When the average particle size of the recycled powder is detected to be less than 50 μm or greater than 150 μm, the system immediately issues a warning, prompting the operator to check and adjust the recycling process to ensure the quality of the recycled powder.

[0051] In the powder cycle monitoring module, a correlation model between the powder cycle times and performance degradation is established. Through the analysis of a large amount of experimental data, it is found that the performance degradation of the powder is approximately linearly related to the cycle times, which can be expressed as , where is the performance index of the powder (such as the fluidity, purity, etc. of the powder) after cycles, is the initial performance index of the powder, is the performance degradation coefficient (related to powder material characteristics, recycling process, etc.), is the powder cycle times. For example, for a certain aluminum alloy powder, the initial fluidity is (indicating the time required for 50 g of powder to flow out of a standard funnel), and the performance degradation coefficient is experimentally determined to be . When the cycle times times, according to the formula, the fluidity of the powder at this time can be calculated.

[0052] The particle swarm optimization algorithm is used to dynamically adjust the ratio of new and old powders. The particle swarm optimization algorithm is an optimization algorithm that simulates the foraging behavior of bird flocks. In this algorithm, each particle represents a combination scheme of the ratio of new and old powders. Through continuous iteration, the algorithm makes the particles fly towards the direction of the optimal solution. With the goal of improving the comprehensive performance of the powder and reducing costs, according to the current performance state of the powder and the operation of the recycling system, the speed and position of the particles are dynamically adjusted to obtain the optimal ratio of new and old powders. For example, when it is found that the performance of the recycled powder has declined, the algorithm will appropriately increase the proportion of new powder to ensure that the performance of the mixed powder meets the requirements of additive manufacturing. Finally, based on the performance decay model and the optimized ratio of new and old powders, the maximum cycle number threshold is generated. When the powder cycle number reaches this threshold, a prompt is given to indicate that special treatment or replacement of the powder is required to ensure the quality and stability of the additive manufacturing of the rocket engine thrust chamber.

[0053] In practical applications, assume that after multiple experiments, the performance decay coefficient of a certain copper alloy powder is determined (for the purity index, unit: % / time), and the initial purity . During the recycling process of the powder, the particle swarm optimization algorithm determines that the optimal ratio of new powder to old powder at the current stage is 3:7. As the number of cycles increases, when the calculated number of cycles times, according to the formula , at this time the powder purity . If this copper alloy powder is used for additive manufacturing of the rocket engine thrust chamber, a purity lower than 90% will affect the product quality. Then, it can be determined that the maximum cycle number threshold is 12 times. When the powder cycle number approaches or reaches 12 times, the system gives a prompt, and the operator can decide whether to purify the powder or supplement new powder according to the actual situation to ensure the smooth progress of the additive manufacturing process and the reliability of the product quality.

[0054] Example 6: This example details the specific implementation methods of the exception handling module and the parameter adjustment instruction generation process. Specifically, it includes: In the anomaly handling module, a powder defect pattern library based on an autoencoder is constructed. An autoencoder is a deep learning model composed of an encoder and a decoder. First, a large amount of powder sample data containing various defect situations is collected, such as powder samples with impurities, abnormal particle sizes, irregular shapes, etc. These sample data are input into the encoder part of the autoencoder, and the encoder will compress the high-dimensional original data into a low-dimensional feature vector, which contains the key feature information of the powder samples. Then, the decoder restores the low-dimensional feature vector to reconstructed data. During the training process, the parameters of the autoencoder are continuously adjusted to minimize the difference between the reconstructed data and the original data. After sufficient training, the low-dimensional feature vectors corresponding to powder samples of different defect types are stored in the powder defect pattern library.

[0055] During the recycling process, wavelet packet transform is used to extract transient anomaly signals. Wavelet packet transform is a time-frequency analysis method that can perform fine analysis on signals in different frequency bands and time periods. The signals collected by various sensors in the recycling system (such as the pressure signal of the negative pressure adsorption device, the vibration signal of the screening device, etc.) are processed by wavelet packet transform. For example, for the pressure signal of the negative pressure adsorption device, under normal circumstances, its waveform is relatively stable, and when an anomaly occurs (such as a blockage in the adsorption pipeline), the pressure signal will show a sudden change. Through wavelet packet transform, these transient changes can be accurately captured and converted into time-frequency domain feature information. The extracted feature information is compared and matched with the data in the powder defect pattern library to generate the identification result of the powder pollution type. For example, if the feature vector corresponding to impurity pollution is matched, it is determined that there is an impurity pollution problem in the powder, and the possible types and degrees of pollution of the impurities are further determined, providing a basis for subsequent targeted treatment measures.

[0056] In terms of generating parameter adjustment instructions, a mixed-integer programming model with time-varying constraints is constructed, with the equipment life and recycling efficiency as the boundary conditions. The equipment life (unit: hours) refers to the duration that the key equipment in the recycling system can continuously operate under normal operating conditions, and the recycling efficiency (unit: %) represents the percentage of the powder mass recycled per unit time to the original powder mass. The constraint conditions of the model also include the operating power limit and processing capacity limit of the equipment. For example, the maximum operating power of the vibrating screening device is (kW), and the maximum powder processing capacity per hour is (kg). The decision variables of this model include the sieve aperture (μm) and the air flow velocity (m / s), etc.

[0057] The branch and bound method is used to solve the multi-stage decision-making problem and generate a feasible solution set for the screen aperture and air flow velocity. The branch and bound method is an effective algorithm for solving integer programming problems. It continuously decomposes the problem into sub-problems, evaluates and screens the solutions of each sub-problem, gradually narrows the range of feasible solutions, and finally obtains the optimal solution or approximate optimal solution. During the solution process, according to the characteristics and constraints of the problem, reasonable branching and bounding are performed on variables such as the screen aperture and air flow velocity . For example, assume that the value range of the screen aperture is 50 - 500 μm, and the value range of the air flow velocity is 1 - 10 m / s. These value ranges are divided into multiple sub-intervals at a certain interval. Then, the variable combinations within each sub-interval are calculated and evaluated to determine whether they meet the constraints, and the objective function values (such as recovery cost, powder quality, etc.) are calculated. The variable combinations that meet the constraints and have better objective function values are retained to form a feasible solution set.

[0058] The comprehensive priority of the solution set is evaluated through the grey relational analysis method. The grey relational analysis method is a multi-factor statistical analysis method. It evaluates the similarity and influence degree between factors by calculating the correlation degree between various factors. In this embodiment, multiple factors such as the recovery cost, powder quality, and equipment wear corresponding to each solution in the feasible solution set are used as analysis objects, and the correlation degree between them and the ideal solution (i.e., the solution with all indicators reaching the optimal) is calculated. The higher the correlation degree, the closer the solution is to the ideal solution and the higher the comprehensive priority. According to the results of the grey relational analysis, the feasible solution set is sorted, and a real-time control instruction sequence is output. For example, if it is found through analysis that a certain solution has a lower recovery cost, better powder quality, and less equipment wear, and its correlation degree is the highest, then the parameters such as the screen aperture and air flow velocity corresponding to this solution are sent as real-time control instructions to each device of the recovery system to achieve precise control of the recovery system, ensure its operation in the best state, and improve the recovery quality and efficiency of metal powder.

[0059] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0060] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A metal powder recovery system for additive manufacturing of rocket engine thrust chamber, characterized in that: The system comprises: The powder collection module is used to capture the residual metal powder in the additive manufacturing equipment through a multi-stage negative pressure adsorption device, and simultaneously collect the powder particle size distribution, humidity and impurity content data to generate a powder state data set; A screening and grading module, for dynamically separating the powder state data set by using a vibration screening and airflow sorting combined algorithm to generate a graded powder set; A purification processing module is used to construct a multi-stage purification model based on high-temperature degassing and chemical cleaning, remove the oxide layer and decompose organic matter from the classified powder set, and generate a purified powder sequence; A particle size adjustment module, used to design a composite control strategy based on ball milling and spray drying, adjust the particle size distribution of the purified powder sequence according to a preset particle size threshold, and generate a regenerated powder sample; The intelligent control module is used to establish a multi-objective optimization model based on dynamic programming and generate parameter adjustment instructions according to the real-time powder recovery efficiency and energy consumption constraints.

2. The metal powder recovery system according to claim 1, characterized in that: The vibration screening and airflow sorting combined algorithm includes: Construct a multi-layer screen structure of the vibrating screening device and monitor the vibration frequency and amplitude of the screen through an acceleration sensor; Design the vortex intensity adjustment equation of the airflow separator and generate the stratified airflow velocity gradient according to the powder density difference; The fuzzy control algorithm is used to dynamically match the screening efficiency and airflow sorting accuracy, and the multi-level particle size interval division results are output.

3. The metal powder recovery system according to claim 1, characterized in that: The construction of the multi-stage purification model includes: A gradient heating curve was set during the high-temperature degassing stage, and the temperature distribution on the powder surface was monitored using an infrared thermal imager; During the chemical cleaning stage, an acid-base alternating reaction tank is configured, and a conductivity sensor is used to detect the ion concentration of the cleaning liquid in real time; The suspended impurities are removed by a combined process of residue sedimentation and centrifugal separation to generate a purified powder quality assessment report.

4. The metal powder recovery system according to claim 3, characterized in that: The design of the composite control strategy includes: The rotation speed and grinding time of the ball mill were defined as the first control variables, and a correlation model between ball mill energy consumption and powder crushing efficiency was established; The atomization pressure and hot air temperature of the spray drying tower are defined as the second control variable, and the optimization function of powder sphericity and fluidity is constructed; The process parameter combination of ball milling and spray drying is solved by collaborative optimization algorithm to generate the particle size distribution curve of the regenerated powder.

5. The metal powder recovery system according to claim 1, characterized in that: The establishment of the multi-objective optimization model includes: The state space is defined as the combination vector of powder recovery rate, energy consumption index and equipment wear rate; Construct a constrained non-dominated sorting genetic algorithm framework, and use elite retention strategy and crossover mutation operation to optimize control parameters; The objective function is designed as the weighted difference between the recovery cost and the powder reuse value, and the optimal operating parameter configuration scheme is output.

6. The metal powder recovery system according to claim 1, characterized in that: The system further comprises: Powder property analysis module, used to characterize the crystal structure and morphology of the recycled powder by X-ray diffractometer and scanning electron microscope; The test data is input into the powder performance database, and the abnormal particle size warning signal is triggered through difference comparison.

7. The metal powder recovery system according to claim 1, characterized in that: The generation of the powder state data set includes: Synchronously collect the powder flow rate of the negative pressure adsorption device, the humidity sensor readings and the impurity distribution map of the image recognition system; The principal component analysis algorithm is used to reduce the dimension of multidimensional data and construct a low-dimensional feature matrix.

8. The metal powder recovery system according to claim 1, characterized in that: The system further comprises: Powder cycle monitoring module, used to establish a correlation model between powder cycle times and performance decay; The particle swarm optimization algorithm is used to dynamically adjust the ratio of new and old powders and generate the maximum cycle number threshold.

9. The metal powder recovery system according to claim 1, characterized in that: The generation of the parameter adjustment instruction includes: A mixed integer programming model with time-varying constraints is constructed, with equipment life and recycling efficiency as boundary conditions; the branch-and-bound method is used to solve the multi-stage decision problem and generate feasible solution sets for screen aperture and airflow velocity; The comprehensive priority of the solution set is evaluated through grey correlation analysis, and a real-time control instruction sequence is output.

10. The metal powder recovery system according to claim 1, characterized in that: The system further comprises: An exception handling module for building a powder defect pattern library based on an autoencoder; Wavelet packet transform is used to extract transient abnormal signals in the recycling process and generate powder pollution type identification results.

Images