Multi-energy-field regulation and control process method and system for prolonging service life of aerospace complex component

Abstract

The invention discloses a multi-energy field regulation and control process method and system for prolonging service life of an aerospace complex component, and belongs to the field of aerospace complex component manufacturing. The method comprises the steps that the aerospace complex component is placed in a magnetic suspension polishing device integrating a magnetic suspension magnetic field, an ultrasonic field, a pressure field and a temperature field; surface treatment and destressing optimization are conducted on the component through the synergistic effect of a composite energy field, the destressing effect and polishing uniformity are remarkably improved, a surface layer structure with high compactness and smoothness is formed, and closing of internal microdefects is effectively promoted; the method comprises the following steps: acquiring technological parameters of a magnetic suspension magnetic field, an ultrasonic field, a pressure field and a temperature field, performing normalization preprocessing, constructing a hidden space dimension reduction model based on a variational auto-encoder, extracting a nonlinear coupling relationship among complex technological parameters, and dynamically adjusting key technological parameters in combination with a Hemma optimization algorithm on the premise of remarkably improving the optimization efficiency. And the optimal process scheme is quickly and accurately obtained, and the service performance of the aerospace complex component is comprehensively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of aerospace complex component manufacturing, and particularly relates to a multi-energy field regulation process method and system for improving the service life of aerospace complex components. Background Technique

[0002] Aerospace complex components are important structural components used in aircraft and spacecraft in the aerospace field, and have the characteristics of complex structure, thin-walled and weak rigidity, high performance requirements, and great manufacturing process difficulty. Their manufacturing often uses light alloys, composite materials or superalloys, and is completed through advanced processes such as precision machining, additive manufacturing or superplastic forming, and is widely used in key parts such as aircraft wings, spacecraft cabins and engine blades. The design and manufacturing of aerospace complex components directly affect the performance, reliability and cost of aerospace products, and are an important symbol to measure the aerospace technology level. These components usually have complex internal cavities and flow channels in design, such as cooling channels in engine blades, heat system channels in spacecraft, and fuel pipelines inside aircraft wings. These designs not only help to reduce weight, but also improve efficiency or meet special functional requirements. At the same time, aerospace complex components often also involve complex geometric shapes and tiny support structures.

[0003] Traditional magnetic polishing technology can effectively improve the irregular microscopic morphology (roughness) of the surface of ordinary components. However, for the manufacturing of aerospace complex components, the requirements are extremely strict, and the internal defects and stress distribution are highly random and highly unpredictable. Traditional surface polishing technology is difficult to effectively remove the internal and surface microscopic defects and residual stress of components at the same time, and cannot meet the comprehensive quality requirements of high precision and high performance. Compared with traditional magnetic polishing technology, magnetic levitation polishing technology uses the principle of magnetic levitation to make the component float in the container, so that the abrasive is more evenly and stably distributed around the workpiece; it has higher precision and stronger adaptability, can accurately control the polishing position and force, can flexibly adjust parameters according to the shape, size and material of the workpiece, and has more advantages in dealing with workpieces with special shapes or complex curved surfaces; and the damage is smaller, the workpiece floats without direct mechanical contact, reducing the risk of wear and scratch, and producing a better flexible treatment effect on the surface of the component, which is suitable for workpieces that are easy to deform, fragile or surface-sensitive.

[0004] However, magnetic levitation polishing technology has little effect on the internal microscopic defects and residual stress of components. The existence of internal microscopic defects and residual stress causes the components to easily appear failure phenomena such as deformation and cracking during subsequent use, greatly shortening the service life of the components. To sum up, in the face of the strict requirements of aerospace complex components for high reliability and long life, it is necessary to break through the limitations of traditional processes. Summary of the Invention

[0005] The present invention aims to solve the deficiencies of the prior art and provides the following solutions:

[0006] A multi-energy field control process method for improving the service life of complex aerospace components comprises the following steps:

[0007] S1. Place the complex aerospace component to be processed in a magnetic suspension polishing machine for magnetic field-ultrasound-pressure-high temperature multi-physical field coordinated grinding and polishing, collect the process parameters and component performance of the processed complex aerospace component and perform normalization processing to obtain normalized process parameters and normalized component performance;

[0008] S2. Using a variational autoencoder to extract the latent space structure between the normalized process parameters, and mapping the high-dimensional parameters to a low-dimensional latent space to obtain low-dimensional latent variables, and using the low-dimensional latent variables to initialize the individual positions of the population in the Hippo optimization algorithm;

[0009] S3. Using the initialized Hippo optimization algorithm to search for the optimal parameter combination in the compression space, update the individual position, take the normalized process parameters as input, and the component performance index as the objective function to construct a comprehensive objective function;

[0010] S4. Taking the comprehensive objective function as the target, calculating the function value through the process parameters, optimizing the iterative process parameter combination until the comprehensive objective function converges, extracting the function value and mapping it to component performance, and introducing the mapped component performance as a feedback control quantity into the optimization iteration process;

[0011] S5. In each iteration of the optimization process in step S4, the process parameters are adjusted using the variational autoencoder and the Hippo optimization algorithm and the corresponding objective function value is calculated. If the change in the objective function value of this round is less than the threshold, the algorithm stops and the process parameters are optimized. Otherwise, proceed to step S2;

[0012] S6. Repeat steps S2 to S5 until the global optimal solution of the process parameter combination is obtained.

[0013] Preferably, the method of using the magnetic suspension polishing machine to perform multi-physical field coordinated grinding and polishing processing includes:

[0014] Ensure that the magnetic suspension polishing machine is in a closed state and the interior is clean and free of debris, place the complex aerospace component to be processed in the closed box of the magnetic suspension polishing machine, and after placement, close the box to ensure airtightness, then turn on the power, and start the magnetic suspension polishing machine through the console;

[0015] The process parameters of the magnetic suspension magnetic field, ultrasonic field, pressure field and temperature field are set in the control console, and then the AC magnetic coil generates the alternating magnetic suspension magnetic field, the bottom sound wave generating device generates the ultrasonic field, the pressurizing device generates the pressure field, and the heating device generates the temperature field;

[0016] Under the action of the magnetic levitation magnetic field, the grinding tool generates relative movement with the surface of the aerospace complex component, and through the synergistic action of the magnetic levitation magnetic field, the ultrasonic field, the pressure field and the temperature field, the surface of the aerospace complex component is micro-ground and polished.

[0017] Preferably, the normalized process parameters include: the magnetic field strength H' of the magnetic levitation polishing machine, the magnetic field frequency f m ', the ultrasonic power W', the ultrasonic frequency f u ', the container pressure P', the internal cavity heating temperature T', the magnetic field action time t m ', the ultrasonic field action time t u ', the pressure field action time t p ', and the temperature field action time t t ';

[0018] The normalized component properties include: surface roughness R', residual stress σ', internal defect closure degree D', and fatigue life L'.

[0019] Preferably, the method for initializing the population individual positions in the hippopotamus optimization algorithm includes:

[0020] ,

[0021] where X t represents the position of the current individual, X rand represents an individual randomly selected from the population, X best represents the global optimal individual in the current population, R(τ)= σ(τ)·(X rand -Xτ) represents the random perturbation term, σ(τ) represents the perturbation factor that changes dynamically with time, r 1 , r 2 and r 3 represent random numbers between 0 and 1, LB represents the lower bound of the search space, UB represents the upper bound of the search space, q represents the random threshold of the control strategy, ω 1 (τ), ω 2 (τ), ω 3 (τ) and ω 4 (τ) represent the dynamic weight adjustment factors, φ represents the nonlinear adjustment coefficient, ω i,min represents the minimum value of the weight, ω i,max represents the maximum value of the weight, τ represents the current iteration number, represents the maximum iteration number, α(τ) represents the first correction factor, α 0 represents the first initial correction factor value, γ represents the parameter controlling the attenuation speed of the first correction factor.

[0022] Preferably, the method for updating the individual location includes:

[0023] ,

[0024] Among them, X prey represents the current global optimal solution, r represents a random number between 0 and 1, E represents the escape energy, and E 0 represents the initial escape energy, δ represents the parameter that controls the nonlinear attenuation of the escape energy, Ψ(τ) represents the dynamically adjusted convergence factor, η(τ) represents the penalty term weight, and C penalty represents the penalty factor of constrained optimization, β(τ) represents the second correction factor, β 0 represents the second initial correction factor value, and μ controls the parameter of the decay speed of the second correction factor.

[0025] Preferably, the comprehensive objective function includes:

[0026] ,

[0027] Among them, R' min Indicates the minimum value of surface roughness, R' max Indicates the maximum value of surface roughness, σ' min Represents the minimum value of residual stress, σ' max Represents the maximum value of residual stress, D' min Indicates the minimum value of the internal defect closure degree, D' max Indicates the maximum value of the internal defect closure degree, L' min Indicates the minimum value of fatigue life, L' max represents the maximum value of fatigue life, w 1 、w 2 、w 3 and w 4 Represents the weight of each objective function.

[0028] The present invention also provides a multi-energy field control process system for improving the service life of complex aerospace components. The system applies any of the above methods, including: a component processing module, an algorithm initialization module, a function construction module, an iterative optimization module and a threshold comparison module;

[0029] The component processing module is used to place the complex aerospace component to be processed in a magnetic suspension polishing machine for magnetic field-ultrasound-pressure-high temperature multi-physical field coordinated grinding and polishing, and collect the process parameters and component performance of the processed complex aerospace component and perform normalization processing to obtain normalized process parameters and normalized component performance;

[0030] The algorithm initialization module extracts the latent space structure between the normalized process parameters by using a variational autoencoder, maps the high-dimensional parameters into a low-dimensional latent space to obtain low-dimensional latent variables, and initializes the population individual positions in the hippopotamus optimization algorithm by using the low-dimensional latent variables;

[0031] The function construction module uses the initialized hippopotamus optimization algorithm to search for the optimal parameter combination in the compressed space, updates the individual positions, takes the normalized process parameters as the input, and the component performance index as the objective function to construct a comprehensive objective function;

[0032] The iterative optimization module takes the comprehensive objective function as the goal, calculates the function value through the process parameters, optimizes and iterates the process parameter combination until the comprehensive objective function converges, extracts the function value and maps it to the component performance, and introduces the mapped component performance as a feedback control quantity into the optimization iteration process;

[0033] In each iteration of the optimization process of the threshold comparison module, the variational autoencoder and the hippopotamus optimization algorithm are used to adjust the process parameters and calculate the corresponding objective function value. If the change amount of the objective function value in this round is less than the threshold, the algorithm stops, and the process parameters at this time are the optimization result. Otherwise, the algorithm initialization module is restarted;

[0034] Repeat the working processes of the algorithm initialization module, the function construction module, the iterative optimization module and the threshold comparison module until the global optimal solution of the process parameter combination is obtained.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] The present invention solves the problem that conventional polishing techniques are difficult to rapidly, effectively and comprehensively improve the service performance of aerospace complex components. The aerospace complex components are placed in a magnetic levitation polishing machine with the coupling of a magnetic levitation magnetic field, an ultrasonic field, a pressure field and a temperature field. Through the multi-energy field synergistic effect, the surface roughness and finish of the components are significantly improved, the residual stress level is effectively reduced and homogenized, various defects inside the material are greatly eliminated, and the overall durability and service life of the components are comprehensively improved; combining the variational autoencoder and the hippopotamus optimization algorithm to scientifically, rapidly and accurately seek the best process condition combination, and finally giving a multi-energy field optimal collaborative control scheme for comprehensively and deeply improving the surface quality, reliability and service life of aerospace complex components. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the present invention, the drawings required to be used in the embodiments are briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0038] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention;

[0039] Figure 2 It is a schematic structural diagram of the magnetic levitation polishing machine according to an embodiment of the present invention;

[0040] Explanation of reference numerals:

[0041] 1. Pressurizing device; 2. Magnetic levitation polishing machine box body; 3. Multi-field sensor; 4. Console; 5. Alternating current magnetic coil; 6. Heating device; 7. Aerospace complex component; 8. Magnetic levitation generating device; 9. Acoustic wave generating device; 10. Alternating current power supply; 11. Bracket. Detailed implementation manners

[0042] 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. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0044] First, introduce the technology involved in this application:

[0045] Aerospace complex components are core components widely used in aerospace aircraft, usually having complex geometric shapes, thin-walled and weak rigidity, lightweight, high strength, high precision, anti-fatigue, resistance to extreme environments, high reliability, and multi-functional integration characteristics. Their manufacturing processes include casting, forging, additive manufacturing, spinning, stretch bending, roll bending, explosive forming, numerical control machine tool processing, special processing, welding, riveting, screwing, superplastic forming-diffusion bonding, etc.

[0046] Magnetic levitation polishing technology: Magnetic levitation polishing technology enables the workpiece to levitate in the polishing space through the magnetic force generated by the magnetic levitation magnetic field. High-hardness abrasives with uniform particle size are used to move around the workpiece to achieve efficient and precise surface machining. Commonly used abrasives include alumina, silicon carbide, diamond, etc. By adjusting the magnetic field strength and frequency, the movement trajectory and speed of the abrasives can be precisely controlled to ensure that the abrasives evenly cover the surface of the workpiece, significantly improving the surface roughness accuracy. Since the workpiece levitates without mechanical contact, it can effectively avoid scratches and deformations caused by traditional polishing, especially suitable for aerospace complex components with thin walls, easy deformation, and high surface quality requirements. For components with complex curved surfaces or irregular geometric shapes, such as aeroengine blades and spacecraft cabins, this technology can dynamically adjust the magnetic levitation magnetic field parameters according to geometric features, enabling the abrasives to act precisely on the target parts, significantly improving the surface quality and comprehensive performance.

[0047] Magnetic aging treatment technology: Magnetic aging treatment technology is a technical means of using a magnetic field to process materials to change properties such as residual stress. It relies on the interaction between the magnetic field and the material. When the material is in a magnetic field environment, the magnetic field affects the movement of charged particles, dislocation behavior, magnetic domain structure, and atomic diffusion within the material, thereby influencing the residual stress. The alternating magnetic field causes changes in the internal microstructure of complex components, resulting in relative microscopic movement between molecules. This microscopic movement helps to break the uneven structures such as lattice defects and dislocations that may exist within the components, promoting the homogenization and stability improvement of the microstructure. On the other hand, the periodic change of the alternating magnetic field generates a small magnetostrictive effect within the components, that is, the material undergoes a small deformation under the action of the magnetic levitation magnetic field. This deformation helps to improve the internal stress distribution of the material at the microscopic level, reduce stress concentration areas, thereby enhancing the overall mechanical properties and stability of the components, and further improving the polishing effect.

[0048] Ultrasonic aging technology: Ultrasonic aging technology is a processing method that uses ultrasonic energy to process materials. Its principle is that when ultrasonic waves propagate in materials, they cause internal vibrations and high-frequency alternating stresses, which can activate dislocations and promote their movement and proliferation, relax and redistribute residual stress, refine the microstructure, increase the grain boundaries by refining grains, hinder dislocation movement, improve the strength and toughness of the material, and can also repair or close microscopic defects, improving the density and integrity of the material.

[0049] Pressure Compaction Enhancement Technology: Pressure compaction enhancement technology is a technique that uses pressurization to close internal micro-defects in components. By installing a pressurization device in a closed container of a magnetic levitation polishing machine, a controllable pressure is applied to the component, causing the internal micro-defects in the material to gradually close under the action of pressure, thereby improving the material density. As the density increases, the atomic arrangement inside the material becomes more compact, the atomic binding force increases, and the continuity and overall mechanical properties (strength, hardness, toughness) of the material are significantly improved. At the same time, the pressure equalizes the internal stress distribution, effectively relieves local stress concentration, reduces the risk of deformation and cracking, and comprehensively improves the stability, reliability, and service life of the component.

[0050] The ultrasonic field and pressure field act on the component. The ultrasonic equipment generates ultrasonic waves that propagate in the component, causing high-frequency vibrations of the component medium, resulting in periodic tensile and compressive deformations of the internal microstructure of the material, breaking the stress concentration areas in the microstructure, redistributing and releasing the residual stress, and ultimately making the surface of the component smoother after polishing and obtaining better polishing quality. At the same time, the pressure field changes the distance and arrangement pattern between atoms. Under high-pressure conditions, internal micro-pores, cracks, and other defects in aerospace complex components will close or deform under the action of pressure, reducing the discontinuity in the microstructure, improving material density, and promoting the closure of internal defects.

[0051] Thermal aging technology: During the processing, the complex components are first placed in a hot temperature field and heated to reach the corresponding temperature level. After that, the temperature is maintained for a period of time, that is, the insulation process is carried out. Its basic working principle is that the yield strength of the material shows a certain correlation with the temperature. Specifically, as the temperature continues to rise, the yield strength of the material will decrease accordingly. When the temperature rises to a certain level, the residual stress originally existing inside the workpiece will exceed the yield limit of the material itself. At this time, local plastic deformation will occur inside the workpiece. In this process, the residual strain is released, and the residual stress is reduced. Moreover, there is a significant synergistic gain effect between the thermal field and other energy fields. The thermal field can enhance the polishing effect of the magnetic levitation magnetic field. By intensifying the thermal motion of atoms, the polishing medium driven by the magnetic levitation magnetic field can act more efficiently on the surface of the component, accurately remove microscopic protrusions, fill depressions, and reduce surface roughness. The thermal field and the ultrasonic field work together to accelerate atomic migration and rearrangement, promote the relaxation and release of residual stress and even distribution, and reduce the risk of component deformation and cracking. The thermal field can also help the pressure field to close defects. On the one hand, it reduces the yield strength of the material, making it easier for the material around the defect to flow plastically under pressure. On the other hand, it enhances the diffusion capacity of atoms, uses dislocation movement to even out stress, avoids stress concentration, and further promotes defect closure, effectively improving the density of the material and the overall performance of the component. The temperature field has a thermal aging effect on complex components. For the internal microstructure of complex aerospace components, high temperature helps the diffusion and migration of atoms, allowing microscopic defects to recombine and repair, thereby promoting the closure and optimization of the microstructure; at the same time, the yield strength of the material is reduced under high temperature conditions, and the internal residual stress is released and redistributed, which effectively reduces stress concentration, improves the dimensional stability and fatigue resistance of the component, and further enhances the improvement effect of the component surface quality during polishing.

[0052] Embodiment 1

[0053] In this embodiment, if Figure 1 As shown, a multi-energy field control process method for improving the service life of complex aerospace components includes the following steps:

[0054] S1. Place the complex aerospace component 7 to be processed in a magnetic levitation polishing machine for magnetic field-ultrasound-pressure-high temperature multi-physical field coordinated grinding and polishing treatment, and collect the process parameters and component performance of the processed complex aerospace component 7 and normalize them to obtain normalized process parameters and normalized component performance.

[0055] In this embodiment, the complex aerospace component 7 to be processed is placed in a magnetic suspension polishing machine assisted by multi-energy field coupling. Figure 2As shown in the figure, a multi-field sensor 3 is arranged inside the magnetic levitation polishing machine housing 2. The magnetic levitation polishing machine housing 2 is installed on the bracket 11. An alternating current is supplied to the device through the alternating current power supply 10. The alternating current magnetic coil 5 generates an alternating magnetic field. A magnetic levitation generating device 8 is arranged inside. An acoustic wave generating device 9 is installed at the bottom of the device. A pressurizing device 1 and a heating device 6 are connected to the device, so as to generate a pressure field and a temperature field inside the sealed container. The entire machine is controlled by the console 4 to operate, and multi-field assisted polishing treatment is carried out on the aerospace complex component 7. It should be noted that the multi-field sensor 3 is internally provided with a dedicated physical field detection module, which involves the detection of multiple physical fields such as magnetic levitation magnetic field, ultrasonic field, pressure field, and temperature field.

[0056] The usage steps of the magnetic levitation polishing machine are as follows: First, ensure that the magnetic levitation polishing machine is in the off state and the inside of the housing is clean without debris. Place the aerospace complex component 7 to be processed in the sealed housing of the magnetic levitation polishing machine. After placing it, close the housing tightly to ensure airtightness. Then, connect the power supply and start the device through the console 4. Set the process parameters related to the magnetic levitation magnetic field, ultrasonic field, pressure field, and temperature field on the console 4, and the device starts to run. At this time, the alternating current magnetic coil 5 generates an alternating magnetic field, the acoustic wave generating device 9 at the bottom generates an ultrasonic field, the pressurizing device 1 and the heating device 6 generate a pressure field and a temperature field respectively. The abrasive tool makes a relative movement with the surface of the component under the action of the magnetic levitation magnetic field, and the abrasive grains perform micro-cutting and grinding on the surface of the component. After polishing is completed, first stop the device through the console 4, wait for a period of time to ensure the stability of the internal environment, and then open the housing and take out the polished aerospace complex component 7. Through the synergistic action of the magnetic levitation magnetic field, ultrasonic field, pressure field, and temperature field, the surface quality of the material is effectively improved, the residual stress distribution is improved, the internal micro-defects are eliminated, and the overall comprehensive performance and service life of the aerospace complex component 7 are comprehensively improved in all aspects. Based on precise detection and data analysis technology, the surface roughness R, residual stress σ, internal defect closure degree D, and fatigue life L of the aerospace complex component 7 after polishing are characterized.

[0057] In this embodiment, the magnetic field strength H and magnetic field frequency f are extracted from the collected original data m , ultrasonic power W, ultrasonic frequency f u , container pressure P, internal cavity heating temperature T, record the magnetic field action time t m , ultrasonic field action time t u , pressure field action time t p , temperature field action time t t and other process parameters. At the same time, collect the parameters of surface roughness R, residual stress σ, internal defect closure degree D, and fatigue life L. The minimum-maximum normalization method is used to map the data to the [0,1] interval. The formula is as follows:

[0058] ,

[0059] Among them, x represents the original data, and x = [H, f m , W, f u , P, T, t m , t u , t p , t t , R, σ, D, L], x min represents the minimum value of the original data, x max represents the maximum value of the original data, x norm represents the data after normalization. The process parameters after normalization include: the magnetic field strength H' of the magnetic levitation polishing machine, the magnetic field frequency f m ', the ultrasonic power W', the ultrasonic frequency f u ', the container pressure P', the inner cavity heating temperature T', the magnetic field action time t m ', the ultrasonic field action time t u ', the pressure field action time t p ' and the temperature field action time t t '; the component performances after normalization include: the surface roughness R', the residual stress σ', the internal defect closure degree D' and the fatigue life L'.

[0060] S2. Use the variational autoencoder to extract the latent space structure between the normalized process parameters, map the high-dimensional parameters to the low-dimensional latent space to obtain low-dimensional latent variables, and use the low-dimensional latent variables to initialize the population individual positions in the hippopotamus optimization algorithm.

[0061] In this embodiment, the variational autoencoder (VAE) is used to extract the latent space structure between the process parameters, and the high-dimensional parameter space is mapped to the low-dimensional latent space to reduce the optimization calculation complexity. The specific structure is as follows: The VAE consists of an encoder, a decoder, and a latent variable generation network; the encoder maps the input process parameters to the low-dimensional latent space and obtains the latent variables by learning the latent distribution; the decoder then generates the output data from the latent variables and reconstructs the input features; the model input layer includes the magnetic field strength H', the magnetic field frequency f m ' of the magnetic levitation polishing machine after normalization processing, the ultrasonic power W', the ultrasonic frequency f u ', the container pressure P', the inner cavity heating temperature T', the magnetic field action time t m ', the ultrasonic field action time t u ', the pressure field action time t p ' and the temperature field action time t t'Process parameters such as etc. The encoder network extracts the latent space features of the input parameters. The mean and variance of the latent space are used to generate latent variables. The output data generated by the decoder based on the latent variables is used to evaluate the reconstruction effect of the original process parameters. VAE optimizes the network parameters by maximizing the similarity between the input and the reconstructed output, thereby effectively capturing the non-linear relationship of the process parameters and providing a more accurate low-dimensional feature representation for the subsequent process parameter optimization based on the Hippopotamus Optimization Algorithm (HHO); taking the above low-dimensional latent variables extracted by the Variational Autoencoder (VAE) as the input of the Hippopotamus Optimization Algorithm (HHO), initializing the population individual positions, and capturing the potential non-linear associations between the process parameters through the latent variables to reduce the complexity of the optimization calculation; in the exploration stage, the methods for initializing the population individual positions in the Hippopotamus Optimization Algorithm include:

[0062] ,

[0063] where X t represents the position of the current individual, X rand represents an individual randomly selected from the population, X best represents the global optimal individual in the current population, R(τ)= σ(τ)·(X rand -Xτ) represents the random perturbation term, σ(τ) represents the perturbation factor that changes dynamically with time, r 1 , r 2 and r 3 represent random numbers between 0 and 1, LB represents the lower bound of the search space, UB represents the upper bound of the search space, q represents the random threshold of the control strategy, ω 1 (τ), ω 2 (τ), ω 3 (τ) and ω 4 (τ) represent the dynamic weight adjustment factors, φ>1 represents the non-linear adjustment coefficient, ω i,min represents the minimum value of the weight, ω i,max represents the maximum value of the weight, τ represents the current iteration number, represents the maximum iteration number, α(τ) represents the first correction factor, α 0 represents the first initial correction factor value, γ represents the parameter controlling the decay rate of the first correction factor.

[0064] S3. Using the initialized Hippopotamus Optimization Algorithm to search for the optimal parameter combination in the compressed space, updating the individual positions, taking the normalized process parameters as the input and the component performance index as the objective function to construct the comprehensive objective function.

[0065] In the exploitation stage, with the dynamic escape energy E as the core, combining the prey position and the constraint optimization mechanism, the methods for updating the individual positions include:

[0066] ,

[0067] Among them, X prey represents the current global optimal solution, r represents a random number between 0 and 1, E represents the escape energy, and E 0 represents the initial escape energy, δ > 1 represents the parameter controlling the non-linear decay of the escape energy, Ψ(τ) represents the dynamically adjusted convergence factor, η(τ) represents the penalty term weight, and C penalty represents the penalty factor for constrained optimization, defined as , and g(X) represents the independent process parameter constraint, , λ represents the penalty coefficient, β(τ) represents the second correction factor, and β 0 represents the second initial correction factor value, and μ is the parameter controlling the decay rate of the second correction factor. The algorithm iteratively optimizes based on the performance feedback through an automatic adjustment mechanism, using the normalized process parameters as the input and the component performance index as the objective function to construct the comprehensive objective function:

[0068] ,

[0069] Among them, R' min represents the minimum value of the surface roughness, and R' max represents the maximum value of the surface roughness, σ' min represents the minimum value of the residual stress, and σ' max represents the maximum value of the residual stress, D' min represents the minimum value of the internal defect closure degree, and D' max represents the maximum value of the internal defect closure degree, L' min represents the minimum value of the fatigue life, and L' max represents the maximum value of the fatigue life, w 1 , w 2 , w 3 and w 4 represent the weights of each objective function.

[0070] S4. Taking the comprehensive objective function as the goal, calculate the function value through the process parameters, optimize and iterate the process parameter combination until the comprehensive objective function converges, extract the function value and map it to the component performance, and introduce the mapped component performance as the feedback control quantity into the optimization iteration process.

[0071] S5. In each iteration of the optimization process in step S4, use the variational autoencoder and the hippopotamus optimization algorithm to adjust the process parameters and calculate the corresponding objective function value. If the change amount of the objective function value in this round is less than the threshold, the algorithm stops, and the process parameters at this time are the optimization results; otherwise, go to step S2.

[0072] In this embodiment, the threshold k is a dynamically adaptive parameter, and its initial value range is set to 0.0001 to 0.01, and it is dynamically adjusted according to the convergence trend of the objective function value during the iteration process. When the relative error of the change amount of the objective function value for three consecutive iterations is less than k, it is determined that the algorithm converges and the optimization stops, and the current process parameter combination is output as the optimal solution;

[0073] The value of the threshold k is determined in the following ways: (1) In the initial optimization stage, k is set to 0.01 to allow a large range of searches and quickly approach the global optimal region; (2) When the rate of decrease of the objective function value is lower than the preset threshold (the decrease amplitude is less than 5% every 10 iterations), k is adjusted to 0.001 to improve the local search accuracy; (3) If the objective function value does not change significantly (the change amount is less than 0.1%) after two consecutive adjustments, then k is further compressed to 0.0001 to ensure the convergence accuracy of the final solution; This dynamic adjustment mechanism combines the high-dimensional non-linear characteristics of the multi-energy field coupling optimization of aerospace complex components 7. Through the cooperation of the rough adjustment and fine adjustment stages, it not only avoids premature convergence and falling into local optima, but also reduces the redundant calculation amount, significantly improving the optimization efficiency and result reliability.

[0074] S6. Repeat steps S2 to S5 until the global optimal solution of the process parameter combination is obtained.

[0075] Embodiment 2

[0076] In this embodiment, a multi-energy field regulation process system for improving the service life of aerospace complex components includes: a component processing module, an algorithm initialization module, a function construction module, an iterative optimization module, and a threshold comparison module.

[0077] The component processing module is used to place the aerospace complex component 7 to be processed in a magnetic levitation polishing machine for grinding and polishing treatment with the cooperation of multiple physical fields of magnetic field - ultrasound - pressure - high temperature, and collect the process parameters and component performance of the processed aerospace complex component 7 and perform normalization processing to obtain the normalized process parameters and normalized component performance.

[0078] The algorithm initialization module uses a variational autoencoder to extract the latent space structure between the normalized process parameters, maps the high-dimensional parameters to a low-dimensional latent space to obtain low-dimensional latent variables, and initializes the population individual positions in the hippopotamus optimization algorithm using the low-dimensional latent variables.

[0079] The function construction module uses the initialized hippopotamus optimization algorithm to search for the optimal parameter combination in the compressed space, updates the individual positions, takes the normalized process parameters as the input, and the component performance index as the objective function to construct a comprehensive objective function.

[0080] The iterative optimization module aims at the comprehensive objective function, calculates the function value through the process parameters, optimizes and iterates the process parameter combination until the comprehensive objective function converges, extracts the function value and maps it to the component performance, and introduces the mapped component performance as a feedback control quantity into the optimization iteration process.

[0081] In each iteration of the optimization process of the iterative optimization module, the threshold comparison module uses the variational autoencoder and the hippopotamus optimization algorithm to adjust the process parameters and calculate the corresponding objective function value. If the change amount of the objective function value in this round is less than the threshold, the algorithm stops, and the process parameters at this time are the optimization result. Otherwise, the algorithm initialization module is restarted.

[0082] Repeat the working processes of the algorithm initialization module, the function construction module, the iterative optimization module and the threshold comparison module until the global optimal solution of the process parameter combination is obtained.

[0083] The embodiments described above are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solution of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A multi-energy field control process method for improving the service life of complex aerospace components, characterized in that: The following steps are involved: S1. Place the complex aerospace component to be processed in a magnetic suspension polishing machine for magnetic field-ultrasound-pressure-high temperature multi-physical field coordinated grinding and polishing, collect the process parameters and component performance of the processed complex aerospace component and perform normalization processing to obtain normalized process parameters and normalized component performance; S2. Using a variational autoencoder to extract the latent space structure between the normalized process parameters, and mapping the high-dimensional parameters to a low-dimensional latent space to obtain low-dimensional latent variables, and using the low-dimensional latent variables to initialize the individual positions of the population in the Hippo optimization algorithm; S3. Using the initialized Hippo optimization algorithm to search for the optimal parameter combination in the compression space, update the individual position, take the normalized process parameters as input, and the component performance index as the objective function to construct a comprehensive objective function; S4. Taking the comprehensive objective function as the target, calculating the function value through the process parameters, optimizing the iterative process parameter combination until the comprehensive objective function converges, extracting the function value and mapping it to component performance, and introducing the mapped component performance as a feedback control quantity into the optimization iteration process; S5. In each iteration of the optimization process in step S4, the process parameters are adjusted using the variational autoencoder and the Hippo optimization algorithm and the corresponding objective function value is calculated. If the change in the objective function value of this round is less than the threshold, the algorithm stops and the process parameters are optimized. Otherwise, proceed to step S2; S6. Repeat steps S2 to S5 until the global optimal solution of the process parameter combination is obtained.

2. According to claim 1, a multi-energy field control process method for improving the service life of complex aerospace components is characterized in that: The method for performing multi-physical field coordinated grinding and polishing processing using the magnetic suspension polishing machine includes: Ensure that the magnetic suspension polishing machine is in a closed state and the interior is clean and free of debris, place the complex aerospace component to be processed in the closed box of the magnetic suspension polishing machine, and after placement, close the box to ensure airtightness, then turn on the power, and start the magnetic suspension polishing machine through the console; The process parameters of the magnetic suspension magnetic field, ultrasonic field, pressure field and temperature field are set in the control console, and then the AC magnetic coil generates the alternating magnetic suspension magnetic field, the bottom sound wave generating device generates the ultrasonic field, the pressurizing device generates the pressure field, and the heating device generates the temperature field; The grinding tool generates relative movement with the surface of the complex aerospace component under the action of the magnetic levitation magnetic field, and the surface of the complex aerospace component is micro-grinded and polished through the synergistic effect of the magnetic levitation magnetic field, the ultrasonic field, the pressure field and the temperature field.

3. According to claim 1, a multi-energy field control process method for improving the service life of complex aerospace components is characterized in that: The normalized process parameters include: the magnetic field intensity H', the magnetic field frequency f m ', ultrasonic power W', ultrasonic frequency f u ', container pressure P', inner cavity heating temperature T', magnetic field action time t m ', Ultrasonic field action time t u ', pressure field action time t p ' and temperature field action time t t '; The normalized component properties include: surface roughness R', residual stress σ', internal defect closure degree D' and fatigue life L'.

4. According to claim 1, a multi-energy field control process method for improving the service life of complex aerospace components is characterized in that: Methods for initializing the positions of individuals in the Hippo optimization algorithm include: ， Among them, X t represents the current individual position, X rand represents a randomly selected individual in the population, X best represents the global optimal individual in the current population, R(τ)= σ(τ)·(X rand -Xτ) represents the random disturbance term, σ(τ) represents the disturbance factor that changes dynamically with time, r1, r2 and r3 represent random numbers between 0 and 1, LB represents the lower bound of the search space, UB represents the upper bound of the search space, q represents the random threshold of the control strategy, ω1(τ), ω2(τ), ω3(τ) and ω4(τ) represent dynamic weight adjustment factors, φ represents the nonlinear adjustment coefficient, ω i,min represents the minimum value of the weight, ω i,max represents the maximum value of the weight, τ represents the current number of iterations, represents the maximum number of iterations, α(τ) represents the first correction factor, α0 represents the first initial correction factor value, and γ represents the parameter controlling the decay speed of the first correction factor.

5. According to claim 4, a multi-energy field control process method for improving the service life of complex aerospace components is characterized in that: Methods for updating individual locations include: ， Among them, X prey represents the current global optimal solution, r represents a random number between 0 and 1, E represents the escape energy, E0 represents the initial escape energy, δ represents the parameter that controls the nonlinear attenuation of the escape energy, Ψ(τ) represents the dynamically adjusted convergence factor, η(τ) represents the penalty term weight, and C penalty represents the penalty factor of constrained optimization, β(τ) represents the second correction factor, β0 represents the second initial correction factor value, and μ controls the parameter of the decay speed of the second correction factor.

6. According to claim 1, a multi-energy field control process method for improving the service life of complex aerospace components is characterized in that: The comprehensive objective function includes: ， Among them, R' min Indicates the minimum value of surface roughness, R' max Indicates the maximum value of surface roughness, σ' min Represents the minimum value of residual stress, σ' max Represents the maximum value of residual stress, D' min Indicates the minimum value of the internal defect closure degree, D' max Indicates the maximum value of the internal defect closure degree, L' min Indicates the minimum value of fatigue life, L' max represents the maximum value of fatigue life, w1, w2, w3 and w4 represent the weights of each objective function.

7. A multi-energy field control process system for improving the service life of complex aerospace components, the system using the method described in any one of claims 1 to 6, characterized in that: include: Component processing module, algorithm initialization module, function construction module, iterative optimization module and threshold comparison module; The component processing module is used to place the complex aerospace component to be processed in a magnetic suspension polishing machine for magnetic field-ultrasound-pressure-high temperature multi-physical field coordinated grinding and polishing, and collect the process parameters and component performance of the processed complex aerospace component and perform normalization processing to obtain normalized process parameters and normalized component performance; The algorithm initialization module uses a variational autoencoder to extract the latent space structure between the normalized process parameters, and maps the high-dimensional parameters to a low-dimensional latent space to obtain low-dimensional latent variables, and uses the low-dimensional latent variables to initialize the population individual positions in the Hippo optimization algorithm; The function construction module uses the initialized Hippo optimization algorithm to search for the optimal parameter combination in the compression space, updates the individual positions, takes the normalized process parameters as input and the component performance index as the objective function, and constructs a comprehensive objective function; The iterative optimization module takes the comprehensive objective function as the target, calculates the function value through the process parameters, optimizes the iterative process parameter combination until the comprehensive objective function converges, extracts the function value and maps it to component performance, and introduces the mapped component performance as a feedback control quantity into the optimization iteration process; The threshold comparison module adjusts the process parameters and calculates the corresponding objective function values ​​using the variational autoencoder and the Hippo optimization algorithm during each iteration of the optimization process of the iterative optimization module. If the change in the objective function value of this round is less than the threshold, the algorithm stops and the process parameters are the optimization results. Otherwise, the algorithm initialization module is restarted. The workflows of the algorithm initialization module, the function construction module, the iterative optimization module and the threshold comparison module are repeated until a global optimal solution of the process parameter combination is obtained.

Images