Shot peening process parameter reverse solving method, device and medium based on global search

Abstract

The present application relates to the technical field of metal material surface strengthening process, and discloses a shot blasting process parameter reverse solving method, equipment and medium based on global search, comprising: establishing a shot blasting intensity prediction model and a coverage prediction model; the shot blasting intensity prediction model outputs the predicted shot blasting intensity; the coverage prediction model outputs the predicted coverage; setting the target shot blasting intensity and the target coverage, and the corresponding allowable error; discretizing the value range of each shot blasting process parameter to form a parameter space containing a limited number of discrete parameter combinations; traversing each discrete parameter combination in the parameter space, respectively calculating the predicted shot blasting intensity and the predicted coverage corresponding to the current parameter combination; screening the feasible solutions to form a feasible solution set; determining the final optimal process parameter combination from the feasible solution set. The present application traverses the entire discretized parameter space through systematic exhaustive search, fundamentally overcoming the problem that the experience method and the local optimization method are easy to fall into local optimization.

Description

Technical Field

[0001] This invention relates to the field of surface strengthening technology for metallic materials, specifically to a method, equipment, and medium for inversely solving shot peening process parameters based on global search. Background Technology

[0002] Shot peening is a core process for surface strengthening of critical components such as aero-engine blades and automotive gears. Its effectiveness is mainly characterized by two key indicators: shot peening intensity (Almen intensity) and coverage. During the process design phase, engineers need to determine the control parameters of the shot peening equipment in reverse engineering based on the required strength and coverage of the target workpiece. These parameters mainly include air pressure (p), flow rate (q), the moving speed of the shot peening equipment (v), and the distance (h) between the peening gun and the workpiece surface.

[0003] Currently, the main methods for determining these process parameters have significant shortcomings:

[0004] 1. Experience-based trial and error method: Over-reliance on engineers' personal experience for parameter debugging requires multiple production-inspection cycles for verification, which is extremely inefficient, yields unstable results, and makes it difficult to accumulate and pass on process knowledge.

[0005] 2. Single-objective optimization method: Existing optimization studies often only perform positive modeling and optimization for a single index such as shot peening intensity or coverage, failing to systematically coordinate parameter conflicts between two objectives. This method fails when both objectives need to be satisfied simultaneously.

[0006] 3. Lack of systematic reverse engineering capability: Existing technologies lack a systematic method that can reverse and completely solve for all feasible combinations of process parameters starting from given and determined target values ​​of intensity and coverage.

[0007] 4. Difficulty in process transfer: When the same material is produced on different equipment or in different batches, due to the lack of quantitative model guidance, a large number of process experiments often need to be carried out again, which is costly.

[0008] Therefore, developing a method for solving process parameters that can achieve synergistic, inverse, and global optimization of both shot peening intensity and coverage is of great significance for improving the scientific nature, efficiency, and quality stability of shot peening process design. Summary of the Invention

[0009] To address the problems of existing technologies, such as reliance on experience, inability to systematically perform dual-objective inverse solutions, and susceptibility to local optima, this invention provides a method, equipment, and medium for inverse solution of shot peening process parameters based on global search. This method constructs an accurate dual-objective forward prediction model and combines it with a systematic exhaustive search of the parameter space to ensure that all globally feasible solutions that meet the process requirements are found. Then, based on a multi-objective decision-making method, the optimal process scheme is selected from these solutions.

[0010] This invention is achieved through the following technical solution:

[0011] In a first aspect, the present invention provides a method for reverse solving of shot peening process parameters based on global search, comprising the following steps:

[0012] S1. Establish a shot peening intensity prediction model and a coverage prediction model; the shot peening intensity prediction model takes the air pressure (p) and flow rate (q) in the shot peening process parameters as input and outputs the predicted shot peening intensity (PI); the coverage prediction model takes the air pressure (p), flow rate (q), moving speed (v) and distance (h) in the shot peening process parameters as input and outputs the predicted coverage (C).

[0013] S2, Set the target shot peening intensity ( ) and target coverage ( ), and the corresponding allowable error ( , );

[0014] S3. Discretize the value range of each shot peening process parameter to form a parameter space containing a finite number of discrete parameter combinations. ;

[0015] S4. Traverse the parameter space For each discrete parameter combination, the predicted shot peening intensity is calculated using the shot peening intensity prediction model and the coverage prediction model. ) and predicted coverage ( );

[0016] S5. Parameter combinations that meet the following conditions are selected as feasible solutions: the absolute value of the difference between the predicted shot peening intensity and the target shot peening intensity corresponding to the parameter combination does not exceed the corresponding allowable error, and the absolute value of the difference between the predicted coverage and the target coverage corresponding to the parameter combination does not exceed the corresponding allowable error; all feasible solutions constitute a feasible solution set. );

[0017] S6. From the feasible solution set ( The final optimal combination of process parameters is determined in the process. ).

[0018] Further, in step S6, determining the final optimal combination of process parameters from the feasible solution set specifically includes:

[0019] S6.1 Construct a comprehensive evaluation function F(x), which is used to calculate the evaluation value corresponding to any feasible solution. The evaluation value is calculated based at least on the error between the predicted shot peening intensity and the target shot peening intensity, and the error between the predicted coverage and the target coverage.

[0020] S6.2 Calculate the evaluation value corresponding to each feasible solution in the feasible solution set, and select the process parameter combination corresponding to the feasible solution with the best evaluation value (usually the minimum value) as the optimal process parameter combination.

[0021] Furthermore, the comprehensive evaluation function F(x) is also calculated based on process parameters that reflect production efficiency. Preferably, the process parameter reflecting production efficiency is the moving speed (v).

[0022] Furthermore, a specific expression for the comprehensive evaluation function F(x) is as follows:

[0023] ;

[0024] Where PI represents the predicted shot peening intensity. For the target shot peening intensity, C represents the allowable error for shot peening intensity; C represents the predicted coverage. For target coverage, The coverage tolerance is denoted by G; G is a parameter related to process efficiency. , , These are the weight coefficients for the corresponding terms, and .

[0025] Furthermore, a preferred expression for the parameter G related to process efficiency is: Where v is the movement speed in the current parameter combination, This represents the upper limit of the range of the moving speed parameter. This means that, provided the accuracy requirements are met, the larger the moving speed v (the higher the production efficiency), the smaller this value will be, which is more beneficial for reducing the overall score F(x), thereby guiding the system to select the most efficient parameter combination.

[0026] Furthermore, in step S3, the specific process of discretizing the value range of each shot peening process parameter involves discretizing each parameter at equal intervals within its feasible range to generate a series of discrete levels, such as... And so on.

[0027] Furthermore, the shot peening intensity prediction model is established by fitting shot peening process experimental data using a multivariate nonlinear regression method; the coverage prediction model is established by fitting shot peening process experimental data using the response surface methodology (RSM). These models are the foundation for the accurate prediction achieved by this invention.

[0028] Further, in step S4, the parameter space is traversed. Each discrete parameter combination in the solution is implemented through an exhaustive search method to ensure that all discrete parameter combinations in the parameter space are traversed, avoiding the omission of any potential feasible solutions, thereby guaranteeing the completeness and global optimality of the solution.

[0029] In a second aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the aforementioned method for reverse solving of shot peening process parameters based on global search.

[0030] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned method for reverse solving of shot peening process parameters based on global search.

[0031] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0032] This invention overcomes the problem of empirical methods and local optimization methods easily getting trapped in local optima by systematically exhaustively searching the entire discretized parameter space, ensuring that all feasible solutions that meet the process requirements can be found, and providing sufficient choices for decision-making.

[0033] This invention provides for the first time a systematic method for solving process parameters in reverse and systematically, starting from given target values ​​of shot peening intensity and coverage, thus solving a long-standing technical problem in this field.

[0034] This invention not only pursues process precision, but also incorporates production efficiency (such as movement speed) into the decision-making system through a comprehensive evaluation function, thereby achieving synergistic optimization of multiple objectives such as quality and efficiency, which meets actual production needs.

[0035] This invention performs rapid calculations based on an experimentally verified accurate model, reducing the process development cycle from several days to minutes, significantly reducing the number of trial and error experiments, and saving material and time costs.

[0036] This invention transforms process knowledge into reusable models and algorithms, reducing reliance on personnel experience and facilitating stable process transfer across different equipment or production batches, thereby improving the first-pass yield. Attached Figure Description

[0037] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0038] Figure 1 This is a flowchart of a reverse solution method for shot peening process parameters based on global search, as described in this invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0040] This embodiment 1 provides a method for inversely solving shot peening process parameters based on global search, such as... Figure 1 As shown, there are 6 steps in total, including S1-S6. Next, the implementation process of the 6 steps in this invention will be described in detail.

[0041] S1. Establish a shot peening intensity prediction model and a coverage prediction model; the shot peening intensity prediction model takes the air pressure (p) and flow rate (q) in the shot peening process parameters as input and outputs the predicted shot peening intensity (PI); the coverage prediction model takes the air pressure (p), flow rate (q), moving speed (v) and distance (h) in the shot peening process parameters as input and outputs the predicted coverage (C).

[0042] This step is fundamental to achieving accurate reverse engineering in this invention. It requires establishing quantitative mathematical relationships between two key process indicators (shot peening intensity and coverage) and controllable process parameters, i.e., a forward prediction model.

[0043] S1.1 Shot peening intensity prediction model:

[0044] Model inputs: air pressure (p) and flow rate (q) from the shot peening process parameters.

[0045] Model output: Predicted shot peening intensity (PI).

[0046] Methodology: A multivariate nonlinear regression method was used, based on experimental data from previous shot peening processes, to establish the model. Its general mathematical form can be expressed as:

[0047] ;

[0048] Where p is air pressure and q is flow rate. , , , , These are the model coefficients determined through regression analysis based on experimental data. The model reflects the combined effects of air pressure and flow rate (and their interaction and nonlinear effects) on shot peening intensity. Furthermore, during the process tests, the gun movement speed was kept no greater than the speed required for saturation intensity to ensure that the shot peening strengthening process achieved a defined saturation intensity.

[0049] S1.2 Coverage Prediction Model:

[0050] Model inputs: air pressure (p), flow rate (q), moving speed (v), and distance (h) in the shot peening process parameters.

[0051] Model output: Predicted coverage (C).

[0052] Methodology: Response surface methodology (RSM) was employed, and the model was established based on experimental data from shot peening. This is an effective statistical method suitable for modeling and optimizing multivariate systems. Its general mathematical form can be expressed as a quadratic polynomial model containing linear terms, interaction terms, and square terms.

[0053] ;

[0054] in, , is a vector containing four process parameters. For constant terms, The coefficients of the linear term, The coefficient of the interaction term. These are the coefficients for the squared terms. All coefficients were determined based on experimental data through response surface analysis fitting. This model can characterize the complex effects of independent actions, pairwise interactions, and nonlinear effects on coverage.

[0055] Furthermore, the shot peening intensity prediction model is established based on shot peening process experimental data using a multivariate nonlinear regression method; the coverage prediction model is established based on shot peening process experimental data using the response surface methodology (RSM). These models form the computational basis for achieving accurate prediction and subsequent inverse solving.

[0056] S2, Set the target shot peening intensity ( ) and target coverage ( ), and the corresponding allowable error ( , ).

[0057] Based on the technical requirements of the workpiece (such as a TC17 titanium alloy blade), the target shot peening intensity to be achieved in this process design is clearly defined. ) and target coverage ( At the same time, set a reasonable allowable error (). , This is used to define the boundaries of the "feasible solution" in subsequent steps. The allowable error can be determined based on measurement accuracy and process tolerance.

[0058] For example, the first set of objectives: mmN, Allowable error: mmN, Second group objective: mmN, Allowable error: mmN, .

[0059] S3. Discretize the value range of each shot peening process parameter to form a parameter space containing a finite number of discrete parameter combinations. .

[0060] Based on the actual working range and process feasibility of the shot peening equipment, determine the continuous value range of each process parameter (p, q, v, h). For example, the value range of air pressure p is... (For example: 0.6 bar to 1.5 bar); the value range of flow rate q is... (For example: 0.6 kg / min to 0.95 kg / min); the range of the moving speed v is... (For example: 30 mm / min to 120 mm / min); the range of distance h is... (For example: 5mm to 25mm).

[0061] To transform a continuous, infinite parameter space into a finite search space that can be processed by a computer system, the range of values ​​for each parameter needs to be discretized. This invention preferably employs an equal-interval discretization method, which uniformly divides the continuous interval of each parameter into several segments, discretizing the continuous parameter space into finite grid points to ensure coverage of the entire feasible region. Each segment point is taken as the discretization level of the parameter, and its mathematical expression is as follows:

[0062] ;

[0063] ;

[0064] ;

[0065] ;

[0066] The set of corresponding parameters is as follows:

[0067] (Gas pressure collection)

[0068] (Traffic set)

[0069] (Velocity set)

[0070] (Distance set)

[0071] , , , These are the interval lengths for air pressure, flow rate, speed, and distance, respectively.

[0072] Through this discretization process, the original continuous four-dimensional parameter space is transformed into a discrete parameter space consisting of finite grid points. It is the Cartesian product of four discrete parameter sets:

[0073] ;

[0074] This space Includes A set of discrete parameter combinations (i.e., grid points) represents all candidate process schemes to be examined.

[0075] For example, if each parameter has 5 levels, specifically: the air pressure p is discretized as {0.6, 0.8, 1.0, 1.2, 1.5}, in bar; the flow rate q is discretized as {0.6, 0.7, 0.8, 0.9, 0.95}, in kg / min; the velocity v is discretized as {30, 50, 80, 100, 120}, in mm / min; and the distance h is discretized as {5, 10, 15, 20, 25}, in mm.

[0076] but This involves a total of 5×5×5×5=625 points. This step ensures that the search scope systematically covers the entire process feasible region, laying the foundation for subsequent exhaustive searches.

[0077] It is understood that the discretization method described in step S3 is not limited to equal-interval discretization. Those skilled in the art can adjust the discretization strategy according to actual computational resources and accuracy requirements. For example, a non-uniform discretization strategy can be adopted: in regions where, based on experience or preliminary analysis, the process effect (intensity, coverage) is judged to be more sensitive to parameter changes, a smaller discretization length (i.e., a denser grid) can be used; in regions where the effect changes gradually, a larger discretization length (i.e., a sparser grid) can be used. This strategy helps to allocate computational resources more intelligently while controlling the total computational load, improving search efficiency, and it can also achieve the goal of a global search traversing all discrete points.

[0078] S4. Traverse each discrete parameter combination in the parameter space, and use the shot peening intensity prediction model and the coverage prediction model to calculate the predicted shot peening intensity corresponding to the current parameter combination. ) and predicted coverage ( ).

[0079] This step is the core of the global search. Specifically, it is implemented through exhaustive search, that is, systematically and without omission access to the parameter space. All 625 discrete points in the dataset. For each parameter combination... :

[0080] 1. Take one of them and Substitute the shot peening intensity prediction model established in step S1 The predicted shot peening intensity under this combination was calculated. .

[0081] 2. Combine the complete set Substitute into the coverage prediction model established in step S1 Calculate the predicted coverage under this combination. .

[0082] This process ensures that every possible combination of parameters is evaluated using a bi-objective model.

[0083] S5. Parameter combinations that meet the following conditions are selected as feasible solutions: the absolute value of the difference between the predicted shot peening intensity and the target shot peening intensity corresponding to the parameter combination does not exceed the corresponding allowable error, and the absolute value of the difference between the predicted coverage and the target coverage corresponding to the parameter combination does not exceed the corresponding allowable error; all feasible solutions constitute a feasible solution set. ).

[0084] Based on the target and allowable error set in step S2, each parameter combination calculated in the previous step is evaluated. The evaluation criteria are as follows:

[0085] and ;

[0086] That is, a combination of parameters that simultaneously satisfies the absolute error between the predicted shot peening intensity and the target value being within the allowable range, and the absolute error between the predicted coverage value and the target value also being within the allowable range, is considered a "feasible solution".

[0087] Collect all feasible solutions to form a feasible solution set. , .

[0088] The following details the specific algorithmic implementation process of steps S4 (traversal calculation) and S5 (selecting feasible solutions). This process involves systematically enumerating the parameter space. All discrete parameter combinations are selected and filtered using bi-objective constraints.

[0089] The algorithm flow is as follows:

[0090] 1. Initialization: Create an empty feasible solution set. It is used to store all parameter combinations that meet the process requirements and their calculation results.

[0091] 2. Traversal and Calculation: A multi-layered loop structure is adopted to traverse each discrete level in the pressure set P, flow rate set Q, speed set V, and distance set H in turn.

[0092] For each specific combination of parameters Perform the following calculations:

[0093] a. Intensity Prediction: Based on the current air pressure and traffic Substitute into the shot peening intensity prediction model The predicted shot peening intensity value was calculated. .

[0094] b. Coverage prediction: Combining the current complete parameter set Substitute into the coverage prediction model The predicted coverage value is calculated. .

[0095] c. Error Calculation: Calculate the absolute error between the predicted value and the target value respectively.

[0096] Shot peening intensity error ;

[0097] Coverage error ;

[0098] 3. Conditional Judgment and Filtering: Determine whether the current parameter combination simultaneously meets the dual-target accuracy requirements:

[0099] Judgment criterion: Strength error Less than or equal to the preset strength tolerance And coverage error Less than or equal to the preset coverage tolerance error .

[0100] If the conditions are met, then combine the parameters. Together with its calculation results , and error , Added to the feasible solution set as a complete record. If the condition is not met, ignore the combination and continue to check the next parameter combination.

[0101] 4. Loop Termination and Output: The algorithm terminates after all parameter combinations (a total of m×n×k×l) have been traversed and evaluated. At this point, the feasible solution set is... It contains all possible combinations of process parameters that can simultaneously satisfy the target shot peening intensity and target coverage requirements. This feasible solution set is output as input for subsequent multi-objective decision-making.

[0102] This process is essentially a specific application of exhaustive search or brute-force search in shot peening process parameter optimization. Its core advantage lies in the completeness of the search, meaning that a solution that meets the requirements exists in the discrete parameter space. In this way, the method can find the optimal combination, thus avoiding the problem of traditional optimization methods getting trapped in local optima. Although the number of combinations to be checked may be large, the overall solution efficiency is much higher than that of traditional physical trial-and-error methods because the calculation speed of a single model prediction is extremely fast.

[0103] It is understood that the traversal of the parameter space described in this invention aims to emphasize the systematic and complete nature of the search, and its specific implementation algorithm is not limited to the basic exhaustive search method. Within the framework of this invention, those skilled in the art can employ various methods to ensure the accuracy and completeness of the parameter space. This is achieved through a systematic search strategy. The core principle is that, regardless of the strategy employed, the discrete parameter space must ultimately be ensured. Each discrete parameter combination (grid point) is evaluated once by the shot peening intensity prediction model and the coverage prediction model to meet the completeness requirement of the global search.

[0104] Based on this principle, to better balance the globality of the search and computational efficiency under complex conditions, a more efficient hybrid search strategy can be adopted. One feasible implementation is to combine global filtering with local fine-grained search. Specifically:

[0105] Phase 1: Global Coarse Screening. First, in step S3, a relatively sparse initial discrete parameter space is constructed. (For example, using a larger equal spacing or a larger non-uniform step size). Then, steps S4 and S5 are performed in this sparse space. A full traversal and filtering process is performed to obtain a preliminary set of feasible solutions or to identify key regions containing feasible solutions.

[0106] Phase Two: Local Fine-Search. Within the key regions identified in the previous phase, a more refined secondary discrete parameter space is constructed. (For example, narrowing the parameter range and using a smaller offset step size). Then, again in this refined space... The global traversal and filtering in steps S4 and S5 are then performed. Regions deemed non-critical in the global coarse screening are not subject to further fine-grained searching.

[0107] This two- or multi-stage search strategy, progressing from coarse to fine, essentially allocates computational resources by adjusting the discretization granularity. It performs intensive searches in critical regions to ensure accuracy, while conducting coarse searches in non-critical regions to improve efficiency, ultimately achieving an equivalent systematic coverage of the entire original parameter space Ω. This represents an equivalent optimization and variation within the core framework of the proposed method for systematically searching the discrete parameter space based on a forward model to screen for feasible solutions with dual objectives.

[0108] S6. From the feasible solution set ( The final optimal combination of process parameters is determined in the process. ).

[0109] When there are multiple solutions in the feasible solution set, it is necessary to determine the optimal solution for production. This is achieved through a multi-objective decision-making process:

[0110] 1. Construct a comprehensive evaluation function: Define a function F(x) to quantitatively evaluate the overall merits of each feasible solution x. This function should at least consider process accuracy, for example:

[0111] ;

[0112] in, , This is a weighting coefficient that reflects the degree of importance attached to the accuracy of intensity and coverage.

[0113] To enhance its practicality in engineering applications, this function can also incorporate terms reflecting production efficiency, such as terms related to the movement speed v. The goal is to select the faster processing speed while still meeting the required accuracy. The final function form is: ;and .

[0114] 2. Calculate scores and select the best solution: Divide the feasible solution set... Substituting the parameters and predicted PI and C values ​​of each solution into the comprehensive evaluation function above, a comprehensive score F is calculated. The process parameter combination corresponding to the solution with the smallest comprehensive score F (representing the best overall performance) is selected as the final optimal process parameter combination. .

[0115] The method of the present invention will be described next through specific examples.

[0116] The shot peening process for the titanium alloy blades of a certain type of aero-engine TC17 requires that specific shot peening intensity and coverage indicators be met simultaneously.

[0117] (1) Forward model establishment and validation:

[0118] Implementation steps:

[0119] Based on 29 sets of system experimental data (see Table 1), a specific prediction model was established through multivariate nonlinear regression fitting.

[0120] Table 1. Parameters of different shot peening processes and shot peening intensity and coverage (partial key data)

[0121]

[0122] Shot peening intensity model (based on air pressure p and flow rate q):

[0123] ;

[0124] Model validation:

[0125] Coefficient of determination ;

[0126] Mean absolute error (MAE) = 0.008 mmN;

[0127] Maximum error: 2.5% (Group 23);

[0128] Coverage model (based on full parameters p, q, v, h):

[0129] ;

[0130] Model validation:

[0131] Coefficient of determination ;

[0132] Mean absolute error (MAE) = 18.5%;

[0133] Maximum error: 12% (Group 9).

[0134] (2) Exhaustive search parameter settings:

[0135] Discrete parameter space:

[0136] =[0.6,0.8,1.0,1.2,1.5]#5 barometric pressure levels;

[0137] =[0.6,0.7,0.8,0.9,0.95]#5 flow levels;

[0138] =[30,50,80,100,120]#5 speed levels;

[0139] =[5,10,15,20,25]#5 horizontal distances;

[0140] Total number of parameter combinations: 5×5×5×5=625.

[0141] (3) Reverse solution process:

[0142] First target: PI = 0.27 ± 0.02 mmN, C = 200% ± 25%;

[0143] Exhaustive search results statistics:

[0144] Total number of search combinations: 625;

[0145] Strength requirements met: 58;

[0146] Simultaneously satisfying dual objectives: 15;

[0147] Search time: <20 seconds;

[0148] The feasible solution set (the first 5 are sorted by comprehensive score) is shown in Table 2:

[0149] Table 2

[0150]

[0151] Second set of targets: PI = 0.21 ± 0.02 mmN, C = 100% ± 25%

[0152] Exhaustive search results statistics:

[0153] Total number of search combinations: 625;

[0154] Strength requirements met: 42;

[0155] Simultaneously satisfying dual objectives: 8;

[0156] Search time < 20 seconds;

[0157] The feasible solution set (the first 5 are sorted by comprehensive score) is shown in Table 3:

[0158] Table 3

[0159]

[0160] (4) Multi-objective decision making and verification

[0161] Weighting configuration (precision priority):

[0162] weights = {

[0163] 'w1': 0.40, # Strength accuracy

[0164] 'w2': 0.40, # Coverage precision

[0165] 'w3': 0.2, # Efficiency

[0166] }

[0167] Group 1:

[0168] Optimal solution: x* = (p=1.2, q=0.75, v=80, h=20)

[0169] Experimental verification:

[0170] Measured shot peening intensity: 0.270 mmN (error 0.4%);

[0171] Actual coverage: 200% (0% error);

[0172] Group 2:

[0173] Optimal solution: x*=(p=0.8,q=0.65,v=120,h=25);

[0174] Experimental verification:

[0175] Measured shot peening intensity: 0.209 mmN (error 0.5%);

[0176] Actual coverage: 100% (0% error);

[0177] (5) Comparison with traditional methods;

[0178] The results are shown in Table 4:

[0179] Table 4

[0180] Performance indicators Traditional trial and error method Local optimization method This invention Search combination number 15-25 50-100 625 (Comprehensive) Find a feasible solution 1-2 3-5 15 / 8 Optimal solution quality Local Optimum Local Optimum Global Optimum Solution time 2-3 days 2-4 hours <5 minutes Success rate 65-75% 80-85% 100%

[0181] Engineering application results:

[0182] Process development cycle: shortened from 5 days to 2 hours;

[0183] Material savings: 28 fewer experiments were conducted, resulting in an 85% cost reduction;

[0184] First-pass yield increased from 70% to 98%;

[0185] Process stability: The optimal solution can be found quickly under different target requirements;

[0186] In summary, the present invention has the following effects:

[0187] Global optimal guarantee: Through exhaustive search of the system, the system ensures that the globally optimal combination of process parameters is found;

[0188] Solution set completeness: Obtaining all feasible solutions that meet the requirements, providing sufficient choices for multi-objective decision-making;

[0189] Highly practical for engineering applications: Based on an experimentally validated, accurate model, ensuring the reliability of prediction results;

[0190] High computational efficiency: Although the number of search combinations is large, the single calculation is fast, and the total time is significantly better than the trial and error method;

[0191] Multi-objective coordination: Simultaneously optimizing multiple objectives such as intensity, coverage, and efficiency;

[0192] Highly adaptable: Suitable for optimizing process parameters for different materials and equipment;

[0193] High degree of standardization: Eliminates the influence of human factors and ensures consistency of process results;

[0194] The inverse solution method for dual-objective process parameters of shot peening intensity and coverage established in this invention, based on exhaustive enumeration, fundamentally solves the problems of local optima and reliance on experience in traditional methods by systematically traversing the entire parameter space. This provides a systematic solution for the precise design and optimization of shot peening strengthening processes, significantly improving process development efficiency and quality stability. This method can not only be used for new process development but also guide the optimization and improvement of existing processes, possessing significant practical value for ensuring product quality and reducing production costs. Based on this method, when the same material is shot peened on the same equipment, the shot peening process parameters required to meet different shot peening intensity and coverage requirements can be directly calculated.

[0195] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for inversely solving shot peening process parameters based on global search, characterized in that, include: Establish a shot peening intensity prediction model and a coverage prediction model; the shot peening intensity prediction model takes the air pressure and flow rate in the shot peening process parameters as input and outputs the predicted shot peening intensity; the coverage prediction model takes the air pressure, flow rate, moving speed and distance in the shot peening process parameters as input and outputs the predicted coverage. Set the target shot peening intensity and target coverage, as well as the corresponding allowable error; The range of values ​​for each shot peening process parameter is discretized to form a parameter space containing a finite number of discrete parameter combinations. For each discrete parameter combination in the parameter space, the predicted shot peening intensity and predicted coverage are calculated respectively using the shot peening intensity prediction model and the coverage prediction model. The parameter combinations that meet the following conditions are selected as feasible solutions: the absolute value of the difference between the predicted shot peening intensity and the target shot peening intensity corresponding to the parameter combination does not exceed the corresponding allowable error, and the absolute value of the difference between the predicted coverage and the target coverage corresponding to the parameter combination does not exceed the corresponding allowable error; all feasible solutions constitute a feasible solution set. The final optimal combination of process parameters is determined from the feasible solution set, and the specific process is as follows: A comprehensive evaluation function is constructed to calculate the evaluation value corresponding to any feasible solution. The evaluation value is calculated based at least on the error between the predicted shot peening intensity and the target shot peening intensity, and the error between the predicted coverage and the target coverage. The comprehensive evaluation function F is: ； Where PI represents the predicted shot peening intensity. For the target shot peening intensity, C represents the allowable error for shot peening intensity; C represents the predicted coverage. For target coverage, The coverage tolerance is denoted by G; G is a parameter related to process efficiency. , , These are the weight coefficients for the corresponding terms, and ; Calculate the evaluation value corresponding to each feasible solution in the feasible solution set, and select the process parameter combination corresponding to the feasible solution with the best evaluation value as the optimal process parameter combination.

2. The method for reverse solving of shot peening process parameters based on global search according to claim 1, characterized in that, The comprehensive evaluation function is also calculated based on process parameters that reflect production efficiency, and the process parameter reflecting production efficiency is the moving speed.

3. The method for reverse solving of shot peening process parameters based on global search according to claim 1, characterized in that, The expression for the parameter G related to process efficiency is: ； Where v is the movement speed in the current parameter combination, This represents the upper limit of the movement speed parameter range.

4. The method for reverse solving of shot peening process parameters based on global search according to claim 1, characterized in that, The specific process of discretizing the value range of each shot peening process parameter is to discretize the value range of each shot peening process parameter at equal intervals.

5. The method for reverse solving of shot peening process parameters based on global search according to claim 1, characterized in that, The shot peening intensity prediction model is established by fitting the shot peening process experimental data using a multivariate nonlinear regression method; the coverage prediction model is established by fitting the shot peening process experimental data using the response surface methodology.

6. The method for reverse solving of shot peening process parameters based on global search according to claim 1, characterized in that, The traversal of each discrete parameter combination in the parameter space is specifically achieved through an exhaustive search method to traverse all discrete parameter combinations in the parameter space.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the reverse solution method for shot peening process parameters based on global search as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the reverse solution method for shot peening process parameters based on global search as described in any one of claims 1 to 6.

Images