Spraying process parameter optimization method and system based on big data

Abstract

The application discloses a spraying process parameter optimization method and system based on big data, and the method comprises data acquisition, calculation of parameter influence value, calculation of quality index correlation degree, reconstruction of correlation matrix, construction of neighborhood similarity matrix, optimization of membership matrix, and real-time process parameter optimization.The application belongs to the field of data processing, and specifically refers to a spraying process parameter optimization method and system based on big data.The scheme defines parameter node influence value, describes the level and attenuation characteristics of parameter influence, introduces a time factor to define process parameter combination similarity, objectively reflects the internal correlation of quality indexes based on quality index correlation degree, completes unverified correlation by reconstructing the correlation matrix, introduces a dynamic attenuation term, dynamically adjusts the attenuation degree according to the spraying process combination similarity, and avoids information loss, dynamically adjusts the similarity influence, and optimizes the neighborhood similarity weight, thereby improving the process parameter optimization effect.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for optimizing spraying process parameters based on big data. Background Technology

[0002] Optimization methods for spray coating process parameters involve data analysis to explore the correlation patterns of spray coating quality indicators and determine the optimal parameter combination to achieve stable coating quality and improved production efficiency. However, general optimization methods for spray coating process parameters often neglect the attenuation of the transmission of process parameter influences, leading to distorted assessments of indirect effects and superficial analysis of quality indicator correlations, resulting in low reliability of parameter optimization. Furthermore, general optimization methods often waste information due to unverified correlations, leading to missing key information and rigid handling of the influence of similar process parameters and quality, ultimately resulting in poor optimization effects. Summary of the Invention

[0003] To address the above issues and overcome the shortcomings of existing technologies, this invention provides a method and system for optimizing spraying process parameters based on big data. Addressing the problems of general spraying process parameter optimization methods neglecting the transmission and attenuation of process parameter influences, leading to distorted assessments of indirect effects and superficial analysis of quality index correlations, resulting in low reliability of parameter optimization, this solution defines parameter node influence values ​​to characterize the hierarchy and attenuation characteristics of parameter influences. It quantifies the total influence of process parameters based on the total influence value, more comprehensively reflecting the true total impact of process combinations on indicators. Furthermore, it introduces a time-effect factor to define the similarity of process parameter combinations, aligning with the influence of environmental changes on spraying effects. Finally, it quantifies the degree of overlap of parameter combinations based on quality index correlations, providing an objective... This approach reflects the inherent correlation of quality indicators, thereby making process parameter optimization more closely aligned with the actual characteristics of the spraying process and improving the reliability of process parameter optimization. Addressing the issue of unverified correlations wasting information in general spraying process parameter optimization methods, leading to missing key information and rigid handling of the influence of similar process parameters and quality, resulting in poor optimization performance, this solution reconstructs the correlation matrix to complete unverified correlations and introduces a dynamic attenuation term. The attenuation level is dynamically adjusted based on the similarity of spraying process combinations, making the mutual influence of similar process combinations stronger and avoiding information loss. Based on the dynamic adjustment of similarity influence, the correlation accuracy is improved, and subjective errors are reduced by optimizing the neighborhood similarity weights, thus improving the optimization effect of process parameters.

[0004] The technical solution adopted by this invention is as follows: The method for optimizing spraying process parameters based on big data provided by this invention includes the following steps:

[0005] Step S1: Data Acquisition;

[0006] Step S2: Calculate the influence value of the parameters;

[0007] Step S3: Calculate the correlation degree of quality indicators;

[0008] Step S4: Reconstruct the correlation matrix;

[0009] Step S5: Construct a neighborhood similarity matrix;

[0010] Step S6: Membership matrix optimization;

[0011] Step S7: Real-time process parameter optimization.

[0012] Furthermore, in step S1, the data acquisition involves collecting historical spraying process data, constructing a directed acyclic hierarchical graph, and performing normalization processing.

[0013] Furthermore, in step S2, the calculated parameter influence value is the defined parameter node influence value, which is the influence of the parameter on the quality index, and is dynamically calculated based on the influence value of the child nodes; the calculated parameter total influence value is the total influence of the parameter on all relevant quality indicators.

[0014] Furthermore, in step S3, the calculation of the correlation degree of quality indicators is to define the similarity degree of process parameter combinations, quantify it by the degree of overlap of influence paths, and add a time factor; calculate the similarity degree between the process parameter combinations and the set of quality indicators, and define the correlation degree of quality indicators.

[0015] Further, in step S4, the reconstruction of the correlation matrix involves setting up a correlation matrix between the combination of process parameters and the quality index, defining a parameter neighborhood matrix, and introducing a dynamic attenuation term to dynamically adjust the attenuation degree based on the process similarity of the spraying process parameter combination; defining a quality neighborhood optimization matrix, and finally reconstructing the correlation matrix.

[0016] Further, in step S5, constructing the neighborhood similarity matrix involves constructing a parameter neighborhood similarity matrix and a quality neighborhood similarity matrix.

[0017] Further, in step S6, the membership matrix optimization involves constructing a parameter transition matrix and a quality transition matrix; initializing the membership matrix; and outputting the optimized membership matrix of the process parameter combination on the quality index through bidirectional membership propagation, thereby selecting the optimal process parameter combination.

[0018] Furthermore, in step S7, the real-time process parameter optimization is based on optimizing the membership matrix and dynamically adjusting the process parameters in conjunction with real-time production data.

[0019] The big data-based spraying process parameter optimization system provided by this invention includes a data acquisition module, a parameter influence value calculation module, a quality index correlation calculation module, a correlation matrix reconstruction module, a neighborhood similarity matrix construction module, a membership matrix optimization module, and a real-time process parameter optimization module.

[0020] The data acquisition module collects and normalizes historical spraying process data to construct a directed acyclic hierarchical graph.

[0021] The parameter influence value calculation module is based on a directed acyclic hierarchical graph, defines the influence value of parameter nodes, and calculates the total influence value of parameters.

[0022] The quality index correlation calculation module calculates the similarity of process parameter combinations and then defines the quality index correlation.

[0023] The correlation matrix reconstruction module completes the correlation matrix and constructs a reconstructed correlation matrix through the parameter neighborhood matrix and the quality neighborhood optimization matrix.

[0024] The neighborhood similarity matrix construction module obtains the similarity matrix by minimizing the reconstruction error correction weight coefficient;

[0025] The membership matrix optimization module obtains the optimized membership matrix of process parameter combinations on quality indicators through bidirectional membership propagation and iterative optimization.

[0026] The real-time process parameter optimization module optimizes process parameters in real time based on the optimized membership matrix and real-time data.

[0027] The beneficial effects achieved by the present invention using the above solution are as follows:

[0028] (1) To address the problem that general spraying process parameter optimization methods neglect the transmission and attenuation of the influence of process parameters, leading to distorted assessment of indirect influences and superficial analysis of the correlation of quality indicators, resulting in low reliability of parameter optimization, this solution defines the influence value of parameter nodes to characterize the hierarchy and attenuation characteristics of parameter influence, quantifies the total influence of process parameters based on the total influence value, and more comprehensively reflects the true total influence of process combinations on indicators; introduces an aging factor to define the similarity of process parameter combinations, which conforms to the influence law of environmental changes on spraying effect; quantifies the degree of overlap of parameter combinations based on the correlation of quality indicators, objectively reflecting the intrinsic correlation of quality indicators; thus, the optimization of process parameters is more in line with the actual characteristics of the spraying process, improving the reliability of process parameter optimization.

[0029] (2) To address the problem that general spraying process parameter optimization methods waste information due to unverified associations, resulting in missing key information and rigid handling of the influence of similar process parameters and quality, which leads to poor optimization results, this solution completes the unverified associations by reconstructing the association matrix and introduces a dynamic attenuation term. The attenuation degree is dynamically adjusted according to the similarity of spraying process combinations, making the mutual influence of similar process combinations stronger and avoiding information loss. Based on the dynamic adjustment of similarity influence, the association accuracy is improved, and subjective errors are reduced by optimizing the neighborhood similarity weight, thereby improving the optimization effect of process parameters. Attached Figure Description

[0030] Figure 1 A flowchart illustrating the big data-based method for optimizing spraying process parameters provided by this invention;

[0031] Figure 2 This is a schematic diagram of the big data-based spraying process parameter optimization system provided by the present invention.

[0032] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0033] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0034] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the system or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0035] Example 1, see Figure 1 The present invention provides a method for optimizing spraying process parameters based on big data, which includes the following steps:

[0036] Step S1: Data acquisition, collect historical spraying process data and normalize it to construct a directed acyclic hierarchical graph;

[0037] Step S2: Calculate the parameter influence value. Based on the directed acyclic hierarchical graph, define the parameter node influence value and calculate the total parameter influence value.

[0038] Step S3: Calculate the correlation degree of quality indicators, calculate the similarity degree of process parameter combinations, and then define the correlation degree of quality indicators;

[0039] Step S4: Reconstruct the correlation matrix, complete the correlation matrix, and construct the reconstructed correlation matrix through the parameter neighborhood matrix and the quality neighborhood optimization matrix;

[0040] Step S5: Construct a neighborhood similarity matrix, and obtain the similarity matrix by minimizing the reconstruction error and correcting the weight coefficients;

[0041] Step S6: Membership matrix optimization. Through bidirectional membership propagation and iterative optimization, the optimized membership matrix of process parameter combinations on quality indicators is obtained.

[0042] Step S7: Real-time process parameter optimization. Based on the optimized membership matrix and real-time data, real-time process parameter optimization is achieved.

[0043] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, data acquisition involves collecting historical spraying process data and constructing a directed acyclic hierarchical graph. The historical spraying process data includes spraying distance, spray gun speed, paint viscosity, spraying pressure, atomized particle diameter, paint deposition rate, coating thickness, adhesion, and roughness. Normalization is performed to eliminate the influence of dimensions. The directed acyclic hierarchical graph uses quality indicators as root nodes, intermediate variables as intermediate nodes, and process parameters as leaf nodes to construct a hierarchical influence graph.

[0044] Data Acquisition:

[0045] Spraying distance is the vertical distance between the spray gun nozzle and the surface of the workpiece being sprayed, which is measured by a laser rangefinder.

[0046] Spray gun speed refers to the linear velocity of the spray gun moving along the workpiece surface, which is collected by a speed sensor;

[0047] Coating viscosity, a parameter reflecting the flow resistance of the coating, is measured using a rotational viscometer;

[0048] Spraying pressure refers to the compressed air pressure at the inlet of the spray gun, which is collected by a pressure sensor on the pipeline.

[0049] Atomized particle diameter, the average diameter of the paint particles after being atomized by the spray gun, is measured by a laser particle size analyzer at a distance from the nozzle.

[0050] Paint deposition rate, the ratio of the actual mass of paint adhering to the workpiece surface during the spraying process to the total mass of sprayed paint, is calculated by weighing.

[0051] The coating thickness was measured at 10 points evenly selected on the workpiece surface using an eddy current thickness gauge, and the average value was taken.

[0052] Adhesion, which reflects the bonding strength between the coating and the substrate, is tested using the cross-cut test.

[0053] Roughness, which reflects the microscopic unevenness of the coating surface, is measured using a surface roughness meter;

[0054] Hierarchical influence diagram construction: Determine node types, including root nodes, intermediate nodes, and leaf nodes; root nodes are quality indicators, including thickness, adhesion, and roughness; intermediate nodes are intermediate variables, including atomized particle diameter and deposition rate; leaf nodes are process parameters, including distance, velocity, viscosity, and pressure; establish causal relationships as the direction of edges, including: process parameters → intermediate variables, intermediate variables → quality indicators, and direct correlations; root nodes are located at the top layer, intermediate nodes at the middle layer, and leaf nodes at the bottom layer, with arrows indicating the direction of influence.

[0055] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, the calculated parameter influence value is the defined parameter node influence value, which is the influence of the parameter on the quality index. It is dynamically calculated based on the influence value of the child nodes, and is expressed as follows: ; It is a combination of process parameters Impact value on nodes; It is the transmission attenuation coefficient, reflecting the attenuation of influence along the path; the total influence value of the calculated parameter is the total impact of the parameter on all relevant quality indicators. , is the sum of its own influence value and the influence values ​​of all its ancestor nodes, expressed as: ; It is the set of nodes affected by the combination of process parameters; It is the total influence value of the combination of process parameters.

[0056] Example 4, see Figure 1 This embodiment is based on the above embodiment. In step S3, calculating the correlation degree of quality indicators is to define the similarity degree of process parameter combinations. By quantifying the degree of overlap in the impact paths, the more nodes that are jointly affected by the two sets of process parameters, the stronger the impact on these nodes, and the more similar the optimization directions. Considering the influence of external factors such as environment and equipment, an aging factor is added. The closer the sampling dates, the closer the sampling times of the two sets of process parameter combinations, and the higher the similarity, as expressed as: ; ;in, and It is a combination of two sets of process parameters; r is the node index; It is the attenuation coefficient; It is the sampling time difference between two sets of process parameter combinations; calculate the similarity between the process parameter combinations and the quality index set. , is represented as: Define the correlation of quality indicators , is represented as: ;in, It is the kth combination of process parameters; It affects quality indicators A set of process parameter combinations; It affects quality indicators The set of process parameter combinations, P and They are and The combination of process parameters.

[0057] By performing the above operations, this solution addresses the problems of general spraying process parameter optimization methods, which neglect the transmission and attenuation of process parameter influence, leading to distorted assessments of indirect effects, superficial analysis of quality index correlations, and consequently low reliability of parameter optimization. This solution defines parameter node influence values ​​to characterize the hierarchy and attenuation characteristics of parameter influence, quantifies the total influence of process parameters based on the total influence value, and more comprehensively reflects the true total impact of process combinations on indicators. It introduces an aging factor to define the similarity of process parameter combinations, aligning with the influence of environmental changes on spraying effects. Furthermore, it quantifies the overlap of parameter combinations based on quality index correlations, objectively reflecting the inherent correlations of quality indicators. Ultimately, this makes process parameter optimization more closely aligned with the actual characteristics of the spraying process, improving the reliability of process parameter optimization.

[0058] Example 5, see Figure 1 This embodiment is based on the above embodiment. In step S4, the reconstruction of the correlation matrix is ​​as follows: let Y be the correlation matrix between process parameter combinations and quality indicators, where Y(i,j)=1 indicates that process parameter combination i affects quality indicator j, and Y(i,j)=0 indicates that it has not been verified and needs to be completed; define the parameter neighborhood matrix. Furthermore, a dynamic attenuation term is introduced to dynamically adjust the attenuation level based on the process similarity of the spraying process parameter combinations. This strengthens the mutual influence between similar process parameter combinations and avoids excessive weakening of similar but lower-ranked neighboring regions by a fixed attenuation, expressed as: Define the quality neighborhood optimization matrix. To avoid excessive weakening of similar but lower-ranked neighbors by a fixed decay, it is represented as: ;in, is the neighborhood normalization coefficient; K is the number of neighborhoods, and np is the neighborhood index; It is the attenuation base. It is the decay index; Is with The most recent nth combination of process parameters; It corresponds to Y The entire line; Is with The most recent nth quality metric; It corresponds to Y A whole column; finally, construct the correlation matrix, represented as: ; , and These are the reconstruction of the correlation matrix. , and The element in row u and column v.

[0059] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S5, constructing the neighborhood similarity matrix is ​​to construct the parameter neighborhood similarity matrix W. p And the similarity matrix W of the quality neighborhood q The parameter neighborhood similarity matrix is ​​a combination of process parameters. Represented as eigenvectors Define reconstruction error , is represented as: ; It is a weight vector. These are weighting coefficients; to minimize reconstruction error, the weighting coefficients are corrected; specifically, this includes: adjusting the weighting coefficients... Initialize the similarity of process parameter combinations Construct the inner product matrix ,Depend on The neighborhood feature vectors are calculated, and the matrix elements are the inner products between neighborhood vectors; the target vector b is defined as... The inner product with neighboring samples; after expanding the reconstruction error into a quadratic form, for... Taking the derivative and setting it to zero, we get I is the identity matrix; It is the regularization parameter; solving the system of equations yields the optimal weight vector. Each element These are the elements of the parameter neighborhood similarity matrix; similarly, the quality neighborhood similarity matrix represents the quality index as an eigenvector, and the weight coefficients are initialized as the correlation degree of the quality index.

[0060] Example 7, see Figure 1 This embodiment is based on the above embodiment. In step S6, the membership matrix optimization is achieved by bidirectional membership propagation, outputting the optimized membership of the process parameter combination to the quality index, and screening the optimal process parameter combination; the specific operation is as follows: constructing the parameter transition matrix and the quality transition matrix; ; ;in, These are elements of the parameter transition matrix, representing the membership degree from the e-th process parameter combination to the i-th process parameter combination; These are elements of the mass transfer matrix; and It is a matrix element; This involves summing the similarity scores of all identical process parameter combinations for the e-th process parameter combination; and initializing the membership matrix. ; Yes Summing all elements for normalization; performing bidirectional iteration, parameter → mass direction: Mass → Parameter Direction: ;in, and It is the propagation membership degree after the (t+1)th iteration; These are the iteration coefficients; the final optimized membership matrix is ​​expressed as: ; It is the membership matrix of the (t+1)th iteration; P(i,e) represents the combination of process parameters S. i Optimize quality index q e The higher the membership degree, the better; when the iteration converges or the maximum number of iterations is reached, the final optimized membership degree matrix is ​​obtained; the process parameter combination with the highest P(i,e) is selected to adjust the process.

[0061] Example 8, see Figure 1 This embodiment is based on the above embodiment. In step S7, the real-time process parameter optimization is based on the optimized membership matrix output in S6. Combined with real-time production data, the process parameters are dynamically adjusted to ensure that the coating quality indicators consistently meet the standards. The offline model is continuously optimized through feedback data. Based on the process requirements, the target values ​​and allowable deviation ranges of the quality indicators are set, and the deviations from the target values ​​of the online detectable quality indicators are calculated in real time. , is represented as: ; and These are real-time quality indicators and target quality indicators. If the deviation of any quality indicator exceeds the allowable range, a parameter adjustment process is triggered. Based on the optimized membership matrix of S6 and combined with the current process parameter combination, an adjustment plan is quickly generated; otherwise, the current parameters are maintained. The adjustment plan involves matching the real-time collected process parameter combinations with existing process parameter combinations to find the three most similar process parameter combinations. 雷同1 S 雷同2 and S 雷同3 Based on the average process parameter combination S 雷同 Calculate the fine-tuning amount k is an adjustment factor; and They are S 雷同 The process parameter settings and the current process parameter settings.

[0062] By performing the above operations, this solution addresses the problem of unverified associations being wasted in general spraying process parameter optimization methods, leading to missing key information, rigid handling of the influence of similar process parameters and quality, and consequently poor optimization results. This solution reconstructs the association matrix to complete unverified associations and introduces a dynamic attenuation term. The attenuation level is dynamically adjusted based on the similarity of spraying process combinations, making the mutual influence of similar process combinations stronger and avoiding information loss. Based on the dynamic adjustment of similarity influence, the association accuracy is improved, and subjective errors are reduced by optimizing the neighborhood similarity weights, thereby improving the optimization effect of process parameters.

[0063] Example 9, see Figure 2 This embodiment is based on the above embodiments. The big data-based spraying process parameter optimization system provided by the present invention includes a data acquisition module, a parameter influence value calculation module, a quality index correlation calculation module, a correlation matrix reconstruction module, a neighborhood similarity matrix construction module, a membership matrix optimization module, and a real-time process parameter optimization module.

[0064] The data acquisition module collects and normalizes historical spraying process data to construct a directed acyclic hierarchical graph.

[0065] The parameter influence value calculation module is based on a directed acyclic hierarchical graph, defines the influence value of parameter nodes, and calculates the total influence value of parameters.

[0066] The quality index correlation calculation module calculates the similarity of process parameter combinations and then defines the quality index correlation.

[0067] The correlation matrix reconstruction module completes the correlation matrix and constructs a reconstructed correlation matrix through the parameter neighborhood matrix and the quality neighborhood optimization matrix.

[0068] The neighborhood similarity matrix construction module obtains the similarity matrix by minimizing the reconstruction error correction weight coefficient;

[0069] The membership matrix optimization module obtains the optimized membership matrix of process parameter combinations on quality indicators through bidirectional membership propagation and iterative optimization.

[0070] The real-time process parameter optimization module optimizes process parameters in real time based on the optimized membership matrix and real-time data.

[0071] It should be noted that, in this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0072] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0073] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for optimizing spraying process parameters based on big data, characterized in that: The method includes the following steps: Step S1: Data acquisition, collect historical spraying process data and normalize it to construct a directed acyclic hierarchical graph; Step S2: Calculate the parameter influence value. Based on the directed acyclic hierarchical graph, define the parameter node influence value and calculate the total parameter influence value. Step S3: Calculate the correlation degree of quality indicators, calculate the similarity degree of process parameter combinations, and then define the correlation degree of quality indicators; Step S4: Reconstruct the correlation matrix, complete the correlation matrix, and construct the reconstructed correlation matrix through the parameter neighborhood matrix and the quality neighborhood optimization matrix; Step S5: Construct a neighborhood similarity matrix, and obtain the similarity matrix by minimizing the reconstruction error and correcting the weight coefficients; Step S6: Membership matrix optimization. Through bidirectional membership propagation and iterative optimization, the optimized membership matrix of process parameter combinations on quality indicators is obtained. Step S7: Real-time process parameter optimization. Based on the optimized membership matrix and real-time data, real-time process parameter optimization is achieved.

2. The method for optimizing spraying process parameters based on big data according to claim 1, characterized in that: In step S2, the calculated parameter influence value is the defined parameter node influence value, which is the influence of the parameter on the quality index, and is dynamically calculated based on the influence value of the child nodes; the calculated parameter total influence value is the total influence of the parameter on all relevant quality indicators.

3. The method for optimizing spraying process parameters based on big data according to claim 2, characterized in that: In step S3, the calculation of the correlation degree of quality indicators is defined by the similarity of process parameter combinations, which is quantified by the degree of overlap of the influence paths, and a time factor is added. Calculate the similarity between the combination of process parameters and the set of quality indicators, and define the correlation degree of the quality indicators.

4. The method for optimizing spraying process parameters based on big data according to claim 3, characterized in that: In step S4, the reconstruction of the correlation matrix involves setting up a correlation matrix between the combination of process parameters and quality indicators, defining a parameter neighborhood matrix, and introducing a dynamic attenuation term to dynamically adjust the attenuation degree based on the process similarity of the spraying process parameter combination; defining a quality neighborhood optimization matrix, and finally reconstructing the correlation matrix.

5. The method for optimizing spraying process parameters based on big data according to claim 4, characterized in that: In step S5, constructing the neighborhood similarity matrix involves constructing a parameter neighborhood similarity matrix and a quality neighborhood similarity matrix.

6. The method for optimizing spraying process parameters based on big data according to claim 5, characterized in that: In step S6, the membership matrix optimization involves constructing a parameter transition matrix and a quality transition matrix; initializing the membership matrix; and outputting the optimized membership matrix of the process parameter combination on the quality index through bidirectional membership propagation to select the optimal process parameter combination.

7. The method for optimizing spraying process parameters based on big data according to claim 6, characterized in that: In step S7, the real-time process parameter optimization is based on optimizing the membership matrix and dynamically adjusting the process parameters in conjunction with real-time production data.

8. A big data-based spraying process parameter optimization system, used to implement the big data-based spraying process parameter optimization method as described in any one of claims 1-7, characterized in that: It includes a data acquisition module, a parameter influence value calculation module, a quality index correlation calculation module, a correlation matrix reconstruction module, a neighborhood similarity matrix construction module, a membership matrix optimization module, and a real-time process parameter optimization module; The data acquisition module collects and normalizes historical spraying process data to construct a directed acyclic hierarchical graph. The parameter influence value calculation module is based on a directed acyclic hierarchical graph, defines the influence value of parameter nodes, and calculates the total influence value of parameters. The quality index correlation calculation module calculates the similarity of process parameter combinations and then defines the quality index correlation. The correlation matrix reconstruction module completes the correlation matrix and constructs a reconstructed correlation matrix through the parameter neighborhood matrix and the quality neighborhood optimization matrix. The neighborhood similarity matrix construction module obtains the similarity matrix by minimizing the reconstruction error correction weight coefficient; The membership matrix optimization module obtains the optimized membership matrix of process parameter combinations on quality indicators through bidirectional membership propagation and iterative optimization. The real-time process parameter optimization module optimizes process parameters in real time based on the optimized membership matrix and real-time data.

Images