Laser cladding coating parameter and performance correlation analysis method based on artificial intelligence

Abstract

The invention discloses a laser cladding coating parameter and performance correlation analysis method based on artificial intelligence, and relates to the technical field of computer correlation analysis systems. Comprising the steps of constructing an artificial intelligence mapping relation model, taking parameter data features as input, taking performance data features as output, obtaining a plurality of predicted mapping relations, marking a data set where parameter data and performance data are interacted according to the characteristics of the mapping relations, and obtaining an artificial intelligence mapping relation model; the mapping relation is positioned with the parameter data and the performance data according to the data set. According to the method, the updating and adjusting rule based on the time series is constructed, the performance data of each time series is associated with the performance data threshold, the adjusting data is screened out to update and adjust the data set variable, and then the characteristics of the mapping relation are updated according to the change of the data set variable in the associated packet. The method can dynamically adapt to continuous changes of the relationship between parameters and performance in the laser cladding process.

Description

Technical Field

[0001] The present invention relates to the technical field of computer correlation analysis systems, and specifically to a method for correlative analysis of laser cladding coating parameters and performance based on artificial intelligence. Background Technique

[0002] In the process of the laser cladding technology, there is a close and complex interaction relationship between the coating parameters and performance. This complex interaction relationship makes it difficult to predict the final performance of the coating. Therefore, there is an urgent need to construct a method for correlative analysis of laser cladding coating parameters and performance based on artificial intelligence technology.

[0003] After retrieval, the Chinese invention patent with the publication number "CN114580637 A" discloses "a method for correlating Internet of Things devices and parameters based on a knowledge graph". This application uses the knowledge graph to correlate Internet of Things devices and their associated items, spaces, and types to construct a network-shaped knowledge base composed of nodes and edges; it can be used for building Internet of Things parameter management, control logic, fault diagnosis, energy efficiency analysis, energy consumption analysis, and life analysis. It improves the correlation between various items. When a project change occurs, it directly matches the associated Internet of Things devices, improves efficiency, realizes efficient project implementation, reduces duplicate configuration work, and can lower the threshold of related work.

[0004] In addition, the Chinese invention patent with the publication number "CN115481757 A" discloses "a federated data correlation analysis method, a data recommendation method, and a device". This application determines a first frequent item set in the sample data and the support degree corresponding to the first frequent item set; then receives a second frequent item set sent by a second participant, and the second frequent item set is obtained by the second participant according to the sample data of the second participant; and determines the confidence degree of the second frequent item set when the first frequent item set appears, so as to determine the association rule according to the confidence degree. At this time, the first participant only receives the frequent item sets sent by other participants without exchanging the corresponding support degrees, realizing federated correlation analysis based on privacy protection among multiple participants, and being able to obtain highly practical and accurate association rules under the conditions of diverse data composition and complex association relationships; and it is not necessary to obtain all the source data to break the state of data fragmentation among participants, which not only ensures data privacy and security but also reduces the requirements of the association analysis processing process for hardware configuration, thereby reducing the analysis cost and improving the analysis efficiency.

[0005] However, in the actual application process, due to the uncertainty of the laser cladding process, it is required to continuously associate parameter data with performance data during the cladding completion stage to ensure the stability of the cladding process and the quality of the final product. However, this process requires the computer system to continuously allocate computing resources to provide support, which to a certain extent limits the processing ability of the computer system for other tasks. Nevertheless, in actual operation, the above-mentioned disclosed existing patents and similar technologies usually only analyze static data sets and cannot adapt to the dynamic changes in the relationship between parameters and performance during the laser cladding process. Therefore, there is an urgent need to develop a more intelligent and efficient method that can real-time and dynamically associate laser cladding coating parameters with performance and minimize the impact on the processing ability of other tasks of the computer system. Summary of the Invention

[0006] The purpose of the present invention is to provide an artificial intelligence-based method for analyzing the correlation between laser cladding coating parameters and performance to solve the problems raised in the above background technology.

[0007] To achieve the above purpose, the present invention provides the following technical solution: An artificial intelligence-based method for analyzing the correlation between laser cladding coating parameters and performance, including:

[0008] Construct an artificial intelligence mapping relationship model, using parameter data features as input and performance data features as output to obtain multiple predicted mapping relationships;

[0009] According to the characteristics of the mapping relationship, mark the data set where parameter data and performance data interact, and the mapping relationship is located according to the data set with respect to parameter data and performance data respectively;

[0010] According to the type of data set, assign an association package outside the corresponding parameter data or performance data, and the association package stores the data set variables between parameter data and performance data;

[0011] Construct an update and adjustment rule based on time series, associate the performance data of each time series with the performance data threshold, and screen out the adjustment data, and then update and adjust the data set variables;

[0012] Update the characteristics of the mapping relationship based on the change of the data set variables in the association package.

[0013] As a further preference of this technical solution, the construction method of the artificial intelligence mapping relationship model includes:

[0014] According to the instruction command, obtain the performance data feature interval;

[0015] Use the positioning method to extract parameter data features and several corresponding performance data features;

[0016] Taking the parameter data feature as the starting node and combining several performance data features as the ending nodes, construct the corresponding linear connection relationship;

[0017] Obtain the priority acting on the performance data feature according to the instruction command;

[0018] Based on the priority, screen the corresponding mapping relationship in the linear connection relationship.

[0019] As a further optimization of this technical solution, the positioning method includes:

[0020] Extract the character attributes of the parameter data feature;

[0021] According to the character attributes, classify the parameter data feature and clarify the data type corresponding to the character attributes;

[0022] According to the data type, retrieve the known performance data features matching the parameter data feature in the database;

[0023] Use the similarity calculation formula to evaluate the similarity between the parameter data feature and the known performance data feature;

[0024] Select the known performance data feature with the highest similarity as the positioning result and establish a corresponding relationship with the parameter data feature.

[0025] As a further optimization of this technical solution, the similarity calculation formula includes:

[0026]

[0027] Where R represents the Pearson correlation coefficient between the parameter data feature and the known performance data feature, which is used to measure the linear similarity between the parameter data feature and the known performance data feature. The closer the Pearson correlation coefficient is to 1, the stronger the positive correlation between the two features; the closer it is to -1, the stronger the negative correlation; and being close to 0 indicates no obvious linear relationship;

[0028] x i and y i respectively represent the observed values of the parameter data feature and the known performance data feature on the i-th attribute;

[0029] and respectively represent the means of the parameter data feature and the known performance data feature on all attributes;

[0030] n represents the number of observed values, that is, the number of attributes in the parameter data feature and the known performance data feature;

[0031] The sum of products representing the deviation between the parameter data characteristics and the known performance data characteristics is used to measure the synchronization of the deviations of the parameter data characteristics and the known performance data characteristics in each attribute, that is, when one characteristic is above the average level in a certain attribute, whether the other characteristic is also above the average level in the same attribute;

[0032] The product of the standard deviations representing the deviations of the parameter data characteristics and the known performance data characteristics is used for the amplitude of the deviations of the parameter data characteristics and the known performance data characteristics in each attribute, that is, the degree of dispersion of the characteristic values.

[0033] As a further preference of this technical solution, the marking method of the data set includes:

[0034] Obtain the total data transmission volume between the parameter data and the performance data within the mapping relationship based on the time series;

[0035] Obtain the primary data, where the primary data is the average value of the total data transmission volume based on the time series;

[0036] Obtain the secondary data, where the secondary data is any one of the total data transmission volume;

[0037] Compare the primary data and the secondary data to obtain the data difference;

[0038] Based on the positive or negative of the data difference, determine the data transmission trend under the current time series;

[0039] Based on the transmission trend, obtain the data in the same direction as the marked data set.

[0040] As a further preference of this technical solution, the construction method of the update adjustment rule includes:

[0041] Set the performance index threshold according to the performance data characteristic interval;

[0042] Based on the time series, conduct a comparative analysis of the performance data and the performance index threshold;

[0043] If the performance data exceeds the established threshold interval, it is marked as primary adjustment data;

[0044] According to the characteristics of the primary adjustment data, retrieve and locate the data items of the data set variables in the associated package;

[0045] Interact the adjustment data and the data values in the data items;

[0046] Determine the adjustment data based on the interaction result.

[0047] As a further preference of this technical solution, the update method of the mapping relationship characteristics includes:

[0048] Monitor the changes in dataset variables in the associated package;

[0049] Evaluate the impact of changes in dataset variables on the overall structure and distribution of the dataset;

[0050] Based on the evaluation results, analyze the impact of changes on each data item in the associated data packet;

[0051] According to the degree of influence, obtain the change trend of the association strength between dataset variables;

[0052] According to the change trend of the association strength, update the characteristics of the mapping relationship.

[0053] As a further optimization of this technical solution, the method for updating the characteristics of the mapping relationship includes:

[0054] Obtain the positions of the data items corresponding to the dataset variables in the dataset;

[0055] According to the positions of the data items, obtain the spatial proximity between each data item in the dataset;

[0056] Based on the spatial proximity, analyze the degree of mutual influence between each data item in the dataset;

[0057] Evaluate the impact of the degree of influence on the association strength between dataset variables;

[0058] Combined with the changes in dataset variables, update the values between data items.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] This method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence stores dataset variables between parameter data and performance data through an associated package, and assigns an associated package outside the corresponding data according to the type of dataset, achieving a clear sorting of data relationships;

[0061] At the same time, by constructing an update and adjustment rule based on time series and associating the performance data of each time series with the performance data threshold, screening out adjustment data to update and adjust dataset variables, and then updating the characteristics of the mapping relationship according to the changes in dataset variables in the associated package, this method can dynamically adapt to the continuous changes in the relationship between parameters and performance during the laser cladding process. Compared with the prior art, it realizes that there is no need for a computer system to continuously consume a large amount of scattered computing resources to maintain the correlation between parameter data and performance data, so that the computer system can still effectively handle other tasks while processing this analysis of parameter data and performance data. Brief Description of the Drawings

[0062] Figure 1 It is a step diagram of the method disclosed in the present invention;

[0063] Figure 2 This is the auxiliary explanatory diagram for step S502 of the present invention;

[0064] Figure 3 This is the construction step diagram of the artificial intelligence mapping relationship model in step S100 of the present invention;

[0065] Figure 4 This is the step diagram of the method for marking the data set in step S200 of the present invention. Detailed implementation manners

[0066] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the 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.

[0067] Before understanding the technical solutions proposed in the present invention application, it should be clear that in the prior art, the laser power parameter, scanning speed parameter, powder feeding rate parameter, and spot size parameter all affect the final performance of the coating to varying degrees. These parameters are common clad coating parameters in the prior art, that is, the parameter data characteristics proposed in the present invention. In addition, the coating performance in the prior art is mainly reflected in hardness, wear resistance, and corrosion resistance. Specifically in the present invention, it is the performance data characteristics. In addition, it should be added that since the correlation relationship between the parameter data characteristics and the performance data characteristics is not fixed, but affected by various factors, when analyzing the correlation between the parameter data and the performance data, a large amount of computing resources and time are required, and it is difficult to reflect the dynamic change relationship between the parameters and the performance in real time. Based on this scenario, the present invention proposes an artificial intelligence-based method for analyzing the correlation between laser clad coating parameters and performance.

[0068] Refer to Figure 1 It can be seen that an artificial intelligence-based method for analyzing the correlation between laser clad coating parameters and performance proposed by the present invention includes steps S100 - S500.

[0069] Step S100: Construct an artificial intelligence mapping relationship model.

[0070] It should be clear that in step S100, the artificial intelligence mapping relationship model takes the parameter data characteristics as the input and the performance data characteristics as the output to obtain multiple predicted mapping relationships.

[0071] Specifically, refer to Figure 3 It can be seen that the construction method of the artificial intelligence mapping relationship model in step S100 includes: steps S101 - S105.

[0072] Step S101: Obtain the characteristic interval of performance data according to the instruction command.

[0073] It should be clear that in the present invention, step S101 is used to determine the range of performance data characteristics, that is, the characteristic interval of performance data. The characteristic interval of performance data is preset according to the requirements of the actual application scenario. For example, if the hardness, wear resistance or corrosion resistance of the coating needs to reach a value determined based on the instruction command, the corresponding performance data characteristics should fall within the preset interval. This step provides a performance target orientation for the subsequent artificial intelligence mapping relationship model, enabling the model to specifically learn the correlation between parameters and performance during the training process.

[0074] Specifically in the application scenario of the present invention, when the instruction command is specifically to make the Vickers hardness (HV) of the coating be in the range of HV500 - HV800, the performance data characteristic region is HV500 - HV800.

[0075] Step S102: Adopt a positioning method to extract parameter data characteristics and several corresponding performance data characteristics.

[0076] It should be clear that in the present invention, step S102 is used to extract the parameter data characteristics and their corresponding coating performance data characteristics during the laser cladding process. It should be noted that the parameter data characteristics include: laser power, scanning speed, powder feeding rate, and spot size, while the performance data characteristics correspond to the performance targets determined in step S101, such as the hardness, wear resistance, and corrosion resistance of the coating. In addition, the present invention adopts a positioning method to screen out the parameter data characteristics and performance data characteristics closely related to the coating performance from a large amount of experimental data, providing accurate data support for the subsequent artificial intelligence mapping relationship model.

[0077] It should be supplemented that the operation of the positioning method in step S102 includes: steps S102.1 - S102.5.

[0078] Step S102.1: Extract the character attributes of the parameter data characteristics.

[0079] It should be clear that in the present invention, step S102.1 is used to extract the information of the parameter data characteristics during the laser cladding process. These information are presented in the form of character attributes, including the numerical range of laser power, the specific value of scanning speed, the unit of powder feeding rate, and the size description of the spot size. It should be supplemented that the extraction of character attributes helps to convert unstructured experimental data into structured information, facilitating subsequent data processing and analysis.

[0080] Step S102.2: Classify the parameter data features according to the character attributes, and clarify the data types corresponding to the character attributes.

[0081] It should be clear that in the present invention, Step S102.2 is used to classify the character attributes of the extracted parameter data features in detail, and establish the data types corresponding to various character attributes. For example, the numerical range of the laser power is classified as continuous numerical data, the specific value of the scanning speed is regarded as discrete numerical data, the unit of the powder feeding rate belongs to text data, and the size description of the spot size is classified as interval numerical data. It should be added that such classification and clarification of data types lay a solid foundation for subsequent data cleaning, standardization, and the establishment of an artificial intelligence mapping relationship model.

[0082] Step S102.3: Retrieve the known performance data features that match the parameter data features in the database according to the data types.

[0083] It should be clear that in the present invention, Step S102.3 is used to quickly retrieve and match the known performance data features that are consistent with the current parameter data features in the database based on network interconnection according to the already clarified data types. It should be added that this process aims to provide a basis for performance prediction for new laser cladding coating experiments by using the existing experimental data and performance manifestations. It should be noted that to improve the efficiency of the present invention during actual operation, the known performance data features in the database are also classified and stored according to data types.

[0084] Step S102.4: Use a similarity calculation formula to evaluate the similarity between the parameter data features and the known performance data features.

[0085] As a preferred implementation method, Step S102.4 discloses a method for matching parameters and performance data based on similarity calculation.

[0086] It should be clear that in the present invention, Step S102.4 aims to use a similarity calculation formula to evaluate the similarity between the parameter data features and the known performance data features. The core function of this Step S102.4 is to analyze the similarity between the current parameter data features and the known performance data features retrieved from the database through the similarity calculation formula. The purpose of this step is to accurately identify the historical performance data features with a high degree of matching with the current parameter data features based on the calculated similarity results, so as to provide a more solid and accurate reference basis for subsequent performance prediction work.

[0087] In addition, it should be added to Step S102.4 that in the present invention, the similarity calculation formula is specifically:

[0088] The similarity calculation formula proposed by the present invention is obtained through in-depth research and integrated innovation on a variety of common similarity measurement algorithms. Specifically, the similarity calculation formula proposed by the present invention combines the characteristics of the Manhattan distance algorithm in considering the absolute differences in each dimension of data features, and the advantages of the Pearson correlation coefficient algorithm in evaluating the linear relationship between data features.

[0089] In addition, it should be supplemented that the similarity calculation formula of the present invention particularly emphasizes comprehensively analyzing the correlation between parameter data features and known performance data features from multiple dimensions. In the actual calculation process, the mutual relationship of each data feature component under different value-taking situations will be carefully considered, so as to obtain the similarity quantization value between the two. Based on the similarity quantization value, the matching degree between the current parameter data feature and a certain known performance data feature in the database is judged, and then more accurate data support with important reference value is provided for the subsequent performance prediction work, which improves the efficiency and accuracy of the method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence to a certain extent.

[0090] In addition, the similarity calculation formula proposed by the present invention still needs to be further improved. Specifically, in the similarity calculation formula, R represents the Pearson correlation coefficient between parameter data features and known performance data features, which is used to evaluate the linear similarity between the two. When the absolute value of the Pearson correlation coefficient is closer to 1, it indicates that the positive correlation between the two features is stronger, and when the absolute value is closer to -1, it indicates that the negative correlation is stronger; if it is close to 0, it means that there is no obvious linear relationship between the two.

[0091] The symbol x i and y i respectively represent the observed values of parameter data features and known performance data features on the i-th attribute;

[0092] The symbol and respectively represent the means of parameter data features and known performance data features on all attributes;

[0093] The symbol n represents the number of observed values, that is, the total number of attributes in parameter data features and known performance data features;

[0094] The symbol represents the product sum of the deviations between parameter data features and known performance data features, which is used to reflect the synchronism of the deviations of parameter data features and known performance data features on each attribute, that is, when one feature is above the average level on a certain attribute, whether the other feature is also above the average level on the same attribute;

[0095] The symbol The product of the standard deviation representing the deviation between the parameter data characteristics and the known performance data characteristics is used to measure the magnitude of the deviation between the parameter data characteristics and the known performance data characteristics in each attribute, that is, the degree of dispersion of the characteristic values.

[0096] In addition, what needs to be further supplemented for the above content is that when the similarity calculation formula disclosed in the present invention is actually running, it is assumed that the observed values of the parameter data characteristics and the parameter data characteristics in the third attribute are 650 (parameter data characteristics) and 700 (known performance data characteristics), the means of the parameter data characteristics and the known performance data characteristics in all attributes are 600 and 680, and the number of observed values, that is, the total number of attributes, is 10. At this time, substituting the values into the similarity calculation formula proposed by the present invention, we get

[0097] Therefore, it can be known that the similarity between the current parameter data characteristics and the known performance data characteristics is 1. This result indicates that there is a strong positive correlation between the two, that is, the coating performance under the current parameter data characteristics is very likely to be close to the performance level represented by the known performance data characteristics.

[0098] Step S102.5: Select the known performance data characteristics with the highest similarity as the positioning result and establish a corresponding relationship with the parameter data characteristics.

[0099] It should be clear that step S102.5 proposed by the present invention is used to accurately locate the coating performance level that the current parameter data characteristics can reach. By selecting the known performance data characteristics with the highest similarity to the current parameter data characteristics and establishing a corresponding relationship, this is crucial for subsequent performance prediction and optimization.

[0100] Step S103: Use the parameter data characteristics as the starting node and combine several performance data characteristics as the ending nodes to construct the corresponding linear connection relationship.

[0101] As a preferred implementation manner, steps S104 and S105 disclose a screening method for the mapping relationship based on priority.

[0102] It should be noted that in step S104, the present invention proposes a screening method for the mapping relationship based on priority. The priority of the performance data characteristics is determined through instruction commands. For example, when the hardness index is HRC50, the hardness is given the highest priority. In step S105, the mapping relationship that best fits the key performance data characteristics is screened out from the linear connection relationship based on the priority.

[0103] Specifically, step S104: Obtain the priority acting on the performance data characteristics according to the instruction command.

[0104] It should be clear that step S104 in the present invention is used to determine which performance data features should be given priority in the performance prediction and optimization process. By sorting the priorities of performance data features, it can be ensured that under limited resources, key performances are preferentially optimized and adjusted, thereby improving the prediction accuracy and optimization efficiency of the overall coating performance.

[0105] It should be supplemented that when the content of the instruction command specifically requires a hardness index of HRC50, the performance data feature of hardness will be given the highest priority and become the primary consideration factor in the performance prediction and optimization process. If the instruction command requires the wear resistance to reach a certain value, then the wear resistance data feature will take the priority position. It should be noted that the setting of step S104 ensures that the analysis method can be flexibly adjusted according to actual needs to achieve the control of key performances.

[0106] Specifically, step S105: Screen the corresponding mapping relationship based on the priority in the linear connection relationship.

[0107] It should be clear that in the present invention, step S105 is used to accurately screen out the corresponding mapping relationship from the already constructed linear connection relationship based on the determined priority. It should be clear that step S105 in the present invention is mainly to be able to select the most suitable mapping relationship from numerous linear connection relationships according to the priority status given to the performance data features. Doing so can make the subsequent analysis of the relationship between laser cladding coating parameters and performance more targeted, closely focus on key performance data features, and then more accurately and efficiently discover the internal relationship between parameters and performance, avoiding being interfered by the mapping relationships corresponding to non-key factors, making the entire analysis process smoother and more effective. For example, if the wear resistance data feature is determined to be a priority consideration factor, then in step S105, the mapping relationship related to wear resistance will be focused on screening for subsequent analysis.

[0108] Step S200: Mark the data set where parameter data and performance data interact according to the characteristics of the mapping relationship.

[0109] It should be clear that in the present invention, the mapping relationship is located according to the data set with parameter data and performance data respectively. In addition, it should be noted that referring to Figure 4 It can be seen that the marking method for the data set in step S200 includes steps S201 - S206.

[0110] Step S201: Obtain the total data transmission volume between parameter data and performance data in the mapping relationship based on the time series.

[0111] It should be clear that in the present invention, step S201 is used to evaluate the data transmission scale during the interaction between parameter data and performance data. By performing time series analysis, the total amount of data exchanged between parameter data and performance data within the mapping relationship during a time period is calculated. This not only helps to understand the frequency of data interaction but also provides a basis for further data processing and analysis.

[0112] Step S202: Obtain primary data, where the primary data is the average value of the total data transmission volume based on the time series.

[0113] It should be clear that in the present invention, step S202 aims to grasp the average level of the interaction between parameter data and performance data as a whole, providing a reference value for subsequent data processing and analysis.

[0114] Step S203: Obtain secondary data, where the secondary data is any one of the total data transmission volume.

[0115] Step S204: Compare the primary data and the secondary data to obtain the data difference.

[0116] It should be clear that in the present invention, step S204 is used to evaluate the abnormal fluctuations during the data interaction process. By comparing the primary data and the secondary data and calculating the difference between them, it can help identify whether there are abnormal or sudden data transmissions in the data interaction. The information that may be contained in these abnormal fluctuations during operation includes equipment failures, parameter adjustments, or sudden changes in experimental conditions, which has important reference value for subsequent performance prediction and optimization.

[0117] Step S205: Based on the positive or negative value of the data difference, determine the data transmission trend under the current time series.

[0118] It should be clear that in the present invention, step S205 is used to judge the overall trend of data interaction, that is, whether the data is increasing or decreasing. When the data difference is positive, it indicates that the data transmission volume has increased compared to the average level, meaning that the interaction between parameter data and performance data is more frequent. When the data difference is negative, it indicates that the data transmission volume is below the average level, indicating that there are limiting factors affecting the data interaction. By further analyzing the potential changes during the experiment through step S205, it provides a deeper insight for subsequent performance prediction and optimization.

[0119] Step S206: Based on the transmission trend, obtain the data in the same direction as the marked data set.

[0120] It should be clear that in step S206 proposed in the present invention, the determination of the same direction is specifically as follows: when the data difference is positive, it is determined that the parameter data is transmitted in the direction of the performance data with a difference, and when the data difference is negative, it is determined that the performance data is transmitted in the direction of the parameter data with a difference. The purpose of this step is to screen out those data with similar transmission trends to form a labeled data set for subsequent analysis and processing. Through in-depth analysis of these data, the correlation between parameters and performance can be further understood, as well as how they affect each other over time, providing more accurate guidance for optimizing the performance of the laser cladding coating.

[0121] Step S300: Assign an association package outside the corresponding parameter data or performance data according to the type of data set.

[0122] It should be clear that in step S300, the association package stores the data set variables between the parameter data and the performance data. In addition, it should be supplemented that in the present invention, the association package is used as a link to establish a mapping relationship.

[0123] As a preferred implementation manner, in step S400, a dynamic data update method based on time series is disclosed. By setting a performance index threshold, the performance data in the time series is compared with the threshold to screen out the adjustment data and update the data set variables.

[0124] Specifically, step S400: Construct an update and adjustment rule based on time series, associate the performance data of each time series with the performance data threshold, screen out the adjustment data, and then update and adjust the data set variables.

[0125] It should be clear that in the present invention, step S400 is used to dynamically adjust the performance data in the mapping relationship, thereby adjusting the mapping relationship to ensure the timeliness and accuracy of the analysis method. Specifically, step S400 first sets a threshold range for the performance data. The threshold range is determined manually based on historical data and practical experience and can reflect the key indicators of the coating performance. Subsequently, the performance data in the time series is compared with the set threshold to identify the performance data that exceeds the normal range or requires special attention. These screened-out data are called adjustment data. Once the adjustment data is identified, the corresponding data set variables will be updated to reflect the latest performance state. It should be supplemented that the setting of step S400 not only helps to capture the changes in the coating performance in a timely manner, but also provides more accurate data support for subsequent performance prediction and optimization. By continuously updating and adjusting the data set variables, the present invention can ensure that the mapping relationship always keeps pace with the actual coating performance, avoiding the continuous allocation of resources by the computer system.

[0126] In addition, it should be further noted that the method for constructing the update and adjustment rules in step S400 includes steps S401 - S406.

[0127] Step S401: Set the performance index threshold based on the performance data characteristic interval.

[0128] It should be clear that step S401 is the basis for constructing the adjustment rules. By analyzing the performance data and determining the normal fluctuation range and abnormal critical values of different performance indexes through historical data and practical experience, the setting of these thresholds not only depends on the statistical analysis of historical data but also combines the knowledge and experience of domain experts, ensuring the rationality and practicality of the thresholds.

[0129] Step S402: Based on the time series, conduct a comparative analysis of the performance data and the performance index threshold.

[0130] It should be noted that the comparison and analysis between the performance data and the performance index threshold in step S402 are specifically achieved by comparing the difference between the data volume of the performance data and the performance index threshold.

[0131] Step S403: If the performance data exceeds the established threshold range, mark it as primary adjustment data.

[0132] Step S404: According to the characteristics of the primary adjustment data, retrieve and locate the data items of the dataset variables in the associated package.

[0133] It should be clearly pointed out that in the present invention, the method for locating the data items of the dataset variables in the associated package in step S404 is to identify and extract the corresponding items of the difference between the data volume in the performance data and the performance index threshold and determine the data items of the dataset variables in the associated package and the data items with matching characters based on the character characteristics of this corresponding item.

[0134] Step S405: Interact the adjustment data and the data values in the data items.

[0135] Step S406: Determine the adjustment data based on the interaction result.

[0136] It should be clear that the content represented by step S405 and step S406 is to determine the adjustment data by interacting the adjustment data and the data values in the data items.

[0137] As a preferred implementation method, step S500 presents a method for updating the dataset variables based on the associated package. By monitoring the changes in the dataset variables in the associated package, evaluating their impact on the overall structure and distribution of the dataset, and updating the characteristics of the mapping relationship.

[0138] Specifically, step S500: Update the characteristics of the mapping relationship based on the changes in the dataset variables in the associated package.

[0139] It should be clear that the method for updating the mapping relationship characteristics in step S500 includes steps S501 - S505.

[0140] Step S501: Monitor the change status of the dataset variables in the associated package.

[0141] It should be clear that in the present invention, step S501 is used to monitor in real time the changes in the dataset variables in the associated package, including the increase or decrease in the data volume, the change in the data value, and the addition or deletion of data items, to ensure a comprehensive control of the dataset variables in the associated package.

[0142] Step S502: Evaluate the impact of the change in the dataset variables on the overall structure and distribution of the dataset.

[0143] As a preferred implementation manner, in this implementation manner, in the present invention, step S500 is used to analyze the degree of change in the dataset variables relative to the original overall structure of the dataset, and how this change affects the distribution characteristics of the dataset.

[0144] Specifically, in the context set by Figure 2 assuming that the overall structure of the original dataset is composed of four different values, each of which corresponds to specific laser cladding coating parameters or performance indicators and other related elements, when there is a change in the dataset variables in the associated package, just like one of the values changes, whether it increases, decreases, or a data item is missing, it will inevitably cause a corresponding change in the original overall structure of the dataset.

[0145] Step S503: Based on the evaluation results, analyze the impact of the change on each data item in the associated data packet.

[0146] It should be clear that in the present invention, step S503 is used to analyze how the change in the dataset variables specifically affects each data item in the associated data packet, which includes whether the association relationship between data items changes, whether the value of the data item generates a new trend due to the change, and whether these changes may trigger an adjustment of the internal logic or structure of the associated data packet. Combining Figure 2 with the given scenario, it can be known that the degree of impact is one - quarter.

[0147] Step S504: According to the degree of impact, obtain the change trend of the association strength between dataset variables.

[0148] It should be clear that in the present invention, step S504 is used to, according to the change in the association strength between dataset variables, combining Figure 2 with the given scenario, it can be known that the changed dataset volume has changed by one - quarter relative to the original dataset volume, and at this time, the change trend can be set as a weak change trend.

[0149] In addition, during actual use, referring to the Figure 2 scenario, the number set of changes has two more data items than the original data set. At this time, it is determined as a strong change trend. In addition, Figure 2 It should be explained that Figure 2 in the data set items within, Y, M, F, R, B, and X respectively represent specific items in the laser cladding coating parameter data characteristics or performance parameter characteristics. Therefore, the specific meanings of Y, M, F, R, B, and X are not limited.

[0150] Step S505: Update the characteristics of the mapping relationship according to the change trend of the association strength.

[0151] As a specific implementation manner of step S505, it should be clear that in the present invention, the ways in which the strong change trend and the weak change trend affect the update of the mapping relationship include:

[0152] Step S505.1: Obtain the positions of the data items corresponding to the data set variables within the data set.

[0153] It should be clear that in the present invention, step S505.1 is used to determine which data items are directly associated with the changing data set variables and find the specific positions of these data items in the data set.

[0154] Step S505.2: Obtain the spatial proximity between each data item within the data set according to the data item positions.

[0155] Step S505.3: Analyze the change in the numerical difference between each data item within the data set based on the spatial proximity.

[0156] Step S505.4: Update the values of the data items based on the magnitude of the difference change.

[0157] It should be clear that in the present invention, for the acquisition of spatial proximity involved in step S505.1 and step S505.5, the proposed similarity calculation formula is used to determine the association degree between different items in the data set, so as to evaluate the spatial proximity.

[0158] Specifically, when the association strength shows a weak change trend, the values of the data items in the data set before and after the change are substituted into the publicly disclosed similarity formula to calculate the value of R. Subsequently, the value of R is multiplied by the influence degree, and addition and subtraction operations are respectively performed on R according to the characteristics of the influence degree, namely the weak change trend and the strong change trend, to obtain the final similarity value.

[0159] It should be added that the final similarity value represents the spatial proximity between data items in the dataset. In addition, after obtaining the spatial proximity, by comparing the difference in the similar values of the spatial proximity between the variable items and other items in the dataset at different time periods, if the change range of the difference in other items is within the set threshold range, the value of the data item is updated.

[0160] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended embodiments and their equivalents.

Claims

1. A method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence, characterized in that, Including: Construct an artificial intelligence mapping relationship model, taking parameter data features as input and performance data features as output, and obtaining multiple predicted mapping relationships; According to the characteristics of the mapping relationship, mark the dataset where parameter data and performance data interact, and the mapping relationship is located with the parameter data and performance data respectively based on the dataset; According to the dataset type, assign an associated package outside the corresponding parameter data or performance data, and the associated package stores the dataset variables between the parameter data and the performance data; Construct an update and adjustment rule based on time series, associate the performance data of each time series with the performance data threshold, and filter out the adjustment data, and then update and adjust the dataset variables; Based on the changes in the dataset variables in the associated package, update the characteristics of the mapping relationship.

2. The method for analyzing the correlation between the parameters and performance of a laser cladding coating based on artificial intelligence according to claim 1, wherein: The construction method of the artificial intelligence mapping relationship model includes: According to the instruction command, obtain the performance data feature interval; Adopt a positioning method to extract parameter data features and several corresponding performance data features; Taking the parameter data feature as the starting node and combining several performance data features as the ending nodes, construct the corresponding linear connection relationship; According to the instruction command, obtain the priority acting on the performance data features; Based on the priority, filter the corresponding mapping relationships in the linear connection relationship.

3. The method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence according to claim 2, wherein: The said positioning method includes: Extract the character attributes of the parameter data features; According to the character attributes, classify the parameter data features and clarify the data types corresponding to the character attributes; According to the data type, retrieve the known performance data features matching the parameter data features in the database; Use the similarity calculation formula to evaluate the similarity between the parameter data features and the known performance data features; Select the known performance data feature with the highest similarity as the positioning result and establish a corresponding relationship with the parameter data feature.

4. A method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence according to claim 3, characterized in that: The said similarity calculation formula includes: Where R represents the Pearson correlation coefficient between the parameter data feature and the known performance data feature, which is used to measure the linear similarity between the parameter data feature and the known performance data feature. The closer the Pearson correlation coefficient is to 1, the stronger the positive correlation between the two features; the closer it is to -1, the stronger the negative correlation; and close to 0 indicates no obvious linear relationship; x i and y i respectively represent the observed values of the parameter data feature and the known performance data feature on the i-th attribute; and respectively represent the means of the parametric data characteristics and the known performance data characteristics over all attributes; n represents the number of observations, that is, the number of attributes in the parameter data feature and the known performance data feature; The sum of products representing the deviation between the parameter data feature and the known performance data feature is used to measure the synchronism of the deviation of the parameter data feature and the known performance data feature in each attribute, that is, when one feature is above the average level in a certain attribute, whether the other feature is also above the average level in the same attribute; The product of the standard deviation representing the deviation between the parameter data characteristics and the known performance data characteristics, which is used for the magnitude of the deviation between the parameter data characteristics and the known performance data characteristics in each attribute, that is, the degree of dispersion of the characteristic values.

5. The method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence according to claim 1, wherein: The marking method of the dataset includes: Based on the time series, obtain the total data transmission volume between the parameter data and the performance data in the mapping relationship; Obtain the primary data, and the primary data is the average value of the total data transmission volume based on the time series; Obtain the secondary data, and the secondary data is any one of the total data transmission volume; Compare the primary data and the secondary data to obtain the data difference; Based on the positive or negative of the data difference, judge the data transmission trend under the current time series; Based on the transmission trend, obtain the data in the same direction as the marked dataset.

6. The method for analyzing the correlation between the parameters and performance of a laser cladding coating based on artificial intelligence according to claim 2, wherein: The construction method of the update and adjustment rule includes: According to the performance data feature interval, set the performance index threshold; Based on the time series, conduct a comparative analysis of the performance data and the performance index threshold; If the performance data exceeds the established threshold interval, it is marked as primary adjustment data; Retrieve and locate the data items of the dataset variables within the associated package according to the characteristics of the primary adjustment data; Interact with the adjustment data and the data values within the data items; Determine the adjustment data based on the interaction result.

7. The method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence according to claim 1, wherein: The method for updating the mapping relationship characteristics includes: Monitor the change status of the dataset variables in the associated package; Evaluate the impact of the change in the dataset variables on the overall structure and distribution of the dataset; Based on the evaluation result, analyze the impact of the change on each data item within the associated data package; According to the degree of impact, obtain the change trend of the association strength between the dataset variables; Update the characteristics of the mapping relationship according to the change trend of the association strength.

8. A method for analyzing the correlation between laser cladding coating parameters and performance based on artificial intelligence according to claim 1, characterized in that: The method for updating the mapping relationship characteristics includes: Obtain the positions of the data items corresponding to the dataset variables within the dataset; According to the data item positions, obtain the spatial proximity between each data item within the dataset; Based on the spatial proximity, analyze the change in the numerical difference between each data item within the dataset; Update the numerical values of the data items based on the amplitude of the difference change.

Images