Waviness measurement accuracy correction method and system based on artificial intelligence

Abstract

The present application relates to the field of artificial intelligence engineering technology, and in particular to a method and system for correcting the accuracy of corrugation measurement based on artificial intelligence. The method includes obtaining the dynamic process parameters of the machine tool and the corrugation characteristic data of the workpiece when the machine tool processes the workpiece, and constructing a causal structure map; collecting environmental impact data and the surface corrugation data of the workpiece in real time, and judging whether the surface corrugation data exceeds a preset corrugation threshold. If so, abnormal corrugation data is obtained; the abnormal corrugation data is input into the causal structure map to obtain the influence path of the main process factors and the influence path of the secondary process factors; a corrugation correction value is generated to obtain a corrugation correction value. By constructing a causal structure map, the present application provides a solid foundation for subsequent correction. By comprehensively considering the environmental impact data, the correction result is more consistent with the corrugation conditions under the actual processing environment, thereby enhancing the practicality and adaptability of the correction method.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence engineering technology, and in particular to a method and system for correcting waviness measurement accuracy based on artificial intelligence. Background Art

[0002] Surface waviness refers to the periodic or quasi-periodic geometric deviation between macroscopic shape error and microscopic roughness that occurs on a workpiece after machining. It is a key indicator of machining quality. Accurately correcting waviness measurement accuracy can more accurately reflect the actual surface quality of a workpiece. This helps machining companies strictly control product quality, ensuring that parts meet design requirements and usage standards, reducing defective products due to unqualified waviness, and improving overall product performance and durability. This is crucial for the production of high-precision mechanical parts, aerospace components, and other applications in the mechanical manufacturing industry, directly impacting product reliability and safety.

[0003] However, traditional methods usually simply eliminate or smooth the measurement data afterwards, which cannot accurately identify the main process factors and causal influence paths that lead to anomalies, making it difficult to take effective corrective measures from the source. Therefore, it is easy for the corrected waviness data to still deviate from the actual situation, affecting the accurate evaluation and control of the machining process. Summary of the Invention

[0004] Based on this, it is necessary to provide an artificial intelligence-based waviness measurement accuracy correction method and system that can improve data processing efficiency in response to the above technical problems.

[0005] The technical solutions of the present invention are as follows:

[0006] A method for correcting waviness measurement accuracy based on artificial intelligence, the method comprising:

[0007] Acquiring dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, and constructing a causal structure graph based on the dynamic process parameters and the waviness characteristic data;

[0008] Collecting environmental impact data and surface waviness data of the workpiece in real time, determining whether the surface waviness data exceeds a preset waviness threshold, and obtaining waviness abnormality data if the determination is yes;

[0009] Inputting the abnormal waviness data into a causal structure map, analyzing the abnormal waviness data based on the causal structure map, and obtaining the influencing path of the main process factors and the influencing path of the secondary process factors;

[0010] A waviness correction value is generated according to the main process factors and the influencing paths of the main process factors, and a waviness correction value is calculated according to the environmental impact data and the waviness correction value.

[0011] Specifically, the abnormal waviness data is input into a causal structure map, the abnormal waviness data is analyzed based on the causal structure map, and the influencing path of the main process factors and the influencing path of the secondary process factors are obtained; including:

[0012] Inputting the abnormal waviness data into a causal structure map, and performing reverse reasoning using a graph convolutional network method in the causal structure map to obtain a process parameter responsibility factor;

[0013] The Bayesian inference method is used to screen the process parameter responsibility factors and obtain the main process factors and secondary process factors;

[0014] By querying the causal structure map, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively.

[0015] Specifically, generating a waviness correction value according to the main process factors and the influencing paths of the main process factors includes:

[0016] According to the main process factors and the influencing paths of the main process factors, a main factor correction value is generated by an adaptive projection algorithm;

[0017] According to the secondary process factors and the influencing paths of the secondary process factors, a secondary factor correction value is calculated and generated by an adaptive projection algorithm;

[0018] A waviness correction value is calculated based on the primary cause correction value and the secondary cause correction value.

[0019] Specifically, the waviness correction value is calculated based on the environmental impact data and the waviness correction value, including:

[0020] generating an environmental impact coefficient according to the environmental impact data;

[0021] Calculating an environmental compensation value according to the degree of influence of the environmental impact coefficient on the waviness;

[0022] A waviness correction value is calculated based on the waviness correction value and the environmental compensation value.

[0023] Specifically, generating an environmental impact coefficient according to the environmental impact data includes:

[0024] Standardizing the environmental impact data and generating standardized environmental data;

[0025] An environmental impact coefficient is generated based on a preset Gaussian process regression and the standardized environmental data.

[0026] Specifically, the dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece are obtained when the machine tool processes the workpiece, and a causal structure graph is constructed according to the dynamic process parameters and the waviness characteristic data; including:

[0027] The dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece are obtained when the machine tool is processing the workpiece. The correlation between the dynamic process parameters and the waviness characteristic data is calculated using the multi-head graph attention layer method to obtain the edge weight coefficient;

[0028] A causal structure graph is constructed according to the dynamic process parameters, the waviness characteristic data and the edge weight coefficients.

[0029] Specifically, determining whether the surface waviness data exceeds a preset waviness threshold value further includes:

[0030] If the judgment is no, then the surface waviness data is determined to be normal.

[0031] Specifically, a waviness measurement accuracy correction system based on artificial intelligence, the system comprising:

[0032] A cause-effect graph construction module is used to obtain dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, and to construct a cause-effect structure graph based on the dynamic process parameters and the waviness characteristic data;

[0033] an abnormal data acquisition module, for collecting environmental impact data and surface waviness data of a workpiece in real time, and determining whether the surface waviness data exceeds a preset waviness threshold; and if so, obtaining abnormal waviness data;

[0034] an abnormal data analysis module, configured to input the abnormal waviness data into a causal structure map, analyze the abnormal waviness data based on the causal structure map, and obtain an influencing path of a primary process factor and an influencing path of a secondary process factor;

[0035] The environmental impact correction module is used to generate a waviness correction value according to the main process factors and the influence path of the main process factors, and to calculate the waviness correction value according to the environmental impact data and the waviness correction value.

[0036] Specifically, the abnormal data analysis module is also used to: input the corrugation abnormality data into the causal structure map, use the graph convolution network method to perform reverse reasoning in the causal structure map to obtain the process parameter responsibility factor; use the Bayesian reasoning method to screen the process parameter responsibility factor to obtain the main process factor and the secondary process factor; by querying the causal structure map, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively.

[0037] Specifically, the environmental impact correction module is also used to: generate a main cause correction value through an adaptive projection algorithm according to the main process factor and the influence path of the main process factor; generate a secondary cause correction value through an adaptive projection algorithm according to the secondary process factor and the influence path of the secondary process factor; and calculate a corrugation correction value based on the main cause correction value and the secondary cause correction value.

[0038] Specifically, the environmental impact correction module is also used to: generate an environmental impact coefficient based on the environmental impact data; calculate an environmental compensation value based on the degree of influence of the environmental impact coefficient on the corrugation; and calculate a corrugation correction value based on the corrugation correction value and the environmental compensation value.

[0039] Specifically, the environmental impact correction module is further used to: standardize the environmental impact data and generate standardized environmental data; and generate an environmental impact coefficient based on a preset Gaussian process regression and the standardized environmental data.

[0040] Specifically, the causal graph construction module is also used to: obtain the dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece when the machine tool processes the workpiece, calculate the correlation between the dynamic process parameters and the waviness characteristic data through the multi-head graph attention layer method, and obtain the edge weight coefficient; construct a causal structure graph based on the dynamic process parameters, the waviness characteristic data and the edge weight coefficient.

[0041] Specifically, the abnormal data acquisition module is further configured to determine whether the surface waviness data exceeds a preset waviness threshold, and then, if the determination is no, determine that the surface waviness data is normal.

[0042] Optionally, a computer device is also provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned artificial intelligence-based waviness measurement accuracy correction method when executing the computer program.

[0043] Optionally, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned artificial intelligence-based waviness measurement accuracy correction method are implemented.

[0044] The present invention relates to deep learning and deep neural network technology, and its technical effects are as follows:

[0045] The above-mentioned artificial intelligence-based corrugation measurement accuracy correction method and system successively obtains the dynamic process parameters of the machine tool and the corrugation characteristic data of the workpiece when the machine tool processes the workpiece, and constructs a causal structure map based on the dynamic process parameters and the corrugation characteristic data; collects environmental impact data and surface corrugation data of the workpiece in real time, and determines whether the surface corrugation data exceeds a preset corrugation threshold. If so, abnormal corrugation data is obtained; the abnormal corrugation data is input into the causal structure map, and the abnormal corrugation data is analyzed based on the causal structure map, and the influence path of the main process factor and the influence path of the secondary process factor are obtained; a corrugation correction value is generated based on the main process factor and the influence path of the main process factor, and the corrugation correction value is calculated based on the environmental impact data and the corrugation correction value. On the one hand, this application obtains rich dynamic process parameters and waviness feature data, uses advanced algorithms such as the multi-head graph attention layer method, accurately calculates the correlation between process parameters and waviness feature data, constructs a causal structure map, and provides a solid foundation for subsequent correction. It can effectively reduce measurement errors, improve the accuracy of waviness measurement, and make the measurement results closer to the actual surface waviness of the workpiece. On the other hand, by accurately determining the waviness anomaly, and starting from the abnormal node, through graph convolution operations and Bayesian reasoning, in-depth analysis is carried out to find the main and secondary process factors that cause the anomaly and their causal influence paths, and targeted calculation of the main and secondary cause corrections, to achieve effective correction of waviness anomalies, improve the stability and reliability of the production process, and reduce production interruptions and defective rates caused by waviness problems. In addition, by comprehensively considering environmental impact data, including humidity, cutting fluid properties and viscosity, calculating the environmental impact coefficient and combining it with the correction value, it fully compensates for the shortcomings of traditional methods that ignore environmental factors, making the correction results more consistent with the waviness conditions in the actual processing environment, and enhancing the practicality and adaptability of the correction method. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 1 is a flow chart of a method for correcting waviness measurement accuracy based on artificial intelligence in one embodiment;

[0047] Figure 2 The figure is a structural block diagram of an artificial intelligence-based waviness measurement accuracy correction system in one embodiment. DETAILED DESCRIPTION

[0048] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0049] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0050] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0051] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0052] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0053] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0054] In one embodiment, a terminal is provided, which is used to: obtain dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, and construct a causal structure map based on the dynamic process parameters and the waviness characteristic data; collect environmental impact data and surface waviness data of the workpiece in real time, and judge whether the surface waviness data exceeds a preset waviness threshold, and if so, obtain abnormal waviness data; input the abnormal waviness data into the causal structure map, analyze the abnormal waviness data based on the causal structure map, and obtain the influence path of the main process factor and the influence path of the secondary process factor; generate a corrugation correction value based on the main process factor and the influence path of the main process factor, and calculate the corrugation correction value based on the environmental impact data and the corrugation correction value.

[0055] The terminal may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices.

[0056] In one embodiment, Figure 1 As shown, a method for correcting waviness measurement accuracy based on artificial intelligence is provided, the method comprising:

[0057] Step S100: acquiring dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, and constructing a causal structure graph based on the dynamic process parameters and the waviness characteristic data;

[0058] Step S200: collecting environmental impact data and surface waviness data of the workpiece in real time, determining whether the surface waviness data exceeds a preset waviness threshold, and if so, obtaining waviness abnormality data;

[0059] Step S300: inputting the abnormal waviness data into a causal structure map, analyzing the abnormal waviness data based on the causal structure map, and obtaining the influencing paths of the main process factors and the influencing paths of the secondary process factors;

[0060] Step S400: generating a waviness correction value according to the main process factors and the influencing paths of the main process factors, and calculating a waviness correction value according to the environmental impact data and the waviness correction value.

[0061] In this embodiment, the dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece are obtained in sequence when the machine tool processes the workpiece, and a causal structure map is constructed according to the dynamic process parameters and the waviness characteristic data; environmental impact data and surface waviness data of the workpiece are collected in real time, and it is determined whether the surface waviness data exceeds a preset waviness threshold, and if so, abnormal waviness data is obtained; the abnormal waviness data is input into the causal structure map, and the abnormal waviness data is analyzed based on the causal structure map, and the influence path of the main process factor and the influence path of the secondary process factor are obtained; a corrugation correction value is generated according to the main process factor and the influence path of the main process factor, and a corrugation correction value is calculated according to the environmental impact data and the corrugation correction value. On the one hand, this application obtains rich dynamic process parameters and waviness feature data, uses advanced algorithms such as the multi-head graph attention layer method, accurately calculates the correlation between process parameters and waviness feature data, constructs a causal structure map, and provides a solid foundation for subsequent correction. It can effectively reduce measurement errors, improve the accuracy of waviness measurement, and make the measurement results closer to the actual surface waviness of the workpiece. On the other hand, by accurately determining the waviness anomaly, and starting from the abnormal node, through graph convolution operations and Bayesian reasoning, in-depth analysis is carried out to find the main and secondary process factors that cause the anomaly and their causal influence paths, and targeted calculation of the main and secondary cause corrections, to achieve effective correction of waviness anomalies, improve the stability and reliability of the production process, and reduce production interruptions and defective rates caused by waviness problems. In addition, by comprehensively considering environmental impact data, including humidity, cutting fluid properties and viscosity, calculating the environmental impact coefficient and combining it with the correction value, it fully compensates for the shortcomings of traditional methods that ignore environmental factors, making the correction results more consistent with the waviness conditions in the actual processing environment, and enhancing the practicality and adaptability of the correction method.

[0062] In one embodiment, step S100: obtaining dynamic process parameters of a machine tool and waviness characteristic data of a workpiece when the machine tool processes the workpiece, and constructing a causal structure graph based on the dynamic process parameters and the waviness characteristic data; including:

[0063] Step S110: obtaining dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, calculating the correlation between the dynamic process parameters and the waviness characteristic data by using a multi-head graph attention layer method, and obtaining an edge weight coefficient;

[0064] Step S120: constructing a causal structure graph according to the dynamic process parameters, the waviness characteristic data and the edge weight coefficients.

[0065] In this embodiment, the dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece are first obtained when the machine tool processes the workpiece, and the correlation between the dynamic process parameters and the waviness characteristic data is calculated through the multi-head graph attention layer method to obtain the edge weight coefficient; then, according to the dynamic process parameters, the waviness characteristic data and the edge weight coefficient, a causal structure map is constructed to calculate the correlation between the process parameters and the waviness characteristic data, and construct a causal structure map, which provides a solid foundation for subsequent correction, can effectively reduce measurement errors, improve the accuracy of waviness measurement, and make the measurement results closer to the actual surface waviness condition of the workpiece.

[0066] Specifically, in step S110, the dynamic process parameters include tool surface wear, dominant vibration frequency, feed rate and spindle thermal elongation.

[0067] The tool flank wear is monitored in real time using contact displacement sensors such as laser probes. A piezoelectric or capacitive accelerometer is selected and tightly mounted near the vibration source, such as the spindle bearing seat, to convert the vibration mechanical quantity into an electrical signal. The signal is acquired at high speed using a data acquisition card, and then a fast Fourier transform is performed on the signal using spectrum analysis software such as MATLAB or LabVIEW to separate the different frequency components and determine the dominant vibration frequency. The real-time feed rate of the machine tool CNC system is read, and the feed rate includes the original speed curve of the acceleration and deceleration stages, including but not limited to trapezoidal acceleration and deceleration and S-shaped acceleration and deceleration.

[0068] The thermal elongation of the spindle is measured by deploying thermocouple sensors on the machine tool spindle box and workpiece fixture, combined with a laser displacement sensor. Measuring the thermal elongation of the spindle is an existing technology, and the specific test method is as follows: The thermal elongation of the spindle is the axial length expansion caused by the temperature increase during the operation of the spindle system. Thermocouple sensors are installed on the machine tool spindle box and workpiece fixture, and can measure the temperature change of the spindle during the processing process in real time; the working principle of the thermocouple is based on the Seebeck effect, that is, the contact point of two different metals will generate an electric potential difference when the temperature changes, and the temperature value can be obtained by measuring this potential difference; the laser displacement sensor measures the displacement of the spindle by emitting a laser beam to the spindle surface and receiving the reflected light. By establishing a temperature-displacement relationship model, the thermal elongation of the spindle can be calculated according to the temperature change.

[0069] The waviness characteristic data includes wave height, wavelength and waviness distribution.

[0070] By scanning the workpiece surface with a white light interferometer, a grid point cloud (spatial resolution less than or equal to 10 ), calculate the regional waviness characteristic data. The wave height H represents the vertical distance between the highest point and the lowest point of the surface in a certain measurement area, that is, the peak-to-valley difference in the local area r, where the local area r is the measurement area of ​​the white light interferometer, and the unit of the peak-to-valley difference is .

[0071] The wavelength L is the wavelength corresponding to the dominant ripple period extracted by two-dimensional Fourier transform; the waviness distribution (x, y): the wave height value and wavelength value of each grid point are marked according to the processing time to form a time-space matrix.

[0072] In this embodiment, in order to ensure the accuracy of the waviness feature data and not affect subsequent processing, the waviness feature data needs to undergo time series alignment, missing value processing, data cleaning, and standardization.

[0073] Specifically, the timing alignment is to align the waviness feature data of each workpiece with the processing time period [t start ,t end ] Bind and extract the mean / extreme value of dynamic factors in the corresponding time period to form one-to-one corresponding data.

[0074] Missing value processing involves checking for missing values ​​in the data. During spindle vibration modal data collection, some data may be missing due to reasons such as temporary sensor failure. If the missing value ratio is small, linear interpolation, polynomial interpolation, or other methods can be used to fill the missing values ​​based on the time series characteristics of the data. If the missing value ratio is large, technical personnel will consider the importance and integrity of the data and decide whether to discard the data or re-collect it.

[0075] Data cleaning is to detect abnormal values ​​in various types of collected data. For example, in the tool wear data, if the wear increases sharply in a short period of time and exceeds the reasonable range, it may be caused by sensor failure or other abnormal factors and needs to be marked and processed. For such abnormal values, technicians in this field can use the following methods: Statistical methods such as are used to determine the value. For example, if a data point deviates from the mean by more than three standard deviations, it is considered an abnormal value. For abnormal values, those skilled in the art can choose to delete them or replace them with the mean of adjacent time points as a reasonable value, depending on the actual situation.

[0076] After completing the data cleaning, in order to eliminate the differences in dimensions and orders of magnitude of different data features and make each feature have equal importance in subsequent analysis and model training, the data is standardized. That is, when the correlation between dynamic process parameters and waviness feature data is calculated by the multi-head graph attention layer method in step S110 and the edge weight coefficient is obtained, the data has already been standardized.

[0077] In one embodiment, in step S110, the correlation between the dynamic process parameters and the waviness feature data is calculated by a multi-head graph attention layer method to obtain an edge weight coefficient, which is as follows:

[0078] In the multi-head graph attention layer method, node feature encoding: the node parameters (dynamic process parameters and waviness characteristics) into high-dimensional vectors , for example, when considering the wear amount in the first two time steps, ,in, is a high-dimensional vector related to dynamic process parameters, Represents the values ​​of dynamic process parameters (such as tool wear, dominant vibration frequency, feed rate and thermal deformation influence cycle) at different time points t, is the value of the dynamic process parameter in the previous time step, are the values ​​of the dynamic process parameters in the first two time steps.

[0079] Attention coefficient calculation: node parameters Neighbor nodes Calculate its importance, the specific formula is as follows:

[0080] ,

[0081] in, is the attention coefficient; the symbols i‌ and j‌ are common mathematical notations for nodes, is a trainable attention weight vector responsible for capturing the correlation between dynamic process parameters and waviness feature data; and are the eigenvector of the node and the eigenvector of the neighboring nodes, Indicates that the node and The feature vectors of are spliced ​​along the feature dimension to fuse the information of the two. The nodes are multiplied by matrix and The feature vector of is linearly mapped from d-dimensional space to d-dimensional space, with the purpose of feature enhancement and expression improvement; LeakyReLU is an activation function used to introduce nonlinearity to avoid gradient disappearance. The formula is ( is a small constant, such as 0.2), where k is the index of the neighbor node and W is a learnable weight matrix. The learnability of W is one of the core designs of the graph attention layer, which is to adaptively adjust the node features through linear transformation to better calculate attention and aggregate neighbor information. In this formula, the numerator is obtained by The function amplifies the difference, the denominator is All neighbor nodes of Normalize and finally express right The relative importance of .

[0082] Next, the weights are updated, specifically by calculating the edge weights based on multi-head attention fusion. Assuming there are h attention heads, the updated edge weights are as follows:

[0083] ,

[0084] in, Represents the final edge weight between nodes i and j, which is the value that measures the strength of the association between the two nodes after multi-head attention fusion; Represents the number of "heads" in the multi-head attention mechanism, that is, the number of attention sub-models calculated in parallel. is the summation symbol, indicating The value ranges from 1 to All items are accumulated; The superscript indicates the Attention head, subscript It represents the attention coefficient of node i to node j, reflecting the relative importance of node j to i in the head.

[0085] During the training process, for high frequency impact links (such as , which represents the effect of tool wear on wave height), and its corresponding attention coefficient Large, after softmax calculation, the weight It will also be significantly higher than secondary links (such as those with other parameters that have less influence on wave height). This allows the model to focus more on high-frequency influencing links, highlighting the impact of key factors on the results, and improving the accuracy and effectiveness of causal relationship modeling. For example, in actual machining, tool wear often has a critical impact on wave height. This weight update method can give this link a more prominent position in the model, thereby more accurately capturing and expressing this causal relationship.

[0086] Next, construct the causal structure graph as follows:

[0087] Among them, the definition of G(V, E) nodes and edges is: the node set V represents the process parameter nodes and the waviness feature nodes. The specific process parameter nodes include: V W (tool wear), V A (vibration frequency), V V (feed rate), V D (thermal deformation). The waviness characteristic nodes are V corresponding to wave height and wavelength respectively. H 、VL .

[0088] E is a set of directed edges, and the initial edges are preset by the physical mechanism as follows:

[0089]

[0090] in, (Tool wear → wave height): In turning, as the tool cuts continuously (e.g., processing 100 parts), the tool flank wear increases. When the tool is worn and blunt, the cutting force increases and becomes unstable, resulting in the wave height of the machining marks on the workpiece surface. Significant increase.

[0091] (Vibration frequency → wavelength): During iron cutting, if the spindle vibration frequency Sudden increase due to equipment failure (such as from 50 Increased to 200 ), the relative motion rhythm between the tool and the workpiece is disrupted. At this time, the wavelength of the ripples on the machined surface is There will be obvious fluctuations, the originally uniform The wavelengths become different in length, and some areas are shortened to , some areas are extended to .

[0092] (Feed rate → wave height): In CNC machining, if the feed rate The sudden acceleration causes the cutting force to increase instantly and the cutting process to become unstable, which results in high wave height on the workpiece surface. Increased dramatically,

[0093] (Thermal deformation → wave height): After a long period of milling, the machine tool spindle will produce thermal deformation due to continuous operation. Thermal deformation changes the relative position of the tool and the workpiece, resulting in uneven cutting depth and ultimately causing high wave height on the workpiece surface. Abnormally elevated.

[0094] Edge attributes include: edge weight , represents the impact strength, optimized by the graph attention mechanism, and takes values ​​[0, 1]. It also includes the lag time, whose parameter changes to the delay time of the waviness anomaly.

[0095] In one embodiment, step S200: collecting environmental impact data and surface waviness data of a workpiece in real time, determining whether the surface waviness data exceeds a preset waviness threshold, and if so, obtaining waviness abnormality data, specifically including:

[0096] First, collect real-time workpiece surface waviness data, use a high-resolution topography measurement system such as a 3D laser scanner or white light interferometer to perform grid scanning on the workpiece surface, collect real-time waviness data, and obtain the data for each grid point. Wave height at ,wavelength , spatial resolution m.

[0097] Then, threshold comparison and signal processing are performed. Among them, wave height anomaly: the preset wave height threshold is m, if m, marked as abnormal wave height; abnormal wavelength: set the target wavelength ,like , marked as wavelength anomaly; spatial filtering: through Gaussian filtering or bandpass filter Separate high-frequency noise to obtain a pure ripple signal.

[0098] If an anomaly is detected, the waviness anomaly data is output. The spatial coordinates of the abnormal area are expressed as grid coordinates or geometric contours, such as a rectangular area. Indicates; Abnormal type label: clearly marked as "wave height abnormality", "wavelength abnormality" or "combined abnormality"; quantitative indicators such as abnormal wave height value: (like ); Abnormal wavelength value is expressed as .

[0099] In one embodiment, in step S200 , it is determined whether the surface waviness data exceeds a preset waviness threshold, and the method further includes: if the determination is no, determining that the surface waviness data is normal.

[0100] In one embodiment, step S300: inputting the abnormal waviness data into a causal structure map, analyzing the abnormal waviness data based on the causal structure map, and obtaining the influencing paths of the main process factors and the influencing paths of the secondary process factors; includes:

[0101] Step S310: Inputting the abnormal waviness data into a causal structure map, and performing reverse reasoning using a graph convolutional network method in the causal structure map to obtain a process parameter responsibility factor;

[0102] Step S320: Using the Bayesian inference method to screen the process parameter responsibility factors to obtain the main process factors and the secondary process factors;

[0103] Step S330: by querying the cause-effect structure graph, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively.

[0104] In one embodiment, the corrugation abnormality data is input into a causal structure graph in turn, and reverse reasoning is performed in the causal structure graph using a graph convolution network method to obtain a process parameter responsibility factor; the process parameter responsibility factor is screened using a Bayesian reasoning method to obtain a main process factor and a secondary process factor; by querying the causal structure graph, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively, thereby achieving accurate determination of the corrugation abnormality, and starting from the abnormal node, through graph convolution operations and Bayesian reasoning, in-depth analysis is performed to find out the main and secondary process factors that cause the abnormality and their causal influence paths, and targeted calculation of the main cause and secondary cause correction amounts, thereby achieving effective correction of the corrugation abnormality, improving the stability and reliability of the production process, and reducing production interruptions and defective rates caused by corrugation problems.

[0105] First, the abnormal waviness data is input into the causal structure map, and reverse reasoning is performed in the causal structure map G(V,E) to locate the process parameter responsible factors that cause the abnormality. Among them, the abnormal node activation is to map the abnormal type label to the corresponding node of the causal structure map (such as wave height abnormal activation, wavelength abnormal activation), and the initial abnormal intensity of the abnormal node is set to , the remaining nodes The abnormal signal is propagated layer by layer to the upstream process parameter node through the graph convolutional network (GCN). Upstream process parameter nodes with incoming edge connections , calculate its abnormal intensity, the specific formula is as follows:

[0106]

[0107] in, It is abnormal intensity, is the edge weight, is the initial abnormal strength of the abnormal node, Points to abnormal nodes The edge set of For nodes The number of outgoing edges, which is used for normalization to avoid score overflow;

[0108] The process parameter responsibility factor is derived from the process parameter node , including V W (tool wear), V A (vibration frequency), V V (feed rate), V D (thermal deformation).

[0109] Causal path screening and prioritization are performed by path compression, Bayesian reasoning and posterior probability calculation. Among them, path compression is to remove weights Weakly related links, reducing redundant reasoning (such as only keeping Bayesian reasoning is to combine abnormal intensity data and historical data to calculate the posterior probability of each process parameter node and screen out the top two factors with the highest contribution to the abnormality.

[0110] The posterior probability is calculated as follows:

[0111]

[0112] in, is the abnormal intensity, i refers to the i-th process factor node, and is the counting symbol. is the edge weight, is the weight from node n to node j, calculated by the multi-head graph attention layer, representing the process parameters Waviness characteristics the intensity of the impact; It is a priori probability, based on historical data statistics, such as the priori probability of abnormal tool wear , the prior probability of abnormal spindle vibration frequency wait; The numerator represents the observed joint probability, which is the prior probability multiplied by the likelihood probability; The denominator is a summation operation, which satisfies ( Represents A set of related nodes) and The sum of all nodes with (edge ​​weight threshold) is the total probability observed under all assumptions, which can also be understood as marginal probability. is the number of nodes, is the abnormal intensity of the nth node, is the prior probability of the nth node.

[0113] According to the posterior probability Sort in descending order and select the first two factors as the main and secondary process factors. (highest), then is the main factor; if (Second highest), then is a secondary factor.

[0114] The main process factor obtained is tool wear (posterior probability ), the secondary process factor is the main shaft vibration frequency (posterior probability ). The output result is tool wear exceeding the limit (corresponding node ,weight , lag time 15 minutes); spindle vibration frequency abnormality (corresponding node ,weight , lag time 5 minutes); by querying the causal structure map, the influence path of the main process factor and the influence path of the secondary process factor, as well as the weight and weights The specific value of .

[0115] The impact path of the main process factors: :Tool wear → abnormal wave height, path description "Tool wear leads to blunting of cutting edge, increased fluctuation of cutting force, causing wave height to exceed standard", weight By edge weight Get the specific value;

[0116] The impact path of secondary process factors: :Vibration frequency → wavelength abnormality, path description "spindle vibration intensifies, leading to unstable relative motion between tool and workpiece, and periodic wavelength fluctuations", weight By edge weight Get the specific value.

[0117] In one embodiment, in step S400, generating a waviness correction value according to the main process factors and the influencing paths of the main process factors includes:

[0118] Step S411: generating a main factor correction value through an adaptive projection algorithm according to the main process factors and the influencing paths of the main process factors;

[0119] Step S412: generating a secondary factor correction value by calculating the secondary process factor and the influencing path of the secondary process factor using an adaptive projection algorithm;

[0120] Step S413: Calculate a waviness correction value based on the primary correction value and the secondary correction value.

[0121] In one embodiment, in the above description, the main process factor (tool wear) and the secondary process factor (spindle vibration frequency) have been identified, and now the abnormal data is corrected in a coordinated manner based on the two. The correction of the main process factor is done through the main process factor and the influence path of the main process factor. , the main cause correction value is calculated by adaptive projection algorithm.

[0122] Calculate correction amount :

[0123]

[0124] in, is the main cause correction amount, which is the correction amount of tool wear to wave height. is the weight of the influence of tool wear on wave height, is the weight of the influence of vibration frequency on wavelength. is the standard wave height value, is the coordinate The abnormal wave height value is at . Among them, the influence weight of vibration frequency on wavelength is introduced The purpose is to achieve a reasonable distribution of the correction amount through the weight ratio relationship, reflect the relative influence of the weights of the primary and secondary factors, and ensure that the correction calculation is in line with the balance of the overall weight system.

[0125] According to the main cause correction amount and abnormal data, the main cause correction value is calculated:

[0126]

[0127] in, is the main cause correction value, is the coordinate The abnormal wave height value, is the main cause correction amount, the main cause dynamic scaling factor , the higher the main cause weight, The larger the value is, the stronger the correction of the main cause will be. The specific value is set by those skilled in the art according to actual conditions, and this application does not provide examples or limitations.

[0128] Secondary factor (spindle vibration frequency ) correction is based on the secondary process factor spindle vibration frequency and the impact path of secondary process factors , the secondary factor correction value is calculated by the adaptive projection algorithm. The specific calculation formula is as follows:

[0129]

[0130] in, This is the secondary correction amount, is the weight of the influence of tool wear on wave height, The weight is the influence of the vibration frequency on the wavelength; is the standard target wavelength (such as 2mm). is the coordinate Abnormal wavelength value at ; is the influence weight of tool wear on wave height.

[0131] ,in, This is the secondary correction value, is the coordinate The abnormal wavelength value of This is the secondary correction amount, is the dynamic scaling factor of the secondary factor, and the lower the secondary factor weight, the smaller it is. The smaller the value, the less likely secondary factors are to interfere excessively with the correction result. Similarly, the specific value is set by those skilled in the art based on actual conditions, and this application does not provide examples or limitations.

[0132] The adaptive superposition correction formula is used in the waviness correction process to make independent corrections for different influencing factors (such as process parameters, environmental factors, etc.) to obtain parallel correction values, as follows:

[0133] ,

[0134] in, is the waviness correction value, is the main cause correction value, is the main dynamic scaling factor, This is the correction value of the secondary cause; it is iteratively adjusted according to the residual error after correction. This is due to the dynamic scaling factor, is the main cause correction amount, This correction method distinguishes the main cause from the secondary cause, combines their respective weights and dynamic scaling factors, and more accurately detects abnormal areas. Perform multi-factor collaborative correction to improve the pertinence and accuracy of error correction.

[0135] In one embodiment, in step S400, calculating the waviness correction value according to the environmental impact data and the waviness correction value includes:

[0136] Step S421: generating an environmental impact coefficient according to the environmental impact data;

[0137] Specifically, step S421: generating an environmental impact coefficient according to the environmental impact data, including:

[0138] Step S4211: standardizing the environmental impact data and generating standardized environmental data;

[0139] Step S4212: Generate an environmental impact coefficient based on a preset Gaussian process regression and the standardized environmental data.

[0140] The environmental impact data include humidity, cutting fluid performance and viscosity. For humidity, the relative humidity of the environment is collected by the humidity sensor, and the effect of humidity on the hygroscopic expansion of the workpiece material is analyzed (for example, the linear expansion coefficient of aluminum alloy increases by 10% when the humidity increases by 10%). ) and the impact of the cutting fluid's evaporation rate. For cutting fluid performance, a conductivity meter is used to measure the concentration of water-based cutting fluids. Insufficient concentration (e.g., 5% below the standard value) can lead to insufficient lubrication and increased tool wear. For viscosity, a rotational viscometer is used to test the cutting fluid's viscosity. Abnormal viscosity (e.g., 20% above the standard value) can affect chip removal and cause spindle vibration.

[0141] Humidity, cutting fluid properties and viscosity are standardized and converted into dimensionless standardized values. Specifically, humidity is standardized as follows:

[0142]

[0143] in, is the standardized humidity, is the standard humidity, The maximum humidity, Minimum humidity.

[0144] Concentrations were normalized as follows:

[0145]

[0146] in, is the normalized concentration, is the standard concentration, is the maximum concentration, is the minimum concentration.

[0147] Viscosity is normalized as follows:

[0148]

[0149] in, is the normalized viscosity, is the standard viscosity, is the maximum viscosity, is the minimum viscosity.

[0150] Next, Gaussian process regression (GPR) is used to model the nonlinear relationship between factors and output the environmental impact coefficient:

[0151]

[0152] in, is the environmental impact coefficient, ; is the normalized concentration; is the standardized humidity, () represents the Gaussian process regression function;

[0153] like , indicating that the influence of environmental factors (humidity, concentration, viscosity, etc.) on waviness is almost negligible, and the waviness correction is mainly dominated by process parameters (such as tool wear, vibration frequency, etc.).

[0154] like , indicating that the impact of environmental factors on the corrugation is the greatest, and the comprehensive effect of environmental factors must be fully considered when correcting the corrugation.

[0155] like , reflecting that environmental factors have a certain degree of influence on the waviness. The larger the value, the more significant the influence of environmental factors. For example, in a humid environment (high ), abnormal cutting fluid concentration (deviate Standard value) or viscosity change ( Fluctuations), through Gaussian process regression and squared exponential kernel function The nonlinear correlation of these normalized environmental factors is modeled and the output is It will increase accordingly, suggesting the need to strengthen compensation and correction for environmental factors.

[0156] If EIC>1 occurs, it is a data anomaly or model failure. Technical personnel in this field should check whether the measured values ​​of environmental factors exceed the limit and whether the standardized parameters are reasonable, and make adaptive adjustments based on actual conditions. This application does not provide examples or explanations.

[0157] It should be understood that The comprehensive influence of environmental factors on waviness is quantified, providing a basis for calculating the environmental compensation amount in subsequent waviness correction, making the correction more suitable for actual processing scenarios.

[0158] Step S422: Calculating an environmental compensation value according to the degree of influence of the environmental influence coefficient on the waviness;

[0159] Step S423: Calculate a waviness correction value according to the waviness correction value and the environmental compensation value.

[0160] In one embodiment, a nonlinear compensation model is first constructed based on the degree of influence of the Environmental Impact Coefficient (EIC) on the corrugation. Taking into account the different degrees of influence of different EIC intervals on the corrugation, a piecewise function is used to achieve more accurate compensation.

[0161] ,

[0162] in, yes The environmental compensation value when is the environmental impact coefficient, is the waviness correction value; The value of k1, obtained by those skilled in the art through extensive historical data fitting and experimental verification, represents the coefficient of environmental influence on waviness within this range. It should be understood that when environmental influence is low, the compensation amount increases linearly with increasing EIC and is proportional to the waviness correction value.

[0163] ,

[0164] yes The environmental compensation value when is the environmental impact coefficient, is the waviness correction value; The value of k2, obtained by those skilled in the art through extensive historical data fitting and experimental verification, represents the coefficient of environmental influence on waviness within this range. This method allows for more effective waviness correction when environmental influences are moderate, as the impact of environmental factors can be more complex at this time, and a quadratic function can better capture this variation.

[0165] ,

[0166] in, yes The environmental compensation value when , under the condition of strong environmental influence, the waviness can be fully adjusted to ensure the final measurement accuracy.

[0167] Combining the waviness correction value and the environmental compensation value, the final waviness correction value is calculated:

[0168]

[0169] in, is the waviness correction value, is the waviness correction value, is the environmental compensation value of the i-th case.

[0170] The final waviness correction value is output in matrix format, corresponding to the coordinates of each measured point on the workpiece surface. This ensures consistency with the previously measured data structure, facilitating subsequent analysis and use. The Environmental Impact Factor (EIC) value, along with the compensation range and formula used in the calculation, is also output, facilitating traceability and verification of the results. For EIC values ​​greater than 0.7, the result is marked with "Significant environmental impact, correction value significantly affected by environmental factors," reminding personnel to pay attention to the impact of environmental factors on measurement results.

[0171] Therefore, the waviness measurement accuracy correction method described in this application collects dynamic process parameters and waviness characteristic data, applies advanced algorithms to construct a causal structure map, accurately analyzes abnormal waviness data, identifies primary and secondary process factors and causal influence paths, and calculates corrections using an adaptive projection algorithm. Furthermore, environmental impact data is considered and environmental impact coefficients are calculated, ultimately resulting in an accurate waviness correction value. This method comprehensively and systematically addresses the issue of waviness measurement accuracy, fully accounting for the influence of multiple factors, including process parameters, abnormal data processing, and environmental factors.

[0172] In one embodiment, Figure 2 As shown, a waviness measurement accuracy correction system based on artificial intelligence is also provided, the system comprising:

[0173] A cause-effect graph construction module is used to obtain dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, and to construct a cause-effect structure graph based on the dynamic process parameters and the waviness characteristic data;

[0174] an abnormal data acquisition module, for collecting environmental impact data and surface waviness data of a workpiece in real time, and determining whether the surface waviness data exceeds a preset waviness threshold; and if so, obtaining abnormal waviness data;

[0175] an abnormal data analysis module, configured to input the abnormal waviness data into a causal structure map, analyze the abnormal waviness data based on the causal structure map, and obtain an influencing path of a primary process factor and an influencing path of a secondary process factor;

[0176] The environmental impact correction module is used to generate a waviness correction value according to the main process factors and the influence path of the main process factors, and to calculate the waviness correction value according to the environmental impact data and the waviness correction value.

[0177] In one embodiment, the abnormal data analysis module is further used to: input the corrugation abnormality data into a causal structure map, use the graph convolution network method to perform reverse reasoning in the causal structure map to obtain the process parameter responsibility factor; use the Bayesian reasoning method to screen the process parameter responsibility factor to obtain the main process factor and the secondary process factor; by querying the causal structure map, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively.

[0178] In one embodiment, the environmental impact correction module is also used to: generate a main cause correction value through an adaptive projection algorithm based on the main process factor and the influence path of the main process factor; generate a secondary cause correction value through an adaptive projection algorithm based on the secondary process factor and the influence path of the secondary process factor; and calculate a corrugation correction value based on the main cause correction value and the secondary cause correction value.

[0179] In one embodiment, the environmental impact correction module is further used to: generate an environmental impact coefficient based on the environmental impact data; calculate an environmental compensation value based on the degree of influence of the environmental impact coefficient on the corrugation; and calculate a corrugation correction value based on the corrugation correction value and the environmental compensation value.

[0180] In one embodiment, the environmental impact correction module is further used to: standardize the environmental impact data and generate standardized environmental data; and generate an environmental impact coefficient based on a preset Gaussian process regression and the standardized environmental data.

[0181] In one embodiment, the causal graph construction module is also used to: obtain the dynamic process parameters of the machine tool and the waviness characteristic data of the workpiece when the machine tool processes the workpiece, calculate the correlation between the dynamic process parameters and the waviness characteristic data through the multi-head graph attention layer method, and obtain the edge weight coefficient; construct a causal structure graph based on the dynamic process parameters, the waviness characteristic data and the edge weight coefficient.

[0182] In one embodiment, the abnormal data acquisition module is further configured to determine whether the surface waviness data exceeds a preset waviness threshold, and if the determination is no, determine that the surface waviness data is normal.

[0183] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned artificial intelligence-based waviness measurement accuracy correction method when executing the computer program.

[0184] In one embodiment, a computer-readable storage medium is further provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned artificial intelligence-based waviness measurement accuracy correction method are implemented.

[0185] It should be noted that the information interaction, execution process and other contents between the above modules are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0186] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0187] It should be noted that the information interaction, execution process and other contents between the above modules are based on the same concept as the method embodiment of this application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0188] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0189] An embodiment of the present application also provides a network device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps of any of the above-mentioned method embodiments when executing the computer program.

[0190] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned various method embodiments can be implemented.

[0191] An embodiment of the present application provides a computer program product. When the computer program product is run on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned various method embodiments when executing the computer program product.

[0192] If the integrated unit is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the process steps in the above-mentioned method embodiments by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a camera / terminal device, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, removable hard drives, magnetic disks, or optical disks. In some jurisdictions, based on legislation and patent practice, computer-readable media cannot be electric carrier signals or telecommunication signals.

[0193] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0194] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0195] In the embodiments provided in this application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely illustrative. For example, the division of the modules or units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0196] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0197] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

[0198] An embodiment of the present application also provides a computer device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any embodiment of the above method when executing the computer program.

[0199] The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the above description is an example of a computer device and does not limit the computer device. The computer device may include more or fewer components than described above, or a combination of certain components, or different components. For example, the computer device may also include input / output devices, network access devices, etc.

[0200] The processor may be a central processing unit (CPU), and the processor 0 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0201] In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped with the computer device. Furthermore, the memory may include both an internal storage unit of the computer device and an external storage device. The memory is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory may also be used to temporarily store data that has been output or is about to be output.

[0202] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0203] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and such modifications and improvements are intended to fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for correcting waviness measurement accuracy based on artificial intelligence, characterized in that: The method comprises: Acquire dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, calculate the correlation between the dynamic process parameters and the waviness characteristic data using a multi-head graph attention layer method, and obtain edge weight coefficients; construct a causal structure graph based on the dynamic process parameters, the waviness characteristic data, and the edge weight coefficients; Collecting environmental impact data and surface waviness data of the workpiece in real time, determining whether the surface waviness data exceeds a preset waviness threshold, and obtaining waviness abnormality data if the determination is yes; Inputting the abnormal waviness data into a causal structure map, analyzing the abnormal waviness data based on the causal structure map, and obtaining the influencing path of the main process factors and the influencing path of the secondary process factors; A waviness correction value is generated according to the main process factors and the influencing paths of the main process factors, and a waviness correction value is calculated according to the environmental impact data and the waviness correction value.

2. The method for correcting waviness measurement accuracy based on artificial intelligence according to claim 1, characterized in that: Inputting the abnormal waviness data into a causal structure map, analyzing the abnormal waviness data based on the causal structure map, and obtaining the influencing path of the main process factors and the influencing path of the secondary process factors; including: Inputting the abnormal waviness data into a causal structure map, and performing reverse reasoning using a graph convolutional network method in the causal structure map to obtain a process parameter responsibility factor; The Bayesian inference method is used to screen the process parameter responsibility factors and obtain the main process factors and secondary process factors; By querying the causal structure map, the influence path of the main process factor and the influence path of the secondary process factor are obtained respectively.

3. The method for correcting waviness measurement accuracy based on artificial intelligence according to claim 1, characterized in that: According to the main process factors and the influencing paths of the main process factors, a waviness correction value is generated, including: According to the main process factors and the influencing paths of the main process factors, a main factor correction value is generated by an adaptive projection algorithm; According to the secondary process factors and the influencing paths of the secondary process factors, a secondary factor correction value is calculated and generated by an adaptive projection algorithm; A waviness correction value is calculated based on the primary cause correction value and the secondary cause correction value.

4. The method for correcting waviness measurement accuracy based on artificial intelligence according to claim 1, characterized in that: The waviness correction value is calculated according to the environmental impact data and the waviness correction value, including: generating an environmental impact coefficient according to the environmental impact data; Calculating an environmental compensation value according to the degree of influence of the environmental impact coefficient on the waviness; A waviness correction value is calculated based on the waviness correction value and the environmental compensation value.

5. The method for correcting waviness measurement accuracy based on artificial intelligence according to claim 4, characterized in that: Generating an environmental impact coefficient according to the environmental impact data includes: Standardizing the environmental impact data and generating standardized environmental data; An environmental impact coefficient is generated based on a preset Gaussian process regression and the standardized environmental data.

6. The method for correcting waviness measurement accuracy based on artificial intelligence according to claim 1, characterized in that: Determining whether the surface waviness data exceeds a preset waviness threshold, and then further comprising: If the judgment is no, then the surface waviness data is determined to be normal.

7. An artificial intelligence-based waviness measurement accuracy correction system, characterized in that: The system comprises: A causal graph construction module is used to obtain dynamic process parameters of the machine tool and waviness characteristic data of the workpiece when the machine tool processes the workpiece, calculate the correlation between the dynamic process parameters and the waviness characteristic data through a multi-head graph attention layer method, and obtain edge weight coefficients; and construct a causal structure graph based on the dynamic process parameters, the waviness characteristic data, and the edge weight coefficients; an abnormal data acquisition module, for collecting environmental impact data and surface waviness data of a workpiece in real time, and determining whether the surface waviness data exceeds a preset waviness threshold; and if so, obtaining abnormal waviness data; an abnormal data analysis module, configured to input the abnormal waviness data into a causal structure map, analyze the abnormal waviness data based on the causal structure map, and obtain an influencing path of a primary process factor and an influencing path of a secondary process factor; The environmental impact correction module is used to generate a waviness correction value according to the main process factors and the influence path of the main process factors, and to calculate the waviness correction value according to the environmental impact data and the waviness correction value.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images