Graphite heater production line quality online monitoring method

Abstract

The application provides a graphite heater production line quality online monitoring method, and relates to the technical field of industrial automation, and comprises the following steps: when a graphite heater blank enters a starting point of a production line, a digital twin file of a preset three-dimensional model is created and bound; after mechanical processing is completed, three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater are acquired; through a non-contact global scanning measurement technology, an intrinsic attribute three-dimensional distribution field is acquired, which is written into the digital twin file as a material attribute parameter; through a non-contact spectral imaging technology, an intrinsic attribute two-dimensional distribution cloud picture is acquired, and is converted into a thermal boundary condition parameter; based on the preset three-dimensional model, the material attribute parameter and the thermal boundary condition parameter in the digital twin file, a physical field simulation model of the graphite heater is constructed and calculated, and a performance prediction report is generated; and multi-source monitoring data in the digital twin file and the performance prediction report are fused.

Description

Technical Field

[0001] This application relates to the field of industrial automation technology, and more specifically, to a method for online quality monitoring of a graphite heater production line. Background Technology

[0002] In the research and development stage of graphite heaters, existing technologies mainly rely on step-by-step offline measurement and uniformity simulation for quality assessment. After sample fabrication, the macroscopic geometric dimensions and average electrical parameters are measured offline. To obtain key material properties, destructive sampling of the sample is required, followed by analysis using equipment such as X-ray diffractometers. The obtained finite number of representative material parameters are then input into finite element software to construct an idealized simulation model assuming uniform material properties, in order to predict its operating temperature field. This method provides fundamental support for design scheme selection and preliminary macroscopic performance estimation, and is currently a standard technical path in the research and development process.

[0003] In R&D scenarios that require studying the intrinsic relationship between process, structure, and performance, the shortcomings of existing methods become apparent when optimizing sintering processes: they struggle to obtain complete information on the actual distribution of the material's microstructure in three-dimensional space without damaging valuable R&D samples. Subtle changes in process parameters often lead to non-uniformity within the material, which directly affects the local performance of electrothermal properties. However, destructive sampling can only reflect individual points, and simulations based on the assumption of uniformity mask this spatial difference. When the measured performance of the sample deviates from the prediction of the uniform model, there is a lack of key data to reveal where the microstructure anomaly caused the performance distortion, making it impossible to accurately trace the root cause of the process. As a result, process optimization still relies to some extent on trial and error, affecting the efficiency and depth of R&D.

[0004] Therefore, there is an urgent need for an online quality monitoring method for graphite heater production lines. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. The first aspect of this application provides a method for online quality monitoring of a graphite heater production line, comprising the following specific steps: creating and binding a digital twin file of a preset three-dimensional model when the graphite heater blank enters the starting point of the production line;

[0006] After machining is completed, the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater are obtained;

[0007] The three-dimensional distribution field of intrinsic properties is obtained through non-contact full-domain scanning measurement technology and written into the digital twin archive as material property parameters. The two-dimensional distribution cloud map of intrinsic properties is obtained through non-contact spectral imaging technology and converted into thermal boundary condition parameters.

[0008] Based on the preset three-dimensional model, material property parameters and thermal boundary condition parameters in the digital twin archive, a physical field simulation model of the graphite heater is constructed and calculated to generate a performance prediction report.

[0009] By integrating multi-source monitoring data and performance prediction reports from digital twin archives, a comprehensive quality assessment of graphite heaters is conducted based on preset rules.

[0010] Preferably, the digital twin archive includes,

[0011] When the graphite heater blank arrives at the starting point of the production line, the identification information of the graphite heater blank arriving at the starting point of the production line is collected.

[0012] A digital twin file containing a preset 3D model is created using the identification information of the graphite heater blank, and the digital twin file of the preset 3D model is bound to the graphite heater blank.

[0013] Preferably, obtaining the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater includes,

[0014] The machined graphite heater is transferred to the measurement station, where a three-dimensional optical scan is performed to generate a point cloud model of the graphite heater.

[0015] The point cloud model of the graphite heater is compared with the preset 3D model to calculate the 3D geometric dimension deviation data of the graphite heater.

[0016] A four-probe tester was used to contact multiple preset points on the surface of the graphite heater to measure the multi-point resistance value of the graphite heater. The three-dimensional geometric deviation data and multi-point resistance value of the graphite heater were recorded as the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater.

[0017] Preferably, obtaining the three-dimensional distribution field of intrinsic attributes includes,

[0018] The graphite heater is transferred to the microwave resonant cavity scanning station, and the microwave resonant cavity is controlled to perform spatial point-by-point scanning on the graphite heater located at the microwave resonant cavity scanning station to collect the microwave resonant cavity parameter dataset.

[0019] The microwave resonance parameter dataset of the microwave resonant cavity is input into the pre-trained microwave response-material microstructure inversion model to calculate the three-dimensional distribution field of graphitization degree of the graphite heater.

[0020] Based on the three-dimensional distribution field of graphitization degree of the graphite heater, the three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are derived.

[0021] The three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are written into the digital twin file of the preset three-dimensional model as material property parameters.

[0022] Preferably, the conversion into thermal boundary condition parameters includes,

[0023] The graphite heater is transferred to the hyperspectral imaging scanning station, and the surface of the graphite heater coating located at the hyperspectral imaging scanning station is subjected to push-broom imaging to obtain the hyperspectral data cube of the graphite heater.

[0024] The hyperspectral data cube of the graphite heater is input into the pre-trained spectral feature-coating physical property inversion model to resolve the coating thickness distribution cloud map and coating porosity distribution cloud map of the graphite heater.

[0025] Based on the cloud map of coating porosity and coating thickness of the graphite heater, the distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is calculated.

[0026] The distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is used as a thermal boundary condition parameter and written into the digital twin file of the preset three-dimensional model.

[0027] Preferably, the performance prediction report includes,

[0028] The three-dimensional distribution fields of electrical conductivity and thermal conductivity in the material property parameters are assigned to the finite element mesh elements corresponding to the preset three-dimensional model to construct the physical field simulation model of the graphite heater.

[0029] The equivalent surface thermal radiation coefficient distribution field in the thermal boundary condition parameters is assigned to the finite element mesh surface element of the physical field simulation model. The rated working voltage excitation is applied to the physical field simulation model and the electro-thermal coupling field equation is solved to calculate the steady-state temperature field of the graphite heater.

[0030] The highest temperature, lowest temperature, maximum temperature difference, and hot spot location are extracted from the steady-state temperature field as key performance indicators to generate a performance prediction report.

[0031] Preferably, the multi-source monitoring data and performance prediction report in the integrated digital twin archive include,

[0032] Logically link the key performance indicators, three-dimensional geometric deviation data, and multi-point resistance values ​​in the performance prediction report;

[0033] Based on the preset quality rule base, multi-level threshold comparisons are performed on the logically associated multi-source data; based on the results of the multi-level threshold comparisons, the comprehensive quality judgment result of the graphite heater is obtained.

[0034] Preferably, the comprehensive quality assessment includes,

[0035] Using the pre-set quality rule library in the digital twin archive of the pre-set 3D model, the comprehensive quality judgment of the multi-source data after logical association is performed based on the quality rule library to obtain the comprehensive quality judgment result of the graphite heater.

[0036] Preferably, the actuator is driven by the comprehensive quality judgment result to sort the graphite heaters accordingly, and archives the digital twin files of the three-dimensional geometric dimension deviation data, multi-point resistance value, three-dimensional conductivity distribution field, three-dimensional thermal conductivity distribution field, equivalent surface thermal radiation coefficient distribution field, performance prediction report and comprehensive quality judgment result to form a traceable product quality data chain.

[0037] A second aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of the above-described online quality monitoring method for a graphite heater production line.

[0038] Compared with existing technologies, the online quality monitoring method for graphite heater production lines provided in this application achieves non-destructive, global online acquisition of the three-dimensional distribution field of graphitization degree through microwave resonant cavity scanning and inversion model. It reveals the spatial non-uniformity of material microstructure caused by process fluctuations, elevating the root cause analysis of processes from empirical speculation based on average values ​​to precise diagnosis based on spatial distribution anomalies, greatly accelerating process iteration. Through hyperspectral imaging and analytical models, it achieves quantitative evaluation of the two-dimensional distribution cloud map of coating thickness and porosity, and transforms it into an equivalent surface thermal radiation coefficient distribution field, providing realistic non-uniform thermal boundary conditions for performance simulation. It realizes quantitative and spatial evaluation of the impact on coating function. Based on these accurate distribution field data, a high-fidelity electro-thermal coupling simulation model is constructed, which can predict the core performance of the sample, such as thermal field uniformity, without powering on during the R&D stage. It realizes the transformation of the R&D mode from post-measurement to pre-simulation. It integrates multi-source data and prediction reports for intelligent judgment and archiving, forming a complete data closed loop to support process optimization and mechanism research, improving the accuracy and efficiency of R&D. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0040] Figure 1 This is a flowchart of an online quality monitoring method for a graphite heater production line according to an embodiment of this application. Detailed Implementation

[0041] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0042] As mentioned in the background section, in R&D scenarios that require studying the intrinsic relationship between process, structure, and performance, the shortcomings of existing methods become apparent when optimizing sintering processes: they are difficult to obtain complete information on the actual distribution of the material's microstructure in three-dimensional space without damaging valuable R&D samples. Subtle changes in process parameters often lead to non-uniformity within the material, which directly affects the local performance of electrothermal properties. However, destructive sampling can only reflect individual points, and simulations based on the assumption of uniformity mask this spatial difference. When the measured performance of the sample deviates from the prediction of the uniform model, there is a lack of key data to reveal where the microstructure anomaly caused the performance distortion, making it impossible to accurately trace the root cause of the process. As a result, process optimization still relies to some extent on trial and error, affecting the efficiency and depth of R&D.

[0043] Figure 1 The flowchart below shows a method for online quality monitoring of a graphite heater production line according to an embodiment of this application, including the following specific steps:

[0044] S1: When the graphite heater blank enters the starting point of the production line, create and bind a digital twin file of the preset three-dimensional model.

[0045] In step S1, the graphite heater blank arrives at the starting point of the production line, and the identification information of the graphite heater blank arriving at the starting point of the production line is collected.

[0046] Furthermore, when the graphite heater blank is conveyed from the previous process to the conveyor belt at the start of the production line and triggers the inlet position sensor, an industrial barcode reader or vision recognition device located above or to the side of the inlet is activated. This device aligns with and reads the QR code or barcode pre-attached to the surface of the graphite heater blank, collecting the identification information of the graphite heater blank arriving at the start of the production line. The industrial barcode reader or vision recognition device then transmits the collected identification information to the central database via an industrial Ethernet or fieldbus network.

[0047] A digital twin file containing a preset 3D model is created using the identification information of the graphite heater blank, and the digital twin file of the preset 3D model is bound to the graphite heater blank.

[0048] Furthermore, after receiving the identification information of the graphite heater blank, the central database queries and matches it in the preset product model-process parameter mapping table, retrieves the standard 3D computer-aided design model file and standard process parameter file associated with the product model corresponding to the identification information, and creates a new digital file entry using the identification information as the unique key. This digital file entry is the digital twin file containing the preset 3D model. After the creation operation is completed, the central database sends an instruction to the production line material tracking system to associate the file number of the digital twin file containing the preset 3D model with the RFID number or physical location number of the tray or carrier currently carrying the graphite heater blank on the conveyor belt, thereby completing the binding of the digital twin file of the preset 3D model with the graphite heater blank.

[0049] S2: After machining is completed, obtain the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater.

[0050] In step S2, the completed graphite heater is transferred to the measurement station, and a three-dimensional optical scan is performed on the graphite heater located at the measurement station to generate a point cloud model of the graphite heater.

[0051] Furthermore, the machined graphite heater is removed from the machining center by a robotic arm or conveyor belt and transported to a designated 3D optical scanning measurement station. At the measurement station, multiple sets of structured light projectors and high-resolution industrial cameras are deployed around the graphite heater at preset angles. The structured light projectors project coded grating stripe patterns onto the surface of the graphite heater located at the measurement station. The high-resolution industrial cameras simultaneously capture stripe images deformed by the surface contour of the graphite heater. The deformed stripe image sequence captured by multiple sets of high-resolution industrial cameras is processed through the multi-view visual triangulation principle, and the data set consisting of dense 3D coordinate points covering the entire outer surface of the graphite heater is calculated and fused to generate a point cloud model of the graphite heater.

[0052] The point cloud model of the graphite heater is compared with the preset 3D model to calculate the 3D geometric dimension deviation data of the graphite heater.

[0053] Furthermore, after obtaining the point cloud model of the graphite heater, the preset 3D model in the digital twin file containing the preset 3D model is called. The point cloud model of the graphite heater and the preset 3D model are spatially aligned and registered in the same coordinate system by the iterative nearest point algorithm. After the registration is completed, the surface of the preset 3D model is discretized by triangulation, and the point with the closest normal distance in the point cloud model of the graphite heater is found for each grid vertex. The Euclidean distance between each corresponding point pair is calculated. This set of distance values ​​constitutes the normal deviation of each discrete position on the surface of the graphite heater relative to the ideal design model. This set of deviation values ​​is the 3D geometric dimension deviation data of the graphite heater.

[0054] The expression for the three-dimensional geometric dimension deviation data is:

[0055] ;

[0056] in, For the first Three-dimensional geometric dimensional deviation data of individual graphite heaters For the first The point cloud model coordinates of the actual graphite heater. For the first Each point has preset 3D model coordinates. For the first Point cloud model coordinates of a graphite heater For the first Each point has preset 3D model coordinates. For the first Point cloud model coordinates of a graphite heater For the first Each point has preset 3D model coordinates. For point indexes.

[0057] A four-probe tester was used to contact multiple preset points on the surface of the graphite heater to measure the multi-point resistance value of the graphite heater. The three-dimensional geometric deviation data and multi-point resistance value of the graphite heater were recorded as the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater.

[0058] Furthermore, the robotic arm moves the four-probe tester to the surface of the graphite heater. According to the preset measurement point coordinates, the four probes of the four-probe tester are pressed against each measurement point on the surface of the graphite heater with constant force. After the probes make stable contact, the internal constant current source of the four-probe tester outputs a standard test current to measure the voltage drop between the probes. Based on the resistivity of each measurement point, the resistivity values ​​of a series of measurement points are recorded as the multi-point resistance values ​​of the graphite heater. The three-dimensional geometric dimension deviation data of the graphite heater and the multi-point resistance values ​​of the graphite heater are stored together in a digital twin file containing a preset three-dimensional model, which serves as the three-dimensional geometric dimension and multi-point resistivity data of the graphite heater used in subsequent processes.

[0059] S3: Using non-contact global scanning measurement technology, the three-dimensional distribution field of intrinsic properties is obtained and written into the digital twin archive as material property parameters. Using non-contact spectral imaging technology, the two-dimensional distribution cloud map of intrinsic properties is obtained and converted into thermal boundary condition parameters.

[0060] In step S3, the graphite heater is transferred to the microwave resonant cavity scanning station, and the microwave resonant cavity is controlled to perform spatial point-by-point scanning on the graphite heater located at the microwave resonant cavity scanning station to collect the microwave resonant cavity microwave resonant parameter dataset.

[0061] Furthermore, the tray or transmission line carrying the graphite heater delivers the graphite heater, which has completed basic testing, into a closed testing unit with electromagnetic shielding. This testing unit is the microwave resonant cavity scanning station. Inside the microwave resonant cavity scanning station, a precise three-axis CNC moving platform receives control commands and drives the graphite heater located on the platform to execute a preset motion trajectory. This allows different parts of the graphite heater to sequentially and precisely pass through the central axis region of the fixedly installed cylindrical microwave resonant cavity. When the graphite heater moves to each preset spatial sampling point, the movement pauses, and the network analyzer connected to the microwave resonant cavity is triggered. The network analyzer feeds a swept-frequency microwave signal with a specific center frequency into the microwave resonant cavity and measures the scattering parameter matrix of the microwave resonant cavity at this time. From the scattering parameter matrix, the measured values ​​of the resonant frequency and quality factor of the microwave resonant cavity under the current graphite heater intervention state are extracted. The resonant frequency offset and quality factor change at this position coordinate are recorded. After all preset spatial sampling points are measured, the spatial coordinates, resonant frequency offset, and quality factor change of all sampling points are summarized to form a structured microwave resonant cavity microwave resonant parameter dataset.

[0062] The microwave resonance parameter dataset of the microwave resonant cavity is input into a pre-trained microwave response-material microstructure inversion model to calculate the three-dimensional distribution field of graphitization degree of the graphite heater.

[0063] Furthermore, the microwave resonant parameter dataset of the microwave resonant cavity is transmitted to a server deployed with a machine learning model via a data interface. The server calls and loads a pre-trained deep learning network, which is the pre-trained microwave response-material microstructure inversion model. The resonant frequency offset, quality factor variation, and corresponding three-dimensional spatial coordinate information in the microwave resonant parameter dataset of the microwave resonant cavity are combined and normalized according to the format required by the pre-trained microwave response-material microstructure inversion model to form an input vector. The input vector is then fed into the pre-trained microwave response-material microstructure inversion model. The pre-trained microwave response-material microstructure inversion model uses its internal fixed weight parameters to perform forward propagation calculations on the input vector, outputting the predicted graphitization degree value corresponding to each input spatial location. All these graphitization degree prediction values ​​at discrete locations are combined with their three-dimensional spatial coordinates, and three-dimensional kriging interpolation or radial basis function interpolation algorithms are used to perform interpolation calculations over the entire solid space of the graphite heater, generating a graphitization degree scalar field continuously distributed in three-dimensional space.

[0064] The expression for the three-dimensional distribution field of graphitization degree is:

[0065] ;

[0066] in, For position The three-dimensional distribution field of graphitization degree at the location, For the pre-trained microwave response-material microstructure inversion model, This is a dataset of microwave resonant parameters for a microwave resonant cavity. These are the weight parameters for the pre-trained microwave response-material microstructure inversion model. This is a spatial position coordinate vector.

[0067] Based on the three-dimensional distribution field of graphitization degree of the graphite heater, the three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are derived.

[0068] Furthermore, after obtaining the three-dimensional distribution field of graphitization degree of the graphite heater, based on the generally accepted physical correlation between graphitization degree and electrical / thermal conductivity in graphite materials science—a correlation typically manifested as a deterministic relationship where electrical and thermal conductivity monotonically increase with increasing graphitization degree—a graphitization degree-electrical conductivity-thermal conductivity correspondence table was pre-stored in the database. This table was obtained by fitting a polynomial or piecewise linear function after experimental measurements of a series of standard material samples with different graphitization degrees. For each three-dimensional voxel in the three-dimensional distribution field of graphitization degree, its graphitization degree value was read. Then, the theoretical electrical conductivity and theoretical thermal conductivity values ​​at the voxel location are obtained by searching the graphitization degree-electrical conductivity-thermal conductivity correspondence table or by calculating them using the corresponding fitting function. All voxels in the graphitization degree three-dimensional distribution field are processed sequentially, and the calculated electrical conductivity and thermal conductivity values ​​are arranged and stored according to the same three-dimensional spatial coordinates to form two new three-dimensional data fields with the same spatial dimension and resolution as the graphitization degree three-dimensional distribution field. These two data fields are the electrical conductivity three-dimensional distribution field and the thermal conductivity three-dimensional distribution field of the graphite heater, respectively.

[0069] The three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are written into the digital twin file of the preset three-dimensional model as material property parameters.

[0070] Furthermore, the two generated three-dimensional data field files—the three-dimensional distribution field data file of electrical conductivity and the three-dimensional distribution field data file of thermal conductivity of the graphite heater—are transmitted over a network to a database server storing a digital twin archive containing a preset three-dimensional model. The database server, based on the unique identifier of the graphite heater, locates the corresponding digital twin archive record containing the preset three-dimensional model and creates or updates a data structure named "Material Property Parameters" in that record. The three-dimensional distribution field data files of electrical conductivity and thermal conductivity of the graphite heater are stored and associated as two fields under this structure, completing the final operation of writing the three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater as material property parameters into the digital twin archive of the preset three-dimensional model.

[0071] The graphite heater is conveyed to the hyperspectral imaging scanning station, and push-broom imaging is performed on the coating surface of the graphite heater located at the hyperspectral imaging scanning station to obtain the hyperspectral data cube of the graphite heater.

[0072] Furthermore, at the hyperspectral imaging scanning station, a set of high-power, highly uniform LED surface light sources provides constant illumination for the graphite heater located at the scanning station. The hyperspectral imager is fixedly installed above the guide rail, and the stage below, which holds the graphite heater, moves at a constant speed in a straight line perpendicular to the slit of the hyperspectral imager under the drive of a stepper motor. At each equidistant step position of the stage's movement, the linear array detector of the hyperspectral imager synchronously collects the light reflected from the surface of the graphite heater coating through the slit, and after splitting the light, it is recorded by the two-dimensional surface array detector to obtain spectral intensity information of hundreds of continuous wavelengths along a spatial line. As the stage moves continuously, the hyperspectral imager collects a series of spectral data of spatial lines. Finally, these line scan spectral data arranged in spatial order are stacked and spliced ​​in the direction of movement to form a three-dimensional data array consisting of two spatial dimensions and one spectral dimension. This three-dimensional data array is the hyperspectral data cube of the graphite heater.

[0073] The hyperspectral data cube of the graphite heater is input into the pre-trained spectral feature-coating physical property inversion model to resolve the coating thickness distribution cloud map and coating porosity distribution cloud map of the graphite heater.

[0074] Furthermore, after acquiring the hyperspectral data cube of the graphite heater, the hyperspectral data cube is transferred to a workstation equipped with spectral processing software. The spectral processing software loads a pre-trained spectral feature-coating physical property inversion model. This model is typically based on partial least squares regression and is trained using a large amount of hyperspectral data from training samples with known coating thickness and porosity. The spectral processing software reads the hyperspectral data cube and extracts the complete reflectance spectrum curve in the wavelength range of 400 nm to 1000 nm for each spatial pixel position in the data cube. Each spectral curve is then subjected to necessary pre-processing. The data undergoes processing, such as denoising and normalization. The preprocessed spectral curve data is then used as an input vector and fed into a pre-trained spectral feature-coating physical property inversion model to obtain the predicted coating thickness and porosity values ​​corresponding to the pixel location. The model iterates through every spatial pixel in the hyperspectral data cube and records the prediction results for all pixels. The thickness prediction values ​​are arranged according to pixel coordinates to generate a two-dimensional grayscale image, which is a cloud map of the coating thickness distribution of the graphite heater. The porosity prediction values ​​are arranged according to the same coordinates to generate another two-dimensional grayscale image, which is a cloud map of the coating porosity distribution of the graphite heater.

[0075] Based on the cloud map of coating porosity and coating thickness of the graphite heater, the distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is calculated.

[0076] Furthermore, after obtaining the coating porosity distribution cloud map and coating thickness distribution cloud map of the graphite heater, the material physical parameter database is called to read the thermal radiation coefficient of the graphite matrix material on which the coating is attached. This coefficient is the thermal radiation coefficient of the matrix material. The coating structure attenuation coefficient obtained through experimental calibration for the current coating material and process is used. For each pixel coordinate in the coating porosity distribution cloud map and coating thickness distribution cloud map, the coating thickness value of that coordinate is read from the coating thickness distribution cloud map of the graphite heater, and the coating porosity value of that coordinate is read from the coating porosity distribution cloud map of the graphite heater. This calculation is repeated for all pixel coordinates in the coating porosity distribution cloud map and coating thickness distribution cloud map, and the calculation results are arranged according to the original coordinate positions to generate a new two-dimensional image. This image is the equivalent surface thermal radiation coefficient distribution field of the graphite heater.

[0077] The equivalent surface thermal radiation coefficient distribution field is expressed as:

[0078] ;

[0079] in, coordinates The equivalent surface thermal radiation coefficient distribution field at that location, The thermal emissivity of the matrix material. The coating structure attenuation coefficient, coordinates Coating thickness distribution cloud map of the graphite heater at the location, coordinates Coating porosity distribution cloud map at the location;

[0080] The distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is used as a thermal boundary condition parameter and written into the digital twin file of the preset three-dimensional model.

[0081] Furthermore, the data file of the equivalent surface thermal radiation coefficient distribution field of the graphite heater is uploaded to the central database through the data interface. The central database finds the corresponding digital twin file record containing the preset three-dimensional model based on the unique identification information of the graphite heater, and creates or updates a data field named thermal boundary condition parameter in the record. The data file of the equivalent surface thermal radiation coefficient distribution field of the graphite heater is stored under this field, thus completing the operation of writing the equivalent surface thermal radiation coefficient distribution field of the graphite heater as a thermal boundary condition parameter into the digital twin file of the preset three-dimensional model.

[0082] S4: Based on the preset 3D model, material property parameters, and thermal boundary condition parameters in the digital twin archive, construct a physical field simulation model of the graphite heater and perform calculations to generate a performance prediction report.

[0083] In step S4, the three-dimensional distribution fields of electrical conductivity and thermal conductivity in the material property parameters are assigned to the finite element mesh elements corresponding to the preset three-dimensional model to construct the physical field simulation model of the graphite heater.

[0084] Furthermore, the simulation workstation reads the preset 3D model file from the digital twin archive containing the preset 3D model, and uses finite element preprocessing software to perform tetrahedral or hexahedral meshing on the preset 3D model, generating a finite element mesh model with the same geometry as the preset 3D model. The simulation workstation loads the 3D conductivity and 3D thermal conductivity distribution field data files of the graphite heater from the material property parameter fields of the digital twin archive containing the preset 3D model. In the finite element mesh model, the 3D spatial coordinates of the center point of each volume mesh element are determined. Based on the center point coordinates of the volume mesh element, 3D spatial interpolation queries are performed in the 3D conductivity and 3D thermal conductivity distribution fields of the graphite heater to obtain the conductivity and thermal conductivity values ​​corresponding to the position of the volume mesh element. The obtained conductivity value is used as the material's electrical conductivity property, and the obtained thermal conductivity value is used as the material's thermal conductivity property, and these are assigned to the corresponding volume mesh elements respectively. After assigning the material properties to all volume mesh elements, the physical field simulation model of the graphite heater is constructed.

[0085] The equivalent surface thermal radiation coefficient distribution field in the thermal boundary condition parameters is assigned to the finite element mesh surface element of the physical field simulation model. The rated working voltage excitation is applied to the physical field simulation model and the electro-thermal coupling field equation is solved to calculate the steady-state temperature field of the graphite heater.

[0086] Furthermore, the equivalent surface thermal radiation coefficient distribution field data file of the graphite heater is loaded from the thermal boundary condition parameter field of the digital twin archive containing the preset 3D model. All surface mesh elements characterizing the outer surface of the graphite heater are identified in the finite element mesh model. The two-dimensional surface coordinates of the center point of each outer surface surface mesh element are determined. Based on the surface coordinates of the center point of the surface mesh element, a two-dimensional spatial interpolation query is performed in the equivalent surface thermal radiation coefficient distribution field of the graphite heater to obtain the equivalent thermal radiation coefficient value corresponding to the location of that surface mesh element. The obtained equivalent thermal radiation coefficient value is used as the surface thermal radiation boundary condition parameter and assigned to the corresponding outer surface surface mesh element. In the physical field simulation model, the rated operating voltage value is applied as a potential load to the specified node set of the simulated electrode connections. The multiphysics coupling solver is invoked to solve the steady-state electro-thermal coupled partial differential equations describing the Joule heating generated by current conduction and the heat conduction and surface radiation / convection heat dissipation. The solver iterates until the temperature field and electric potential field converge, and outputs the converged temperature value distribution defined on all nodes of the finite element mesh. This temperature value distribution is the steady-state temperature field of the graphite heater.

[0087] The steady-state temperature field expression is:

[0088] ;

[0089] in, For position The steady-state temperature field at that location, For position The three-dimensional distribution field of conductivity at that location, For position The three-dimensional distribution field of thermal conductivity at that location, Excitation is performed at the rated operating voltage. For the steady-state solution operator.

[0090] The highest temperature, lowest temperature, maximum temperature difference, and hot spot location are extracted from the steady-state temperature field as key performance indicators to generate a performance prediction report.

[0091] Furthermore, in the solved steady-state temperature field of the graphite heater, the temperature values ​​of all nodes are compared, and the maximum and minimum values ​​are identified and recorded as the highest and lowest temperatures, respectively. The difference between the highest and lowest temperatures is recorded as the maximum temperature difference. The spatial coordinates of the node whose temperature value equals the highest temperature are located; these coordinates are the hotspot locations. These extracted information on the highest temperature, lowest temperature, maximum temperature difference, and hotspot locations, along with an overall overview of the steady-state temperature field of the graphite heater, are organized into a structured document to generate a performance prediction report. The performance prediction report is then stored back in the corresponding fields of a digital twin archive containing a pre-defined 3D model.

[0092] S5: Integrates multi-source monitoring data and performance prediction reports from digital twin archives to make a comprehensive quality judgment on graphite heaters based on preset rules.

[0093] In step S5, the key performance indicators, three-dimensional geometric deviation data, and multi-point resistance values ​​in the performance prediction report are logically correlated.

[0094] Furthermore, the decision server reads performance prediction reports, 3D geometric deviation data, and multi-point resistance values ​​from the digital twin archive containing the preset 3D model. The decision server generates a unique decision data record for the digital twin archive containing the preset 3D model. The decision server combines and aligns the key performance indicators from the performance prediction report—including maximum temperature, minimum temperature, maximum temperature difference, and hotspot locations—with the statistical characteristic values ​​of the 3D geometric deviation data (such as maximum positive deviation, maximum negative deviation, and average deviation) and the statistical characteristic values ​​of the multi-point resistance values ​​(such as average resistance and resistance range) according to a preset data structure format, forming a logically associated multi-source data record. This logical association process ensures that all different dimensions of the quality parameters of the same graphite heater have a corresponding relationship in the decision data record.

[0095] Based on a preset quality rule base, perform multi-level threshold comparisons on logically correlated multi-source data;

[0096] Furthermore, the decision server retrieves a pre-defined quality rule library from a digital twin archive containing a pre-defined 3D model. This pre-defined quality rule library defines the judgment levels, judgment thresholds, and judgment logic for each quality parameter. Based on the pre-defined quality rule library, the decision server checks each logically correlated multi-source data record item by item. For example, the first-level rule of the pre-defined quality rule library specifies comparing the maximum temperature difference in the performance prediction report with the first-level threshold temperature difference in the rule library to determine if the maximum temperature difference is less than the first-level threshold temperature difference. The second-level rule specifies comparing the resistance range in the multi-point resistance values ​​with the second-level threshold resistance range in the rule library to determine if the resistance range is less than the second-level threshold resistance range. The third-level rule specifies comparing the maximum positive deviation in the 3D geometric dimension deviation data with the third-level threshold dimension deviation in the rule library to determine if the maximum positive deviation is less than the third-level threshold dimension deviation. The decision server sequentially executes all multi-level threshold comparison operations defined in the pre-defined quality rule library.

[0097] Based on the results of multi-level threshold comparison, a comprehensive quality assessment result for the graphite heater is obtained.

[0098] Furthermore, the decision server aggregates the results of all multi-level threshold comparisons. Based on the overall judgment logic defined in the preset quality rule base, for example, if all level comparisons pass, the product is judged as superior; if the first and second levels pass but the third level fails, the product is judged as qualified; and if any core level, such as the first level, fails, the product is judged as unqualified. The decision server performs logical operations based on the aggregated multi-level threshold comparison results and the preset quality rule base's overall judgment logic, outputting a clear judgment conclusion. This conclusion yields the comprehensive quality judgment result for the graphite heater, such as outputting "qualified" or "unqualified."

[0099] Using the pre-set quality rule library in the digital twin archive of the pre-set 3D model, the comprehensive quality judgment of the multi-source data after logical association is performed based on the quality rule library to obtain the comprehensive quality judgment result of the graphite heater.

[0100] Furthermore, the decision server writes the overall quality assessment result of the graphite heater into a designated field of a digital twin file containing a preset 3D model, and sends a trigger signal to the actuators on the production line, such as sorting robots or inkjet printers, notifying the actuators to perform corresponding subsequent operations based on the overall quality assessment result of the graphite heater.

[0101] Example 2

[0102] like Figure 1 As shown, this embodiment provides a method for online quality monitoring of a graphite heater production line, including driving an actuator to sort graphite heaters according to the comprehensive quality judgment result, and archiving digital twin files of three-dimensional geometric dimension deviation data, multi-point resistance values, three-dimensional conductivity distribution field, three-dimensional thermal conductivity distribution field, equivalent surface thermal radiation coefficient distribution field, performance prediction report and comprehensive quality judgment result to form a traceable product quality data chain.

[0103] Furthermore, the overall quality assessment result of the graphite heater generated by the decision server is sent to the programmable logic controller (PLC) at the end of the production line via an industrial fieldbus or real-time Ethernet protocol. The PLC parses the received overall quality assessment result of the graphite heater and, according to the pre-programmed sorting logic, drives the corresponding pneumatic sorting valve, robotic arm, or conveyor belt branch to perform actions. For example, when the overall quality assessment result of the graphite heater is excellent, the sorting valve of channel A is controlled to open, guiding the graphite heater to the excellent product storage area; when the overall quality assessment result of the graphite heater is qualified, the sorting valve of channel B is controlled to open, guiding it to the qualified product area; when the overall quality assessment result of the graphite heater is unqualified, it is guided to the return area. The system is designed to process graphite heaters in a repair or scrap area, thereby enabling the execution mechanism to perform corresponding sorting and archiving based on the comprehensive quality judgment results. The server is then activated to extract three-dimensional geometric dimension deviation data, multi-point resistance values, three-dimensional conductivity distribution field, three-dimensional thermal conductivity distribution field, equivalent surface thermal radiation coefficient distribution field, performance prediction report, and comprehensive quality judgment results from the digital twin archive containing the preset three-dimensional model. This data is packaged according to a preset encrypted data structure containing timestamps and unique product identifiers to generate an immutable data packet. This data packet is then uploaded to an enterprise-level cloud database or blockchain storage platform for permanent storage. This data packet is permanently associated with the physical entity of the graphite heater through a unique identifier, forming a graphite heater product quality data chain.

[0104] To quantitatively illustrate the advantages of this invention over traditional methods during the research and development stage, we take the optimization of graphitization sintering process as an example to compare the efficiency and accuracy of the two methods in identifying process-structure-performance correlations and guiding process iteration.

[0105] Traditional R&D methods

[0106] Objective: To optimize the axial temperature gradient of the sintering furnace to improve the axial thermal uniformity of the graphite heater samples.

[0107] Steps: Design three different axial temperature curves for processes A, B, and C, and sinter 3 samples for each group, for a total of 9 samples.

[0108] Evaluation method:

[0109] Destructive sampling: A tiny cylinder was taken from each of the two ends and the center of each sample along its axis and subjected to destructive XRD testing to obtain the graphitization degree values ​​at three discrete points. , and

[0110] Uniformity assumption: Take the average of three points. The degree of graphitization of the sample was used as a representative parameter, and the corresponding average electrical conductivity and thermal conductivity were then retrieved from the database.

[0111] Simulation prediction: Based on the average material properties, a finite element model of a uniform material is constructed to predict the steady-state temperature field of the sample under rated voltage and obtain the maximum axial temperature difference.

[0112] Actual verification: All nine samples were subjected to electric heating tests, and their steady-state temperature field and maximum axial temperature difference were measured.

[0113] Results and Challenges: Data analysis revealed that process B had the highest average graphitization degree and the smallest predicted temperature difference using its uniform model. However, actual measurements showed that two of the three samples from process B had measured temperature differences greater than the predicted values, exhibiting significant dispersion. Since traditional methods only obtain the graphitization degree at three discrete points, they cannot reconstruct the complete axial distribution. Therefore, it is impossible to explain why the process with the best average performance produced unstable and sometimes poor-performing results. It is also impossible to determine which region along the axial direction exhibits the material performance bottleneck, leading to a standstill in process optimization.

[0114] This plan

[0115] This method was applied to nine samples from the same batch, after sintering and before electrical testing.

[0116] Key steps and data acquisition:

[0117] Non-destructive global scanning: Obtain the complete three-dimensional distribution field of graphitization degree through microwave resonant cavity scanning.

[0118] Distribution field analysis: Taking a sample with a large measured temperature difference from process B as an example, its axial distribution curve of graphitization degree shows that there is a significant trough in the central region, where the graphitization degree is about 15% lower than at both ends.

[0119] High-fidelity simulation: Input the electrical conductivity and thermal conductivity distribution fields, which reflect the above non-uniform distribution, into the simulation model, and use hyperspectral scanning to obtain coating uniformity data as boundary conditions.

[0120] Performance prediction: Temperature difference predicted by a high-fidelity model Compared with actual measurement Highly compatible.

[0121] Quantitative comparison and results:

[0122] Improving prediction accuracy: Defining the prediction error rate %.

[0123] The average error rate of the traditional uniform model for 9 samples = 22.5%.

[0124] The average error rate of the high-fidelity model of this invention = 4.8%.

[0125] The formula for relative improvement in prediction accuracy is:

[0126] % = (22.5% -4.8%) / 22.5% × 100% ≈ 78.7%.

[0127] Root cause analysis and process iteration: This solution identifies a brief temperature drop in the central region of the axial temperature curve caused by process B, which is the direct cause of the saddle-shaped graphitization distribution. This is a continuous distribution pattern that traditional three discrete point data cannot reveal. Based on this, the temperature control curve in the central region of process B was precisely adjusted. The samples under the new process have a flat graphitization distribution curve, and the measured thermal uniformity is significantly improved and stable.

[0128] Improved R&D efficiency: Traditional methods require redesigning and sintering multiple sets of experiments to guess the cause when encountering contradictory data. This solution, through a single non-destructive test, directly locates the spatial distribution of microstructural defects, shortening the cycle of root cause analysis from the usual 2-3 iterations to a single iteration.

[0129] in conclusion

[0130] This embodiment demonstrates that, during the research and development phase of graphite heaters, this solution non-destructively obtains the spatial distribution field of the material's intrinsic properties and drives high-fidelity physical field simulation. This solves the problems of missing microstructural spatial information and distorted performance predictions caused by traditional methods that rely on destructive sampling and homogeneity assumptions. Specific quantitative benefits include: reducing the simulation prediction error rate from 22.5% to 4.8%, improving prediction accuracy by approximately 78.7%, and achieving spatial localization of process defects. This significantly shortens the research and development iteration cycle of process-structure-performance correlation analysis, and substantially improves the accuracy and efficiency of process optimization.

[0131] Example 3

[0132] According to one aspect of the present invention, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the methods provided in the various alternative implementations described above.

[0133] In another aspect, the present invention also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to implement the online quality monitoring method for the graphite heater production line described in the above embodiments.

[0134] It should be noted that although several modules or units of the device for performing actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0135] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of the present invention.

[0136] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein.

[0137] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for on-line monitoring of the quality of a graphite heater production line, characterized in that, The specific steps include: creating and binding a digital twin file of a preset 3D model when the graphite heater blank enters the starting point of the production line; After machining is completed, the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater are obtained; The three-dimensional distribution field of intrinsic properties is obtained through non-contact full-domain scanning measurement technology and written into the digital twin archive as material property parameters. The two-dimensional distribution cloud map of intrinsic properties is obtained through non-contact spectral imaging technology and converted into thermal boundary condition parameters. Based on the preset three-dimensional model, material property parameters and thermal boundary condition parameters in the digital twin archive, a physical field simulation model of the graphite heater is constructed and calculated to generate a performance prediction report. By integrating multi-source monitoring data and performance prediction reports from digital twin archives, a comprehensive quality assessment of graphite heaters is conducted based on preset rules.

2. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The digital twin archive includes, When the graphite heater blank arrives at the starting point of the production line, the identification information of the graphite heater blank arriving at the starting point of the production line is collected. A digital twin file containing a preset 3D model is created using the identification information of the graphite heater blank, and the digital twin file of the preset 3D model is bound to the graphite heater blank.

3. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The acquisition of the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater includes, The machined graphite heater is transferred to the measurement station, where a three-dimensional optical scan is performed to generate a point cloud model of the graphite heater. The point cloud model of the graphite heater is compared with the preset 3D model to calculate the 3D geometric dimension deviation data of the graphite heater. A four-probe tester was used to contact multiple preset points on the surface of the graphite heater to measure the multi-point resistance value of the graphite heater. The three-dimensional geometric deviation data and multi-point resistance value of the graphite heater were recorded as the three-dimensional geometric dimensions and multi-point resistivity data of the graphite heater.

4. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The acquisition of the three-dimensional distribution field of intrinsic properties includes... The graphite heater is transferred to the microwave resonant cavity scanning station, and the microwave resonant cavity is controlled to perform spatial point-by-point scanning on the graphite heater located at the microwave resonant cavity scanning station to collect the microwave resonant cavity parameter dataset. The microwave resonance parameter dataset of the microwave resonant cavity is input into the pre-trained microwave response-material microstructure inversion model to calculate the three-dimensional distribution field of graphitization degree of the graphite heater. Based on the three-dimensional distribution field of graphitization degree of the graphite heater, the three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are derived. The three-dimensional distribution fields of electrical conductivity and thermal conductivity of the graphite heater are written into the digital twin file of the preset three-dimensional model as material property parameters.

5. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The parameters converted into thermal boundary conditions include The graphite heater is transferred to the hyperspectral imaging scanning station, and the surface of the graphite heater coating located at the hyperspectral imaging scanning station is subjected to push-broom imaging to obtain the hyperspectral data cube of the graphite heater. The hyperspectral data cube of the graphite heater is input into the pre-trained spectral feature-coating physical property inversion model to resolve the coating thickness distribution cloud map and coating porosity distribution cloud map of the graphite heater. Based on the cloud map of coating porosity and coating thickness of the graphite heater, the distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is calculated. The distribution field of the equivalent surface thermal radiation coefficient of the graphite heater is used as a thermal boundary condition parameter and written into the digital twin file of the preset three-dimensional model.

6. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The performance prediction report includes, The three-dimensional distribution fields of electrical conductivity and thermal conductivity in the material property parameters are assigned to the finite element mesh elements corresponding to the preset three-dimensional model to construct the physical field simulation model of the graphite heater. The equivalent surface thermal radiation coefficient distribution field in the thermal boundary condition parameters is assigned to the finite element mesh surface element of the physical field simulation model. The rated working voltage excitation is applied to the physical field simulation model and the electro-thermal coupling field equation is solved to calculate the steady-state temperature field of the graphite heater. The highest temperature, lowest temperature, maximum temperature difference, and hot spot location are extracted from the steady-state temperature field as key performance indicators to generate a performance prediction report.

7. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The multi-source monitoring data and performance prediction reports in the integrated digital twin archive include, Logically link the key performance indicators, three-dimensional geometric deviation data, and multi-point resistance values ​​in the performance prediction report; Based on a preset quality rule base, perform multi-level threshold comparisons on logically correlated multi-source data; Based on the results of multi-level threshold comparison, a comprehensive quality assessment result for the graphite heater is obtained.

8. The online quality monitoring method for a graphite heater production line according to claim 1, characterized in that, The comprehensive quality assessment includes, Using the pre-set quality rule library in the digital twin archive of the pre-set 3D model, the comprehensive quality judgment of the multi-source data after logical association is performed based on the quality rule library to obtain the comprehensive quality judgment result of the graphite heater.

9. The online quality monitoring method for a graphite heater production line according to claim 8, characterized in that, Based on the comprehensive quality judgment results, the actuator is driven to sort the graphite heaters accordingly. The digital twin files of three-dimensional geometric dimension deviation data, multi-point resistance values, three-dimensional conductivity distribution field, three-dimensional thermal conductivity distribution field, equivalent surface thermal radiation coefficient distribution field, performance prediction report and comprehensive quality judgment results are archived to form a traceable product quality data chain.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the online quality monitoring method for graphite heater production line according to any one of claims 1-8.

Images