A method for real-time evaluation of uniformity of flux in smelting flux production

Abstract

The application discloses a kind of real-time evaluation methods of flux uniformity in smelting flux production, it is related to data processing technical field, comprising: by constructing melt solidification impedance layer distribution in thermal radiation image inside particle;According to melt solidification impedance layer distribution, the composition segregation boundary in the inside of flux particle is identified;Based on composition segregation boundary and thermal radiation image, generate segregation position mark;The time sequence corresponding verification of segregation position mark and the microstructure evolution process of flux particle is carried out, and the cause-effect chain is established;When the cause-effect chain appears chain fracture, according to the recapture adjustment melt solidification impedance layer distribution of thermal conduction shunt in thermal radiation image;Based on adjusted melt solidification impedance layer distribution, determine the composition uniformity of flux particle;The application realizes the high-precision evaluation of flux particle composition uniformity by monitoring the thermal radiation characteristics of flux particle, and accurately analyzing the dynamic evolution of melt solidification impedance layer and composition segregation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a real-time evaluation method for uniformity of flux in smelting flux production. BACKGROUND

[0002] As a key material in the welding process, the composition uniformity of smelting flux directly affects the weld quality, welding stability and the performance of the final product. Especially the smelting flux used in the submerged arc welding process, due to the difference in melting and solidification rate of different components in the smelting process, it is easy to form local component segregation inside the particle. In the prior art, the composition uniformity of the flux particle is mainly evaluated offline through sampling analysis or post-weld metallographic examination, which cannot realize real-time detection. This offline evaluation method has obvious hysteresis and randomness, which cannot timely find and control the factors leading to component segregation in the smelting process of the flux, thereby affecting the consistency and quality stability of the smelting flux production.

[0003] In addition, the existing evaluation method of flux uniformity is mostly based on post-weld macroscopic performance or laboratory metallographic observation, without considering the dynamic evolution process of the microstructure formation of the flux particle in the early stage of solidification, and lacking of fine real-time monitoring and analysis means for the correlation between the thermophysical phenomena and the microstructure composition in the smelting process. This defect makes it difficult for the traditional method to accurately capture the key transient information (such as the change of heat conduction path in the early stage of solidification and the formation mechanism of internal solidification resistance layer) in the solidification process of the flux particle, resulting in insufficient reliability and accuracy of the evaluation results.

[0004] Therefore, there is an urgent need for an evaluation method that can real-time monitor and accurately analyze the solidification process of the internal melt of the flux particle, can quickly locate the component segregation area based on the microstructure solidification characteristics, so as to realize high-precision real-time evaluation of the composition uniformity of the flux particle, guide and optimize the flux production process, and significantly improve the stability, consistency and reliability of the flux quality. SUMMARY

[0005] The present application aims at at least solving one of the technical problems existing in the prior art; for this purpose, the present application provides a real-time evaluation method for uniformity of flux in smelting flux production.

[0006] To achieve the above-mentioned purpose, the present application provides a real-time evaluation method for uniformity of flux in smelting flux production, comprising:

[0007] Collecting a thermal radiation image of the flux particle at the moment of discharging from the furnace; constructing the distribution of the solidification resistance layer of the internal melt of the particle through the heat conduction shunt phenomenon of the particle surface in the thermal radiation image;

[0008] The composition segregation boundary inside the flux particle is identified according to the spatial segmentation mode of the melt solidification impedance layer distribution; and the segregation position marker is generated based on the spatial overlap relationship between the composition segregation boundary and the temperature change track in the thermal radiation image;

[0009] The segregation position marker is time-correspondingly verified with the microstructure evolution process of the flux particle, and a causal chain of composition segregation and microstructure change is established; when the causal chain appears chain fracture, the melt solidification impedance layer distribution is adjusted according to the recapture of heat conduction shunt in the thermal radiation image;

[0010] The composition uniformity grade of the flux particle is determined based on the adjusted melt solidification impedance layer distribution, and a uniformity detection result is output.

[0011] Further, the melt solidification impedance layer distribution inside the particle is constructed by the heat conduction shunt phenomenon of the particle surface in the thermal radiation image, comprising:

[0012] The main flow channel of heat transfer on the surface of the flux particle is identified in the thermal radiation image, and the region with the maximum temperature gradient in the temperature field on the surface of the particle is marked as the main flow channel of heat transfer;

[0013] The heat conduction shunt phenomenon occurring in the propagation process of the main flow channel of heat transfer is tracked, and when the main flow channel is dispersed into multiple conduction branches at a position, the position is marked as a heat conduction shunt point;

[0014] The impedance layer position in the melt solidification process inside the particle is inferred by the spatial distribution rule of the heat conduction shunt point, and each heat conduction shunt point corresponds to a boundary position of a solidification impedance layer;

[0015] The spatial profile of the impedance layer is formed by connecting the boundary positions of adjacent solidification impedance layers, and the melt solidification impedance layer distribution is constructed.

[0016] Further, the identification of the heat conduction shunt phenomenon is based on the geometric shape change of the main flow channel of heat transfer, and when the width of the main flow channel increases and is dispersed into multiple narrow channels in the propagation process, the shape change position is determined as a heat conduction shunt point.

[0017] Further, the generation of the segregation position marker comprises:

[0018] The spatial continuity state of the impedance layer in the melt solidification impedance layer distribution is analyzed, and the position where the impedance layer appears spatial fracture is identified;

[0019] The position of the spatial fracture of the impedance layer is taken as the position where the composition separation occurs in the melt solidification process, and the fracture position corresponds to the boundary line of different composition regions inside the particle;

[0020] A complete boundary line network is formed by connecting adjacent fracture positions, and the boundary line network is determined as the composition segregation boundary inside the flux particle.

[0021] extracting a temperature variation trajectory of a corresponding region in the thermal radiation image, and combining the spatial coordinates of the segregation boundary with the temperature variation trajectory to generate a segregation position marker.

[0022] Further, the identification of the spatial fracture of the impedance layer is based on a geometric discontinuity in the profile of the impedance layer, and when the profile of the impedance layer has a geometric discontinuity or absence during spatial extension, the discontinuity position is determined as the spatial fracture position of the impedance layer.

[0023] Further, the extraction of the temperature variation trajectory is based on a time sequence evolution pattern of the temperature field in the thermal radiation image, and when the temperature of a certain region shows a continuous decline or fluctuation feature in the time sequence, the temperature evolution path of the region is taken as the temperature variation trajectory.

[0024] Further, the time sequence correspondence verification of the segregation position marker with the microstructure evolution process of the flux particles establishes a causal chain of composition segregation and microstructure change, including:

[0025] Obtaining microstructure information of the flux particles at different times during solidification, and identifying the evolution stage of the microstructure;

[0026] Establishing a time correspondence relationship between the formation time of the composition segregation boundary and the time of the microstructure evolution stage, and analyzing the time sequence synchronization of segregation occurrence and microstructure change;

[0027] According to the degree of coincidence of the time sequence synchronization, the causal correlation strength of the composition segregation and the microstructure evolution is evaluated;

[0028] Based on the evaluation results of the causal correlation strength, a causal chain of composition segregation and microstructure change is established.

[0029] Further, the microstructure evolution stage includes primary phase precipitation, eutectic reaction and solid state phase transition, and the microstructure information is obtained by metallographic analysis of the flux particles, and the evolution stage is identified based on the formation and transformation characteristics of different phase structures.

[0030] Further, the adjustment of the melt solidification impedance layer distribution according to the recapture of the heat conduction shunt in the thermal radiation image includes:

[0031] Identifying the time node of chain fracture in the causal chain of composition segregation and microstructure change;

[0032] Reanalyzing the occurrence mode of heat conduction shunt in the thermal radiation image at the time node of chain fracture, and tracking the spatial evolution trajectory of the shunt phenomenon;

[0033] Based on the reanalysis results of the spatial evolution trajectory of the shunt phenomenon, the position distribution of the heat conduction shunt point is corrected;

[0034] Based on the corrected location of the heat conduction diversion point, the spatial profile of the solidification resistance layer is reconstructed, and the distribution of the melt solidification resistance layer is adjusted.

[0035] Furthermore, the identification of chain breakage is based on the degree of mismatch between the formation time of component segregation and the evolution time of microstructure. When the difference between the formation time of the segregation boundary and the occurrence time of the corresponding microstructure evolution stage exceeds the time cycle of the flux particle solidification process, it is determined to be a causal chain breakage.

[0036] Compared with the prior art, the beneficial effects of the present invention are:

[0037] This invention acquires thermal radiation images at the moment flux particles exit the furnace and constructs the melt solidification impedance layer distribution by utilizing the shunting phenomenon of heat conduction on the particle surface. This allows for real-time and accurate reflection of the microstructure evolution process of the melt in the early stage of solidification within the flux particles. It effectively overcomes the problem that traditional offline detection methods cannot capture the initial dynamic characteristics of the solidification process in real time, and realizes real-time monitoring of the flux production process.

[0038] This invention analyzes the geometric continuity interruption characteristics in the spatial distribution of the solidification impedance layer to accurately identify the spatial boundaries of melt component segregation inside flux particles, and combines this with spatial matching based on temperature change trajectories. This effectively improves the accuracy and real-time performance of identifying component segregation regions inside flux particles, overcoming the shortcomings of traditional methods that rely solely on post-conductivity metallographic testing, which leads to detection lag and position misjudgment.

[0039] This invention establishes a temporal causal chain between segregation location markers and microstructure evolution, and dynamically adjusts the spatial distribution of the solidification impedance layer based on the chain breakage. This ensures the consistency and reliability between component segregation boundaries and microstructure changes, thereby significantly improving the accuracy and robustness of real-time evaluation of flux particle composition uniformity. It provides precise data support for optimizing flux smelting production processes and ensuring flux quality stability. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0042] The technical solutions of the present application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0043] Please refer to Figure 1 The present embodiment provides a real-time evaluation method for the uniformity of flux in the production of smelting flux, comprising:

[0044] S101: Collecting the thermal radiation image of the flux particles at the moment of discharging from the smelting furnace; constructing the distribution of the solidification resistance layer of the melt inside the particles through the heat conduction shunt phenomenon of the particle surface in the thermal radiation image;

[0045] It should be noted that in the present embodiment, at the moment of discharging from the smelting furnace, the surface of the flux particles has obvious thermal radiation phenomenon. In the present embodiment, the thermal radiation image of the flux particles at the moment of discharging from the smelting furnace is collected by an infrared thermal imager, which is used for subsequent analysis of the solidification process of the melt inside the particles.

[0046] In specific implementation, the infrared thermal imager should be installed at a position directly above the discharging port of the flux particles, and image collection should be performed immediately after the flux particles enter the field of view of the instrument, so as to accurately capture the real-time thermal radiation distribution on the surface of the particles. The infrared thermal radiation image of the flux particles at the moment of discharging from the smelting furnace is recorded in the form of gray value, and is converted into real temperature field data according to the calibration parameters of the equipment, and the specific conversion formula is:

[0047]

[0048] In the formula: represents the actual temperature value at the coordinate position of the surface of the flux particles; represents the gray value at the position in the infrared thermal radiation image; represents the actual temperature value at the coordinate position of the surface of the flux particles; represents the gray value at the position in the infrared thermal radiation image; and are respectively the calibration coefficients provided by the infrared thermal imager when it is shipped, and the specific values are determined according to the factory calibration manual of the thermal imager.

[0049] In specific implementation, the high-precision temperature field data of the flux particles at the moment of discharging from the smelting furnace is obtained through the above steps, which is used to analyze the heat conduction shunt phenomenon on the surface of the particles.

[0050] In the specific implementation process, the construction of the distribution of the solidification resistance layer of the melt inside the particles through the heat conduction shunt phenomenon of the particle surface in the thermal radiation image comprises:

[0051] Identifying the main flow channel of heat transfer on the surface of the flux particles in the thermal radiation image, and marking the region with the largest temperature gradient in the temperature field of the particle surface as the main flow channel of heat transfer;

[0052] It can be understood that the heat transfer on the surface of the flux particles has obvious spatial directionality. The embodiment first performs spatial gradient calculation on the temperature field data to identify the main flow channel of heat transfer on the surface of the particles, specifically using a two-dimensional temperature field spatial gradient calculation formula:

[0053]

[0054] In the formula: represents the spatial gradient value of the temperature field at the position coordinate; respectively represent the partial derivatives of the temperature field in the horizontal direction (x direction) and the vertical direction (y direction), which are calculated using the central difference method, and the calculation formula is as follows:

[0055] In the formula: ,

[0056] is the actual physical resolution of the image space, that is, the spatial distance between pixels in the image. Exemplarily, after obtaining the temperature field spatial gradient distribution, the embodiment uses a local maximum value search algorithm, such as a non-maximum suppression algorithm, to automatically identify and mark the region with the maximum and continuous temperature gradient, and determine it as the main flow channel of heat transfer on the surface of the flux particles.

[0057] Tracking the heat conduction shunt phenomenon occurring in the propagation process of the main flow channel of heat transfer, when the main flow channel is dispersed into multiple conduction branches at a position, the position is marked as a heat conduction shunt point;

[0058] It should be understood that, in the cooling process of the flux particles, due to the non-uniformity of the internal composition and temperature field of the particles, the main flow channel of heat transfer on the surface may appear dispersion phenomenon in the spatial propagation process. The embodiment uses image thinning technology to perform skeleton analysis on the main flow channel and calculates the local width change of the main flow channel to identify the heat conduction shunt phenomenon. The specific implementation is as follows:

[0059] First, the image thinning process is performed on the main flow channel region identified in the foregoing to obtain a skeleton curve with a single-pixel width; then, the local width of the main flow channel is calculated using a skeleton analysis method, and the specific definition formula is as follows:

[0060] In the formula: represents the local width of the main flow channel; represents the area of the main flow channel in the corresponding region; and L represents the length of the skeleton line of the corresponding region.

[0061] ​​Specifically, the identification of the heat conduction shunt phenomenon is based on the geometric shape change of the heat transfer main flow channel. When the width of the main flow channel increases during propagation and is dispersed into multiple narrow channels, the shape change position is determined as the heat conduction shunt point;

[0062] It should be noted that the present embodiment identifies the heat conduction shunt point by monitoring the local width change of the main flow channel. When the width of the main flow channel (W0) ) increases at a certain position in the spatial propagation direction and splits into two or more narrow branches that can be clearly distinguished, the position is determined as the heat conduction shunt point. Exemplarily, the specific determination condition can be set as: on the propagation path, when the increase of the local width value compared to the initial channel width reaches 20% , and the main flow channel skeleton line after the position splits into at least two independent channel branches in the propagation direction, the spatial position is marked as the heat conduction shunt point, and its spatial coordinate position is recorded;

[0063] Specifically, the judgment standard for channel splitting is that the continuous pixel points of the main flow channel skeleton line no longer form a single path after the position, but at least two path branches appear, and the included angle between any two path branches satisfies (the included angle threshold can be determined according to actual experiments), and the independent extension length of each path branch in space is greater than twice the initial average width of the main flow channel.

[0064] The impedance layer position in the melt solidification process inside the particle is inferred from the spatial distribution of the heat conduction shunt points. Each heat conduction shunt point corresponds to the boundary position of a solidification impedance layer.

[0065] It should be noted that during the melt solidification process inside the flux particle, the composition or temperature difference of different regions will cause the heat transfer path to exhibit obvious heat conduction shunt phenomenon. Through the spatial distribution of the heat conduction shunt points identified in the foregoing, the present embodiment infers the impedance layer position in the solidification process inside the particle.

[0066] In the specific implementation process, first, the spatial position coordinates of each heat conduction shunt point obtained in the foregoing steps are collected as:

[0067]

[0068] In the formula: represents the set of spatial coordinates of all identified heat conduction shunt points; represents the spatial position coordinates of the i-th heat conduction shunt point; n represents the total number of heat conduction shunt points.

[0069] It should be appreciated that each heat conduction diversion point corresponds to a specific position of the boundary of the impedance layer during the solidification process of the melt inside the flux particle, i.e., the melt solidification interface during the solidification process of the melt due to the difference in thermal conductivity or solidification rate in the local region.

[0070] Further, to accurately represent the spatial structure of the melt solidification impedance layer inside the particle, the three-dimensional coordinate distribution of the diversion points is analyzed in the embodiment. Specifically, to obtain the three-dimensional spatial coordinates required for subsequent analysis , the embodiment needs to reconstruct the three-dimensional model of the flux particle based on multi-view infrared thermal radiation images. It should be noted that the specific steps of the multi-view infrared image reconstruction are as follows:

[0071] First, at the moment when the flux particle is discharged, multiple infrared thermal imagers are arranged at different spatial orientations of the particle discharge port, and multiple two-dimensional thermal radiation images at different viewing angles are synchronously collected.

[0072] Then, based on the multi-view two-dimensional thermal radiation images, a three-dimensional surface model of the flux particle is obtained through a three-dimensional reconstruction algorithm (such as spatial geometric projection transformation, three-dimensional spatial coordinate mapping, and image fusion algorithm), and the temperature data in the corresponding two-dimensional thermal radiation images are mapped onto the constructed three-dimensional surface model, thereby determining the three-dimensional temperature field distribution on the surface of the flux particle. The specific three-dimensional temperature field distribution function is represented as:

[0073]

[0074] In the formula: represents the temperature value at the spatial coordinates on the surface of the flux particle; represents the mapping function from the two-dimensional thermal radiation image to the three-dimensional surface temperature field, which can specifically adopt a spatial mapping and interpolation fusion algorithm represents the two-dimensional thermal radiation image data collected at different viewing angles.

[0075] The boundary positions connecting adjacent solidification impedance layers form the spatial profile of the impedance layer, and the distribution of the melt solidification impedance layer is constructed.

[0076] Exemplarily, the three-dimensional spatial coordinates of the heat conduction diversion points obtained above are used to connect the boundary positions of each impedance layer by using an interpolation method, thereby forming the spatial profile of the impedance layer and constructing the complete spatial distribution of the melt solidification impedance layer.

[0077] Specifically, the radial basis function (Radial Basis Function, RBF) interpolation algorithm is used to construct the spatial profile between the heat conduction diversion points. The RBF interpolation formula is as follows:

[0078]

[0079] wherein: is a melt solidification impedance layer spatial profile function constructed by interpolation; is a weight coefficient to be determined in the interpolation calculation, obtained by the undetermined coefficient method; is a shape parameter in the radial basis function, and the specific value is determined according to the actual requirement of the interpolation fitting accuracy; is a position to be interpolated is the Euclidean space distance from the position to the i-th heat conduction shunt point position.

[0080] It can be understood that the spatial profile function constructed by the above interpolation can clearly describe the spatial position distribution of the internal solidification impedance layer of the flux particle, and is used as important basic data for subsequent judgment of the uniformity of the internal composition of the particle.

[0081] S102: identifying the composition segregation boundary inside the flux particle according to the spatial fragmentation pattern of the melt solidification impedance layer distribution; generating a segregation position marker based on the spatial overlap relationship between the composition segregation boundary and the temperature change trajectory in the thermal radiation image;

[0082] It should be noted that the spatial distribution of the melt solidification impedance layer inside the flux particle reflects the difference characteristics of the local composition during the solidification process of the flux particle. When the spatial continuity of the impedance layer appears obvious fragmentation characteristics, the fragmentation position represents the interface position between the melts of different compositions inside the particle, which is called the composition segregation boundary.

[0083] In the specific implementation process, based on the spatial distribution characteristics of the melt solidification impedance layer constructed in the above step S101, the spatial fragmentation position in the distribution of the melt impedance layer is identified to determine the spatial position of the composition segregation boundary, and the spatial overlap analysis is performed in combination with the temperature change trajectory information of the corresponding area in the thermal radiation image, and finally the composition segregation position marker is determined.

[0084] In the specific implementation process, the generation of the segregation position marker includes:

[0085] analyzing the spatial continuity state of the impedance layer in the melt solidification impedance layer distribution, and identifying the position where the impedance layer appears spatial fracture;

[0086] Specifically, the identification of the spatial fracture of the impedance layer is based on the geometric continuity interruption of the impedance layer profile. When the impedance layer profile appears geometric interruption or absence in the spatial extension process, the interruption position is determined as the spatial fracture position of the impedance layer.

[0087] It should be noted that the melt solidification impedance layer spatial profile function Spatial continuity analysis is performed. Specifically, the continuity feature of the impedance layer profile in three-dimensional space is detected by calculating the spatial gradient field of the impedance layer profile function The spatial gradient calculation formula is:

[0088]

[0089] In the formula: are the spatial partial derivatives of the impedance layer spatial profile function in three-dimensional space directions, which are obtained by the central difference method.

[0090] Further, to ensure the consistency of the impedance layer spatial fracture position determination, the embodiment needs to meet the following two conditions when identifying the geometric continuity interruption position of the impedance layer profile:

[0091] Condition one: the absolute value of the spatial gradient of the impedance layer profile function at the above spatial position exceeds the set continuity determination threshold :

[0092]

[0093] Condition two: there is a clear geometric connection interruption or absence along the main propagation direction of the impedance layer profile at the position meeting condition one; in specific implementation, the Euclidean distance between the two adjacent position points along the main propagation direction of the impedance layer profile exceeds the set distance threshold , that is:

[0094]

[0095] In the formula: is the current spatial position; is the next adjacent spatial position along the main propagation direction of the impedance layer profile; is the distance determination threshold of the spatial continuity geometric interruption, and the specific value is determined through actual statistical analysis.

[0096] The position of the impedance layer spatial fracture is taken as the occurrence position of the component separation in the melt solidification process, and the fracture position corresponds to the boundary line of different component regions inside the particle;

[0097] It should be noted that the impedance layer spatial fracture position identified by the above represents the interface of different component melt regions inside the flux particle; the positions are directly taken as the occurrence positions of the component separation in the melt solidification process in the embodiment to clearly define the boundary of different component regions inside the particle.

[0098] Specifically, the spatial coordinate set of the impedance layer spatial fracture position is represented as:

[0099]

[0100] In the formula: is a set of coordinates of the spatial fracture position of the impedance layer; represents the three-dimensional coordinates of the spatial fracture position of the impedance layer; is the total number of the identified spatial fracture positions. It should be understood that the set of spatial fracture positions of the impedance layer determined above provides basic data for the spatial positioning of the melt composition boundary inside the flux particles, and directly reflects the composition segregation interface occurring in the melt solidification process.

[0101] Connecting adjacent fracture positions forms a complete boundary line network, and the boundary line network is determined as the composition segregation boundary inside the flux particles;

[0102] Exemplarily, the embodiment connects the set of identified spatial fracture positions of the impedance layer

[0103] using a three-dimensional interpolation method to form a continuous spatial boundary line network structure, and clearly defines the specific spatial position of the composition segregation boundary inside the particles.

[0104] Specifically, the embodiment selects a three-dimensional spline interpolation algorithm to connect adjacent spatial fracture positions, and the interpolation function is expressed as follows:

[0105]

[0106] In the formula: represents the formed spatial boundary line network function, specifically an interpolation curve network; is a three-dimensional spline interpolation function;

[0107] It should be noted that the spatial boundary line network constructed by the above interpolation is the composition segregation boundary inside the flux particles, and provides accurate spatial structure information, which provides spatial positioning basis for further real-time evaluation of the composition uniformity of the flux particles.

[0108] Extracting the temperature change trajectory of the corresponding region of the composition segregation boundary in the thermal radiation image, combining the spatial coordinates of the segregation boundary with the temperature change trajectory to generate a segregation position marker;

[0109] Specifically, the extraction of the temperature change trajectory is based on the time sequence evolution mode of the temperature field in the thermal radiation image. When the temperature of a region shows a continuous decline or fluctuation feature in the time sequence, the temperature evolution path of the region is taken as the temperature change trajectory.

[0110] Exemplarily, when the temperature of a region always shows a monotonic decreasing trend (the temperature reduction value of each measurement is not less than 0.5℃) in at least 3 consecutive sampling time points, the temperature change trajectory of the region is extracted.​ , or when the temperature data appears periodic or non-monotonic fluctuation (fluctuation amplitude greater than 0.5°C, fluctuation period not less than 2 sampling periods, specific threshold determined by experiment) in more than 3 consecutive sampling time, the temperature change sequence of the region is determined as the temperature change trajectory.

[0111] It should be understood that the embodiment first extracts the temperature change data of the component segregation boundary corresponding region in the time sequence data of the thermal radiation image, which is specifically realized by the following way:

[0112] The temperature change trajectory is defined as:

[0113]

[0114] In the formula: represents the temperature trajectory of the temperature change at the spatial position with time; represents different time points;

[0115] When the temperature change trajectory continuously decreases or has obvious fluctuation in the time sequence, it indicates that there is obvious difference in the melt composition inside the particle.

[0116] Further, the temperature change trajectory is matched with the aforementioned spatial boundary line network to mark and confirm the position of the component segregation inside the flux particle, so as to generate accurate segregation position markers.

[0117] S103: The segregation position marker is verified in time sequence with the microstructure evolution process of the flux particle, and a causal chain of component segregation and microstructure change is established; when the causal chain appears chain fracture, the melt solidification resistance layer distribution is adjusted according to the recapture of heat conduction shunt in the thermal radiation image;

[0118] It should be noted that in order to verify the accuracy of the determined component segregation position, the embodiment strictly analyzes the time sequence correspondence between the segregation position marker generated by the above step S102 and the evolution process of the microstructure of the flux particle in the solidification process, so as to clarify the causal relationship between the component segregation and the microstructure change.

[0119] In the specific implementation process, the verification of the time sequence correspondence between the segregation position marker and the microstructure evolution process of the flux particle, and the establishment of the causal chain of component segregation and microstructure change, include:

[0120] Obtain the microstructure information of the flux particle at different time in the solidification process, and identify the evolution stage of the microstructure;

[0121] ​Specifically, the microstructure evolution stages include primary phase precipitation, eutectic reaction and solid state phase transformation, the microstructure information is obtained by metallographic analysis of the flux particles, and the evolution stages are identified based on the formation and transformation characteristics of different phase structures.

[0122] It can be understood that the embodiment collects microstructure images of different solidification stages of the particles by timing sampling the flux particles and using standard metallographic microscope analysis method. Specifically, different microstructure evolution stages of the particle solidification process are explicitly identified through microstructure morphology characteristics (such as grain size, crystal phase morphology, distribution density, etc.), including: primary phase precipitation stage, eutectic reaction stage and solid state phase transformation stage.

[0123] For example, the solidification process of the flux particles is divided into a plurality of specific time periods (such as 1 sampling per second), the corresponding microstructure evolution image data is recorded respectively, and the time label of the microstructure image data is stored, so as to facilitate subsequent analysis.

[0124] The time corresponding relationship between the formation of the composition segregation boundary and the microstructure evolution stage is established, and the time sequence synchronization of segregation occurrence and microstructure change is analyzed.

[0125] It should be noted that the embodiment further determines the time sequence corresponding relationship between the formation time of the composition segregation boundary and the microstructure characteristic change stage according to the microstructure evolution images and the corresponding time data obtained above.

[0126] Specifically, by corresponding matching the time information of the segregation position marked in the thermal radiation image with the time information of the different stages explicitly identified in the microstructure image, the time sequence synchronization between the two is analyzed. The synchronization is quantitatively expressed as a time difference value .

[0127]

[0128] In the formula: represents the formation time of the composition segregation position determined in the thermal radiation image; represents the starting or transformation time of the microstructure of the flux particles explicitly identified in a certain specific evolution stage.

[0129] The causal correlation strength between the composition segregation and the microstructure evolution is evaluated according to the degree of coincidence of the time sequence synchronization;

[0130] It can be understood that the embodiment quantitatively evaluates the causal correlation strength between the composition segregation and the microstructure evolution based on the time difference value obtained by the above analysis. Specifically, the causal correlation strength evaluation index is defined as:

[0131]

[0132] In the formula: represents the causal correlation strength, and the value range is 0 to 1, and the closer the value is to 1, the stronger the causal correlation is; is an empirical coefficient, and the specific value is determined according to experimental data statistical analysis.

[0133] It should be noted that when the causal correlation strength C is less than a preset threshold , it indicates that there is a mismatch between the causal relationship between the component segregation position and the microstructure evolution stage, that is, the breaking phenomenon of the causal chain occurs.

[0134] Based on the evaluation result of the causal correlation strength, the causal chain between the component segregation and the microstructure change is established;

[0135] It should be understood that the present embodiment establishes the causal chain between the component segregation position and the microstructure change based on the evaluation result of the above-mentioned causal correlation strength evaluation index . When the evaluation index is greater than or equal to the set threshold , it is considered that there is a stable causal relationship, an explicit causal chain is formed, and it is recorded in the database.

[0136] In the specific implementation process, the recapturing adjustment of the melt solidification resistance layer distribution according to the heat conduction shunt in the thermal radiation image includes:

[0137] It should be noted that when the evaluation result of the causal chain between the component segregation and the microstructure change breaks in the present embodiment, the previously unidentified or inaccurate heat conduction shunt point will be captured through reanalysis of the heat conduction shunt phenomenon in the thermal radiation image, and the spatial distribution structure of the melt solidification resistance layer will be adjusted again.

[0138] Identify the time node of the chain break in the causal chain of component segregation and microstructure change;

[0139] Specifically, the identification of the chain break is based on the mismatching degree of the component segregation formation time and the microstructure evolution time. When the difference between the formation time of the segregation boundary and the occurrence time of the corresponding microstructure evolution stage exceeds the time period of the flux particle solidification process, it is determined that the causal chain breaks.

[0140] For example, when the absolute time difference between the two exceeds 20% of the average period of the flux particle solidification process (the average period of the solidification process is determined by actual test, for example, the average length of the particle from the furnace to the complete solidification is taken as the average period of the solidification process), it is determined that the causal chain breaks.

[0141] It should be understood that the time node of the chain break is determined as follows:

[0142] First, record the time when the segregation position marker is formed and the time when the corresponding microstructure evolution stage occurs , the time difference between the two ;

[0143] When the time difference exceeds the reasonable time range of microstructure change during the flux particle solidification process (exemplarily, the range can be defined as a certain proportion of the typical cycle of the particle solidification process, for example, 20%), that is:

[0144]

[0145] Then, the embodiment determines that the causal chain is broken at this time node and records it as the specific time node of chain breakage.

[0146] At the time node of chain breakage, re-analyze the occurrence mode of heat conduction diversion in the thermal radiation image and track the spatial evolution trajectory of the diversion phenomenon.

[0147] It can be understood that, in order to repair the above-mentioned chain breakage problem, the embodiment reanalyzes the heat conduction diversion phenomenon on the surface of the flux particle in the thermal radiation image at the determined chain breakage time node, and the specific implementation process is as follows:

[0148] In the thermal radiation image sequence near the chain breakage time node, the aforementioned main flow channel identification and heat conduction diversion point detection steps are re-executed. Specifically, if the width change of the heat conduction main flow channel ignored in the initial analysis is less than a set threshold (for example, the initial channel width increases by less than 20%, but by more than 10% weak change), or the split angle in the initial identification does not reach the original threshold but is found to reach a new corrected threshold (such as , which is determined by actual test) in the re-analysis, these positions are re-identified as heat conduction diversion points and their spatial coordinate positions are recorded to correct the heat conduction diversion point distribution.

[0149] Further, the spatial evolution trajectory of the heat conduction diversion phenomenon is tracked, that is, the heat conduction diversion points re-identified in the continuous image sequence are reconstructed in the order of their time sequence to obtain the corrected heat conduction diversion spatial trajectory information, which is specifically represented as:

[0150]

[0151] In the formula: represents the position and time node of the jth re-identified heat conduction diversion point in three-dimensional space; k represents the total number of heat conduction diversion points on the trajectory.

[0152] Based on the reanalysis result of the spatial evolution trajectory of the flow splitting phenomenon, the position distribution of the heat conduction flow splitting point is corrected.

[0153] It should be noted that the spatial evolution trajectory of the heat conduction flow splitting phenomenon obtained by re-tracing according to the above steps , the spatial position distribution of the heat conduction flow splitting point is re-determined, and specifically corrected as follows:

[0154] Firstly, the position error or missing points of the flow splitting point in the spatial coordinates are corrected by a spatial trajectory analysis method (such as trajectory fitting or filtering method), and a corrected flow splitting point coordinate set is obtained:

[0155]

[0156] In the formula: each position point is the accurate coordinate position of the flow splitting point after trajectory correction.

[0157] It should be understood that the above corrected flow splitting point position distribution more accurately represents the real spatial position distribution of the melt solidification resistance layer inside the particle than the initially determined flow splitting point.

[0158] According to the corrected heat conduction flow splitting point position, the spatial profile of the solidification resistance layer is re-constructed, and the melt solidification resistance layer distribution is adjusted;

[0159] It can be understood that based on the above corrected heat conduction flow splitting point position set , the spatial profile of the melt solidification resistance layer of the flux particle is re-constructed, and the spatial distribution of the resistance layer is accurately adjusted.

[0160] Specifically, the three-dimensional radial basis function interpolation method described above is used to perform spatial interpolation reconstruction on the corrected flow splitting point position set, and an adjusted solidification resistance layer spatial profile function is constructed:

[0161]

[0162] In the formula: represents the adjusted and re-constructed solidification resistance layer spatial profile function; is the weight coefficient obtained by re-interpolation calculation; is the shape parameter in the re-interpolation process; represents the spatial distance between an arbitrary interpolation space point and the th corrected flow splitting point.

[0163] It should be noted that the solidification resistance layer spatial profile function The real melt solidification process space structure inside the particles can be accurately reflected, and the problem of broken causal chain between composition segregation and structure change is effectively repaired, thereby providing reliable data basis for subsequent uniformity grade evaluation.

[0164] In S104, the composition uniformity grade of the flux particles is determined based on the adjusted melt solidification impedance layer distribution, and a uniformity detection result is output.

[0165] It should be noted that the melt solidification impedance layer space distribution function reconstructed in the above embodiment further clarifies the space structure state of the melt inside the flux particles, thereby accurately evaluating the composition uniformity grade of the flux particles.

[0166] In the specific implementation process, the following steps are adopted to realize the evaluation and determination of the composition uniformity grade of the flux particles:

[0167] Firstly, the space uniformity characteristic parameters of the melt impedance layer inside the flux particles are calculated according to the reconstructed impedance layer distribution space profile . Specifically, the space impedance layer density parameter is adopted to characterize the uniformity in the embodiment, and the definition formula is as follows:

[0168]

[0169] In the formula, is the effective number of the melt solidification impedance layer inside the flux particles, that is, the number of the solidification impedance layers that can be clearly distinguished in the space profile function ; represents the total volume of the particles, and the specific calculation can be obtained based on the three-dimensional model of the particles (obtained by three-dimensional reconstruction of the image).

[0170] Further, the value range of the impedance layer density is divided into multiple grades to form the evaluation standard of the composition uniformity of the flux particles. For example:

[0171] When , it is determined as grade I (excellent, highly uniform composition);

[0172] When , it is determined as grade II (good, relatively uniform composition);

[0173] When , it is determined as grade III (general, with slight segregation);

[0174] When , it is determined as grade IV (poor, with obvious segregation);

[0175] Wherein: The threshold parameters for the uniformity evaluation grade are obtained by a large amount of experimental data statistical analysis.

[0176] It should be appreciated that the above division criteria and the selection of the impedance layer density parameter are determined according to actual data statistics of experimental tests, metallographic structure observation and flux performance analysis, and can effectively and objectively evaluate the composition uniformity of the flux particles.

[0177] Finally, the determined composition uniformity grade is automatically generated into a detection report by a pre-designed data processing software system, and is output and displayed in a visual manner (such as a color schematic diagram or a grade number).

[0178] Exemplarily, the output uniformity detection result includes: the identification number of the flux particles;

[0179] the composition uniformity grade (grade I to grade IV);

[0180] the impedance layer density parameter value ;

[0181] the visual particle impedance layer spatial distribution image.

[0182] Through the above specific embodiments, the composition uniformity of the particles in the production process of the smelting flux can be evaluated in real time, accurately and reliably, and the present application has important industrial application value.

[0183] The above embodiments are only used to illustrate the technical method of the present application but not limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. A method for real-time evaluation of uniformity of flux in smelting flux production, characterized by, The method comprises the following steps: Collecting a thermal radiation image of the flux particles at the moment of discharging from the furnace; Constructing a distribution of melt solidification resistance layers in the flux particles by heat conduction shunt phenomenon on the surface of the particles in the thermal radiation image; Identifying the composition segregation boundary in the flux particles according to the spatial fragmentation mode of the distribution of melt solidification resistance layers, and generating a segregation position marker based on the spatial overlap relationship between the composition segregation boundary and the temperature change trajectory in the thermal radiation image; Verifying the segregation position marker and the microstructure evolution process of the flux particles in time sequence, establishing a causal chain of composition segregation and microstructure change, and adjusting the distribution of melt solidification resistance layers according to the recapture of heat conduction shunt in the thermal radiation image when the causal chain is broken; The adjustment of the distribution of melt solidification resistance layers according to the recapture of heat conduction shunt in the thermal radiation image comprises the following steps: Identifying the time node of the broken chain in the causal chain of composition segregation and microstructure change; Reanalyzing the occurrence mode of heat conduction shunt in the thermal radiation image at the time node of the broken chain, and tracking the spatial evolution trajectory of the shunt phenomenon; Correcting the position distribution of the heat conduction shunt point based on the reanalysis result of the spatial evolution trajectory of the shunt phenomenon; Reconstructing the spatial profile of the solidification resistance layer by spatial interpolation reconstruction of the corrected shunt point position set using a three-dimensional radial basis function interpolation method, thereby adjusting the distribution of melt solidification resistance layers; wherein the adjusted reconstructed solidification impedance layer spatial profile is represented by the function as follows: wherein: represents the adjusted restructured solidification resistance layer spatial profile function; is the weight coefficient obtained for the re-interpolation calculation; is the shape parameter in the re-interpolation process; represents the spatial distance between the arbitrary interpolation spatial point and the first modified shunt point; Determining the composition uniformity grade of the flux particles based on the adjusted distribution of melt solidification resistance layers, and outputting the uniformity detection result.

2. The method of claim 1, wherein the method is characterized by: The construction of the distribution of melt solidification resistance layers in the flux particles by heat conduction shunt phenomenon on the surface of the particles in the thermal radiation image comprises the following steps: Identifying the main flow channel of heat transfer on the surface of the flux particles in the thermal radiation image, and marking the region with the largest temperature gradient in the temperature field on the surface of the particles as the main flow channel of heat transfer; Tracking the heat conduction shunt phenomenon occurring in the propagation process of the main flow channel of heat transfer, and marking the position where the main flow channel is dispersed into multiple conduction branches as a heat conduction shunt point; Infer the position of the resistance layer in the melt solidification process in the flux particles by the spatial distribution rule of the heat conduction shunt point, and each heat conduction shunt point corresponds to a boundary position of the solidification resistance layer; Connecting the boundary positions of adjacent solidification resistance layers to form the spatial profile of the resistance layer, and constructing the distribution of melt solidification resistance layers.

3. The method of claim 2, wherein the method is characterized by: The identification of the heat conduction shunt phenomenon is based on the geometric shape change of the main flow channel of heat transfer, and when the width of the main flow channel increases and is dispersed into multiple narrow channels in the propagation process, the shape change position is determined as a heat conduction shunt point.

4. The method of claim 3, wherein the method is characterized by: The generation of the segregation position marker comprises the following steps: Analyzing the spatial continuity state of the resistance layer in the distribution of melt solidification resistance layers, and identifying the position where the resistance layer appears spatial fracture; Taking the position of the spatial fracture of the resistance layer as the position where the composition separation occurs in the melt solidification process, and the fracture position corresponds to the boundary line of different composition regions in the flux particles; Connecting adjacent fracture positions to form a complete boundary line network, and determining the boundary line network as the composition segregation boundary in the flux particles; The temperature change trajectory of the corresponding area in the thermal radiation image is extracted, and the spatial coordinates of the segregation boundary are combined with the temperature change trajectory to generate a segregation position marker.

5. The method of real-time evaluation of uniformity of flux in production of smelting flux according to claim 4, characterized in that, The identification of the spatial fracture of the impedance layer is based on the geometric discontinuity of the impedance layer profile. When the impedance layer profile appears geometric discontinuity or absence during spatial extension, the discontinuity position is determined as the spatial fracture position of the impedance layer.

6. The method of real-time evaluation of uniformity of flux in production of smelting flux according to claim 5, characterized in that, The extraction of the temperature change trajectory is based on the time sequence evolution mode of the temperature field in the thermal radiation image. When the temperature of a certain area shows a continuous decline or fluctuation feature in the time sequence, the temperature evolution path of the area is taken as the temperature change trajectory.

7. The method of claim 6, wherein the method is characterized by: The time sequence correspondence verification of the segregation position marker and the microstructure evolution process of the flux particles establishes the causal chain of composition segregation and microstructure change, including: Obtaining the microstructure information of the flux particles at different time during the solidification process, identifying the evolution stage of the microstructure; Establishing the time correspondence between the formation time of the composition segregation boundary and the time of the microstructure evolution stage, analyzing the time sequence synchronization of segregation occurrence and microstructure change; According to the coincidence degree of the time sequence synchronization, the causal correlation strength of the composition segregation and the microstructure evolution is evaluated; Based on the evaluation results of the causal correlation strength, the causal chain of composition segregation and microstructure change is established.

8. The method of claim 7, wherein the method is characterized by: The microstructure evolution stage includes primary phase precipitation, eutectic reaction and solid state phase transition, and the microstructure information is obtained by metallographic analysis of the flux particles. The evolution stage is identified based on the formation and transformation characteristics of different phase structures.

9. The method of claim 8, wherein the method is characterized by: The identification of the chain fracture is based on the mismatch degree of the formation time of the composition segregation and the evolution time of the microstructure. When the formation time of the segregation boundary and the occurrence time of the corresponding microstructure evolution stage differ by more than the time period of the flux particle solidification process, it is determined that the causal chain is broken.

Images