Method and device for tracing impurity elements in aluminum alloy ingot

Abstract

The application is suitable for the technical field of aluminum alloy production, and particularly relates to an aluminum alloy ingot impurity element traceability analysis method and equipment. The aluminum alloy ingot impurity element traceability analysis method comprises: obtaining aluminum alloy ingot impurity information and production chain information; wherein the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in multiple aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot; an impurity isomorphism graph is constructed based on the aluminum alloy ingot impurity information; a potential source is determined based on the impurity isomorphism graph; and the final traceability result of the impurity element is determined based on the production chain information and the potential source. The method effectively solves the problems of insufficient detection data integrity, weak impurity correlation analysis and disconnection between production chain information and impurity characteristics in the prior art, realizes multi-dimensional information integration and quantitative correlation analysis of the source of the aluminum alloy ingot impurity element, and significantly improves the accuracy, multi-source differentiation capability and result reliability of impurity traceability.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of aluminum alloy production, and particularly relates to a method and equipment for tracing impurity elements in aluminum alloy ingots. BACKGROUND

[0002] As a core raw material in the fields of high-end equipment such as aerospace, automobile manufacturing and rail transportation, the purity and impurity content of aluminum alloy ingots directly determine the mechanical properties, corrosion resistance and processing stability of downstream products. For example, excessive Fe elements in aluminum alloy ingots will lead to a decrease in material plasticity and increase the risk of subsequent rolling cracking; low-melting-point impurities such as Pb and Cd will cause hot working brittleness, and H elements are prone to form porosity defects, which seriously affect the fatigue life of components. Therefore, accurately tracing the introduction source (such as raw materials, processes, equipment or environment) of impurity elements in aluminum alloy ingots is a key link for controlling product quality, optimizing production processes and reducing costs.

[0003] The current aluminum alloy ingot production process covers multiple processes such as raw material pretreatment (primary bauxite smelting or recycled aluminum scrap sorting), melting, alloying, refining and casting, and the impurity sources present the characteristics of "multiple links and complex types": the impurities of primary aluminum raw materials are affected by the source of origin (such as the Si and Fe content difference of bauxite in different mining areas can reach 0.3%-0.8%); recycled aluminum is prone to mix in exogenous impurities such as Pb, Cu and Cl due to incomplete classification and sorting of scrap aluminum; in the melting process, the wear of cast iron crucibles will introduce Fe elements, insufficient purity of refining gas may bring in gas impurities such as O and N, and leakage of the cooling system may cause cooling water pollution and introduce elements such as Na and K. The complex source system puts forward the core needs of "accurate positioning and multi-source differentiation" for impurity tracing technology.

[0004] In the prior art, a linear matching idea of "single element-single source" is often used, and the correlation between elements is not quantitatively analyzed: for example, only the increase in Fe element content is used to determine the wear of the melting crucible, but the synergistic growth of Fe and Si content may be caused by the mixed mixing of scrap steel and scrap silicon-aluminum alloy in the raw material, leading to misjudgment of the tracing result. This method lacks global correlation analysis and cannot meet the needs of accurate positioning of impurity sources for high-end aluminum alloy materials. SUMMARY

[0005] The embodiments of the application provide a method and equipment for tracing impurity elements in aluminum alloy ingots, which can solve the problem of low accuracy of the tracing result of impurity elements in aluminum alloy ingots.

[0006] In a first aspect, the embodiments of the application provide a method for tracing impurity elements in aluminum alloy ingots, comprising:

[0007] acquire aluminum alloy ingot impurity information and production chain information; wherein the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in multiple aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot;

[0008] construct an impurity isomerism graph based on the aluminum alloy ingot impurity information; wherein the impurity isomerism graph is used to reflect the correlation between impurity elements;

[0009] determine a potential source based on the impurity isomerism graph; wherein the potential source is used to reflect the correlation process of each impurity element;

[0010] determine the final traceability result of the impurity element based on the production chain information and the potential source.

[0011] The technical scheme described above in the embodiments of the present application has at least the following technical effects:

[0012] The aluminum alloy ingot impurity element traceability analysis method provided by the present application acquires aluminum alloy ingot impurity information used to reflect the composition and corresponding content interval of impurities in multiple aluminum alloy ingots and production chain information used to reflect the production process of the aluminum alloy ingot. Then, an impurity isomerism graph used to reflect the correlation between impurity elements is constructed based on the aluminum alloy ingot impurity information, which can convert originally isolated impurity element data into a network structure with quantitative correlation, not only intuitively presenting the co-occurrence characteristics between elements, but also quantifying the element correlation strength through the graph structure to reduce the linear misjudgment of "single element-single source". Then, a potential source used to reflect the correlation process of each impurity element is determined based on the impurity isomerism graph, realizing accurate correlation from "impurity characteristics" to "process source". Finally, the final traceability result of the impurity element is determined based on the production chain information and the potential source, improving the accuracy of impurity element source judgment through cross-validation. This method effectively solves the problems of insufficient detection data integrity, weak impurity correlation analysis and disconnection between production chain information and impurity characteristics in the prior art, realizes multi-dimensional information integration and quantitative correlation analysis of the source of aluminum alloy ingot impurity elements, and significantly improves the accuracy, multi-source differentiation ability and result reliability of impurity traceability.

[0013] In a second aspect, the embodiments of the present application provide an aluminum alloy ingot impurity element traceability analysis system, comprising:

[0014] an acquisition module, configured to acquire aluminum alloy ingot impurity information and production chain information; wherein the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in multiple aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot;

[0015] The construction module is configured to construct an impurity isomorphism graph based on the aluminum alloy ingot impurity information, wherein the impurity isomorphism graph is configured to reflect the correlation between impurity elements.

[0016] The first determination module is configured to determine a potential occurrence source based on the impurity isomorphism graph, wherein the potential occurrence source is configured to reflect a correlated process of each impurity element.

[0017] The second determination module is configured to determine a final traceability result of the impurity element based on the production chain information and the potential occurrence source.

[0018] In a third aspect, an aluminum alloy ingot impurity element traceability analysis device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method of any one of the first aspect when executing the computer program.

[0019] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and the computer program is executable by a processor to implement the method of any one of the first aspect.

[0020] In a fifth aspect, a computer program product is provided, which, when executed on an aluminum alloy ingot impurity element traceability analysis device, causes the aluminum alloy ingot impurity element traceability analysis device to execute the aluminum alloy ingot impurity element traceability analysis method of any one of the first aspect.

[0021] It can be understood that the beneficial effects of the second aspect to the fifth aspect can be referred to the related description of the first aspect, which will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0023] Figure 1 is a flowchart of the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application;

[0024] Figure 2 is a flowchart of the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application;

[0025] Figure 3is a structural schematic diagram of an aluminum alloy ingot impurity element traceability analysis system provided by an embodiment of the present application;

[0026] Figure 4 is a structural schematic diagram of an aluminum alloy ingot impurity element traceability analysis device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0027] In the following description, for the purposes of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the application. However, it will be apparent to those skilled in the art that the application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0028] It should be understood that the term "comprises" when used in this specification and the appended claims indicates the presence of the described features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0029] It should also be understood that the term "and / or" when used in this specification and the appended claims indicates that the associated listed items can be present one or more of the associated listed items, and that the items are not limited to only one of the associated listed items.

[0030] As used in this specification and the appended claims, the term "if" can be interpreted as meaning "when" or "upon" or "in response to determining" or "in response to detecting." Similarly, the phrase "if determined" or "if detected" can be interpreted as meaning "upon determining" or "in response to determining" or "upon detecting" or "in response to detecting" depending on the context.

[0031] In addition, in the description of the application and the appended claims, the terms "first", "second", "third", etc. are used only to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0032] Reference within the specification of this application to "one embodiment" or "some embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," and the like in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily referring to some, but not all, embodiments. The terms "including," "comprising," "having," and variations thereof are meant to encompass the items listed thereafter, but do not exclude other items from also being present. The terms "a" and "an," as used herein in the specification, mean "one or more."

[0033] In the prior art, the linear matching idea of "single element-single source" is mostly used, and the correlation between elements is not quantitatively analyzed: for example, only through the increase of Fe element content to determine the wear of smelting crucible, but the synergistic growth of Fe and Si contents may be caused by the mixed mixing of scrap steel and scrap silicon-aluminum alloy in raw materials, resulting in misjudgment of the traceability result. This method lacking global correlation analysis cannot meet the demand of high-end aluminum alloy materials for precise positioning of impurity sources.

[0034] To solve the above problems, the embodiments of the present application provide an aluminum alloy ingot impurity element traceability analysis method and equipment. In the method, aluminum alloy ingot impurity information reflecting the composition and corresponding content interval of impurities in a plurality of aluminum alloy ingots and production chain information reflecting the production process of the aluminum alloy ingot are obtained; then a impurity isomorphism graph reflecting the correlation between impurity elements is constructed based on the aluminum alloy ingot impurity information, which can convert the originally isolated impurity element data into a network structure with quantitative correlation, not only can intuitively present the co-occurrence characteristics between elements, but also can quantify the element correlation strength through the graph structure, reduce the linear misjudgment of "single element-single source"; based on the impurity isomorphism graph, the potential occurrence source reflecting the correlation process of each impurity element is determined, realizing the precise correlation from "impurity characteristics" to "process source"; finally, based on the production chain information and the potential occurrence source, the final traceability result of the impurity element is determined, and the precision of the impurity element source judgment is improved through cross-validation. The method effectively solves the problems of incomplete detection data, weak impurity correlation analysis and disconnection between production chain information and impurity characteristics in the prior art, realizes the multi-dimensional information integration and quantitative correlation analysis of the impurity element sources of the aluminum alloy ingot, and significantly improves the precision, multi-source differentiation ability and result reliability of impurity traceability.

[0035] The aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application can be applied to an aluminum alloy ingot impurity element traceability analysis device, and in this case, the aluminum alloy ingot impurity element traceability analysis device is the execution subject of the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application. The embodiments of the present application do not make any limitation on the specific type of the aluminum alloy ingot impurity element traceability analysis device.

[0036] For example, the aluminum alloy ingot impurity element traceability analysis device can be a handheld device, a computing device or other processing device connected to a wireless modem, an Internet of Things terminal, a computer, a laptop computer, a handheld communication device, a handheld computing device, a satellite wireless device, a wireless modem card and / or other devices for communicating on a wireless system, and a next-generation communication system, such as a mobile terminal in a 5G network or a mobile terminal in a future evolved Public Land Mobile Network (PLMN), etc., which has a wireless communication function on a terminal device such as a mobile phone, a tablet computer, a wearable device, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), a desktop computer, a smart large screen, a smart television, etc.

[0037] In order to better understand the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application, the specific implementation process of the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application is exemplarily introduced as follows.

[0038] Figure 1 And Figure 2 A schematic flowchart of the aluminum alloy ingot impurity element traceability analysis method provided by the embodiments of the present application is shown, and the aluminum alloy ingot impurity element traceability analysis method comprises:

[0039] S100, obtaining aluminum alloy ingot impurity information and production chain information; wherein the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in a plurality of aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot.

[0040] It can be understood that the impurities of the aluminum alloy ingot generally include constant impurities (such as Fe, Si, Cu and other content elements) and trace impurities (such as Pb, Cd, Hg and other content elements), the constant impurities can be detected by inductively coupled plasma optical emission spectrometry (ICP-OES), and the detection accuracy can reach 0.001%; the trace impurities are captured by inductively coupled plasma mass spectrometry (ICP-MS), and the lower limit of detection can be as low as 0.001 ppb. The content interval is obtained from the collection of detection data of a plurality of aluminum alloy ingots. The production chain information specifically includes raw material links (mineral source of primary bauxite, composition report, classification standard of recycled aluminum scrap, sorting record and batch composition detection data), process parameters of each process (temperature curve, holding time, flux addition amount of smelting link; alloying agent type, addition ratio and stirring rate of alloying link; gas flow, vacuum degree, filter medium type of refining link; cooling rate, mold temperature of casting link), equipment operation data (material quality, use frequency, maintenance record of smelting crucible; filter core replacement cycle of refining and filtering equipment; mold wear degree detection report of casting equipment). Time sequence alignment needs to be ensured during data collection, for example, accurately matching the impurity detection data of a batch of aluminum alloy ingots with the smelting temperature record and raw material batch information during the corresponding production period.

[0041] In S200, an impurity isomorphism graph is constructed based on the impurity information of the aluminum alloy ingot; wherein the impurity isomorphism graph is used to reflect the correlation between impurity elements.

[0042] It can be understood that constructing an impurity isomorphism graph based on the impurity information of the aluminum alloy ingot is a core link for realizing quantitative analysis of the correlation between impurity elements, and its essence is to convert scattered impurity data into a visual and calculable network model with a topological structure, solving the limitations of "single element isolated analysis" in traditional methods. Unlike traditional linear correlation analysis (such as only calculating the correlation coefficient of two elements), the impurity isomorphism graph comprehensively describes the complex correlation between impurity elements through the combination of multiple types of nodes and weighted edges.

[0043] Exemplarily, the characteristics of impurity elements can be extracted from the aluminum alloy ingot impurity information, then connection nodes are established through the impurity elements, and edges between the nodes are constructed through the characteristics of the impurity elements, to generate a plurality of heterogeneous graph structures, and the impurity heterogeneous graph is generated based on the heterogeneous graph structures; or the aluminum alloy ingot impurity information can be processed in layers first, classified according to the chemical properties (such as metal elements, non-metal elements) and content orders (constant impurities, trace impurities) of the impurity elements, a sub-heterogeneous graph is constructed for each type of impurity element, the nodes in the sub-heterogeneous graph contain the specific components and content characteristics of the elements of this type, and the weights of the edges are calculated based on the content collaborative change trend between elements of the same type (such as the fluctuation coefficient of the content ratio of Fe to Cu in metal impurities), then the sub-heterogeneous graphs are connected through “inter-class association edges” (such as the complex inclusion formation probability of the O element in non-metal impurities and the Fe element in metal impurities), and finally integrated into a complete impurity heterogeneous graph, and the like, but not limited thereto.

[0044] In a possible implementation, in step S200, the impurity heterogeneous graph is constructed based on the aluminum alloy ingot impurity information, including:

[0045] S210, extracting multi-element feature parameters based on the aluminum alloy ingot impurity information; wherein the multi-element feature parameters include content ratio between impurity elements and impurity element content fluctuation rate.

[0046] It can be understood that the content ratio between impurity elements is not simply calculated as the content ratio of any two impurity elements, but is based on the aluminum alloy production process mechanism and the characteristics of impurity sources, and the element combination with "homology" or "process synergy" is screened. For example, in the production of secondary aluminum, Pb and Cd are often mixed together due to incomplete sorting of waste aluminum, and the content ratio (Pb / Cd) of the two has a stable statistical correlation, which needs to be extracted first; Fe and Si are easy to form Al-Fe-Si composite inclusions in the smelting process, and the content ratio (Fe / Si) directly reflects the formation tendency of inclusions, which is also a key extraction object; and for elements such as Cu and Zn which have no significant process correlation and large differences in sources (Cu mainly comes from alloying agent, and Zn mainly comes from raw material impurities), there is no need to forcibly calculate the ratio. The core of the fluctuation rate of the content of impurity elements is to quantify the stability of the content of impurities in the production process, rather than simply calculating the standard deviation. Specifically, it is executed in three steps of "time window division-statistical quantity calculation-exception marking": first, according to the aluminum alloy production process cycle (such as the total cycle of smelting + casting is 8 hours), a reasonable time window (such as every 4 hours as a window, covering 5-8 production batches) is set, which can make the data in the window reflect short-term fluctuations and have statistical representativeness; secondly, for a certain impurity element in each time window, the "coefficient of variation" (fluctuation rate = (standard deviation / mean) x 100%) of its content is calculated, for example, the content of Fe element in a certain window is 0.07% and the standard deviation is 0.009%, then the fluctuation rate = (0.009 / 0.07) x 100% ≈ 12.86%.

[0047] S220, constructing a heterogeneous graph structure based on the multi-element feature parameters; wherein the heterogeneous graph structure includes multiple types of nodes and weighted edges.

[0048] Exemplarily, the node types and attribute dimensions can be generated based on the multi-element characteristic parameters, then the weighted edges between nodes are calculated based on the node types and attribute dimensions, and then the selective connection between nodes is screened based on the weighted edges to obtain a heterogeneous graph structure; or the correlation dimensions (such as “element synergy dimension” and “content fluctuation dimension”) can be divided according to the physical meaning of the multi-element characteristic parameters, a sub-network is constructed for each dimension, and then the sub-networks are integrated into a complete heterogeneous graph structure through cross-dimension association rules, that is, in the “element synergy dimension”, the impurity elements are taken as nodes, the edge weight is constructed based on the time series similarity of the “content ratio between elements” (such as the similarity of the Fe / Si ratio sequence calculated by the dynamic time warping algorithm is 0.89), and an element synergy sub-network is formed; in the “content fluctuation dimension”, the impurity element nodes and the content interval nodes are taken as cores, the edge weight is calculated based on the “impurity element content fluctuation rate” and the interval distribution probability (such as the distribution probability of Pb element in the exceeding interval is 0.15, and the fluctuation rate is 22%), and a content fluctuation sub-network is formed. Then, the multiple sub-networks are connected through cross-dimension association rules (such as the strong association of Fe-Si in the element synergy sub-network needs to be synchronously associated to the “melting temperature” node in the process response sub-network), and finally a heterogeneous graph structure is formed.

[0049] In a possible implementation, in step S220, the heterogeneous graph structure is constructed based on the multi-element characteristic parameters, including:

[0050] S221, generating node types and attribute dimensions based on the multi-element characteristic parameters; wherein the node types include impurity element nodes, content interval nodes and process sensitive nodes.

[0051] It can be understood that the generation of the impurity element nodes is based on the content ratio between elements, since the content ratio reflects the synergy or antagonism between different elements, therefore each impurity element participating in the ratio calculation needs to be taken as an independent node, for example, if Fe / Si, Cu / Mg and Pb / Cd ratios are calculated, then Fe, Si, Cu, Mg, Pb and Cd six impurity element nodes need to be generated, and the node quantity needs to cover all the detected key impurity elements. The generation of the content interval nodes corresponds to the impurity element content fluctuation rate characteristic, and the interval needs to be divided according to the element quality standard and fluctuation risk; the generation of the process sensitive nodes needs to associate the characteristic parameter anomaly with the potential association of the process link, for example, when the Fe / Si ratio anomaly is often accompanied by melting temperature fluctuation, and the number of crucible uses in the batch with the Fe content fluctuation rate exceeding the standard is all > 50 heats, then two process sensitive nodes “melting temperature fluctuation” and “crucible wear” are generated, and the nodes cover the key processes in the whole process such as raw material pretreatment, melting, refining and casting.

[0052] S222, calculating the cross-dimension association strength of the multi-element characteristic parameters based on the node types and attribute dimensions.

[0053] It can be understood that the cross-dimension association strength refers to the connection strength between different types of nodes, that is, the weighted edge, and the core value thereof lies in breaking the dimension barriers of "element-content-process", and converting the dispersed multi-element characteristic parameters into the associated relationship with process interpretation by quantifying the association degree of different types of nodes. Exemplarily, the association strength between different nodes can be calculated according to the node type and the corresponding attribute dimension, and then a complete cross-dimension association strength system is formed by integration, or a "feature-node-attribute" mapping matrix can be constructed to clearly indicate the corresponding relationship between each multi-element characteristic parameter and the node type and the attribute dimension, and then the cross-dimension association strength is calculated based on the mapping matrix by using a multi-model fusion algorithm, etc., but not limited thereto.

[0054] In a possible implementation, in step S222, the cross-dimension association strength of the multi-element characteristic parameter is calculated based on the node type and the attribute dimension, comprising:

[0055] S2221, based on the node type and the attribute dimension, the first association strength between the impurity element nodes, the second association strength between the impurity element nodes and the content interval nodes, and the third association strength between the impurity element nodes and the process sensitive nodes are calculated.

[0056] It can be understood that the calculation of the three types of association strengths (first, second and third association strengths) based on the node type and the attribute dimension is the core disassembly step for realizing the accurate quantification of the cross-dimension association strength, and the essence thereof is to design differentiated calculation logic according to the association characteristics of different node combinations (chemical synergy between elements, belonging stability of elements and content, and response influence of elements and process), so as to avoid the distortion of the association strength caused by the "one-size-fits-all" calculation. This link closely follows the corresponding relationship between the attribute dimension of each type of node (such as the chemical attribute of the element node, the probability attribute of the content interval node, and the contribution attribute of the process sensitive node) and the multi-element characteristic parameter (content ratio between elements, element content fluctuation rate), so that each type of association strength not only conforms to the statistical law, but also has clear process interpretation, thereby providing accurate basis for the construction of the weighted edge of the heterogeneous graph.

[0057] In a possible implementation, in step S2221, the first association strength between the impurity element nodes is calculated based on the node type and the attribute dimension, comprising:

[0058] S22211a, time series data is generated based on the impurity element nodes in the node type and the content ratio between impurity elements in the attribute dimension; wherein the time series data is used to reflect the change of the content ratio between impurity elements of a plurality of aluminum alloy ingots.

[0059] It can be understood that generating time series data based on impurity element nodes and content ratio between impurity elements is a key step of converting "element correlation" from static statistics to dynamic trend analysis, and the core is to present the change of element content ratio in multiple aluminum alloy ingots through time series, capture the synergistic or antagonistic rules between elements during the production process, and provide continuous and traceable data basis for subsequent linear correlation coefficient calculation. For example, 30 aluminum alloy ingot data are collected during aluminum alloy ingot production, and the horizontal axis of the time series is "serial number (1-30) + production time", and the vertical axis is the element content ratio of the corresponding batch. Taking Fe and Si as an example, the Fe content of the first aluminum alloy ingot is 0.07%, the Si content is 0.09%, and the Fe / Si ratio is approximately 0.78; the Fe content of the second aluminum alloy ingot is 0.068%, the Si content is 0.085%, and the Fe / Si ratio is approximately 0.80; and the time series curve of 30 data points is generated in this way, that is, the time series data.

[0060] S22212a, calculating a linear correlation coefficient based on the time series data; wherein the linear correlation coefficient is used to reflect the synergistic trend of the impurity element content ratio with the production of the aluminum alloy ingot.

[0061] It can be understood that calculating a linear correlation coefficient based on time series data is a core statistical method for quantifying the synergistic trend of impurity element content ratio with the production process, and its essence is to measure the linear correlation degree of two element content ratio time series through mathematical methods. For example, if the time series of two element content ratios both conform to normal distribution (which can be verified by Shapiro-Wilk test with a significance level a=0.05), the Pearson correlation coefficient is calculated, for example, the Fe / Si ratio sequence (mean 0.78, standard deviation 0.03) and the Cu / Mg ratio sequence (mean 1.2, standard deviation 0.05) both conform to normal distribution, and the correlation coefficient is 0.85, indicating that the two are strongly positively correlated; if the sequence does not conform to normal distribution (such as the Pb / Cd ratio sequence appears skewed distribution due to individual abnormal batch), the Spearman rank correlation coefficient is used to calculate the correlation degree by converting the ratio data into ranking.

[0062] S22213a, generating a correction factor based on the formation tendency of the impurity element; wherein the formation tendency of the impurity element is used to indicate the thermodynamic possibility of the impurity element forming a complex inclusion during aluminum alloy melting and solidification.

[0063] It can be understood that the formation tendency of impurity elements is the key step to integrate the "chemical mechanism" into the correlation strength calculation, and the core is to correct the linear correlation coefficient based on data by quantifying the thermodynamic possibility of impurity elements forming complex inclusions in the smelting and solidification process of aluminum alloy, and reducing the misjudgment of correlation strength caused by "strong correlation of data but no chemical correlation". Exemplarily, a mapping relationship of "formation tendency-correction factor" can be established: first, the formation tendency is divided into "extremely high, high, medium, low and extremely low" five levels, and different levels correspond to a correction factor. For example, the combination interval of ΔGθ (free energy change) and diffusion coefficient corresponding to "extremely high" level is ΔGθ≤-40 kJ / mol and diffusion coefficient≥1×10 -10 m 2 / s (such as Al and O forming Al2O3 phase), the correction factor can be taken as 1.4~1.5; the combination interval of ΔGθ and diffusion coefficient corresponding to "high" level is ΔGθ=-40~-20 kJ / mol and diffusion coefficient≥5×10 -11 m 2 / s (such as Fe and Si), the correction factor can be taken as 1.2~1.3; the combination interval of ΔGθ and diffusion coefficient corresponding to "medium" level is ΔGθ=-20~0 kJ / mol and diffusion coefficient≥1×10 -11 m 2 / s (such as Cu and Mg forming Al2CuMg phase), the correction factor can be taken as 1.0~1.1; the combination interval of ΔGθ and diffusion coefficient corresponding to "low" level is ΔGθ>0 kJ / mol or diffusion coefficient<1×10 -11 m 2 / s (such as Pb and Zn), the correction factor is taken as 0.8~0.9; the combination interval of ΔGθ and diffusion coefficient corresponding to "extremely low" level is ΔGθ>+10 kJ / mol and diffusion coefficient<5×10 -12 m 2 / s (such as Pb and Al forming intermetallic compound), the correction factor can be taken as 0.6~0.7.

[0064] S22214a, the linear correlation coefficient is weighted based on the correction factor to obtain the first correlation strength between the impurity element nodes.

[0065] It can be understood that the first correlation strength=correction factor×linear correlation coefficient. The first correlation strength is obtained by weighting the linear correlation coefficient based on the correction factor, which is the final step of integrating "data synergy" and "chemical correlation". The essence is to integrate statistical law and process mechanism by mathematical weighting to generate a quantitative index that can reflect the actual correlation degree between elements and has clear physical meaning. The value range is mapped to [0, 1] after normalization, and the higher the value, the stronger the correlation between the impurity element nodes.

[0066] In this way, the first correlation strength between the impurity element nodes is calculated through the four-step process of "generating time series data - calculating correlation coefficient - generating correction factor - weighted calculation", realizing the deep integration of "data statistical law" and "chemical process mechanism", and effectively solving the one-sidedness of "emphasizing data and ignoring mechanism" or "emphasizing theory and ignoring practice" in traditional correlation analysis. Specifically, the generation of time series data improves the dynamics and continuity of correlation analysis, captures the changing trend of element ratio with the production process, and avoids the accidental interference of single batch data; the calculation of linear correlation coefficient provides a quantitative statistical basis for correlation strength, and through the differential algorithm (Pearson / Spearman), it adapts to different data distribution characteristics and ensures the robustness of the results; the introduction of the correction factor is closely related to the thermodynamic and kinetic laws of aluminum alloy melting and solidification, which converts the chemical formation tendency into a calculable weight, reducing the misjudgment of "strong data correlation but no process significance". For example, the correlation strength between Fe and Si is significantly higher than that of other combinations due to high data synergy and high chemical tendency, corresponding to the strong weighted edge in the heterogeneous graph, which intuitively reflects the core correlation of the two in forming composite inclusions; although Pb and Cd have strong data synergy, their correlation strength is slightly lower due to the moderate chemical formation tendency, which is consistent with the actual situation that they often coexist in the form of single elements in aluminum alloys; although Fe and Zn have certain data synergy, their correlation strength is low due to weak chemical correlation, avoiding the misjudgment of unrelated elements as core correlations.

[0067] In one possible implementation, in step S2221, the second correlation strength between the impurity element nodes and the content interval nodes is calculated based on the node type and the attribute dimension, including:

[0068] S22211b, based on the impurity element nodes and the content interval nodes in the node type and the impurity element content fluctuation rate of the attribute dimension, calculate the interval density value; wherein the interval density value is used to reflect the distribution density of the impurity element content in the corresponding content interval.

[0069] It can be understood that the calculation of the interval density value based on the impurity element node, the content interval node and the impurity element content fluctuation rate is the core step of quantifying the association basis of the impurity element and the specific content interval. The essence is to reflect the distribution density of the element in the interval by counting the proportion of the frequency of the impurity element content falling into the target interval in multiple batches of aluminum alloy ingots, and to provide the "distribution basis" parameter for the subsequent second association strength calculation. The content interval boundary can be determined in combination with the quality standard and process characteristics of the impurity element: for key impurities that affect the performance of aluminum alloy (such as Fe and Si), the interval division needs to be fine (such as Fe can be divided into 0~0.02%, 0.02%~0.1% and >0.1%); for trace harmful elements (such as Pb and Cd), the safety threshold can be highlighted (such as Pb is divided into ≤0.003%, 0.003%~0.005% and >0.005%); for impurities with less influence (such as Zn), it can be simplified into "normal" and "over standard" two intervals. When counting, take multiple batches (≥20 batches) of impurity content data as samples, judge the interval of the element content batch by batch and accumulate the frequency, for example, there are 23 Fe contents in 30 batches of aluminum alloy ingots falling into the "0.02%~0.1%" interval, then the corresponding interval density value = 23 / 30 ≈ 0.77; there are 3 falling into the " >0.1%" interval, the corresponding density value = 0.1, and so on.

[0070] S22212b, generating a fluctuation coefficient based on the impurity element content fluctuation rate.

[0071] It can be understood that the generation of the fluctuation coefficient based on the impurity element content fluctuation rate is the key step of quantifying the "content stability" into a calculable parameter. The core is to generate a correction factor (value range 0~1) reflecting the data stability through the fluctuation degree of the element content in a specific interval. The smaller the fluctuation (the more stable the content), the closer the coefficient to 1; the larger the fluctuation (the less stable the content), the closer the coefficient to 0.

[0072] S22213b, calculating the second association strength of the impurity element node and the content interval node based on the interval density value and the fluctuation coefficient.

[0073] It can be understood that the second association strength = interval density value × fluctuation coefficient. The calculation of the second association strength based on the interval density value and the fluctuation coefficient is the core step of integrating "distribution density" and "content stability". The essence is to generate a quantitative index that can reflect the belonging relationship of the element and the interval and also reflect the reliability of the relationship by integrating the distribution density of the impurity element in a specific interval and the content stability. The value range is mapped to [0, 1] after normalization, and the higher the value, the closer the association and the more stable.

[0074] With such a setting, the second correlation strength is calculated by "calculating interval density value-generating fluctuation coefficient-fusing calculation strength", realizing the organic combination of "distribution characteristics" and "stability characteristics", and effectively solving the limitation problem that the traditional analysis only pays attention to "whether the element is in the interval" and ignores "whether the distribution is stable". This setting can also support quality risk assessment: for example, the strong correlation (normalized strength 1.0) between Fe and the "0.02%~0.1%" interval indicates that its content is stable in the normal range, and the process control is effective; If the medium correlation (such as strength 0.5) appears between Pb and ">0.005%", even if the density value is not high (such as 0.2), but because the fluctuation coefficient is high (such as 0.7), attention should be paid to the trend that it is stable beyond the standard; The weak correlation (strength 0.3) between a certain element and the over-standard interval can be used as a low-priority attention object. This hierarchical quantitative correlation strength provides a clear basis for the design of weighted edges between elements and content intervals in the heterogeneous graph, making the heterogeneous graph not only able to intuitively display the content distribution characteristics of the elements, but also able to reflect the distribution stability through the thickness of the edges, providing "content-stability" dual-dimensional correlation clues for the subsequent positioning of potential sources, which is especially suitable for the scene of fine control of impurity content in aluminum alloy ingot production.

[0075] In a possible implementation, in step S2221, based on the node type and the attribute dimension, a third correlation strength between the impurity element node and the process sensitive node is calculated, including:

[0076] S22211c, based on the process sensitive node in the node type, extracting the key process parameter data of the process sensitive node associated process, and based on the impurity element node in the node type, extracting the corresponding impurity content change data.

[0077] It can be understood that extracting key process parameter data based on process sensitive node and extracting impurity content change data based on impurity element node are core steps for establishing the correlation between process fluctuation and impurity change. The essence is to accurately match process data and impurity data of the same production period to provide "homogeneous and synchronous" analysis samples for subsequent quantification of the correlation degree between the two, avoiding distortion of the correlation analysis due to data time sequence misalignment or parameter mismatch. For example, if the process sensitive node is "smelting crucible wear" (related process: smelting), and the attribute label is "key parameters: crucible use frequency, furnace temperature uniformity, flux addition amount", then the corresponding three parameter data of the smelting process need to be extracted from the production database: crucible use frequency is recorded "once per batch" (e.g. 10 furnace times for the first batch, 30 furnace times for the 20th batch); furnace temperature uniformity is collected "once every 30 minutes", and the temperature standard deviation of each batch is calculated (e.g. temperature standard deviation of ±4°C for the 5th batch); the flux addition amount is recorded "addition weight per batch" (e.g. 2 kg for the 10th batch). For the "smelting crucible wear" process sensitive node, the target impurity element is Fe (because Fe is easily introduced by crucible wear), and the Fe content data of the corresponding production period of the associated process (smelting) needs to be extracted: if the smelting process lasts for 1 hour per batch, and a total of 30 batches are produced, then the Fe content detection values of the 30 batches are extracted (e.g. 0.07% for the 1st batch, 0.068% for the 2nd batch,..., 0.085% for the 30th batch), and the content change amount of adjacent batches is calculated (e.g. a decrease of 0.002% for the 2nd batch compared to the 1st batch, an increase of 0.003% for the 30th batch compared to the 29th batch), forming a "batch-Fe content-change amount" three-dimensional data table.

[0078] S22212c, calculating the entropy correlation degree of the key process parameter data and the impurity content change data according to the mutual information entropy algorithm.

[0079] It can be understood that calculating the entropy correlation degree of the key process parameter data and the impurity content change data according to the mutual information entropy algorithm is a core statistical method for quantifying the non-linear correlation degree between "process-impurity". The essence is to measure the information sharing degree between process parameter fluctuation and impurity content change through the concept of mutual information in information theory. The higher the entropy correlation degree, the stronger the non-linear correlation between the two; otherwise, the weaker. First, the key process parameter data and the impurity content change data can be discretized, dividing the continuous data into several intervals (e.g. crucible use frequency is divided into "1-20 furnace times", "21-40 furnace times", "41-60 furnace times", and ">60 furnace times"; Fe content change is divided into "decrease >0.005%", "decrease 0-0.005%", "no change", "increase 0-0.005%", and "increase >0.005%").

[0080] Then the joint probability distribution P(X, Y) of the process parameter data (X) and the impurity content change data (Y) and the marginal probability distribution P(X), P(Y) are calculated: for example, in 30 batches of data, the sample number of "crucible use frequency 41-60 times (X3)" and "Fe content increase 0-0.005% (Y4)" is 8, the joint probability P(X3, Y4) = 8 / 30 ≈ 0.267; the sample number of "crucible use frequency 41-60 times" is 10, the marginal probability P(X3) = 10 / 30 ≈ 0.333; the sample number of "Fe content increase 0-0.005%" is 9, the marginal probability P(Y4) = 9 / 30 = 0.3.

[0081] Finally, the entropy association degree is calculated according to the mutual information entropy formula: I(X, Y) = ΣΣP(X, Y) × log2[P(X, Y) / (P(X) × P(Y))]. The contribution term of X3 and Y4 is calculated by substituting the above data: 0.267 × log2(0.267 / (0.333 × 0.3)) ≈ 0.267 × log2(2.67) ≈ 0.267 × 1.42 ≈ 0.379; the contribution terms of all interval combinations are calculated and summed, and the final mutual information value (such as I(X, Y) = 1.2) is obtained. In order to facilitate subsequent calculation, the mutual information value is normalized to the interval [0, 1], and the normalization formula is: entropy association degree = I(X, Y) / log2n (n is the sample number, such as 30 batches n = 30, log230 ≈ 4.91), then the normalized entropy association degree = 1.2 / 4.91 ≈ 0.244.

[0082] S22213c, generating an impurity element process influence weight based on the process sensitive node; wherein the process influence weight is used to reflect the impurity introduction contribution degree of the process in the aluminum alloy ingot production process.

[0083] It can be understood that the process of generating the process influence weight of the impurity element based on the process sensitive node is a key step of quantifying the "process contribution degree" as a calculable parameter, and its essence is to evaluate the relative importance of the process sensitive node associated process to the introduction of impurities in the whole process of aluminum alloy ingot production, generate a weight coefficient with a value range of [0, 1], the higher the weight (closer to 1), the greater the contribution of the process to the introduction of the target impurity element; otherwise, it is smaller. Different impurity elements correspond to a process influence weight corresponding to the process. It can be obtained in the production database, or the corresponding process influence weight can be output by a machine learning model, etc., but not limited to this. The production database refers to the database containing the process influence weight corresponding to the different process sensitive nodes corresponding to different impurity elements. These data can be obtained through laboratory experiments, field measurements and monitoring, and past experience, etc. After obtaining the data, the collected data is sorted, classified and archived, useful information and rules are extracted, and related data is saved to the database to form the production database. The machine learning model is trained by multiple sets of training data, each set of training data in the multiple sets of training data includes: the key parameters of the process sensitive node (such as the sorting accuracy of "recycled aluminum sorting", the number of uses of "smelting crucible", etc. represented by continuous values or discrete levels) and the impurity element content (such as the detection value of Pb and Fe) in the aluminum alloy ingot produced by the corresponding process; and the known contribution of the process to the introduction of impurities.

[0084] S22214c, calculating the third association strength between the impurity element node and the process sensitive node based on the process influence weight and the entropy association degree.

[0085] It can be understood that the third association strength = process influence weight x entropy association degree. The third association strength calculated based on the process influence weight and the entropy association degree is the final step of integrating "nonlinear association degree" and "process contribution importance", and its essence is to fuse the statistical level entropy association degree and the process level process weight through the product model to generate a quantitative index that can reflect both the "process-impurity" association tightness and the relative importance of the process. The value range is mapped to [0, 1] after normalization, and the higher the value, the stronger the association between the impurity element node and the process sensitive node, and the more likely the process is a key link for the introduction of impurities.

[0086] In this way, the third correlation strength is calculated by "extracting synchronous data, calculating entropy, generating process weight, and fusing and calculating intensity", realizing the deep fusion of "nonlinear data correlation" and "process contribution importance", and effectively solving the problems of "linear correlation cannot capture complex process correlation" or "only relying on experience to judge the importance of process lacks data support" in traditional analysis. For example, the core correlation (strength 1.0) of "recycled aluminum sorting-Pb" directly shows that this process is the main source of Pb element introduction, and the sorting control needs to be strengthened; the important correlation (strength 0.743) of "smelting crucible wear-Fe" prompts that the wear condition of the crucible needs to be checked regularly; and the weak correlation (strength 0) of "casting cooling rate-Si" can exclude the influence of this process on the fluctuation of Si content. This hierarchical and quantitative correlation strength provides a clear basis for the design of the weighted edges of the impurity isomerism graph, and enables the isomerism graph to intuitively display the key process links through the thickness and color of the edges, and provides "data + process" double support clues for the subsequent determination of potential sources.

[0087] S2222, obtaining a cross-dimension correlation strength based on the first correlation strength, the second correlation strength, and the third correlation strength.

[0088] It can be understood that the cross-dimension correlation strength is a collection of the first correlation strength, the second correlation strength, and the third correlation strength.

[0089] In this way, by first calculating the first correlation strength between impurity element nodes based on node type and attribute dimension classification (focusing on element chemical synergy and data synergy), the second correlation strength between impurity element nodes and content interval nodes (focusing on element distribution density and content stability), and the third correlation strength between impurity element nodes and process sensitive nodes (focusing on process-impurity nonlinear correlation and process contribution), and then integrating the three types of correlation strengths to obtain the cross-dimension correlation strength, the system designs differentiated calculation logic for the correlation characteristics of different node combinations, reduces the distortion of the correlation strength caused by "one-size-fits-all", and eliminates the magnitude difference and physical meaning difference of various correlations through systematic integration, forming a unified correlation system covering all dimensions of "element-element", "element-content", and "element-process". This technical solution not only retains the unique process meaning of various correlations (such as element chemical synergy and direct introduction of process to impurities), but also highlights the key correlations in the impurity tracing core scenario (such as the strong response relationship between process and element) through global correlation quantification, effectively solving the problems of missed judgment and misjudgment in traditional single-dimension correlation analysis, providing accurate and comparable quantitative basis for the construction of weighted edges of the impurity isomerism graph, assisting in the accurate positioning of potential sources, and significantly improving the comprehensiveness and reliability of aluminum alloy ingot impurity element tracing.

[0090] S223, constructing an isomerism graph structure based on the cross-dimension correlation strength according to the corresponding node type.

[0091] It can be understood that for node construction, differentiated design is made according to node types, and the identification, attributes and functions of each node are clearly distinguished: for the "impurity element node", the element symbol (such as Fe, Si, Pb) is taken as the core identification, and the node attributes (such as atomic number, common compound, multi-batch content average) are marked, for example, the Fe node is marked "Fe, atomic number 26, Al3Fe / Al-Fe-Si phase, average 0.07%", and the node size can be set according to the average content proportion of the element in the aluminum alloy (such as Fe content is higher than Pb, Fe node size is larger); for the "content interval node", the "element-interval" combination can be named (such as "Fe-0.02%-0.1%" "Pb->0.005%"), and the interval attributes (mass risk level, historical occurrence probability) are marked, for example, the "Pb->0.005%" node is marked "high risk, occurrence probability 12%", and the node color is distinguished according to the risk level (over-standard interval red, normal interval green, controllable interval yellow); for the "process sensitive node", the "process-key parameter" can be named (such as "melting-crucible wear" "recycled aluminum sorting-accuracy"), and the process attributes (process influence weight, parameter standard range) are marked, for example, the "melting-crucible wear" node is marked "weight 0.8, use frequency ≤50 heats", and the node shape adopts a special icon (such as a crucible icon representing a melting equipment related node). All nodes can be laid out according to "functional division", for example, impurity element nodes are concentrated in the center of the graph, content interval nodes are distributed around the corresponding element nodes, and process sensitive nodes are distributed on the edge of the graph.

[0092] For the construction of weighted edges, the cross-dimension association strength is the core basis, and the correlation degree can be quantified through the thickness and color depth of the edges, and the style of the edges is distinguished according to the node combination type: for the edges between "impurity element nodes-impurity element nodes", the solid line style is adopted, the thickness of the edge is positively correlated with the first correlation strength (such as strength 0.9 corresponding to thick solid line, strength 0.4 corresponding to thin solid line), and the color adopts the gray system (the higher the strength, the deeper the color), for example, the edge between Fe and Si nodes presents a dark gray thick solid line due to the first correlation strength of 0.92; for the edges between "impurity element nodes-content interval nodes", the dashed line style is adopted, the thickness of the edge is positively correlated with the second correlation strength, and the color is consistent with the risk color of the content interval node (such as the green dashed line between Fe and "0.02%-0.1%" normal interval node, and the strength 0.69 corresponds to the medium thickness); for the edges between "impurity element nodes-process sensitive nodes", the dot-dash line style is adopted, the thickness of the edge is positively correlated with the third correlation strength, and the color adopts the blue system (the higher the strength, the deeper the color), for example, the edge between Pb and "recycled aluminum sorting-accuracy" node presents a dark blue thick dot-dash line due to the third correlation strength of 0.665. At the same time, by setting the correlation strength threshold (such as 0.3), only the edges with strength greater than or equal to 0.3 are retained, and the weakly associated edges (such as the element-process edge with strength 0.2) are removed, to obtain the heterogeneous graph structure.

[0093] In this way, by generating nodes of types including impurity element nodes, content interval nodes and process sensitive nodes based on multiple feature parameters, calculating the cross-dimension association strength of the multiple feature parameters, and finally constructing a heterogeneous graph structure based on the strength according to the node type, the conversion from scattered feature parameters to a systematic correlation network is realized; the technical scheme breaks the association barriers between different types of nodes through the quantification of the cross-dimension association strength, and finally constructs a heterogeneous graph that intuitively presents the complex association between impurity elements and content intervals and process links in the form of a visual network, which not only retains the unique information of each dimension feature (such as element chemical properties and process contribution), but also highlights the core association path through the topological structure.

[0094] S230, performing clustering analysis on the heterogeneous graph structure to generate impurity association modules with hierarchical structure; wherein each impurity association module corresponds to a combination of impurity elements with co-occurrence characteristics.

[0095] It can be understood that the clustering analysis of the heterogeneous graph structure to generate the impurity association module with hierarchical structure is a key step for mining the core impurity combination rule from the complex association network. The essence is to identify the node cluster with the characteristics of "high internal association and low external association" through the graph clustering algorithm, and to aggregate the scattered multi-type nodes into hierarchical modules according to the co-occurrence characteristics and process association. For the heterogeneous graph containing "element-interval-process" three types of nodes, the Louvain algorithm based on the optimization of modularity can be used to realize the automatic division of the community (module) by iteratively calculating the "modularity increment of a node into the adjacent community". For example, the increment ΔQ of the Fe element node into the "Si element + smelting process" community is 0.32 (> 0), indicating that the two should be aggregated into the same module; while the increment ΔQ of the Pb element node into the community is -0.15 (< 0), which forms a module alone. In order to strengthen the process association, the "process sensitive node preferential aggregation" constraint needs to be added to the algorithm, that is, if two element nodes are strongly associated with the same process sensitive node (the strength is greater than or equal to 0.7), the clustering tendency of the two is forced to be improved (such as increasing the ΔQ weight by 0.2), for example, Fe and Si are strongly associated with the "smelting temperature" node, and the two are preferentially divided into the same module when clustering.

[0096] The hierarchical structure construction can form multi-level modules according to the "core-periphery" relationship: the first-level module is the "element-process" core cluster (such as "Fe-Si-smelting process"), which contains directly associated impurity elements and corresponding process sensitive nodes; the second-level module incorporates related content interval nodes (such as "Fe-Si-smelting process-Fe normal interval-Si normal interval") on the basis of the first-level module, reflecting the typical content characteristics of the core elements; the third-level module further associates the indirectly affected process nodes or secondary elements (such as adding the "raw material pretreatment" node and the Mn element), forming a complete impurity association network.

[0097] S240, generating an impurity heterogeneous graph based on the impurity association module.

[0098] It can be understood that the generation of the impurity isomerism graph based on the impurity correlation module is the final step of converting the clustering analysis result into a structured and interpretable final network model, and its essence is to integrate the dispersed node clusters into a global network with clear logical relationships through the visual presentation of the module level and the highlighting of the core correlation path. According to the principle of "core module priority and correlation strength orientation", each impurity correlation module is connected: the primary core module (such as "Fe-Si-smelting process") is identified as the center of the isomerism graph, and the secondary modules are connected through the correlation edges between modules (based on the average correlation strength between nodes in the module), for example, the correlation strength between the "Fe-Si-smelting process" module and the "raw material pretreatment" module = the average correlation strength of all node pairs in the two modules (such as Fe and raw material pretreatment node strength 0.6, Si and raw material pretreatment node strength 0.5, average 0.55), to determine the weight of the edge between modules.

[0099] In this way, by extracting the content ratio between each impurity element and the fluctuation rate of the impurity element content from the impurity information of the aluminum alloy ingot impurity ingot as multi-feature parameters, constructing an isomerism graph structure containing multiple types of nodes and weighted edges based on these parameters, and then performing clustering analysis on the structure to generate impurity correlation modules with hierarchical structure and corresponding co-occurrence feature impurity element combinations, and finally generating an impurity isomerism graph based on the modules, the systematic transformation from the original impurity information to the structured correlation network is realized; this technical solution improves the physical meaning and accuracy of the correlation analysis by extracting multi-feature parameters related to the process mechanism (such as content ratio reflecting element homology and fluctuation rate reflecting content stability), realizes the visual presentation of multi-type node correlation through the isomerism graph structure, highlights the core impurity combination rule through the hierarchical impurity correlation module extracted by clustering analysis, and finally generates an impurity isomerism graph that not only retains microscopic quantitative correlation information (such as weighted edge strength) but also presents macroscopic module coordination rules, effectively solving the problem of impurity correlation fragmentation and interpretation difficulty in traditional tracing, providing an analysis model combining "micro-macro" for impurity tracing, which can quickly lock the potential source of the process link and refine the specific reasons, significantly improving the systematicness, accuracy and efficiency of aluminum alloy ingot impurity element tracing.

[0100] S300, determining a potential source based on the impurity isomerism graph; wherein the potential source is used to reflect the correlation process of each impurity element.

[0101] It can be understood that determining the potential source based on the impurity isomerism graph is the core step of transforming the visual association network into specific process improvement targets, and its essence is to locate the process node that contributes most to the introduction of impurities from the "element-process" association network by tracking the process directionality of the strong association path and the core module in the isomerism graph, solving the problem of "knowing the phenomenon but not knowing the reason" (only knowing that the impurity exceeds the standard but unable to lock the specific process link) in traditional tracing. In the tracking process, the high-risk impurity element node can be taken as the starting point, and the strong association path with a weighted edge strength ≥ 0.7 is extended to the process sensitive node to form a "impurity element-content interval-process sensitive node" tracing chain: for example, in the isomerism graph, the edge strength between the Pb element node and the "Pb->0.005%" exceeding interval node is 0.8 (strong association), the edge strength between the interval node and the "regenerated aluminum sorting-accuracy" process node is 0.85 (strong association), and the direct edge strength between the Pb element and the process node is 0.95 (core association), so it can be preliminarily determined that "regenerated aluminum sorting" is the potential source of Pb exceeding; if the Fe element node is simultaneously strongly associated with the "melting-crucible wear" (strength 0.8) and "raw material-waste aluminum purity" (strength 0.65) process nodes, the weights of the two in the core module can be further compared - because the Fe element is located in the "Fe-Si-melting process" core module (the weight of the melting process node in the module is 0.45), which is higher than the weight (0.3) in the "raw material related" secondary module, so "melting-crucible wear" is listed as the main potential source. The key process is determined by the aggregation degree and influence weight of the process sensitive node in the module: for example, in the "Fe-exceeding interval-melting crucible-melting temperature" module, there are 2 process sensitive nodes (crucible wear, temperature fluctuation), among which the process influence weight of the "crucible wear" node is 0.8, the association strength with the Fe element is 0.85, and the connectivity (number of connected nodes) in the module is 5, which is higher than the connectivity 3 of the "temperature fluctuation" node, indicating that "crucible wear" is the core process node of the module; if a module contains "Pb-Cd-regenerated aluminum sorting-manual sorting" 4 nodes, and Pb and Cd both form strong association (strength 0.9 and 0.88 respectively) with the "regenerated aluminum sorting" node, it can be determined that this process is the common potential source of Pb and Cd, which is consistent with the process characteristics of "same source introduction" of the two.

[0102] S400, determining the final tracing result of the impurity element based on the production chain information and the potential source.

[0103] Exemplarily, the potential occurrence source can be verified, the possibility of the potential occurrence source can be determined, and the final traceability result can be determined according to the possibility through the actual operation record in the production chain information; the core cause can also be locked through the mode of "production chain information time sequence matching + multi-dimensional cross verification + root source contribution quantification", for example, for the potential occurrence source "regenerated aluminum sorting process" (associated with Pb and Cd exceeding the standard), first, the production batches of Pb and Cd exceeding the standard are time-aligned with the "waste aluminum raw material entry record", "sorting equipment operation log" and "manual sorting sampling data" in the production chain information in the same period: from the raw material dimension, it is found that the Pb content detection value of the 3rd batch of waste aluminum (supplier B) entering the factory in this period is 0.008% (exceeding the standard by 2.6 times), which is completely coincident with the Pb exceeding standard batch; from the equipment dimension, the magnetic field strength of the magnetic separator of the sorting equipment is stable at 8500Gs (the standard value is greater than or equal to 12000Gs) in this period, and the equipment parameter abnormal record is synchronous with the exceeding standard period. Subsequently, non-core factors are excluded through cross verification: after replacing the qualified waste aluminum raw material (supplier A, Pb content 0.002%), even if the equipment parameter is not adjusted, the Pb content is reduced to 0.003% (up to the standard), which shows that the raw material problem is the basic cause; under the premise of using qualified raw materials, the magnetic field strength of the magnetic separator is calibrated to 12000Gs, and the Pb content is further reduced to 0.0025%, which proves that the equipment parameter and the operation specification are auxiliary causes. Finally, the proportion of each factor is quantified through the "root source contribution model" (raw material problem contribution 70%, equipment abnormality contribution 20%), and the "3rd batch of waste aluminum raw material Pb exceeding the standard + sorting equipment magnetic field strength deficiency" is determined as the final traceability result, and is marked as "raw material problem is the core root source, and needs to be rectified in priority", and the like, but not limited thereto.

[0104] In a possible implementation, in step S400, the final traceability result of the impurity element is determined based on the production chain information and the potential occurrence source, including:

[0105] S410, multi-dimensional verification data of each potential occurrence source corresponding process is extracted based on the production chain information; wherein the multi-dimensional verification data includes process actual process parameters, raw material full cycle data and equipment operation record.

[0106] It can be understood that based on the production chain information, the multi-dimensional verification data of the corresponding process of each potential source is the core step of providing "actual production evidence chain" for the potential source, and its essence is to directly mine the whole process data related to the potential source (such as recycled aluminum sorting and smelting crucible wear). For example, if the potential source is "recycled aluminum sorting process" (associated with Pb and Cd exceeding the standard), "process actual process parameters" need to extract real-time running data of sorting equipment (such as magnetic field strength of magnetic separator, photoelectric sorting identification accuracy, 1 record every 10 minutes, which needs to cover the period from 2 hours before Pb exceeds the standard to the end of the standard); "raw material whole cycle data" needs to extract batch information of waste aluminum entering the factory (supplier, purchase date), incoming component test report (focus on Pb and Cd content, which needs to include test time, test method such as ICP-OES), raw material storage record (whether there is cross contamination caused by mixing different batches of waste aluminum); "equipment running record" needs to extract maintenance log of sorting equipment (the latest maintenance time, replaced parts such as sorting conveyor belt), fault alarm record (such as whether the abnormal alarm time of magnetic field strength sensor coincides with the exceeding standard period).

[0107] S420, generating a comprehensive traceability sequence based on the multi-dimensional verification data and each potential source.

[0108] It can be understood that based on the multi-dimensional verification data and each potential source, the comprehensive traceability sequence is a key step of converting scattered verification data into "time-sequenced and causal" traceability clues, and its essence is to connect the process parameter fluctuation, raw material change, equipment state and other data corresponding to a potential source into a continuous sequence according to "time axis + causal logic", and intuitively present the complete causal chain of "abnormal data occurrence → impurity exceeding standard". Exemplarily, for each potential source, the raw material matching degree, equipment contribution degree and process deviation quantitative value can be extracted from the multi-dimensional verification data, and then the raw material matching degree, equipment contribution degree and process deviation quantitative value of each impurity element are weighted to obtain a traceability score sequence, and the potential sources are arranged based on the traceability score sequence to obtain a comprehensive traceability sequence.

[0109] The sequence can also be constructed in a four-layer structure of "time node-exception type-impact degree-causal verification", for example, for the potential source of "smelting crucible wear (associated with Fe exceeding the standard)", first divide the shaft section with a time granularity of 10 minutes (covering 3 hours before Fe exceeds the standard to 2 hours after exceeding the standard), and then label the multi-dimensional verification data exceptions at each time node: for example, "T-180min (3 hours before exceeding the standard), the equipment operation record shows that the crucible has been used for 49 furnace times (close to the upper limit of 50 furnace times), and is labeled as 'critical abnormality of equipment state, low impact degree (weight 0.2)'; "T-60min (1 hour before exceeding the standard), the raw material full cycle data record shows that the incoming waste aluminum contains Mn impurities (content 0.12%), and is labeled as 'raw material contains high hardness impurities, accelerates crucible wear, medium impact degree (weight 0.3)'; "T-0min (start of exceeding the standard), the actual process parameter of the process shows that the smelting temperature is locally overheated to 750°C (standard ≤730°C), and is labeled as 'temperature anomaly accelerates crucible corrosion, Fe content exceeds the standard to 0.11% for the first time, high impact degree (weight 0.5)'; "T+60min (1 hour after exceeding the standard), the equipment offline detection shows that the inner wall wear of the crucible reaches 2.1mm (exceeding the allowable value of 1mm), and is labeled as 'equipment wear confirmation, Fe content continues to exceed the standard, causal verification is established'"; finally, all nodes are connected in time sequence to form a comprehensive traceability sequence of "critical equipment → raw material acceleration → process trigger → wear confirmation".

[0110] In one possible implementation, in step S420, a comprehensive traceability sequence is generated based on the multi-dimensional verification data and each potential source, including:

[0111] S421, for each potential source, extracting the raw material matching degree, equipment contribution degree and process deviation quantitative value from the multi-dimensional verification data; wherein the raw material matching degree is used to reflect the difference between the content of impurity elements and the raw material admission standard, the equipment contribution degree is used to reflect the contribution of the equipment to the introduction of impurity elements, and the process deviation quantitative value is used to reflect the deviation degree of the actual process parameter of the process from the standard process parameter of the process.

[0112] It can be understood that the raw material matching degree focuses on the relevance of impurity elements and raw materials, which is calculated by the standardization formula of (actual impurity content of raw material - raw material access standard) / raw material access standard (result ≥ 0, the larger the value, the more serious the raw material exceeds the standard), for example, the actual content of Pb in a batch of waste aluminum is 0.008%, and the access standard is 0.003%, then the raw material matching degree = (0.008-0.003) / 0.003 ≈ 1.67; if the raw material does not exceed the standard (such as actual 0.002%), the matching degree is 0, which avoids misjudging qualified raw materials as risk sources. The equipment contribution degree can be quantified by combining equipment parameters and impurity introduction process mechanism. For “smelting crucible wear” type equipment, it can be calculated by “(actual use furnace times - standard furnace times) / standard furnace times x equipment influence weight” (for example, the standard is 50 furnace times, the actual is 60 furnace times, and the equipment influence weight is 0.8, then the contribution degree = (60-50) / 50 x 0.8 = 0.16); for “sorting equipment”, it can be calculated by “(standard parameter - actual parameter) / standard parameter x separation efficiency of equipment to impurities” (for example, the standard magnetic field is 12000Gs, the actual is 9000Gs, and the separation efficiency coefficient is 0.9, then the contribution degree = (12000-9000) / 12000 x 0.9 = 0.225), the larger the value, the greater the contribution of the equipment to the introduction of impurities. The process deviation quantitative value is calculated for the deviation degree of key process parameters, which adopts the normalization formula of “(actual parameter - standard parameter) / standard parameter of allowable fluctuation range” (the result is mapped to [0, 1], 0 is no deviation, and 1 is serious deviation), for example, the standard smelting temperature is 720±10℃, and the actual is 750℃, then the deviation quantitative value = (750-730) / 10 = 2 (when exceeding the range, it is forced to take 1), which intuitively reflects the out-of-control degree of process parameters.

[0113] S422, weighting the raw material matching degree, equipment contribution degree and process deviation quantitative value of each impurity element to obtain a traceability score sequence, and arranging the potential occurrence sources based on the traceability score sequence to obtain a comprehensive traceability sequence.

[0114] It can be understood that the raw material matching degree, equipment contribution degree and process deviation quantitative value of each impurity element are weighted to obtain a traceability score sequence, and the potential occurrence sources are arranged based on the sequence, which is a key step of systematized integration of multi-dimensional verification data. The essence is to convert the quantitative indicators of three dimensions into unified traceability scores through reasonable weight distribution, and then sort the comprehensive traceability sequence according to the score.

[0115] The weight can be allocated according to the process importance of different potential sources in each dimension. For example, for a raw material and equipment dominant process such as "recycled aluminum sorting", the raw material matching degree weight can be set to 0.4 (whether the raw material is qualified is the basis), the equipment contribution degree weight is 0.4 (the efficiency of the sorting equipment directly affects the impurity separation), and the process deviation quantization value weight is 0.2 (the operation deviation has a relatively small influence), for example, the raw material matching degree of a certain potential source is 1.67, the equipment contribution degree is 0.225, and the process deviation quantization value is 0.3, then the traceability score = 1.67x0.4 + 0.225x0.4 + 0.3x0.2≈0.668 + 0.09 + 0.06 = 0.818; for a device and process dominant process such as "smelting", the equipment contribution degree weight is 0.4 (such as crucible wear), the process deviation quantization value weight is 0.4 (such as temperature control), and the raw material matching degree weight is 0.2 (the influence of raw material has been reflected in the previous process), for example, the equipment contribution degree is 0.16, the process deviation quantization value is 1, and the raw material matching degree is 0.2, then the traceability score = 0.16x0.4 + 1x0.4 + 0.2x0.2 = 0.064 + 0.4 + 0.04 = 0.504. The higher the score, the greater the comprehensive influence of the potential source on the impurity exceeding the standard.

[0116] In this way, the systematic integration and priority sorting of multi-dimensional verification data are realized, that is, the objective quantization of the influence of each dimension is ensured through the accurate extraction of the raw material matching degree, the equipment contribution degree, and the process deviation quantization value, and the influence of the process mechanism on the importance of each dimension is reflected through targeted weight allocation (adjusting the weight according to the process characteristics), thereby reducing the false judgment caused by a single index and the like.

[0117] S430, screening based on the comprehensive traceability sequence to obtain the final traceability result of the impurity element.

[0118] It can be understood that by setting scientific screening rules (such as score threshold, causal verification, and process rationality verification), weakly associated and false positive potential sources are eliminated from the comprehensive traceability sequence, and core processes and specific causes that have a direct and key influence on the impurity exceeding the standard are retained, thereby solving the problem of "multiple source interference and primary and secondary confusion" that may exist in the comprehensive traceability sequence.

[0119] First, set the traceability score threshold (usually take 0.5-0.6 according to production accuracy requirements), only keep the potential source of score ≥ threshold, for example, the score of "recycled aluminum sorting" in the comprehensive traceability sequence is 0.818 (≥0.6), the score of "smelting refining" is 0.35 (<0.6), and the score of "raw material storage" is 0.12 (<0.6), so "recycled aluminum sorting" is preliminarily screened out as the core candidate; secondly, the cause and effect closed loop verification is carried out, whether the abnormal data of the candidate source forms a complete closed loop of "abnormal preposition→synchronous occurrence→elimination and recovery", such as "recycled aluminum sorting" of raw material exceeding standard (T-60min)→insufficient magnetic field (T-30min)→Pb exceeding standard (T+0min)→changing raw material+calibrating equipment (T+120min)→Pb meeting standard (T+180min), the cause and effect chain is complete and has no breakpoint, and the verification is passed; finally, the process rationality verification is carried out, and it is confirmed that the inducement of the candidate source conforms to the aluminum alloy production mechanism, such as "recycled aluminum sorting" of raw material Pb exceeding standard and insufficient magnetic field leading to Pb not being separated, which conforms to the process logic of "raw material impurity→equipment sorting failure→impurity residue", and if the candidate source is "casting cooling" (score 0.55), but the cooling speed has no direct correlation with Pb content, it is determined that the process is not reasonable, and it is removed.

[0120] For example, in the comprehensive traceability sequence of synchronous exceeding of Fe and Si, "melting temperature control" scores 0.62 for Fe and 0.58 for Si (both ≥0.6), and the inducement is "temperature overtemperature leading to element dissolution", which is determined as the co-cause traceability result; "melting crucible wear" scores 0.59 for Fe (close to the threshold) and 0.3 for Si (<0.6), and only Fe is directly related to the crucible material, which is determined as the special cause traceability result of Fe, and finally the combined conclusion of "co-cause: melting temperature overtemperature; Fe special cause: crucible wear" is screened out.

[0121] In this way, the final traceability result is determined through the three-step process of "extracting multi-dimensional verification data-generating comprehensive traceability sequence-screening core causes", realizing the full-chain closed loop of "data extraction-clue integration-accurate positioning". The multi-dimensional verification data provides solid production evidence for the potential source of occurrence (avoiding relying only on correlation network inference), the comprehensive traceability sequence converts scattered data into time-sequenced and causal clues (solving the problem of data fragmentation), and finally the scientific screening rule eliminates interference items and locks the core causes (solving the problem of primary and secondary confusion). Specifically, the multi-dimensional verification data ensures the comprehensiveness of the evidence, the comprehensive traceability sequence embodies the logicality of the clues, and the screening link guarantees the accuracy of the conclusion. This layer-by-layer progressive process not only avoids the subjectivity of "judging by experience" in traditional traceability, but also reduces the false positive results of "strong data correlation but meaningless process", so that the final traceability result is not only clear about "which process", but also accurate to "specific causes (such as excessive raw materials in a certain batch, abnormal equipment parameters)", and has three reliabilities of "score support, cause and effect closed loop, and reasonable process", providing direct and implementable basis for targeted rectification of impurity elements in aluminum alloy ingots, and significantly improving the practical value and production guiding significance of the traceability result.

[0122] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0123] Corresponding to the aluminum alloy ingot impurity element traceability analysis method of the above embodiment, the embodiments of the present application also provide an aluminum alloy ingot impurity element traceability analysis system, and each module of the system can realize each step of the aluminum alloy ingot impurity element traceability analysis method. Figure 3 The structure block diagram of the aluminum alloy ingot impurity element traceability analysis system provided by the embodiments of the present application is shown, and only the parts related to the embodiments of the present application are shown for ease of description.

[0124] Referring to Figure 3 The aluminum alloy ingot impurity element traceability analysis system comprises:

[0125] The acquisition module is configured to acquire aluminum alloy ingot impurity information and production chain information; wherein the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in a plurality of aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot.

[0126] The construction module is configured to construct an impurity isomorphism graph based on the aluminum alloy ingot impurity information; wherein the impurity isomorphism graph is used to reflect the correlation between impurity elements.

[0127] The first determining module is used to determine potential sources of occurrence based on the impurity isomorphism diagram; wherein, the potential sources of occurrence are used to reflect the associated processes of each impurity element.

[0128] The second determination module is used to determine the final source tracing result of impurity elements based on production chain information and potential sources.

[0129] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0130] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described module division is merely an example. In practical applications, the above functions can be assigned to different modules as needed, that is, the internal structure of the system can be divided into different modules to complete all or part of the functions described above. The modules in the embodiments can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0131] This application also provides an aluminum alloy ingot impurity element traceability analysis device. Figure 4 This is a schematic diagram of the structure of an aluminum alloy ingot impurity element tracing analysis device 6 provided in an embodiment of this application. Figure 4 As shown, the aluminum alloy ingot impurity element tracing and analysis device 6 of this embodiment includes: at least one processor 60 ( Figure 4 Only one is shown in the image), at least one memory 61 ( Figure 4 (Only one is shown in the image) and a computer program 62 stored in the at least one memory 61 and executable on the at least one processor 60. When the processor 60 executes the computer program 62, it causes the aluminum alloy ingot impurity element traceability analysis device 6 to implement the steps in any of the above-described aluminum alloy ingot impurity element traceability analysis method embodiments, or causes the aluminum alloy ingot impurity element traceability analysis device 6 to implement the functions of each module in the above-described system embodiments.

[0132] Exemplarily, the computer program 62 can be divided into one or more modules / units, which are stored in the memory 61 and executed by the processor 60 to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program 62 in the aluminum alloy ingot impurity element traceability analysis device 6.

[0133] The aluminum alloy ingot impurity element traceability analysis device 6 can be a desktop computer, a notebook, a palm computer, a cloud server and the like. The aluminum alloy ingot impurity element traceability analysis device can include, but is not limited to, a processor 60, a memory 61. Those skilled in the art can understand that, Figure 4 The aluminum alloy ingot impurity element traceability analysis device 6 is only an example and does not constitute a limitation on the aluminum alloy ingot impurity element traceability analysis device 6, and can include more or fewer components than those shown, or combine certain components, or different components, for example, can also include input / output devices, network access devices, buses and the like.

[0134] The processor 60 can be a central processing unit (CPU), and the processor 60 can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0135] The memory 61 can be an internal storage unit of the aluminum alloy ingot impurity element traceability analysis device 6 in some embodiments, such as a hard disk or a memory of the aluminum alloy ingot impurity element traceability analysis device 6. The memory 61 can also be an external storage device of the aluminum alloy ingot impurity element traceability analysis device 6 in other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like equipped on the aluminum alloy ingot impurity element traceability analysis device 6. Further, the memory 61 can include both the internal storage unit and the external storage device of the aluminum alloy ingot impurity element traceability analysis device 6. The memory 61 is used to store an operating system, an application program, a boot loader, data, and other programs, such as program codes of the computer program, and the like. The memory 61 can also be used to temporarily store data that has been output or is to be output.

[0136] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the steps in any of the above method embodiments.

[0137] The embodiments of the present application provide a computer program product. When the computer program product is run on the aluminum alloy ingot impurity element traceability analysis device, the aluminum alloy ingot impurity element traceability analysis device implements the steps in any of the above method embodiments.

[0138] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the embodiments of the present application implement all or part of the processes in the above method embodiments, which can be completed by instructing related hardware through a computer program. The computer program can be stored in a computer readable storage medium. The computer program is executed by a processor to implement the steps in each of the above method embodiments. The computer program includes computer program codes, which can be in the form of source codes, object codes, executable files, or some intermediate forms. The computer readable medium at least includes any entity or device capable of carrying the computer program codes to the aluminum alloy ingot impurity element traceability analysis device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunications signal, and a software distribution medium. For example, a U disk, a mobile hard disk, a magnetic disk or an optical disk, and the like.

[0139] In the above embodiments, the description of each embodiment is focused on, and the part not described or recorded in a certain embodiment can be referred to the relevant description of other embodiments.

[0140] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0141] In the embodiments provided in the present application, it should be understood that the disclosed aluminum alloy ingot impurity element traceability analysis device and method can be implemented in other ways. For example, the above-described aluminum alloy ingot impurity element traceability analysis device embodiments are only illustrative, for example, the division of the modules is only a logical functional division, and actual implementation can have another division method, for example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed can be indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0142] The modules described as separate components can or can not be physically separated, and the components shown as modules can or can not be physical modules, that is, they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiments according to actual needs.

[0143] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method of trace analysis of impurity elements in an aluminum alloy ingot, characterized by, The method comprises the following steps: obtaining aluminum alloy ingot impurity information and production chain information; wherein, the aluminum alloy ingot impurity information is used to reflect the composition and corresponding content interval of impurities in multiple aluminum alloy ingots, and the production chain information is used to reflect the production process of the aluminum alloy ingot; constructing an impurity isomorphism graph based on the aluminum alloy ingot impurity information; wherein, the impurity isomorphism graph is used to reflect the correlation between impurity elements; determining a potential source based on the impurity isomorphism graph; wherein, the potential source is used to reflect the correlation process of each impurity element; determining the final traceability result of the impurity element based on the production chain information and the potential source; the method of constructing an impurity isomorphism graph based on the aluminum alloy ingot impurity information comprises: extracting multi-element feature parameters based on the aluminum alloy ingot impurity information; wherein, the multi-element feature parameters include the content ratio between impurity elements and the impurity element content fluctuation rate; constructing an isomorphism graph structure based on the multi-element feature parameters; wherein, the isomorphism graph structure includes multiple types of nodes and weighted edges; performing cluster analysis on the isomorphism graph structure to generate impurity correlation modules with hierarchical structure; wherein, each impurity correlation module corresponds to an impurity element combination with co-occurrence characteristics; generating an impurity isomorphism graph based on the impurity correlation module; the method of constructing an isomorphism graph structure based on the multi-element feature parameters comprises: generating node types and attribute dimensions based on the multi-element feature parameters; wherein, the node types include impurity element nodes, content interval nodes and process sensitive nodes, and the process sensitive nodes refer to the process links corresponding to the abnormal characteristics of the impurity elements; calculating the cross-dimension correlation strength of the multi-element feature parameters based on the node types and the attribute dimensions; constructing an isomorphism graph structure according to the corresponding node types based on the cross-dimension correlation strength; the method of calculating the cross-dimension correlation strength of the multi-element feature parameters based on the node types and the attribute dimensions comprises: based on the node types and the attribute dimensions, calculating the first correlation strength between the impurity element nodes, the second correlation strength between the impurity element nodes and the content interval nodes, and the third correlation strength between the impurity element nodes and the process sensitive nodes; obtaining the cross-dimension correlation strength based on the first correlation strength, the second correlation strength and the third correlation strength.

2. The method of trace analysis of impurity elements in aluminum alloy ingots of claim 1, wherein, the method of calculating the first correlation strength between the impurity element nodes based on the node types and the attribute dimensions comprises: generating time series data based on the impurity element content ratio in the node types of the impurity element nodes and the attribute dimensions; wherein, the time series data is used to reflect the change of the impurity element content ratio in multiple aluminum alloy ingots; calculating a linear correlation coefficient based on the time series data; wherein, the linear correlation coefficient is used to reflect the synergistic trend of the impurity element content ratio with the production of aluminum alloy ingots; generating a correction factor based on the formation tendency of the impurity element; wherein, the formation tendency of the impurity element is used to indicate the thermodynamic possibility of the impurity element forming complex inclusions in the aluminum alloy smelting and solidification process. The linear correlation coefficient is weighted based on the correction factor to obtain a first correlation strength between the impurity element nodes.

3. The method of trace analysis of impurity elements in aluminum alloy ingots of claim 1, wherein, The second correlation strength between the impurity element node and the content interval node is calculated based on the node type and the attribute dimension, including: An interval density value is calculated based on the impurity element content fluctuation rate of the impurity element node and the content interval node in the node type and the attribute dimension, wherein the interval density value is used to reflect the distribution density of the impurity element content corresponding to the content interval; A fluctuation coefficient is generated based on the impurity element content fluctuation rate; The second correlation strength between the impurity element node and the content interval node is calculated based on the interval density value and the fluctuation coefficient.

4. The method of trace analysis of impurity elements in aluminum alloy ingots of claim 1, wherein, The third correlation strength between the impurity element node and the process sensitive node is calculated based on the node type and the attribute dimension, including: Key process parameter data of the associated process of the process sensitive node is extracted based on the process sensitive node in the node type, and corresponding impurity content change data is extracted based on the impurity element node in the node type; wherein the associated process of the process sensitive node refers to the production process to which the process sensitive node belongs; An entropy correlation degree of the key process parameter data and the impurity content change data is calculated according to the mutual information entropy algorithm; A process influence weight of the impurity element is generated based on the process sensitive node; wherein the process influence weight is used to reflect the contribution degree of the process to the introduction of impurities in the production process of aluminum alloy ingots; The third correlation strength between the impurity element node and the process sensitive node is calculated based on the process influence weight and the entropy correlation degree.

5. The method of trace analysis of impurity elements in aluminum alloy ingots of claim 1, wherein, The final traceability result of the impurity element is determined based on the production chain information and the potential occurrence source, including: Multi-dimensional verification data of the corresponding process of each potential occurrence source is extracted based on the production chain information; wherein the multi-dimensional verification data includes actual process parameters of the process, raw material full cycle data, and equipment operation records; A comprehensive traceability sequence is generated based on the multi-dimensional verification data and each potential occurrence source; The final traceability result of the impurity element is obtained by screening based on the comprehensive traceability sequence.

6. The method of trace analysis of impurity elements in an aluminum alloy ingot of claim 5 wherein, The comprehensive traceability sequence is generated based on the multi-dimensional verification data and each potential occurrence source, including: For each potential occurrence source, raw material matching degree, equipment contribution degree, and process deviation quantization value are extracted from the multi-dimensional verification data; wherein the raw material matching degree is used to reflect the difference between the content of the impurity element and the raw material admission standard, the equipment contribution degree is used to reflect the contribution of the equipment to the introduction of the impurity element, and the process deviation quantization value is used to reflect the deviation degree of the actual process parameters of the process from the standard process parameters of the process; The raw material matching degree, the equipment contribution degree, and the process deviation quantization value of each impurity element are weighted to obtain a traceability score sequence, and the potential occurrence sources are arranged based on the traceability score sequence to obtain a comprehensive traceability sequence.

7. An apparatus for trace analysis of impurity elements in an aluminum alloy ingot, characterized by comprising: A computer program product comprising a storage medium to store the program code of a computer program, the program code dynamicall y executable by a processor to cause the processor to carry out the method according to any one of claims 1 to 6. A computer program for performing the method according to any one of claims 1 to 6 when the computer program is executed by a processor. A computer program for performing the method according to any one of claims 1 to 6 when the computer program is executed by a processor. A computer program for performing the method according to any one of claims

Images