Rare earth permanent magnet material magnetic performance consistency manufacturing process control method and system

Abstract

The invention relates to the technical field of rare earth, and provides a manufacturing process control method and system for magnetic performance consistency of a rare earth permanent magnet material. The method comprises the following steps: analyzing and generating a rare earth permanent magnet material preparation scheme according to a rare earth raw material and demand information; multi-scale magnetic performance consistency variation prediction is carried out according to the rare earth permanent magnet material preparation scheme, and a magnetic performance variation prediction result is generated; the abrupt change results are subjected to cause tracing mining through smelting powder making, orientation forming and sintering heat treatment schemes, and corresponding cause tracing vectors are established; and performing optimization adjustment on the preparation scheme according to the three types of tracing vectors to obtain a preparation scheme optimization result. According to the method, the technical problem of magnetic performance consistency fluctuation caused by insufficient cooperative control of multiple process links in the existing rare earth permanent magnet material preparation process is solved, and accurate identification of magnetic performance variation and dynamic optimization adjustment of a preparation scheme are realized through multi-scale prediction and process stage tracing analysis; and the magnetic performance consistency of the rare earth permanent magnet material is improved.

Description

Technical Field

[0001] The present application relates to the field of rare earth technology, and in particular to a manufacturing process control method and system for ensuring consistency of magnetic properties of rare earth permanent magnet materials. Background Art

[0002] Currently, rare earth permanent magnets, as high-performance magnetic functional materials, are widely used in a variety of high-end manufacturing applications, including new energy vehicles, elevator traction, high-speed motors, wind power generation, and consumer electronics. To meet the high consistency requirements of core magnetic performance metrics such as magnetic energy product, coercivity, and remanence in downstream applications, the manufacturing process places higher demands on process control, including raw material purity, component ratio, grain orientation, and dense sintering. However, existing manufacturing processes lack the ability to dynamically predict the evolution of magnetic properties. The coupling mechanism between parameters in the entire process—melting, pulverizing, orientation forming, and sintering heat treatment—is unclear, leading to significant batch variability in product magnetic performance parameters (such as remanence, coercivity, and maximum energy product), making it difficult to meet the stringent requirements for magnetic performance consistency in high-end applications. Furthermore, existing control methods often rely on empirical trial and error, lacking a data-driven quantitative control mechanism. Especially in a multi-physics coupling environment, dynamic linkages between key nodes such as melting, pulverizing, orientation forming, and sintering heat treatment cannot be established. This results in process parameter adjustments lagging behind the evolution of material properties, making it difficult to achieve closed-loop control of magnetic performance consistency throughout the entire process. Therefore, there is an urgent need for a manufacturing process control method that is driven by the coordinated efforts of raw material information and magnetic performance targets, which can achieve consistent control of magnetic properties throughout the entire process from smelting and powder making, orientation molding to sintering heat treatment, and ensure the manufacturing quality of highly consistent rare earth permanent magnet materials. Summary of the Invention

[0003] This application provides a manufacturing process control method and system for the consistency of magnetic properties of rare earth permanent magnet materials, aiming to solve the technical problem of insufficient coordinated control of multiple process links in the existing rare earth permanent magnet material preparation process leading to fluctuations in magnetic property consistency.

[0004] The first aspect disclosed in the present application provides a manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials, the method comprising: performing multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, the rare earth permanent magnet material preparation plan comprising a smelting and pulverizing plan, an orientation molding plan, and a sintering heat treatment plan; performing multi-scale magnetic property consistency variation prediction based on the rare earth permanent magnet material preparation plan to generate a magnetic property variation prediction result; performing causal mining on the magnetic property variation prediction result based on the smelting and pulverizing plan to establish a smelting and pulverizing causal vector; performing causal mining on the magnetic property variation prediction result based on the orientation molding plan to establish an orientation molding causal vector; performing causal mining on the magnetic property variation prediction result based on the sintering heat treatment plan to establish a sintering heat treatment causal vector; performing optimal adjustment on the rare earth permanent magnet material preparation plan based on the smelting and pulverizing causal vector, the orientation molding causal vector, and the sintering heat treatment causal vector to obtain a preparation plan optimization result.

[0005] Another aspect disclosed in the present application provides a manufacturing process control system for the consistency of magnetic properties of rare earth permanent magnet materials, the system comprising: a control analysis module: performing multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, the rare earth permanent magnet material preparation plan including a smelting and powdering plan, an orientation molding plan, and a sintering heat treatment plan; a variation prediction module: performing multi-scale magnetic property consistency variation prediction based on the rare earth permanent magnet material preparation plan to generate a magnetic property variation prediction result; a powdering traceability mining module: performing multi-scale magnetic property consistency variation prediction based on the smelting and powdering plan to generate a magnetic property variation prediction result; The magnetic property variation prediction results are traced back to establish a smelting and pulverizing traceability vector; the forming traceability mining module: the magnetic property variation prediction results are traced back to establish an orientation forming traceability vector according to the orientation forming scheme; the sintering traceability mining module: the magnetic property variation prediction results are traced back to establish a sintering heat treatment traceability vector according to the sintering heat treatment scheme; the optimization adjustment module: the rare earth permanent magnet material preparation scheme is optimized and adjusted according to the smelting and pulverizing traceability vector, the orientation forming traceability vector and the sintering heat treatment traceability vector to obtain the preparation scheme optimization result.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials mentioned above first conducts multi-node process analysis based on the actual information of rare earth raw materials and the performance requirements of the target material, and formulates an overall preparation plan covering the three key process stages of smelting and powder making, orientation forming and sintering heat treatment. Subsequently, by predicting the consistency of magnetic properties of the preparation plan at multiple scales, the risk links where performance abnormalities may occur are identified. For the predicted results of magnetic property variations, the cause tracing analysis is conducted in combination with the plans of each process stage, and the tracing vectors of the three aspects of smelting and powder making, orientation forming and sintering heat treatment are constructed. Finally, based on this tracing information, the original preparation plan is optimized and adjusted to form a better process configuration plan to improve the consistency and stability of the material's magnetic properties and achieve closed-loop control of the process quality.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0009] Figure 1 The figure is a flow chart of a method for controlling the manufacturing process of the magnetic property consistency of rare earth permanent magnet materials in one embodiment.

[0010] Figure 2 This is an architecture diagram of a control system for a manufacturing process of magnetic property consistency of rare earth permanent magnet materials in one embodiment.

[0011] Explanation of the reference numerals: control analysis module 11 , abnormality prediction module 12 , pulverizing cause mining module 13 , forming cause mining module 14 , sintering cause mining module 15 , optimization and adjustment module 16 . DETAILED DESCRIPTION

[0012] The embodiments of the present application provide a manufacturing process control method and system for the consistency of magnetic properties of rare earth permanent magnet materials, thereby solving the technical problem of insufficient coordinated control of multiple process links in the existing rare earth permanent magnet material preparation process, which leads to fluctuations in magnetic property consistency.

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0015] Example 1, as Figure 1 As shown, the present application provides a manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials, the method comprising: Multi-node preparation control analysis is performed based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, which includes a smelting and powdering plan, an orientation molding plan, and a sintering heat treatment plan.

[0016] In the embodiments of the present application, first, rare earth raw material information is obtained, including but not limited to the raw material type (such as rare earth metals such as Nd, Pr, Dy, Tb), purity level (such as not less than 99.5%), oxygen content, impurity content (such as Fe, O, C, N, etc.), particle size distribution (such as D50 value), batch consistency and other chemical and physical parameters. This information is usually obtained through quality reports and factory test data provided by the raw material supplier. Subsequently, based on the demand information for rare earth permanent magnet materials, the magnetic performance indicators (such as coercivity ≥1000kA / m, remanence ≥1.2T, magnetic energy product ≥40MGOe) and dimensional tolerance (such as within ±0.02mm) required for the final product are clarified. Based on the understanding of raw materials and target performance, multi-node preparation control analysis is carried out, that is, the entire preparation process is divided into three key nodes, namely, melting and powder making, orientation molding and sintering heat treatment, and parameter planning is carried out for each node respectively. In this process, historical records of each node are extracted, and the best matching records are obtained from the historical records through matching maximization search to form melting and powder making schemes, orientation molding schemes and sintering heat treatment schemes. Among them, the melting and powder making scheme includes melting method (such as induction melting, vacuum arc melting), melting temperature (usually 1 The sintering heat treatment plan includes the sintering heating rate (e.g., 10°C / min), sintering temperature (typically 1050°C to 1120°C), holding time (e.g., 2 to 6 hours), and atmosphere (e.g., hydrogen, argon, or vacuum sintering). Finally, the obtained smelting and powdering plan, sintering heat treatment plan, and sintering heat treatment plan are packaged into a comprehensive rare earth permanent magnet material preparation plan, which serves as the basis for subsequent magnetic property prediction and process traceability. This plan has the advantages of a clear structural hierarchy, complete process parameters, and strong traceability, and can be directly applied to process modeling and optimization iterations.

[0017] Furthermore, the present application provides a method for performing multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, including: According to the rare earth raw material information and the rare earth permanent magnet material requirements, the smelting and pulverizing node control analysis is performed to obtain the smelting and pulverizing plan; according to the rare earth raw material information and the rare earth permanent magnet material requirements, the orientation molding node control analysis is performed to obtain the orientation molding plan; according to the rare earth raw material information and the rare earth permanent magnet material requirements, the sintering heat treatment node control analysis is performed to obtain the sintering heat treatment plan; the smelting and pulverizing plan, the orientation molding plan and the sintering heat treatment plan are packaged into the rare earth permanent magnet material preparation plan.

[0018] Preferably, after obtaining rare earth raw material information and rare earth permanent magnet material demand, the smelting and pulverizing node is first controlled and parsed based on this information. During this process, multiple historical smelting and pulverizing records are retrieved from historical logs. Each smelting and pulverizing record includes historical rare earth raw material information, historical rare earth permanent magnet material demand, and historical smelting and pulverizing plans. Subsequently, the historical rare earth raw material information and historical rare earth permanent magnet material demand are vectorized, and the current rare earth raw material information and rare earth permanent magnet material demand are vectorized. The similarity between the current information and the historical information is then calculated using a selected similarity calculation formula (such as Euclidean distance or cosine similarity) to determine the degree of match between the current and historical information. Afterwards, based on the calculated matching degree, a set of historical information with the highest matching degree is obtained from the historical rare earth raw material information and historical rare earth permanent magnet material requirements, and the historical plan corresponding to this set of information is obtained from the historical smelting and powdering plan as the smelting and powdering plan. This smelting and powdering plan includes smelting method, smelting temperature, smelting number, smelting atmosphere, smelting time, etc. to ensure that the powder quality is controllable. In addition, the same method will be used to perform orientation molding node control analysis based on the rare earth raw material information and rare earth permanent magnet material requirements, and sintering heat treatment node control analysis based on the rare earth raw material information and rare earth permanent magnet material requirements, so as to obtain the orientation molding plan (including molding and pressing method, applied magnetic field direction and intensity, molding pressure, pressing time, etc.) and sintering heat treatment plan (sintering heating rate, sintering temperature, holding time, atmosphere, cooling rate, etc.) that are suitable for the current preparation from the historical log. Finally, the smelting and powdering plan, orientation molding plan, and sintering heat treatment plan formed by the above three links are modularly integrated and packaged into a complete rare earth permanent magnet material preparation plan. This solution serves as the main control basis for the manufacturing process and can be used to guide subsequent magnetic property prediction, process adjustment and traceability analysis, thereby achieving consistent magnetic property control driven by data throughout the entire process.

[0019] Furthermore, the present application provides a method for performing smelting and pulverizing node control analysis based on the rare earth raw material information and the rare earth permanent magnet material requirements to obtain the smelting and pulverizing plan, including: Activate the smelting and pulverizing record block, which includes multiple smelting and pulverizing record groups, each smelting and pulverizing record group includes a rare earth raw material information sample, a rare earth permanent magnet material demand sample and a smelting and pulverizing plan sample; use the rare earth raw material information and the rare earth permanent magnet material demand as smelting and pulverizing control constraints, perform a matching maximization search on the smelting and pulverizing record block, and generate the smelting and pulverizing plan.

[0020] Optionally, before performing the smelting and pulverizing node control analysis, the smelting and pulverizing record block corresponding to the smelting and pulverizing node is activated. This smelting and pulverizing record block is a pre-built historical log database that stores a large amount of smelting and pulverizing process data verified by actual production. Each record block consists of multiple independent record groups, each of which contains detailed rare earth raw material information samples, rare earth permanent magnet material demand samples, and smelting and pulverizing plan samples. Subsequently, using the current actual rare earth raw material information and rare earth permanent magnet material demand as key control constraints, non-numeric data is converted using one-hot encoding. For example, the raw material types Nd, Pr, Dy, Tb, etc. each correspond to an independent dimension. The numeric data is then normalized using min-max normalization, thereby limiting the numerical range of each characteristic indicator to between 0 and 1, avoiding calculation errors caused by large differences in the dimensions and orders of magnitude of different characteristic values. Afterwards, these two types of information features are sequentially concatenated into a complete feature vector, which serves as the feature vector to be produced. The same operation is then performed on each record group to obtain the historical feature vector for each record group. By using cosine similarity to calculate the similarity between the feature vector to be produced and each historical feature vector, the matching degree between the feature vector to be produced and each historical feature vector is obtained. The smelting and pulverizing plan samples are then sorted in descending order based on these matching degrees, and the corresponding process parameters of the first smelting and pulverizing plan sample are extracted as a feasible smelting and pulverizing plan. This smelting and pulverizing plan ensures that the target magnetic properties are efficiently achieved during process implementation, reduces trial and error costs, and significantly improves the consistency and stability of rare earth permanent magnet material manufacturing.

[0021] According to the rare earth permanent magnet material preparation plan, a multi-scale magnetic property consistency variation prediction is performed to generate a magnetic property variation prediction result.

[0022] In one embodiment, according to the above-mentioned rare earth permanent magnet material preparation scheme, a group of simulated samples are obtained by computer simulation or experimental small batch preparation. These samples are prepared under the parameters of smelting powder making, orientation molding and sintering heat treatment that fully follow the preparation scheme. Subsequently, the samples obtained by the above-mentioned simulated preparation are divided into multiple scales according to a preset scale, and the magnetic properties consistency of the divided multiple scale simulated materials is compared. For example, the remanence and coercive force are determined by hysteresis loop testing, and the microscale samples are statistically analyzed by image analysis software to calculate indicators such as grain orientation and grain size variance, and then the deviation is calculated with the target performance requirements to obtain specific quantitative magnetic properties comparison results. Afterwards, according to the above-mentioned comparison results, the abnormal features of the performance difference areas are identified, and the identified features are finally integrated to form a complete magnetic properties variation prediction result, which is used to show the inconsistency and variation trend of the magnetic properties of each scale, and guide subsequent tracing analysis and process adjustment optimization.

[0023] Furthermore, the present application provides a method for predicting the consistency variation of multi-scale magnetic properties according to the rare earth permanent magnet material preparation scheme, and generates a magnetic property variation prediction result, including: According to the rare earth permanent magnet material preparation plan, simulated preparation is performed to obtain simulated prepared materials; the simulated prepared materials are divided into multiple scales to obtain multiple scale simulated materials; the magnetic properties consistency of the multiple scale simulated materials are compared according to the target magnetic properties to obtain magnetic property comparison results of each scale; and mutation characteristics are identified according to the magnetic property comparison results of each scale to generate the magnetic property mutation prediction result.

[0024] Preferably, first, based on the aforementioned rare earth permanent magnet material preparation scheme, simulated materials similar to actual production conditions are prepared using process simulation software or a small-scale pilot production process. For example, in 3D simulation software, a vacuum induction furnace model can be used for simulated melting. Melting is performed at a vacuum of 1×10^-3 Pa and a melting temperature of 1350°C for 30 minutes. The powder is then subjected to hydrogen crushing and airflow milling to obtain a powder. This powder is then subjected to orientation shaping in a 2.5T magnetic field, and finally vacuum sintered at 1080°C for 4 hours to obtain a sample. Subsequently, the simulated materials obtained from the simulated preparation are divided according to different preset scales to obtain multiple scale simulated materials. The preset scales can include macroscale (e.g., 10mm×10mm×10mm) and microscale materials (e.g., 10µm×10µm), and can be set according to actual needs. The magnetic properties of the simulated materials at the multiple scales are then compared for consistency, using target magnetic performance indicators (e.g., energy product, coercivity, remanence, orientation angle, grain size, etc.) as a reference benchmark. Specifically, for the macroscale, the hysteresis loop of the macroscale simulated material can be measured and recorded by magnetic performance testing equipment to obtain key macroscale magnetic performance parameters such as remanence, coercive force, and magnetic energy product. The absolute deviations between the actual measured magnetic energy product, coercive force, and remanence and the target values are calculated, and the calculated deviations are added as scale magnetic performance comparison results to the magnetic performance comparison results of each scale. For the microscale, scanning electron microscopy, electron backscatter diffraction, etc. are used to collect data on the sample, and microscopic features such as grain size and grain orientation distribution are obtained through threshold segmentation methods, edge detection, pole figure analysis, and main axis direction extraction. These microscopic features are then compared with the target values to calculate the grain orientation angle deviation, grain size variance, etc., and the calculated deviations are added as scale magnetic performance comparison results to the magnetic performance comparison results of each scale. Then, based on the magnetic property comparison results at each scale, mutation signatures are identified. During this process, the deviations of the features at each scale in the magnetic property comparison results are compared with the target deviations. All scales that do not meet the target deviations and their corresponding features, such as magnetic energy product and orientation angle, are extracted. These features are used as predictions of magnetic property mutations at that scale. Finally, these predicted magnetic property mutation signatures are correlated with the locations at the corresponding scale where mutations are likely to occur (locations with large deviations), generating a magnetic property mutation prediction result. This result, which includes clear mutation signatures and scale locations, can effectively guide subsequent root cause analysis and optimization of the manufacturing process, thereby ensuring the consistency and stability of magnetic properties during material preparation.

[0025] According to the smelting and pulverizing scheme, the magnetic property variation prediction result is traced back to establish a smelting and pulverizing traceability vector.

[0026] In one embodiment, to systematically analyze the causes of abnormal magnetic property consistency in rare earth permanent magnet materials, the coupling relationships of multiple features in the magnetic property variation prediction results were first evaluated to identify which features may have causal or synergistic effects. Subsequently, a decoupling optimization method was applied to structurally decompose the coupled features, clarifying the independence of each variation manifestation and extracting multiple representative magnetic property variation features. Next, a causal path analysis was performed on each individual variation feature, combined with specific smelting and milling parameters (such as smelting temperature, pulverization particle size, and atmosphere control), to identify the core process deviation or material state anomaly causing the feature and generate preliminary single-point causal analysis results. Finally, by aggregating multiple causal analysis results, key influencing factors and their changing trends were extracted, and a complete smelting and milling causal vector was constructed. This vector not only reveals the potential process causes of magnetic property anomalies but also provides data support and adjustment direction for subsequent optimization of the preparation plan, with excellent engineering guidance value and feasibility.

[0027] Furthermore, the present application provides a method for conducting traceable mining on the magnetic property variation prediction results according to the smelting and pulverizing scheme to establish a smelting and pulverizing traceable vector, including: Based on the magnetic property variation prediction results, pairwise coupling evaluation is performed to obtain a variation coupling evaluation set; based on the variation coupling evaluation set, the magnetic property variation prediction results are decoupled and optimized to obtain the first feature of the magnetic property variation, the second feature of the magnetic property variation...the Nth feature of the magnetic property variation, where N is a positive integer; based on the smelting and powdering scheme, the first feature of the magnetic property variation, the second feature of the magnetic property variation...the Nth feature of the magnetic property variation are respectively subjected to causal analysis to obtain the first variation smelting and powdering tracing result, the second variation smelting and powdering tracing result...the Nth variation smelting and powdering tracing result; based on the first variation smelting and powdering tracing result, the second variation smelting and powdering tracing result...the Nth variation smelting and powdering tracing result, the smelting and powdering tracing vector is constructed.

[0028] Preferably, first, a pairwise coupling evaluation is performed on multiple features (such as remanence, coercive force, grain size, etc.) in the magnetic property mutation prediction results. This process uses statistical analysis or modeling methods (such as Pearson correlation coefficient, mutual information or principal component analysis) to determine whether there is a coordinated change trend or causal coupling relationship between each mutation feature. For example, the Pearson correlation coefficient can be used to perform coupling evaluation on each randomly combined feature pair, and the Pearson correlation coefficient of each feature pair can be calculated to measure the correlation strength between the two features. Then, feature pairs with an absolute value of the correlation coefficient greater than or equal to a preset threshold are screened out to form a mutation coupling evaluation set. After obtaining the anomaly coupling evaluation set, methods such as principal component analysis (PCA) and independent component analysis (ICA) are used to decompose the highly coupled features into several independent, dominant anomaly features. Taking PCA as an example, each magnetic property anomaly feature in the coupled feature subset is normalized to a mean of 0 and a variance of 1 to eliminate dimensionality. The covariance matrix of the normalized features is then calculated to quantify the linear correlations between the features. The eigenvalues and corresponding eigenvectors of the covariance matrix are then calculated and sorted by eigenvalue. The top K principal components with cumulative variance contributions greater than a preset contribution are selected and linearly combined to represent these K principal components as new decoupled feature variables. Each variable represents a relatively independent dominant magnetic property anomaly feature, namely the first, second, and Nth magnetic property anomaly features (N is a positive integer). These features represent the primary magnetic property mismatch behavior along different dimensions and have clear engineering significance and physical interpretation. For example, the first feature may represent performance degradation driven by grain size deviation, while the second feature may represent elevated oxygen content caused by atmospheric contamination. Subsequently, for each extracted magnetic property variation feature, combined with the specific smelting and powdering scheme, a pre-constructed smelting and powdering variation traceability tree is used to analyze the possible causes of the variation, analyzing which process parameter deviations or improper control may have caused the variation. For example, the causal relationship between the second feature (high oxygen content) and the failure to promptly seal the powder with inert gas after hydrogen crushing was revealed. Finally, the causal results corresponding to each variation feature are aggregated and sequentially spliced to construct a complete smelting and powdering causal vector. This vector typically includes the corresponding variation feature number, the key process parameters causing the variation, and the deviation between the current parameter value and the recommended value. The establishment of this vector establishes a clear and quantifiable causal chain between the magnetic property anomaly and the specific smelting and powdering process parameters, providing strong technical support for the subsequent optimization and adjustment of the preparation scheme.

[0029] Furthermore, the present application provides a method for performing causal analysis on the first characteristic of magnetic property variation, the second characteristic of magnetic property variation, ... the Nth characteristic of magnetic property variation, respectively, according to the smelting and powder making scheme, including: Abnormal features are extracted based on the magnetic property abnormality event set to obtain multiple magnetic property abnormality feature samples; smelting and pulverizing related factors are identified for each magnetic property abnormality feature sample based on the magnetic property abnormality event set to obtain each abnormal smelting and pulverizing factor set; accident tree learning is performed based on the multiple magnetic property abnormality feature samples and the each abnormal smelting and pulverizing factor set to construct an abnormal smelting and pulverizing causal tree; the first magnetic property abnormality feature and the smelting and pulverizing plan are input into the abnormal smelting and pulverizing causal tree to generate the first abnormal smelting and pulverizing causal result.

[0030] Optionally, first, extract abnormality feature samples based on the magnetic property abnormality event set. The magnetic property abnormality event set is a collection of records containing magnetic property deviations that occurred in historical production batches. Each record is marked with the specific manifestation of the abnormality (such as decreased coercivity, residual magnetization fluctuations, insufficient magnetic energy product, etc.). By extracting these records, multiple magnetic property abnormality feature samples can be obtained. Each sample corresponds to a specific performance abnormality type and its manifestation. For example, one sample may indicate that coarse grains lead to decreased remanence, while another sample may indicate that high oxygen content leads to decreased coercivity. Subsequently, for each abnormal characteristic sample, the possible associated smelting and pulverizing process factors are identified. This process compares and analyzes the differences between these abnormal batches and normal batches in process parameters such as raw material properties, smelting temperature, pulverization atmosphere, and particle size distribution, and screens out the parameters most relevant to this characteristic. This operation can be performed using a similar correlation coefficient analysis as described above, thereby obtaining a set of abnormal smelting and pulverizing factor items corresponding to each sample. For example, a large particle size may correspond to excessively high smelting temperature or slow cooling rate, while an increased oxygen content may correspond to exposure to air for more than 10 minutes after pulverization. Subsequently, based on these multiple characteristic samples and their corresponding factor sets, a systematic smelting and pulverizing process causal tree is constructed using fault tree learning. Specifically, the magnetic property abnormal characteristic is used as the top event, and the corresponding multiple associated process parameters are used as intermediate events. Statistical methods (such as frequency analysis and conditional probability estimation) are used to identify the logical relationships (AND / OR) between the parameters. For example, under the top event of reduced remanence, there may be a joint path of large powder particle size AND high oxygen content, or a single cause path of large smelting temperature fluctuation. Next, an incident tree structure is constructed based on these logical relationships, unfolding layer by layer from top to bottom to build a node system reflecting the causal paths. The incident tree is then trained and pruned using existing historical data, removing low-correlation paths while retaining key factor branches that strongly explain the anomaly. This ultimately forms a traceable and discriminative causal tree for the anomaly in melting and pulverizing. Finally, the first characteristic of the magnetic property anomaly under analysis (such as the decrease in remanence caused by grain coarsening) is input into the constructed causal tree, along with the parameters of the actual melting and pulverizing solution. The causal tree then infers and matches the causal path rules established within the tree, ultimately outputting a set of causal traceability results for the first anomaly in melting and pulverizing for that characteristic. These results clearly indicate which specific process parameters deviated, the direction of the deviation, and whether they represent high-risk factors. These causal traceability results not only provide a clear explanation for the current magnetic property anomaly but also provide a precise basis for subsequent parameter optimization and process path correction.

[0031] According to the orientation molding scheme, the magnetic property variation prediction result is traced back to establish an orientation molding traceability vector.

[0032] In one embodiment, after obtaining the smelting and pulverizing traceability vector, a similar process is used to conduct causal mining for orientation molding. This causal mining differs in that it focuses on the impact of the orientation molding process stages on magnetic property variations. Therefore, using the same approach, pairwise coupling analysis and decoupling optimization are first performed on the variation features to extract multiple independent representative magnetic property variation features. Then, for each variation feature, causal path analysis is performed, combined with existing orientation molding scheme parameters (such as applied magnetic field direction and intensity, pressing method, pressure distribution, molding speed, and demolding stress), to identify which molding steps may cause orientation anomalies, structural stress accumulation, or density unevenness, and obtain the corresponding orientation molding causal results. Finally, the orientation molding causal results for all variation features are integrated to construct an orientation molding causal vector. This vector structure is consistent with the smelting and pulverizing traceability vector and includes the variation feature number and the key process parameters that cause the variation, providing targeted data support for subsequent process optimization.

[0033] Furthermore, the present application provides a method for conducting traceable mining on the magnetic property variation prediction results according to the orientation molding scheme to establish an orientation molding traceable vector, including: According to the magnetic property mutation event set, orientation molding related factors are identified for each magnetic property mutation feature sample to obtain each mutation orientation molding factor set; accident tree learning is performed based on the multiple magnetic property mutation feature samples and the each mutation orientation molding factor set to construct a mutation orientation molding tracing tree; the first feature of the magnetic property mutation and the smelting and powder making plan are input into the mutation orientation molding tracing tree to generate a first mutation orientation molding tracing result; based on the orientation molding plan, the second feature of the magnetic property mutation...the Nth feature of the magnetic property mutation are further subjected to tracing analysis according to the mutation orientation molding tracing tree to obtain a second mutation orientation molding tracing result...the Nth mutation orientation molding tracing result; the orientation molding tracing vector is generated based on the first mutation orientation molding tracing result, the second mutation orientation molding tracing result...the Nth mutation orientation molding tracing result.

[0034] Optionally, based on multiple magnetic property variation feature samples (e.g., decreased coercivity, reduced magnetic anisotropy, and uneven compact density) extracted from the magnetic property variation event set, factors associated with orientation molding are identified for each sample. This identification process is similar to the identification of factors associated with smelting and milling, focusing on analyzing the correlation between parameters such as the orientation magnetic field direction and intensity, the magnitude and uniformity of the molding pressure, the compacting die design, and interference from the demolding process, and magnetic property deviations. This identifies the corresponding orientation molding factor set for each feature. Subsequently, based on these feature samples and their associated process factors, an incident tree learning method is used to construct a cause-effect tree for the abnormal orientation molding. The construction process is consistent with the smelting and milling cause-effect tree described above, using logical structures (e.g., AND / OR gates) to express the causal path between magnetic property anomalies and molding parameters. For example, decreased coercivity may be caused by both deviation in the orientation magnetic field and uneven compacting. The first feature of the magnetic property variation and the parameters of the currently executed orientation molding solution are then fed into the cause-effect tree for the abnormal orientation molding, yielding the first cause-effect result for the abnormal orientation molding, indicating the key risk parameters corresponding to the first feature and their deviations. Similarly, based on the same traceability tree structure, the remaining features (the second to the Nth features) are analyzed one by one to obtain the corresponding traceability results for each feature (the second to the Nth abnormal orientation molding traceability results). Finally, the orientation molding traceability results corresponding to all features are integrated and classified to form a complete orientation molding traceability vector, providing data support for subsequent process adjustments and consistency optimization.

[0035] According to the sintering heat treatment scheme, the magnetic property variation prediction result is traced back to establish a sintering heat treatment traceability vector.

[0036] In one embodiment, similar to the above, multiple magnetic property variation characteristics are first identified as key factors that may be related to the sintering or heat treatment process, such as whether the sintering temperature setting is too high or too low, whether the heating rate is uniform, and whether the holding time is sufficient. Subsequently, using a fault tree modeling method similar to the aforementioned smelting and pulverizing and orientation forming, the multiple magnetic property variation characteristics and the corresponding sintering heat treatment factors are constructed into a variation sintering heat treatment causal tree. Causal logic gates (such as AND and OR) are used to establish the association between magnetic property mismatch and heat treatment process. Each variation characteristic, along with the currently implemented sintering heat treatment plan, is then input into the variation sintering heat treatment causal tree to generate the sintering heat treatment causal results corresponding to that characteristic. For example, if a batch of materials exhibits an abnormal decrease in coercivity after holding at 1080°C for 3 hours, this may indicate insufficient holding time or an excessively fast heating rate. Finally, the sintering heat treatment traceability results of all abnormal characteristics are integrated to form a sintering heat treatment traceability vector. This vector will serve as the input for subsequent process optimization and dynamic adjustment, helping to improve the heat treatment process's ability to ensure magnetic property consistency.

[0037] The rare earth permanent magnet material preparation scheme is optimized and adjusted according to the smelting and powdering traceability vector, the orientation forming traceability vector and the sintering heat treatment traceability vector to obtain a preparation scheme optimization result.

[0038] In one embodiment, to achieve systematic optimization of rare earth permanent magnet material preparation schemes, the previously acquired smelting and pulverizing traceability vectors, orientation-forming traceability vectors, and sintering heat treatment traceability vectors are used as the basis for parameter adjustment, allowing targeted adjustments to be made to each process node scheme. For example, if a wide powder particle size is identified as a key influencing factor in the smelting and pulverizing traceability vector, a set of adjustment combinations of pulverization methods and screening strategies can be formed to form a smelting and pulverizing adjustment group. Similarly, based on the traceability vectors for the orientation-forming and sintering heat treatment stages, an orientation-forming adjustment group and a sintering heat treatment adjustment group are generated, each containing multiple parameter setting schemes and optimization paths. Subsequently, by coupling these three types of adjustment groups, a complete permanent magnet material preparation adjustment group spanning multiple process nodes is constructed. This adjustment group serves as the initial input for iterative optimization. In subsequent steps, it combines consistency variation prediction and consistency loss evaluation to continuously optimize the combined parameters, ultimately achieving a preparation scheme with better performance and more reasonable process paths. The results not only include the adjusted process parameter set, but also its predicted performance in multi-scale magnetic property simulations, which serves as quantitative support for production decisions and ensures that the consistency and repeatability of the final prepared material in various magnetic property indicators are significantly improved.

[0039] Furthermore, the present application provides a method for optimizing and adjusting the rare earth permanent magnet material preparation scheme according to the smelting and pulverizing traceability vector, the orientation forming traceability vector, and the sintering heat treatment traceability vector to obtain a preparation scheme optimization result, including: The smelting and pulverizing scheme is adjusted according to the smelting and pulverizing traceability vector to obtain a smelting and pulverizing adjustment group; the orientation forming scheme is adjusted according to the orientation forming traceability vector to obtain an orientation forming adjustment group; the sintering heat treatment scheme is adjusted according to the sintering heat treatment traceability vector to obtain a sintering heat treatment adjustment group; a multi-node control combination is performed based on the smelting and pulverizing adjustment group, the orientation forming adjustment group and the sintering heat treatment adjustment group to generate a permanent magnet material preparation adjustment group; iterative optimization is performed based on the permanent magnet material preparation adjustment group to generate the preparation scheme optimization result.

[0040] Preferably, first, based on the key influencing parameters identified in the smelting and pulverizing traceability vector (such as melting temperature fluctuations, uneven crushing particle size, and atmosphere control instability), multiple rounds of parameter adjustments are performed on the relevant process settings in the original smelting and pulverizing scheme. For example, different smelting temperature gradients, crushing equipment selection, and screening particle size control methods are set. This generates a number of practical and differentiated smelting and pulverizing adjustment schemes, which constitute the smelting and pulverizing adjustment group. Subsequently, using the same approach, the orientation forming and sintering heat treatment adjustment vectors are used to generate the orientation forming and sintering heat treatment adjustment groups, respectively. For example, the magnetic field strength and direction settings are adjusted during the forming stage, or different combinations of holding time and cooling rate are explored during the sintering stage. Subsequently, these three types of adjustment groups are combined and matched according to the order of the nodes to construct a permanent magnet material preparation adjustment group that includes all process node adjustment combinations. That is, multiple complete, cross-process stage preparation parameter sets. Each combination scheme includes a smelting and powder making adjustment scheme, an orientation forming adjustment scheme, and a sintering heat treatment adjustment scheme, forming multiple candidate permanent magnet material preparation adjustment schemes. Afterwards, based on the preset magnetic performance consistency evaluation index, the above candidate schemes are iteratively optimized. Through simulation prediction, the schemes with better performance are continuously screened, and finally a set of optimal preparation scheme optimization results are generated to guide subsequent actual production process adjustments and achieve improved magnetic performance consistency.

[0041] Furthermore, the present application provides an iterative optimization method based on the permanent magnet material preparation adjustment group to generate the preparation scheme optimization result, including: A multi-scale magnetic property consistency variation prediction is performed on each permanent magnetic material preparation adjustment scheme in the permanent magnetic material preparation adjustment group to obtain the predicted magnetic property variation of each scheme; a magnetic property consistency loss evaluation is performed on the predicted magnetic property variation of each scheme to obtain the loss evaluation coefficient of each scheme; and a magnetic property consistency loss minimization optimization is performed on the permanent magnetic material preparation adjustment group based on the loss evaluation coefficient of each scheme to obtain the optimization result of the preparation scheme.

[0042] Optionally, for each candidate solution in the permanent magnet material preparation adjustment group (i.e., different combinations of smelting and powder making, orientation forming, and sintering heat treatment parameter sets), a complete multi-scale magnetic property consistency variation prediction is performed. The prediction process follows the aforementioned simulation preparation and multi-scale analysis methods to evaluate the magnetic performance of each solution at the macroscopic (such as overall magnetic energy product, coercive force), microscopic (such as grain orientation, orientation angle distribution), and submicroscopic scales (such as magnetic domain structure, grain boundary defects), and identify potential performance fluctuations or abnormal risks. Subsequently, the magnetic property variation characteristics of each predicted solution are quantified. Based on the set target performance range and fluctuation tolerance, the performance deviation of each solution at each scale is calculated using the normalized absolute deviation formula. These performance deviations are then summed according to the preset weights to obtain the loss evaluation coefficient of each solution. The smaller the coefficient, the closer the performance of the solution at multiple scales is to the target and the higher the stability. Finally, consistency optimization screening is performed based on the loss evaluation coefficients of all candidate solutions, and the solution with the smallest loss is selected by sorting as the final preparation solution optimization result. This result ensures that the selected preparation path has highly repeatable and consistent magnetic performance output in actual implementation, thereby meeting the demand for quality stability of high-end rare earth permanent magnet material products.

[0043] In summary, the embodiments of the present application have at least the following technical effects: The embodiment of the present application first performs multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, which includes a smelting and pulverizing plan, an orientation molding plan, and a sintering heat treatment plan; then, a multi-scale magnetic property consistency variation prediction is performed based on the rare earth permanent magnet material preparation plan to generate a magnetic property variation prediction result; thereafter, the magnetic property variation prediction result is traced back to the smelting and pulverizing plan to establish a smelting and pulverizing traceability vector; then, the magnetic property variation prediction result is traced back to the orientation molding plan to establish an orientation molding traceability vector; further, the magnetic property variation prediction result is traced back to the sintering heat treatment plan to establish a sintering heat treatment traceability vector; finally, the rare earth permanent magnet material preparation plan is optimized and adjusted based on the smelting and pulverizing traceability vector, the orientation molding traceability vector, and the sintering heat treatment traceability vector to obtain a preparation plan optimization result. These technical effects jointly solve the technical problem of insufficient coordinated control of multiple process links in the existing preparation process of rare earth permanent magnet materials, which leads to fluctuations in magnetic property consistency. They achieve the technical effect of accurately identifying magnetic property variations and dynamically optimizing and adjusting preparation plans through multi-scale prediction and process stage tracing analysis, thereby improving the magnetic property consistency of rare earth permanent magnet materials.

[0044] Embodiment 2 is based on the same inventive concept as the manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials in the above embodiment. Figure 2 As shown, the present application provides a manufacturing process control system for the consistency of magnetic properties of rare earth permanent magnet materials, the system comprising: a control analysis module 11: performing multi-node preparation control analysis according to rare earth raw material information and rare earth permanent magnet material requirements, obtaining a rare earth permanent magnet material preparation scheme, the rare earth permanent magnet material preparation scheme including a smelting and powdering scheme, an orientation molding scheme and a sintering heat treatment scheme; a variation prediction module 12: performing multi-scale magnetic property consistency variation prediction according to the rare earth permanent magnet material preparation scheme, generating a magnetic property variation prediction result; a powdering tracing mining module 13: performing a multi-scale magnetic property consistency variation prediction according to the smelting and powdering scheme, generating a magnetic property variation prediction result; The mutation prediction results are traced back to establish a smelting and pulverizing traceability vector; the forming traceability mining module 14: according to the orientation forming scheme, the magnetic property mutation prediction results are traced back to establish an orientation forming traceability vector; the sintering traceability mining module 15: according to the sintering heat treatment scheme, the magnetic property mutation prediction results are traced back to establish a sintering heat treatment traceability vector; the optimization adjustment module 16: according to the smelting and pulverizing traceability vector, the orientation forming traceability vector and the sintering heat treatment traceability vector, the rare earth permanent magnet material preparation scheme is optimized and adjusted to obtain the preparation scheme optimization result.

[0045] Furthermore, the control analysis module 11 is further configured to execute the following method: According to the rare earth raw material information and the rare earth permanent magnet material requirements, the smelting and pulverizing node control analysis is performed to obtain the smelting and pulverizing plan; according to the rare earth raw material information and the rare earth permanent magnet material requirements, the orientation molding node control analysis is performed to obtain the orientation molding plan; according to the rare earth raw material information and the rare earth permanent magnet material requirements, the sintering heat treatment node control analysis is performed to obtain the sintering heat treatment plan; the smelting and pulverizing plan, the orientation molding plan and the sintering heat treatment plan are packaged into the rare earth permanent magnet material preparation plan.

[0046] Furthermore, the control analysis module 11 is further configured to execute the following method: Activate the smelting and pulverizing record block, which includes multiple smelting and pulverizing record groups, each smelting and pulverizing record group includes a rare earth raw material information sample, a rare earth permanent magnet material demand sample and a smelting and pulverizing plan sample; use the rare earth raw material information and the rare earth permanent magnet material demand as smelting and pulverizing control constraints, perform a matching maximization search on the smelting and pulverizing record block, and generate the smelting and pulverizing plan.

[0047] Furthermore, the mutation prediction module 12 is further configured to execute the following method: According to the rare earth permanent magnet material preparation plan, simulated preparation is performed to obtain simulated prepared materials; the simulated prepared materials are divided into multiple scales to obtain multiple scale simulated materials; the magnetic properties consistency of the multiple scale simulated materials are compared according to the target magnetic properties to obtain magnetic property comparison results of each scale; and mutation characteristics are identified according to the magnetic property comparison results of each scale to generate the magnetic property mutation prediction result.

[0048] Furthermore, the milling traceability mining module 13 is also used to perform the following method: Based on the magnetic property variation prediction results, pairwise coupling evaluation is performed to obtain a variation coupling evaluation set; based on the variation coupling evaluation set, the magnetic property variation prediction results are decoupled and optimized to obtain the first feature of the magnetic property variation, the second feature of the magnetic property variation...the Nth feature of the magnetic property variation, where N is a positive integer; based on the smelting and powdering scheme, the first feature of the magnetic property variation, the second feature of the magnetic property variation...the Nth feature of the magnetic property variation are respectively subjected to causal analysis to obtain the first variation smelting and powdering tracing result, the second variation smelting and powdering tracing result...the Nth variation smelting and powdering tracing result; based on the first variation smelting and powdering tracing result, the second variation smelting and powdering tracing result...the Nth variation smelting and powdering tracing result, the smelting and powdering tracing vector is constructed.

[0049] Furthermore, the milling traceability mining module 13 is also used to perform the following method: Abnormal features are extracted based on the magnetic property abnormality event set to obtain multiple magnetic property abnormality feature samples; smelting and pulverizing related factors are identified for each magnetic property abnormality feature sample based on the magnetic property abnormality event set to obtain each abnormal smelting and pulverizing factor set; accident tree learning is performed based on the multiple magnetic property abnormality feature samples and the each abnormal smelting and pulverizing factor set to construct an abnormal smelting and pulverizing causal tree; the first magnetic property abnormality feature and the smelting and pulverizing plan are input into the abnormal smelting and pulverizing causal tree to generate the first abnormal smelting and pulverizing causal result.

[0050] Furthermore, the shaping traceability mining module 14 is also used to perform the following method: According to the magnetic property mutation event set, orientation molding related factors are identified for each magnetic property mutation feature sample to obtain each mutation orientation molding factor set; accident tree learning is performed based on the multiple magnetic property mutation feature samples and the each mutation orientation molding factor set to construct a mutation orientation molding tracing tree; the first feature of the magnetic property mutation and the smelting and powder making plan are input into the mutation orientation molding tracing tree to generate a first mutation orientation molding tracing result; based on the orientation molding plan, the second feature of the magnetic property mutation...the Nth feature of the magnetic property mutation are further subjected to tracing analysis according to the mutation orientation molding tracing tree to obtain a second mutation orientation molding tracing result...the Nth mutation orientation molding tracing result; the orientation molding tracing vector is generated based on the first mutation orientation molding tracing result, the second mutation orientation molding tracing result...the Nth mutation orientation molding tracing result.

[0051] Furthermore, the optimization and adjustment module 16 is further configured to execute the following method: The smelting and pulverizing scheme is adjusted according to the smelting and pulverizing traceability vector to obtain a smelting and pulverizing adjustment group; the orientation forming scheme is adjusted according to the orientation forming traceability vector to obtain an orientation forming adjustment group; the sintering heat treatment scheme is adjusted according to the sintering heat treatment traceability vector to obtain a sintering heat treatment adjustment group; a multi-node control combination is performed based on the smelting and pulverizing adjustment group, the orientation forming adjustment group and the sintering heat treatment adjustment group to generate a permanent magnet material preparation adjustment group; iterative optimization is performed based on the permanent magnet material preparation adjustment group to generate the preparation scheme optimization result.

[0052] Furthermore, the optimization and adjustment module 16 is further configured to execute the following method: A multi-scale magnetic property consistency variation prediction is performed on each permanent magnetic material preparation adjustment scheme in the permanent magnetic material preparation adjustment group to obtain the predicted magnetic property variation of each scheme; a magnetic property consistency loss evaluation is performed on the predicted magnetic property variation of each scheme to obtain the loss evaluation coefficient of each scheme; and a magnetic property consistency loss minimization optimization is performed on the permanent magnetic material preparation adjustment group based on the loss evaluation coefficient of each scheme to obtain the optimization result of the preparation scheme.

[0053] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

[0055] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials, characterized in that: include: Perform multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, which includes a smelting and powdering plan, an orientation molding plan, and a sintering heat treatment plan; Perform multi-scale magnetic property consistency variation prediction according to the rare earth permanent magnet material preparation plan to generate magnetic property variation prediction results; Performing traceable mining on the magnetic property variation prediction results according to the smelting and pulverizing scheme to establish a smelting and pulverizing traceable vector; Performing causal mining on the magnetic property variation prediction result according to the orientation molding scheme to establish an orientation molding causal vector; Performing traceability mining on the magnetic property variation prediction results according to the sintering heat treatment scheme to establish a sintering heat treatment traceability vector; The rare earth permanent magnet material preparation scheme is optimized and adjusted according to the smelting and powdering traceability vector, the orientation forming traceability vector and the sintering heat treatment traceability vector to obtain a preparation scheme optimization result.

2. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 1, characterized in that: According to the rare earth permanent magnet material preparation scheme, a multi-scale magnetic property consistency variation prediction is performed to generate a magnetic property variation prediction result, including: Performing simulated preparation according to the rare earth permanent magnet material preparation scheme to obtain simulated prepared material; Performing multi-scale division on the simulated prepared material to obtain multi-scale simulated materials; Performing magnetic property consistency comparison on the multiple scale simulation materials according to the target magnetic properties to obtain magnetic property comparison results at each scale; The abnormality characteristics are identified based on the magnetic property comparison results at each scale to generate the magnetic property abnormality prediction result.

3. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 1, characterized in that: According to the smelting and pulverizing scheme, the magnetic property variation prediction result is traced back to the prediction result, and a smelting and pulverizing traceability vector is established, including: Performing pairwise coupling evaluation based on the magnetic property anomaly prediction results to obtain an anomaly coupling evaluation set; Decoupling and optimizing the magnetic property mutation prediction result according to the mutation coupling evaluation set to obtain a first characteristic of magnetic property mutation, a second characteristic of magnetic property mutation, and an Nth characteristic of magnetic property mutation, where N is a positive integer; According to the smelting and pulverizing scheme, tracing and analyzing the first characteristic of the magnetic property variation, the second characteristic of the magnetic property variation, ... the Nth characteristic of the magnetic property variation, respectively, to obtain a tracing result of the first variation smelting and pulverizing, a tracing result of the second variation smelting and pulverizing, ... an Nth variation smelting and pulverizing; The smelting and pulverizing tracing vector is constructed according to the first abnormal smelting and pulverizing tracing result, the second abnormal smelting and pulverizing tracing result... the Nth abnormal smelting and pulverizing tracing result.

4. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 3, characterized in that: According to the smelting and powder making scheme, the first characteristic of the magnetic property change, the second characteristic of the magnetic property change, ... the Nth characteristic of the magnetic property change are respectively subjected to causal analysis, including: Extract the abnormality features based on the magnetic property abnormality event set to obtain multiple magnetic property abnormality feature samples; Identify smelting and milling related factors for each magnetic property abnormality feature sample according to the magnetic property abnormality event set, and obtain a set of abnormal smelting and milling factors; Conducting fault tree learning based on the plurality of magnetic property variation feature samples and the respective variation smelting and pulverizing factor sets to construct a variation smelting and pulverizing causal tree; The first characteristic of the magnetic property mutation and the smelting and pulverizing plan are input into the mutation smelting and pulverizing tracing tree to generate the first mutation smelting and pulverizing tracing result.

5. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 4, characterized in that: According to the orientation molding scheme, the magnetic property variation prediction result is traced back to the prediction result, and an orientation molding traceability vector is established, including: Identifying orientation-molding related factors for each magnetic property variation characteristic sample according to the magnetic property variation event set to obtain each variation orientation-molding factor set; Performing fault tree learning based on the plurality of magnetic property variation feature samples and the various variation orientation forming factor sets to construct a variation orientation forming causal tree; Inputting the first characteristic of magnetic property variation and the smelting and powder making scheme into the variation orientation molding tracing tree to generate a first variation orientation molding tracing result; Based on the orientation molding scheme, causal analysis is continued on the second characteristic of the magnetic property mutation...the Nth characteristic of the magnetic property mutation according to the abnormal orientation molding causal tree to obtain the second abnormal orientation molding causal result...the Nth abnormal orientation molding causal result; The orientation molding tracing vector is generated according to the first abnormal orientation molding tracing result, the second abnormal orientation molding tracing result...the Nth abnormal orientation molding tracing result.

6. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 1, characterized in that: The preparation scheme of the rare earth permanent magnet material is optimized and adjusted according to the smelting and pulverizing traceability vector, the orientation forming traceability vector, and the sintering heat treatment traceability vector to obtain the preparation scheme optimization result, including: Adjusting the smelting and pulverizing scheme according to the smelting and pulverizing traceability vector to obtain a smelting and pulverizing adjustment group; Adjusting the orientation forming scheme according to the orientation forming traceability vector to obtain an orientation forming adjustment group; Adjusting the sintering heat treatment plan according to the sintering heat treatment traceability vector to obtain a sintering heat treatment adjustment group; Perform multi-node control combination according to the smelting and pulverizing adjustment group, the orientation forming adjustment group and the sintering heat treatment adjustment group to generate a permanent magnet material preparation adjustment group; An iterative optimization is performed according to the permanent magnet material preparation adjustment group to generate the preparation scheme optimization result.

7. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 6, characterized in that: Performing iterative optimization according to the permanent magnet material preparation adjustment group to generate the preparation scheme optimization result includes: Perform multi-scale magnetic property consistency variation prediction according to each permanent magnetic material preparation adjustment scheme in the permanent magnetic material preparation adjustment group to obtain the predicted magnetic property variation of each scheme; Predicting magnetic property changes according to each scheme to evaluate magnetic property consistency loss and obtain loss evaluation coefficients for each scheme; The magnetic performance consistency loss of the permanent magnet material preparation adjustment group is minimized according to the loss evaluation coefficient of each scheme to obtain the optimization result of the preparation scheme.

8. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 1, characterized in that: Based on rare earth raw material information and rare earth permanent magnet material requirements, multi-node preparation control analysis is performed to obtain a rare earth permanent magnet material preparation plan, including: Performing smelting and pulverizing node control analysis based on the rare earth raw material information and the rare earth permanent magnet material requirements to obtain the smelting and pulverizing plan; Performing orientation molding node control analysis according to the rare earth raw material information and the rare earth permanent magnet material requirements to obtain the orientation molding solution; Performing sintering heat treatment node control analysis based on the rare earth raw material information and the rare earth permanent magnet material requirements to obtain the sintering heat treatment plan; The smelting and powder making scheme, the orientation molding scheme and the sintering heat treatment scheme are packaged into the rare earth permanent magnet material preparation scheme.

9. The manufacturing process control method for the consistency of magnetic properties of rare earth permanent magnet materials according to claim 8, characterized in that: Performing smelting and pulverizing node control analysis based on the rare earth raw material information and the rare earth permanent magnet material requirements to obtain the smelting and pulverizing plan includes: Activate the smelting and pulverizing record block, which includes multiple smelting and pulverizing record groups, each of which includes a rare earth raw material information sample, a rare earth permanent magnet material demand sample, and a smelting and pulverizing plan sample; The rare earth raw material information and the rare earth permanent magnet material requirements are used as smelting and pulverizing control constraints, and a matching degree maximization search is performed on the smelting and pulverizing record block to generate the smelting and pulverizing plan.

10. A manufacturing process control system for the consistency of magnetic properties of rare earth permanent magnet materials, characterized in that: The system is used to implement the manufacturing process control method for the magnetic property consistency of rare earth permanent magnet materials according to any one of claims 1 to 9, comprising: Control analysis module: This module performs multi-node preparation control analysis based on rare earth raw material information and rare earth permanent magnet material requirements to obtain a rare earth permanent magnet material preparation plan, which includes a smelting and powdering plan, an orientation molding plan, and a sintering heat treatment plan. Mutation prediction module: performs multi-scale magnetic property consistency mutation prediction based on the rare earth permanent magnet material preparation plan and generates magnetic property mutation prediction results; Pulverizing causal mining module: performs causal mining on the magnetic property variation prediction results according to the smelting and pulverizing scheme, and establishes a smelting and pulverizing causal vector; Molding traceability mining module: performs traceability mining on the magnetic property variation prediction results according to the orientation molding scheme, and establishes an orientation molding traceability vector; Sintering traceability mining module: performs traceability mining on the magnetic property variation prediction results according to the sintering heat treatment scheme, and establishes a sintering heat treatment traceability vector; Optimization and adjustment module: optimizes and adjusts the preparation scheme of the rare earth permanent magnet material according to the smelting and powder making traceability vector, the orientation forming traceability vector and the sintering heat treatment traceability vector to obtain the optimization result of the preparation scheme.

Images