A method and system for preparing a low eddy current loss magnetic steel for high frequency drive motor

Abstract

The application discloses a low eddy current loss magnetic steel preparation method and system for a high-frequency driving motor, and comprises the following steps: receiving a magnetic steel preparation instruction, confirming a preparation environment parameter node and a target magnetic steel performance parameter node based on the magnetic steel preparation instruction, obtaining a magnetic steel raw material proportioning data set, constructing a preparation parameter combination set based on the magnetic steel raw material proportioning data set, a sintering temperature, a magnetic field orientation intensity and a cooling rate, performing an eddy current loss detection operation on each layered magnetic steel sheet in the layered magnetic steel sheet set, obtaining a layered eddy current loss value set, confirming a screened magnetic steel sample and a sample loss characteristic node based on the layered eddy current loss value set and a target eddy current loss threshold value, and confirming a target magnetic steel sample based on the magnetic performance detection node, the sample loss characteristic node, a target magnetic energy product threshold value and a target coercive force threshold value. The application can solve the problem of excessive magnetic steel eddy current loss in the high-frequency driving motor.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of magnetic material preparation, and particularly relates to a low eddy current loss magnetic steel preparation method and system for high-frequency driving motor. BACKGROUND

[0002] In the field of new energy automobile driving system, permanent magnet synchronous motor has become the mainstream power scheme due to its high power density and high efficiency characteristics. As a core component of the motor, the performance of the permanent magnet directly determines the operating efficiency and reliability of the whole machine. With the development of the motor towards high speed and high frequency, the working frequency range has been increased from tens of hertz to hundreds or even thousands of hertz, which puts forward more stringent requirements for the magnetic steel material.

[0003] At present, the preparation of magnetic steel mainly relies on the setting of process parameters driven by experience, and technicians adjust the raw material ratio and sintering parameters according to historical formula and test data. In the face of the demand for eddy current loss control in high-frequency application scenarios, the traditional method often uses single frequency point loss test for quality evaluation, which is difficult to fully reflect the actual performance of the magnetic steel in a wide frequency working environment. At the same time, the uniformity of the internal microstructure of the magnetic steel has a significant influence on the eddy current loss distribution, and the existing detection means mainly measures the whole sample, which cannot effectively identify local defects or structural differences.

[0004] Although the above method can realize the basic preparation and performance detection of the magnetic steel, the traditional process lacks systematic data correlation analysis in the parameter optimization process, and repeated experiments are needed to explore different performance index requirements, which is difficult to form reusable process knowledge accumulation. In addition, the comprehensive performance evaluation of the magnetic steel often focuses on single indicators such as magnetic energy product and coercive force, and fails to establish a balanced evaluation mechanism between eddy current loss characteristics and magnetic properties, resulting in that the prepared magnetic steel is difficult to meet the dual requirements of high frequency and low loss and high magnetic performance, which restricts the performance improvement and application promotion of high-frequency driving motor. Therefore, how to realize the precise preparation and comprehensive performance optimization of low eddy current loss magnetic steel has become a technical problem to be solved. SUMMARY

[0005] The present application provides a low eddy current loss magnetic steel preparation method for high-frequency driving motor, which comprises:

[0006] receiving a magnetic steel preparation instruction, and confirming a preparation environment parameter node and a target magnetic steel performance parameter node based on the magnetic steel preparation instruction, wherein the preparation environment parameter node comprises a sintering temperature, a magnetic field orientation intensity and a cooling rate, and the target magnetic steel performance parameter node comprises a target eddy current loss threshold, a target magnetic energy product threshold and a target coercive force threshold;

[0007] obtain a magnetic steel raw material proportioning dataset, construct a preparation parameter combination set based on the magnetic steel raw material proportioning dataset, a sintering temperature, a magnetic field orientation intensity, and a cooling rate, perform the following operations on each preparation parameter combination in the preparation parameter combination set: perform magnetic steel sample preparation using the preparation parameter combination to obtain an initial magnetic steel sample, perform layered cutting processing on the initial magnetic steel sample to obtain a layered magnetic steel sheet set;

[0008] perform eddy current loss detection operations on each layered magnetic steel sheet in the layered magnetic steel sheet set to obtain a layered eddy current loss value set, and confirm a screened magnetic steel sample and a sample loss feature node based on the layered eddy current loss value set and a target eddy current loss threshold value;

[0009] perform grain boundary optimization processing on the screened magnetic steel sample to obtain an optimized magnetic steel sample, and perform detection on the optimized magnetic steel sample using a pre-constructed magnetic property detection unit to obtain a magnetic property detection node, wherein the magnetic property detection node includes a magnetic energy product detection value and a coercive force detection value;

[0010] confirm a target magnetic steel sample based on the magnetic property detection node, the sample loss feature node, a target magnetic energy product threshold value, and a target coercive force threshold value, and realize preparation of a low eddy current loss magnetic steel.

[0011] Optionally, the constructing of the preparation parameter combination set based on the magnetic steel raw material proportioning dataset, the sintering temperature, the magnetic field orientation intensity, and the cooling rate includes:

[0012] extract a plurality of raw material proportioning nodes from the magnetic steel raw material proportioning dataset, wherein the raw material proportioning node includes a rare earth element content, an iron element content, and an additive content;

[0013] perform the following operations on each raw material proportioning node in the plurality of raw material proportioning nodes:

[0014] perform verification on the raw material proportioning node using a pre-constructed proportioning constraint relationship to obtain a verification result, wherein the verification result is pass or fail;

[0015] after confirming that the verification result is pass, associate the raw material proportioning node, the sintering temperature, the magnetic field orientation intensity, and the cooling rate to obtain an initial preparation parameter combination;

[0016] aggregate the initial preparation parameter combinations to obtain an initial preparation parameter combination set, and screen the initial preparation parameter combination set using a pre-set orthogonal experimental design method to obtain the preparation parameter combination set.

[0017] Optionally, the layered cutting processing on the initial magnetic steel sample to obtain the layered magnetic steel sheet set includes:

[0018] Obtaining a sample thickness parameter of the initial magnetic steel sample, calculating a layer number based on a preset layer thickness threshold and the sample thickness parameter;

[0019] Uniformly cutting the initial magnetic steel sample by using the layer number to obtain a plurality of cut magnetic steel pieces;

[0020] The following operations are performed on each of the plurality of cut magnetic steel pieces:

[0021] Detecting the cut magnetic steel piece by using a pre-constructed surface flatness detection unit to obtain a surface flatness value, comparing the surface flatness value with a preset flatness threshold, and if the surface flatness value is greater than or equal to the flatness threshold, performing grinding processing on the cut magnetic steel piece to obtain a ground magnetic steel piece, and taking the ground magnetic steel piece as a layered magnetic steel piece;

[0022] Otherwise, taking the cut magnetic steel piece as a layered magnetic steel piece, and collecting the layered magnetic steel pieces to obtain a layered magnetic steel piece set.

[0023] Optionally, the eddy current loss detection operation is performed on each of the layered magnetic steel pieces in the layered magnetic steel piece set to obtain a layered eddy current loss value set, including:

[0024] Obtaining a detection frequency sequence for eddy current loss detection, wherein the detection frequency sequence includes a plurality of detection frequencies, and the detection frequency sequence covers a working frequency range of a high-frequency driving motor;

[0025] The following operations are performed on each of the layered magnetic steel pieces in the layered magnetic steel piece set:

[0026] The following operations are performed on each of the detection frequencies in the detection frequency sequence:

[0027] Detecting the layered magnetic steel piece by using the detection frequency and a pre-constructed eddy current loss detection device to obtain a frequency eddy current loss value, associating the frequency eddy current loss value with the detection frequency to obtain a frequency loss node;

[0028] Collecting the frequency loss nodes to obtain a frequency loss node set, and calculating a weighted eddy current loss value based on the frequency loss node set, wherein the weighted eddy current loss value is a weighted average of a plurality of frequency eddy current loss values in the frequency loss node set based on the detection frequency;

[0029] Collecting the weighted eddy current loss values to obtain the layered eddy current loss value set.

[0030] Optionally, the detecting the layered magnetic steel piece by using the detection frequency and the pre-constructed eddy current loss detection device to obtain the frequency eddy current loss value includes:

[0031] Obtaining resistivity parameters and slice thickness parameters of the layered magnetic steel sheet, applying an alternating magnetic field to the layered magnetic steel sheet by using an eddy current loss detection device to obtain a magnetic induction peak value;

[0032] Calculating a frequency eddy current loss value based on the resistivity parameters, the slice thickness parameters, a detection frequency and the magnetic induction peak value, and a calculation formula is as follows:

[0033] ,

[0034] Wherein, represents the frequency eddy current loss value, represents the slice thickness parameters, represents the detection frequency, represents the magnetic induction peak value, represents the resistivity parameters, is a preset circular constant, is a preset magnetic permeability constant, represents a vortex identifier, represents a peak identifier, represents a hyperbolic tangent function.

[0035] Optionally, the confirming the screened magnetic steel sample and the sample loss feature node based on the set of layered eddy current loss values and a target eddy current loss threshold value comprises:

[0036] Calculating a mean value of the set of layered eddy current loss values to obtain a sample eddy current loss mean value, and calculating a variance of the set of layered eddy current loss values to obtain a sample eddy current loss variance;

[0037] Comparing the sample eddy current loss mean value with the target eddy current loss threshold value, and if the sample eddy current loss mean value is less than or equal to the target eddy current loss threshold value, comparing the sample eddy current loss variance with a preset variance threshold value;

[0038] If the sample eddy current loss variance is less than or equal to the variance threshold value, the initial magnetic steel sample is confirmed as the screened magnetic steel sample, and the sample eddy current loss mean value and the sample eddy current loss variance are associated to obtain the sample loss feature node;

[0039] Otherwise, an abnormal eddy current loss value is identified in the set of layered eddy current loss values, wherein an absolute difference value between the abnormal eddy current loss value and the sample eddy current loss mean value is greater than a preset deviation threshold value, the abnormal eddy current loss value is removed from the set of layered eddy current loss values to obtain an updated set of eddy current loss values, the set of layered eddy current loss values is taken as the updated set of eddy current loss values, and the step of calculating the mean value of the set of layered eddy current loss values is returned until the screened magnetic steel sample and the sample loss feature node are obtained.

[0040] Optionally, the step of performing grain boundary optimization processing on the screened magnet samples to obtain optimized magnet samples includes:

[0041] The microstructure images of the grains of the screened magnetic steel samples are obtained. The microstructure images of the grains are analyzed using a pre-built image analysis model to obtain the grain boundary feature nodes. The grain boundary feature nodes include: average grain size, boundary phase thickness and boundary phase distribution uniformity.

[0042] Based on the grain boundary feature nodes and the preset boundary optimization parameter range, the boundary optimization processing parameter nodes are identified, including: diffusion annealing temperature, diffusion annealing time, and diffusion element concentration.

[0043] The grain boundary diffusion process is performed on the screened magnet samples using the boundary optimization processing parameter nodes to obtain diffusion-treated magnet samples. The diffusion-treated magnet samples are then subjected to aging treatment to obtain optimized magnet samples.

[0044] Optionally, the step of identifying boundary optimization processing parameter nodes based on the grain boundary feature nodes and the preset boundary optimization parameter range includes:

[0045] Obtain the historical optimization dataset, which includes multiple historical optimization nodes, and each historical optimization node includes: historical grain boundary feature nodes, historical boundary optimization processing parameters, and historical optimization effect values;

[0046] The average grain size, boundary phase thickness, and boundary phase distribution uniformity are extracted from the grain boundary feature nodes. A feature vector is constructed using the average grain size, boundary phase thickness, and boundary phase distribution uniformity. The feature distance value between the feature vector and the feature vector corresponding to each historical grain boundary feature node in the historical optimization dataset is calculated using a pre-constructed Euclidean distance calculation method, thus obtaining a feature distance value set.

[0047] The feature distance values ​​in the feature distance value set are sorted in ascending order to obtain a feature distance value sequence. The target feature distance value set is then extracted from the feature distance value sequence using a preset extraction quantity.

[0048] Based on the historical boundary optimization parameters and historical optimization effect values ​​corresponding to the target feature distance value set, the recommended boundary optimization parameters are calculated using the weighted average method. The weighting weights of the weighted average method are positively correlated with the historical optimization effect values ​​and negatively correlated with the feature distance values.

[0049] After confirming that the recommended boundary optimization parameters are within the range of boundary optimization parameters, the recommended boundary optimization parameters are used as boundary optimization parameter nodes.

[0050] Optionally, the step of identifying the target magnet sample based on the magnetic property detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold includes:

[0051] The magnetic energy product and coercivity are extracted from the magnetic performance detection node, and the sample eddy current loss mean and sample eddy current loss variance are extracted from the sample loss characteristic node.

[0052] Using the measured magnetic energy product, coercivity, mean eddy current loss, variance of eddy current loss, target magnetic energy product threshold, and target coercivity threshold, the comprehensive magnetic performance evaluation value is calculated. The calculation formula is as follows:

[0053] ,

[0054] in, This represents the comprehensive magnetic performance evaluation value. , , All are preset weighting coefficients, and , This represents the detected value of the magnetic energy product. This represents the target magnetic energy product threshold. This represents the coercivity test value. This represents the target coercivity threshold. This represents the average eddy current loss of the sample. This represents the target eddy current loss threshold. The preset uniformity adjustment coefficient, This represents the variance of the eddy current loss in the sample. Represents the largest identifier. Indicates the threshold identifier. Indicates coercivity identifier, This represents the variance identifier. Indicates the mean identifier. Indicates the preset base;

[0055] Once the comprehensive magnetic performance evaluation value is confirmed to be greater than or equal to the preset evaluation threshold, the optimized magnet sample is confirmed as the target magnet sample.

[0056] To achieve the above objectives, the present invention also provides a system for preparing low eddy current loss magnets for high-frequency drive motors, comprising:

[0057] The parameter node confirmation module is used to receive the magnet preparation instruction and confirm the preparation environment parameter node and the target magnet performance parameter node based on the magnet preparation instruction. The preparation environment parameter node includes: sintering temperature, magnetic field orientation intensity and cooling rate, and the target magnet performance parameter node includes: target eddy current loss threshold, target magnetic energy product threshold and target coercivity threshold.

[0058] The sample preparation and cutting module is used to obtain a magnetic steel raw material ratio dataset. Based on the magnetic steel raw material ratio dataset, sintering temperature, magnetic field orientation intensity and cooling rate, a set of preparation parameter combinations is constructed. For each preparation parameter combination in the preparation parameter combination set, the following operations are performed: magnetic steel sample is prepared using the preparation parameter combination to obtain an initial magnetic steel sample. The initial magnetic steel sample is then subjected to layer cutting to obtain a set of layered magnetic steel sheets.

[0059] The eddy current loss screening module is used to perform eddy current loss detection on each layered magnet sheet in the layered magnet sheet set to obtain a layered eddy current loss value set. Based on the layered eddy current loss value set and the target eddy current loss threshold, the screening magnet sample and the sample loss characteristic node are identified.

[0060] The optimization evaluation and confirmation module is used to perform grain boundary optimization processing on the screened magnetic steel samples to obtain optimized magnetic steel samples. The optimized magnetic steel samples are then tested using a pre-constructed magnetic performance detection unit to obtain magnetic performance detection nodes. The magnetic performance detection nodes include: magnetic energy product detection value and coercivity detection value.

[0061] Based on the magnetic performance detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold, the target magnet sample is identified, thus realizing the preparation of low eddy current loss magnets.

[0062] To address the problems described in the background art, this invention receives a magnet manufacturing instruction and, based on this instruction, identifies manufacturing environment parameter nodes and target magnet performance parameter nodes. The manufacturing environment parameter nodes include sintering temperature, magnetic field orientation strength, and cooling rate, while the target magnet performance parameter nodes include a target eddy current loss threshold, a target magnetic energy product threshold, and a target coercivity threshold. A magnet raw material ratio dataset is obtained. Based on the magnet raw material ratio dataset, sintering temperature, magnetic field orientation strength, and cooling rate, a set of manufacturing parameter combinations is constructed. For each manufacturing parameter combination in the set, the following operations are performed: a magnet sample is prepared using the manufacturing parameter combination to obtain an initial magnet sample; the initial magnet sample is then layered and cut to obtain a set of layered magnet sheets; eddy current loss detection is performed on each layered magnet sheet in the set to obtain a set of layered eddy current loss values; and the target eddy current loss threshold is then used to confirm… The present invention, by introducing a multi-frequency eddy current loss detection and statistical screening mechanism, achieves a comprehensive evaluation and precise screening of the eddy current loss characteristics of magnetic steel samples. Grain boundary optimization processing is performed on the screened magnetic steel samples to obtain optimized magnetic steel samples. A pre-constructed magnetic performance detection unit is used to detect the optimized magnetic steel samples, obtaining magnetic performance detection nodes. These nodes include magnetic energy product detection values ​​and coercivity detection values. The present invention, by introducing an intelligent parameter recommendation mechanism based on historical data, achieves adaptive determination of grain boundary optimization processing parameters. Based on the magnetic performance detection nodes, sample loss characteristic nodes, target magnetic energy product threshold, and target coercivity threshold, the target magnetic steel sample is identified, realizing the preparation of low eddy current loss magnets. Therefore, the present invention, by introducing a multi-dimensional comprehensive evaluation mechanism, achieves the scientific screening and precise preparation of low eddy current loss magnets, providing high-performance magnetic steel materials for high-frequency drive motors. Thus, the present invention can solve the problem of excessive eddy current loss in magnets of high-frequency drive motors. Attached Figure Description

[0063] Figure 1 This is a schematic flowchart of a method for preparing a low-eddy current loss magnet for a high-frequency drive motor according to an embodiment of the present invention.

[0064] Figure 2 This is a functional block diagram of a low eddy current loss magnet manufacturing system for a high-frequency drive motor, provided in an embodiment of the present invention.

[0065] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0066] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0067] Reference Figure 1 The diagram shown is a flowchart illustrating a method for preparing a low-eddy current loss magnet for a high-frequency drive motor according to an embodiment of the present invention. In this embodiment, the method for preparing a low-eddy current loss magnet for a high-frequency drive motor includes:

[0068] S1. Receive the magnet preparation instruction, and confirm the preparation environment parameter node and the target magnet performance parameter node based on the magnet preparation instruction. The preparation environment parameter node includes: sintering temperature, magnetic field orientation intensity and cooling rate, and the target magnet performance parameter node includes: target eddy current loss threshold, target magnetic energy product threshold and target coercivity threshold.

[0069] It should be explained that the magnet manufacturing instruction is issued by technicians who need to manufacture low-eddy-current-loss magnets, and is used to initiate the entire magnet manufacturing process. High-frequency drive motors generate alternating magnetic fields during operation. Traditional magnets are prone to significant eddy-current losses in high-frequency environments, leading to reduced motor efficiency and excessive temperature rise. Therefore, it is necessary to manufacture magnet materials with low eddy-current-loss characteristics. The manufacturing environment parameter nodes are a set of key process parameters that need to be controlled during magnet manufacturing. The sintering temperature is the temperature parameter used to densify the magnet powder compact at high temperatures, typically set and monitored by the temperature control system of the sintering furnace. The magnetic field orientation intensity is the external magnetic field intensity applied during magnet forming, used to align the magnetic powder particles in a specific direction to obtain optimal magnetic properties. The magnetic field orientation intensity is provided by an orientation magnetic field generator. The cooling rate is the speed at which the magnet cools from high temperature to room temperature after sintering, regulated by the control unit of the cooling system. The target magnet performance parameter nodes are the set of performance indicators that the magnet to be prepared needs to achieve. The target eddy current loss threshold is the maximum allowable eddy current loss value of the magnet in a high-frequency working environment. The target magnetic energy product threshold is the minimum magnetic energy product value that the magnet needs to achieve. The target coercivity threshold is the minimum coercivity value that the magnet needs to achieve.

[0070] For example, engineers at a new energy vehicle drive motor manufacturing company need to prepare specialized magnets for a high-speed permanent magnet synchronous motor. The motor operates at frequencies ranging from 500Hz to 2000Hz, and there are strict requirements for the eddy current losses of the magnets. The engineers issue magnet preparation instructions through the production management system and confirm the preparation environment parameters, including a sintering temperature of 1080 degrees Celsius, a magnetic field orientation strength of 2.0 Tesla, and a cooling rate of 5 degrees Celsius per minute. They also confirm the target magnet performance parameters, including a target eddy current loss threshold of 0.8 watts per cubic centimeter, a target magnetic energy product threshold of 42 mega-Oersted, and a target coercivity threshold of 20 kilo-Oersted.

[0071] S2. Obtain the magnetic steel raw material ratio dataset. Based on the magnetic steel raw material ratio dataset, sintering temperature, magnetic field orientation intensity and cooling rate, construct a set of preparation parameter combinations. Perform the following operations on each preparation parameter combination in the preparation parameter combination set: prepare magnetic steel samples using the preparation parameter combination to obtain initial magnetic steel samples. Perform layer cutting processing on the initial magnetic steel samples to obtain a set of layered magnetic steel sheets.

[0072] It should be understood that the aforementioned magnet raw material ratio dataset is a collection of data containing multiple raw material ratio schemes, stored in the raw material database of the production management system. Each ratio scheme specifies the proportional relationship of various raw materials required for preparing the magnet. The aforementioned preparation parameter combination set is a set of parameter combinations formed by associating the raw material ratios with process parameters, used to guide subsequent magnet sample preparation. The aforementioned initial magnet sample is a magnet sample prepared according to a specific combination of preparation parameters and has not yet undergone performance screening. The layered magnet sheet set is a collection of multiple thin sheets obtained by cutting the initial magnet sample along the thickness direction, used for subsequent eddy current loss detection. Since the microstructure of different locations within the magnet may differ, layered cutting allows for a comprehensive evaluation of the eddy current loss characteristics of the entire magnet sample.

[0073] Understandably, the construction of the preparation parameter combination needs to comprehensively consider the matching of raw material ratios and process parameters. An unreasonable combination of preparation parameters may lead to substandard magnet performance or preparation failure. Therefore, the construction of the preparation parameter combination set based on the magnet raw material ratio dataset, sintering temperature, magnetic field orientation intensity, and cooling rate includes:

[0074] Multiple raw material ratio nodes were extracted from the magnetic steel raw material ratio dataset. These raw material ratio nodes include: rare earth element content, iron element content, and additive content.

[0075] Perform the following operation for each of the multiple raw material proportioning nodes:

[0076] The raw material ratio nodes are verified using pre-constructed ratio constraint relationships to obtain verification results, where the verification result is either pass or fail.

[0077] After confirming that the verification result is passed, the initial preparation parameter combination is obtained by associating the raw material ratio node, sintering temperature, magnetic field orientation strength and cooling rate.

[0078] The initial preparation parameter combinations are summarized to obtain an initial preparation parameter combination set. The initial preparation parameter combination set is then screened using a pre-defined orthogonal experimental design method to obtain a final preparation parameter combination set.

[0079] It should be explained that the raw material proportioning node is a data node describing a single raw material proportioning scheme. The rare earth element content is the percentage of rare earth elements (such as neodymium, praseodymium, dysprosium, etc.) in the total mass of the magnet raw material; rare earth elements are key components determining the magnetic properties of magnets. The iron element content is the percentage of iron in the total mass of the magnet raw material; iron is the main matrix component of magnets. The additive content is the percentage of additives (such as copper, aluminum, cobalt, etc.) in the total mass of the magnet raw material; additives are used to improve specific properties of magnets. The proportioning constraint formula is a mathematical formula pre-established based on the principles of magnet materials science, used to determine whether the raw material proportions are within a reasonable range; the proportioning constraint formula is stored in the process knowledge base of the production management system. The verification result is the judgment result obtained after constraining and verifying the raw material proportioning node. The initial preparation parameter combination is a parameter combination formed by associating the verified raw material proportioning node with process parameters. The orthogonal experimental design method is a scientific experimental design method that can use standard orthogonal arrays such as L9 or L16 orthogonal arrays to optimize and screen parameter combinations, and can obtain representative parameter combinations with fewer experiments.

[0080] For example, eight raw material proportioning nodes were extracted from the magnetic steel raw material proportioning dataset. The first node had a rare earth element content of 31%, an iron element content of 67%, and an additive content of 2%. Verification was performed using proportioning constraints, requiring rare earth element content to be between 28% and 35%, iron element content between 63% and 70%, and additive content between 1% and 4%, with the sum of these three being 100%. The first node satisfied all constraints, and the verification result was passed. This node was then correlated with a sintering temperature of 1080 degrees Celsius, a magnetic field orientation strength of 2.0 Tesla, and a cooling rate of 5 degrees Celsius per minute to obtain initial preparation parameter combinations. After verification and correlation of all raw material proportioning nodes, six initial preparation parameter combinations were obtained. The L9 orthogonal array was used to filter the initial preparation parameter combination set, ultimately resulting in a preparation parameter combination set containing four combinations.

[0081] Furthermore, the initial magnet sample is layered and cut to comprehensively evaluate the eddy current loss characteristics of each layer within the magnet. Therefore, the layered cutting process of the initial magnet sample to obtain a set of layered magnet sheets includes:

[0082] Obtain the sample thickness parameters of the initial magnet sample, and calculate the number of layers based on the preset layer thickness threshold and sample thickness parameters;

[0083] The initial magnet sample is uniformly cut using the number of layers to obtain multiple cut magnet pieces;

[0084] Perform the following operation on each of the multiple cutting magnets:

[0085] The surface flatness detection unit is used to detect the cut magnetic steel sheet to obtain the surface flatness value. The surface flatness value is compared with the preset flatness threshold. If the surface flatness value is greater than or equal to the flatness threshold, the cut magnetic steel sheet is ground to obtain a ground magnetic steel sheet. The ground magnetic steel sheet is used as a layered magnetic steel sheet.

[0086] Otherwise, using the cut magnetic steel sheets as layered magnetic steel sheets, the layered magnetic steel sheets are summarized to obtain a set of layered magnetic steel sheets.

[0087] It should be understood that the sample thickness parameter is the total thickness of the initial magnet sample along the cutting direction, which can optionally be obtained by measuring with a thickness measuring instrument. The layer thickness threshold is a pre-set target thickness value for each layer of magnet sheet, stored in the process parameter database. The setting of the layer thickness threshold needs to comprehensively consider the accuracy requirements of eddy current loss detection and the feasibility of the cutting process. The number of layers is the number of cutting layers calculated by dividing the sample thickness parameter by the layer thickness threshold. The cut magnet sheet is a thin sheet obtained by cutting the initial magnet sample using precision cutting equipment such as a wire cutting machine or a diamond cutting machine. The surface flatness detection unit is a detection device used to detect the surface flatness of the magnet sheet. Optionally, it can be achieved using equipment such as a laser displacement sensor or a white light interferometer. The surface flatness value is a numerical value characterizing the surface roughness of the cut magnet sheet; the larger the value, the rougher the surface of the magnet sheet. The flatness threshold is a pre-set maximum allowable surface flatness value, stored in the quality control parameter library. The ground magnetic sheet is a magnetic sheet obtained by grinding cut magnetic sheets with unsatisfactory surface flatness. Optionally, the grinding process can be carried out using a metallographic grinding machine with sandpaper of different grits. The layered magnetic sheet is a thin magnetic sheet that has passed the flatness inspection.

[0088] For example, the initial magnet sample thickness was measured to be 10 mm, the preset layer thickness threshold was 2 mm, and the calculated number of layers was 5. The initial magnet sample was uniformly cut using a wire EDM machine to obtain 5 cut magnet pieces. The surface flatness of the first cut magnet piece was detected using a laser displacement sensor, and the surface flatness value was 0.8 μm. The preset flatness threshold was 1.0 μm. Since 0.8 μm is less than 1.0 μm, the surface flatness is acceptable, and this cut magnet piece was directly used as the layered magnet piece. The third cut magnet piece was tested, and the surface flatness value was 1.5 μm, which is greater than the flatness threshold, requiring grinding. The cut magnet piece was ground using a metallographic grinder with 2000-grit sandpaper, resulting in a ground magnet piece with a surface flatness value reduced to 0.6 μm. This ground magnet piece was used as the layered magnet piece. After processing all the cut magnet pieces, a set of layered magnet pieces containing 5 layers was finally obtained.

[0089] S3. Perform eddy current loss detection on each layered magnetic sheet in the layered magnetic sheet set to obtain a layered eddy current loss value set. Based on the layered eddy current loss value set and the target eddy current loss threshold, identify the screened magnetic sample and the sample loss characteristic node.

[0090] Understandably, eddy current loss is the energy loss of a magnet in an alternating magnetic field due to induced eddy currents. For high-frequency drive motors, eddy current loss is a significant factor affecting motor efficiency. The layered eddy current loss value set is the set of loss values ​​obtained after detecting the eddy current loss of each layered magnet in the layered magnet sheet set. The screened magnet samples are initial magnet samples whose eddy current loss characteristics meet the requirements. The sample loss feature nodes are data nodes describing the eddy current loss characteristics of the screened magnet samples, including statistical features such as the loss mean and loss variance.

[0091] It should be explained that the high-frequency drive motor has a wide operating frequency range, requiring eddy current loss detection at multiple frequency points to comprehensively evaluate the high-frequency characteristics of the magnet. Therefore, the eddy current loss detection operation is performed on each layer of magnet in the set of layered magnet sheets to obtain a set of layered eddy current loss values, including:

[0092] Obtain a detection frequency sequence for eddy current loss detection, wherein the detection frequency sequence includes multiple detection frequencies and covers the operating frequency range of the high-frequency drive motor.

[0093] For each layered magnet in the assembly, perform the following operations:

[0094] For each detection frequency in the detection frequency sequence, perform the following operation:

[0095] The layered magnetic steel sheet is tested using a detection frequency and a pre-constructed eddy current loss detection device to obtain the frequency eddy current loss value. The frequency eddy current loss value is correlated with the detection frequency to obtain the frequency loss node.

[0096] The frequency loss nodes are summarized to obtain a frequency loss node set. The weighted eddy current loss value is calculated based on the frequency loss node set. The weighted eddy current loss value is the weighted average of multiple frequency eddy current loss values ​​in the frequency loss node set based on the detection frequency.

[0097] By summing the weighted eddy current loss values, a layered eddy current loss value set is obtained.

[0098] It should be understood that the detection frequency sequence is a pre-set set of frequency points for eddy current loss detection, stored in the detection parameter database. The setting of the detection frequency sequence is based on the actual operating frequency range of the high-frequency drive motor. The eddy current loss detection device is a dedicated testing device for measuring the eddy current loss of magnetic materials. Optionally, an AC magnetic performance testing system or an iron loss tester can be used for detection. The frequency eddy current loss value is the eddy current loss value measured at a specific detection frequency. The frequency loss node is a data node formed by associating the frequency eddy current loss value with the corresponding detection frequency. The frequency loss node set is a data set containing multiple frequency loss nodes. The weighted eddy current loss value is a comprehensive loss value obtained by weighted averaging of multiple frequency eddy current loss values. The weighting coefficient is determined according to the proportion of each frequency in the actual operation of the motor and is stored in the process parameter database.

[0099] Furthermore, the calculation of eddy current loss needs to consider the physical characteristics of the magnetic steel sheet and the detection conditions. Therefore, the step of using a detection frequency and a pre-constructed eddy current loss detection device to detect the layered magnetic steel sheet and obtain the frequency eddy current loss value includes:

[0100] The resistivity and thickness parameters of the layered magnetic steel sheet are obtained, and an alternating magnetic field is applied to the layered magnetic steel sheet using an eddy current loss detection device to obtain the peak value of the magnetic induction intensity.

[0101] Based on the resistivity parameter, sheet thickness parameter, detection frequency, and peak magnetic induction intensity, the frequency eddy current loss value is calculated using the following formula:

[0102] ,

[0103] in, This represents the frequency eddy current loss value. This indicates the thickness parameter of the sheet. This indicates the detection frequency. This indicates the peak value of the magnetic flux density. This represents the resistivity parameter. The preset value of pi. The preset permeability constant, This indicates the eddy current identifier. Indicates the peak value identifier. This represents the hyperbolic tangent function.

[0104] It should be explained that the resistivity parameter is the resistivity value of the layered magnetic steel sheet, obtained by measurement using a four-probe resistivity tester. The layer thickness parameter is the actual thickness of the layered magnetic steel sheet, obtained by measurement using a micrometer or laser thickness gauge. The peak magnetic induction intensity is the maximum value of the magnetic induction intensity inside the magnetic steel sheet when an alternating magnetic field is applied by the eddy current loss detection device, measured by the fluxmeter of the eddy current loss detection device. The pi is 3.14159, and the permeability constant is the permeability of free space, with a value of [value missing]. Henry per meter. The calculation formula comprehensively considers the effects of classical eddy current loss and skin effect, and can accurately calculate eddy current loss under high-frequency conditions.

[0105] For example, eddy current loss was measured on the first layered magnet sheet. The detection frequency sequence included four frequencies: 500Hz, 1000Hz, 1500Hz, and 2000Hz. The measured resistivity parameter of this layered magnet sheet was 1.5 × 10⁻⁻⁻⁻⁶. 6 The sheet thickness is 2 mm, measured in ohms and 2 mm. At a detection frequency of 500 Hz, an alternating magnetic field is applied using an eddy current loss detection device, and the peak magnetic induction intensity is measured to be 1.2 Tesla. Substituting this into the calculation formula, the frequency eddy current loss value is found to be 0.45 W / cm³. Correlating this frequency eddy current loss value with 500 Hz yields the first frequency loss node. Sequential detections are performed at 1000 Hz, 1500 Hz, and 2000 Hz, yielding frequency eddy current loss values ​​of 0.62 W / cm³, 0.78 W / cm³, and 0.91 W / cm³, respectively. The four frequency loss nodes are then combined to obtain the frequency loss node set. Based on the proportion of each frequency in the motor's operation (500Hz 20%, 1000Hz 35%, 1500Hz 30%, 2000Hz 15%), the weighted eddy current loss value is calculated to be approximately 0.68 watts per cubic centimeter (W / cm³). The same testing operation is performed on all five layered magnets, resulting in a set of layered eddy current loss values: {0.68, 0.71, 0.65, 0.73, 0.69} W / cm³.

[0106] Understandably, the average eddy current loss alone cannot fully assess the quality of a magnet sample; the uniformity of the loss value also needs to be considered. Therefore, the process of identifying the screened magnet samples and sample loss characteristic nodes based on the layered eddy current loss value set and the target eddy current loss threshold includes:

[0107] Calculate the mean of the layered eddy current loss values ​​in the set of layered eddy current loss values ​​to obtain the mean of the sample eddy current loss. Calculate the variance of the layered eddy current loss values ​​in the set of layered eddy current loss values ​​to obtain the variance of the sample eddy current loss.

[0108] The mean value of the sample eddy current loss is compared with the target eddy current loss threshold. If the mean value of the sample eddy current loss is less than or equal to the target eddy current loss threshold, the variance of the sample eddy current loss is compared with a preset variance threshold.

[0109] If the variance of the sample eddy current loss is less than or equal to the variance threshold, the initial magnet sample is confirmed as the screening magnet sample. The mean of the sample eddy current loss and the variance of the sample eddy current loss are correlated to obtain the sample loss feature node.

[0110] Otherwise, abnormal eddy current loss values ​​are identified in the layered eddy current loss value set, wherein the absolute difference between the abnormal eddy current loss value and the average eddy current loss value of the sample is greater than a preset deviation threshold. The abnormal eddy current loss value is removed from the layered eddy current loss value set to obtain an updated eddy current loss value set. The updated eddy current loss value set is used as the layered eddy current loss value set, and the step of calculating the average value of the layered eddy current loss values ​​in the layered eddy current loss value set is returned until the screened magnetic steel sample and the sample loss feature node are obtained.

[0111] It should be understood that the mean eddy current loss of the sample is the arithmetic mean of all layered eddy current loss values ​​in the layered eddy current loss value set, reflecting the overall eddy current loss level of the magnet sample. The variance of the sample eddy current loss is the degree of dispersion of all layered eddy current loss values ​​in the layered eddy current loss value set relative to the mean, reflecting the uniformity of eddy current loss within the magnet sample. The variance threshold is a pre-set maximum allowable value for eddy current loss variance, stored in the quality control parameter library. The variance threshold is set based on the uniformity requirements of the high-frequency drive motor for the magnet. Abnormal eddy current loss values ​​are layered eddy current loss values ​​that deviate excessively from the mean eddy current loss of the sample, possibly caused by local defects or measurement errors. The deviation threshold is a pre-set critical deviation value for judging abnormal eddy current loss values, stored in the quality control parameter library. The updated eddy current loss value set is the set of layered eddy current loss values ​​after removing abnormal eddy current loss values.

[0112] For example, statistical calculations were performed on the layered eddy current loss value set {0.68, 0.71, 0.65, 0.73, 0.69}, yielding a mean eddy current loss of 0.692 watts per cubic centimeter and a variance of 0.00082. The target eddy current loss threshold is 0.8 watts per cubic centimeter. The mean eddy current loss of 0.692 is less than the target threshold of 0.8, satisfying the mean requirement. The preset variance threshold is 0.001. The variance of 0.00082 is less than the variance threshold of 0.001, satisfying the uniformity requirement. Therefore, this initial magnet sample was identified as a screening magnet sample. Correlating the mean eddy current loss of 0.692 with the variance of 0.00082, the sample loss feature node was obtained. Assuming another initial magnetic steel sample has a layered eddy current loss value set of {0.72, 0.75, 0.68, 0.95, 0.71}, the calculated mean eddy current loss of the sample is 0.762 watts per cubic centimeter, and the variance of the sample eddy current loss is 0.0104, which exceeds the threshold. The preset deviation threshold is 0.15 watts per cubic centimeter. The absolute difference between the fourth layered eddy current loss value 0.95 and the mean 0.762 is 0.188, which is greater than the deviation threshold, and is identified as an abnormal eddy current loss value and removed. The eddy current loss value set is updated to {0.72, 0.75, 0.68, 0.71}, and the mean is recalculated to be 0.715 watts per cubic centimeter with a variance of 0.00073, which meets the requirements and is confirmed as a screened magnetic steel sample. This embodiment of the invention, by introducing a multi-frequency eddy current loss detection and statistical screening mechanism, achieves a comprehensive evaluation and accurate screening of the eddy current loss characteristics of magnetic steel samples.

[0113] S4. Perform grain boundary optimization processing on the screened magnet samples to obtain optimized magnet samples. Use a pre-constructed magnetic performance detection unit to detect the optimized magnet samples to obtain magnetic performance detection nodes. The magnetic performance detection nodes include: magnetic energy product detection value and coercivity detection value.

[0114] It should be explained that the grain boundary optimization treatment involves modifying the grain boundaries of the magnet sample through a specific process to improve the overall magnetic properties of the magnet. The grain boundary structure of the magnet has a significant impact on its coercivity and resistance to demagnetization. Optimizing the grain boundaries can improve magnetic properties while maintaining low eddy current losses. The optimized magnet sample is the magnet sample after grain boundary optimization treatment. The magnetic property detection unit is a dedicated detection device for detecting the magnetic property parameters of the magnet. Optionally, a permanent magnet material measuring instrument or a BH hysteresis loop tester can be used for detection. The magnetic property detection node is a data node describing the magnetic properties of the optimized magnet sample. The magnetic energy product detection value is the maximum magnetic energy product measurement value of the magnet, reflecting the magnet's ability to store magnetic energy. The coercivity detection value is the intrinsic coercivity measurement value of the magnet, reflecting the magnet's resistance to demagnetization.

[0115] Understandably, grain boundary optimization requires determining appropriate processing parameters based on the microstructural characteristics of the magnet sample. Therefore, the process of performing grain boundary optimization on the screened magnet sample to obtain an optimized magnet sample includes:

[0116] The microstructure images of the grains of the screened magnetic steel samples are obtained. The microstructure images of the grains are analyzed using a pre-built image analysis model to obtain the grain boundary feature nodes. The grain boundary feature nodes include: average grain size, boundary phase thickness and boundary phase distribution uniformity.

[0117] Based on the grain boundary feature nodes and the preset boundary optimization parameter range, the boundary optimization processing parameter nodes are identified, including: diffusion annealing temperature, diffusion annealing time, and diffusion element concentration.

[0118] The grain boundary diffusion process is performed on the screened magnet samples using the boundary optimization processing parameter nodes to obtain diffusion-treated magnet samples. The diffusion-treated magnet samples are then subjected to aging treatment to obtain optimized magnet samples.

[0119] It should be understood that the grain microstructure image is a microstructure image obtained by observing the screened magnet sample using a scanning electron microscope or metallographic microscope. The image analysis model is a pre-constructed computer vision model for analyzing the magnet microstructure image. Optionally, deep learning-based image segmentation and feature extraction algorithms can be used to perform image analysis. The image analysis model is stored in an image analysis server. The grain boundary feature nodes are data nodes describing the microstructure characteristics of the magnet grain boundaries. The average grain size is the average diameter or equivalent size of the grains in the magnet. The boundary phase thickness is the average thickness of the rare-earth-rich phase at the grain boundary. The boundary phase distribution uniformity is the uniformity of the grain boundary phase distribution in the magnet; a higher value indicates a more uniform distribution. The boundary optimization parameter range is a pre-set allowable range of grain boundary optimization processing parameters, stored in a process parameter database. The boundary optimization processing parameter nodes are a set of process parameters used to guide grain boundary diffusion processing. The diffusion annealing temperature is the heating temperature during grain boundary diffusion processing. The diffusion annealing time is the duration of grain boundary diffusion processing. The diffusion element concentration refers to the concentration of heavy rare earth elements (such as dysprosium or terbium) used for grain boundary diffusion. The diffusion-treated magnet sample is a magnet sample that has undergone grain boundary diffusion treatment. The aging treatment involves holding the diffusion-treated magnet sample at a specific temperature to stabilize its microstructure and magnetic properties.

[0120] Furthermore, the determination of boundary optimization parameters requires reference to historical optimization experience. Therefore, the step of identifying boundary optimization parameter nodes based on the grain boundary feature nodes and the preset boundary optimization parameter range includes:

[0121] Obtain the historical optimization dataset, which includes multiple historical optimization nodes, and each historical optimization node includes: historical grain boundary feature nodes, historical boundary optimization processing parameters, and historical optimization effect values;

[0122] The average grain size, boundary phase thickness, and boundary phase distribution uniformity are extracted from the grain boundary feature nodes. A feature vector is constructed using the average grain size, boundary phase thickness, and boundary phase distribution uniformity. The feature distance value between the feature vector and the feature vector corresponding to each historical grain boundary feature node in the historical optimization dataset is calculated using a pre-constructed Euclidean distance calculation method, thus obtaining a feature distance value set.

[0123] The feature distance values ​​in the feature distance value set are sorted in ascending order to obtain a feature distance value sequence. The target feature distance value set is then extracted from the feature distance value sequence using a preset extraction quantity.

[0124] Based on the historical boundary optimization parameters and historical optimization effect values ​​corresponding to the target feature distance value set, the recommended boundary optimization parameters are calculated using the weighted average method. The weighting weights of the weighted average method are positively correlated with the historical optimization effect values ​​and negatively correlated with the feature distance values.

[0125] After confirming that the recommended boundary optimization parameters are within the range of boundary optimization parameters, the recommended boundary optimization parameters are used as boundary optimization parameter nodes.

[0126] It should be explained that the historical optimization dataset is a collection of historical grain boundary optimization processing records stored in the process knowledge base, containing information such as grain boundary characteristics, optimization processing parameters, and optimization effects of previous magnet samples. The historical optimization node is a data node describing a single historical optimization processing record. The historical grain boundary feature node represents the microstructural characteristics of the grain boundaries of historical magnet samples. The historical boundary optimization processing parameters are the process parameters used in the historical optimization processing. The historical optimization effect value is a quantitative evaluation value of the degree of performance improvement of the magnet after historical optimization processing; a higher value indicates a better optimization effect. The feature vector is a vector composed of three feature values: average grain size, boundary phase thickness, and boundary phase distribution uniformity. The Euclidean distance calculation method is a mathematical method for calculating the Euclidean distance between two feature vectors; a smaller Euclidean distance indicates a more similar grain boundary feature between the two samples (the feature vector of the current sample and the feature vector of the historical sample). The feature distance value is the Euclidean distance between the feature vector of the current sample and the feature vector of the historical sample. The feature distance value sequence is a sequence arranged in ascending order of feature distance values. The number of extractions is a pre-set number of historical optimization nodes for reference, stored in the algorithm parameter library. The target feature distance value set is the set of feature distance values ​​corresponding to multiple historical samples most similar to the current sample, extracted from the feature distance value sequence. The recommended boundary optimization processing parameters are recommended process parameters calculated based on optimization experience of similar historical samples.

[0127] For example, scanning electron microscopy is used to acquire images of the grain microstructure of the screened magnetic steel sample. Image analysis is then performed to obtain grain boundary feature nodes, where the average grain size is 5.2 micrometers, the boundary phase thickness is 3.8 nanometers, and the boundary phase distribution uniformity is 0.72. A feature vector is further constructed as (5.2, 3.8, 0.72). A historical optimization dataset containing 50 historical optimization nodes is obtained from the process knowledge base. The Euclidean distance between the current feature vector and the corresponding feature vector of each historical grain boundary feature node is calculated to obtain a feature distance value set. After sorting the feature distance values ​​in ascending order, with a preset extraction quantity of 5, the 5 historical optimization nodes with the smallest feature distance values ​​are extracted as the target feature distance value set. The feature distance values ​​of the five historical optimization nodes are 0.15, 0.23, 0.31, 0.42, and 0.48, respectively. The corresponding historical boundary optimization parameters are (850°C, 4 hours, 0.8%), (860°C, 3.5 hours, 0.9%), (840°C, 4.5 hours, 0.7%), (855°C, 4 hours, 0.85%), and (845°C, 3.8 hours, 0.75%), respectively. The corresponding historical optimization effect values ​​are 0.92, 0.88, 0.85, 0.90, and 0.86. Based on the weighted average method, the weighting is positively correlated with the historical optimization effect value and negatively correlated with the feature distance value. Therefore, the recommended boundary optimization parameters are calculated as follows: diffusion annealing temperature 851°C, diffusion annealing time 3.9 hours, and diffusion element concentration 0.82%. The preset boundary optimization parameter range is a diffusion annealing temperature of 800 to 900 degrees Celsius, a diffusion annealing time of 2 to 6 hours, and a diffusion element concentration of 0.5% to 1.5%. All recommended parameters fall within this range, thus confirming it as a boundary optimization parameter node. Using this boundary optimization parameter node, the screened magnet samples underwent grain boundary diffusion treatment, followed by aging at 500 degrees Celsius for 2 hours, resulting in optimized magnet samples. The optimized magnet samples were then tested using a permanent magnet material measuring instrument, yielding a magnetic energy product of 45.2 MGO Oersted and a coercivity of 23.5 kilo-Oersted. This embodiment of the invention achieves adaptive determination of grain boundary optimization parameters by introducing an intelligent parameter recommendation mechanism based on historical data.

[0128] S5. Based on the magnetic performance detection node, sample loss characteristic node, target magnetic energy product threshold and target coercivity threshold, the target magnetic steel sample is identified, and the preparation of low eddy current loss magnet is realized.

[0129] Understandably, determining the target magnet sample requires comprehensive consideration of magnetic properties and eddy current loss characteristics. A single indicator cannot fully evaluate the overall performance of the magnet. Therefore, the process of identifying the target magnet sample based on the magnetic property detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold includes:

[0130] The magnetic energy product and coercivity are extracted from the magnetic performance detection node, and the sample eddy current loss mean and sample eddy current loss variance are extracted from the sample loss characteristic node.

[0131] Using the measured magnetic energy product, coercivity, mean eddy current loss, variance of eddy current loss, target magnetic energy product threshold, and target coercivity threshold, the comprehensive magnetic performance evaluation value is calculated. The calculation formula is as follows:

[0132] ,

[0133] in, This represents the comprehensive magnetic performance evaluation value. , , All are preset weighting coefficients, and , This represents the detected value of the magnetic energy product. This represents the target magnetic energy product threshold. This represents the coercivity test value. This represents the target coercivity threshold. This represents the average eddy current loss of the sample. This represents the target eddy current loss threshold. The preset uniformity adjustment coefficient, This represents the variance of the eddy current loss in the sample. Represents the largest identifier. Indicates the threshold identifier. Indicates coercivity identifier, This represents the variance identifier. Indicates the mean identifier. Indicates the preset base;

[0134] Once the comprehensive magnetic performance evaluation value is confirmed to be greater than or equal to the preset evaluation threshold, the optimized magnet sample is confirmed as the target magnet sample.

[0135] It should be explained that the comprehensive magnetic performance evaluation value is a comprehensive performance evaluation index of the magnet calculated by comprehensively considering the magnetic energy product, coercivity, and eddy current loss characteristics. The weighting coefficients... , , These represent the relative importance of magnetic energy product, coercivity, and eddy current loss characteristics in the comprehensive evaluation. They are set by technicians based on the specific application requirements of the high-frequency drive motor and stored in the evaluation parameter database. The uniformity adjustment coefficient... This is a coefficient used to adjust the influence of eddy current loss uniformity on the overall evaluation value, and it is stored in the evaluation parameter library. The evaluation threshold is a pre-set minimum requirement for the overall magnetic performance evaluation value, stored in the quality control parameter library. Only magnetic steel samples whose overall magnetic performance evaluation value reaches this threshold can be identified as target magnetic steel samples. The calculation formula uses a logarithmic function to handle the magnetic energy product and coercivity index, which can smooth the contribution of performance exceeding the standard. The introduction of the reciprocal term of the eddy current loss variance can give extra points to samples with good loss uniformity.

[0136] For example, the magnetic energy product is extracted from the magnetic performance detection node to be 45.2 MGO Oersted, and the coercivity is extracted to be 23.5 kilo-Oersted. The sample eddy current loss is extracted from the sample loss feature node to be a mean of 0.692 watts per cubic centimeter, and the sample eddy current loss variance is 0.00082. The target magnetic energy product threshold is 42 MGO Oersted, the target coercivity threshold is 20 kilo-Oersted, and the target eddy current loss threshold is 0.8 watts per cubic centimeter. Preset weighting coefficients are used. It is 0.35. It is 0.30. The uniformity adjustment coefficient is 0.35. The value is 0.001. Substituting into the calculation formula: Q=0.35×ln(1+45.2 / 42)+0.30×ln(1+23.5 / 20)+0.35×exp(-0.692 / 0.8)×(1+0.001×1219.5)=0.35×0.693+0.30×0.693+0.35×0.421×2.22=0.243+0.208+0.327=0.778. The preset evaluation threshold is 0.70. The comprehensive magnetic performance evaluation value of 0.778 is greater than the evaluation threshold of 0.70. Therefore, the optimized magnet sample is confirmed as the target magnet sample.

[0137] Furthermore, for other combinations of preparation parameters in the set, the same S2 to S5 procedures are performed, potentially yielding multiple target magnet samples. Technicians can select the optimal target magnet sample as the final product based on the comprehensive magnetic performance evaluation value, and record the corresponding preparation parameter combination in the process knowledge base for reference in subsequent mass production. This invention, by introducing a multi-dimensional comprehensive evaluation mechanism, achieves the scientific screening and precise preparation of low-eddy-current-loss magnets, providing high-performance magnet materials for high-frequency drive motors.

[0138] To address the problems described in the background art, this invention receives a magnet manufacturing instruction and, based on this instruction, identifies manufacturing environment parameter nodes and target magnet performance parameter nodes. The manufacturing environment parameter nodes include sintering temperature, magnetic field orientation strength, and cooling rate, while the target magnet performance parameter nodes include a target eddy current loss threshold, a target magnetic energy product threshold, and a target coercivity threshold. A magnet raw material ratio dataset is obtained. Based on the magnet raw material ratio dataset, sintering temperature, magnetic field orientation strength, and cooling rate, a set of manufacturing parameter combinations is constructed. For each manufacturing parameter combination in the set, the following operations are performed: a magnet sample is prepared using the manufacturing parameter combination to obtain an initial magnet sample; the initial magnet sample is then layered and cut to obtain a set of layered magnet sheets; eddy current loss detection is performed on each layered magnet sheet in the set to obtain a set of layered eddy current loss values; and the target eddy current loss threshold is then used to confirm… The present invention, by introducing a multi-frequency eddy current loss detection and statistical screening mechanism, achieves a comprehensive evaluation and precise screening of the eddy current loss characteristics of magnetic steel samples. Grain boundary optimization processing is performed on the screened magnetic steel samples to obtain optimized magnetic steel samples. A pre-constructed magnetic performance detection unit is used to detect the optimized magnetic steel samples, obtaining magnetic performance detection nodes. These nodes include magnetic energy product detection values ​​and coercivity detection values. The present invention, by introducing an intelligent parameter recommendation mechanism based on historical data, achieves adaptive determination of grain boundary optimization processing parameters. Based on the magnetic performance detection nodes, sample loss characteristic nodes, target magnetic energy product threshold, and target coercivity threshold, the target magnetic steel sample is identified, realizing the preparation of low eddy current loss magnets. Therefore, the present invention, by introducing a multi-dimensional comprehensive evaluation mechanism, achieves the scientific screening and precise preparation of low eddy current loss magnets, providing high-performance magnetic steel materials for high-frequency drive motors. Thus, the present invention can solve the problem of excessive eddy current loss in magnets of high-frequency drive motors.

[0139] like Figure 2 The diagram shown is a functional block diagram of a low eddy current loss magnet manufacturing system for a high-frequency drive motor provided in an embodiment of the present invention.

[0140] Depending on the functions implemented, the low eddy current loss magnet preparation system 100 for high frequency drive motors may include a parameter node confirmation module 101, a sample preparation and cutting module 102, an eddy current loss screening module 103, and an optimization evaluation and confirmation module 104.

[0141] The parameter node confirmation module 101 is used to receive the magnet preparation instruction and confirm the preparation environment parameter node and the target magnet performance parameter node based on the magnet preparation instruction. The preparation environment parameter node includes: sintering temperature, magnetic field orientation intensity and cooling rate, and the target magnet performance parameter node includes: target eddy current loss threshold, target magnetic energy product threshold and target coercivity threshold.

[0142] The sample preparation and cutting module 102 is used to acquire a magnetic steel raw material ratio dataset, construct a set of preparation parameter combinations based on the magnetic steel raw material ratio dataset, sintering temperature, magnetic field orientation intensity and cooling rate, and perform the following operations on each preparation parameter combination in the preparation parameter combination set: prepare a magnetic steel sample using the preparation parameter combination to obtain an initial magnetic steel sample, and perform layer cutting on the initial magnetic steel sample to obtain a set of layered magnetic steel sheets.

[0143] The eddy current loss screening module 103 is used to perform eddy current loss detection on each layered magnetic steel sheet in the layered magnetic steel sheet set to obtain a layered eddy current loss value set, and to identify the screened magnetic steel sample and sample loss characteristic node based on the layered eddy current loss value set and the target eddy current loss threshold.

[0144] The optimization evaluation and confirmation module 104 is used to perform grain boundary optimization processing on the screened magnetic steel samples to obtain optimized magnetic steel samples. The optimized magnetic steel samples are then tested using a pre-constructed magnetic performance detection unit to obtain magnetic performance detection nodes. The magnetic performance detection nodes include: magnetic energy product detection value and coercivity detection value.

[0145] Based on the magnetic performance detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold, the target magnet sample is identified, thus realizing the preparation of low eddy current loss magnets.

[0146] In detail, the modules in the low eddy current loss magnet manufacturing system 100 for high-frequency drive motors described in this embodiment of the invention employ the same methods as described above during use. Figure 1 The method used is the same as the one described in the article for preparing low eddy current loss magnets for high frequency drive motors, and can produce the same technical effect, so it will not be repeated here.

[0147] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0148] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0150] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for preparing low-eddy current loss magnets for high-frequency drive motors, characterized in that, The method includes: The system receives a magnetic steel preparation instruction and, based on the instruction, identifies preparation environment parameter nodes and target magnetic steel performance parameter nodes. The preparation environment parameter nodes include sintering temperature, magnetic field orientation strength, and cooling rate, while the target magnetic steel performance parameter nodes include target eddy current loss threshold, target magnetic energy product threshold, and target coercivity threshold. Obtain a dataset of magnetic steel raw material proportions. Based on the dataset of magnetic steel raw material proportions, sintering temperature, magnetic field orientation strength, and cooling rate, construct a set of preparation parameter combinations. Perform the following operations on each preparation parameter combination in the preparation parameter combination set: prepare a magnetic steel sample using the preparation parameter combination to obtain an initial magnetic steel sample. Perform layer cutting on the initial magnetic steel sample to obtain a set of layered magnetic steel sheets. Eddy current loss detection is performed on each layered magnetic sheet in the layered magnetic sheet set to obtain a layered eddy current loss value set. Based on the layered eddy current loss value set and the target eddy current loss threshold, the screened magnetic samples and sample loss characteristic nodes are identified. The grain boundary optimization process is performed on the screened magnetic steel samples to obtain optimized magnetic steel samples. The optimized magnetic steel samples are then tested using a pre-constructed magnetic performance detection unit to obtain magnetic performance detection nodes. The magnetic performance detection nodes include: magnetic energy product detection value and coercivity detection value. Based on the magnetic performance detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold, the target magnet sample is identified, thus realizing the preparation of low eddy current loss magnets.

2. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 1, characterized in that, The set of preparation parameters constructed based on the magnetic steel raw material ratio dataset, sintering temperature, magnetic field orientation strength, and cooling rate includes: Multiple raw material ratio nodes were extracted from the magnetic steel raw material ratio dataset. These raw material ratio nodes include: rare earth element content, iron element content, and additive content. Perform the following operation for each of the multiple raw material proportioning nodes: The raw material ratio nodes are verified using pre-constructed ratio constraint relationships to obtain verification results, where the verification result is either pass or fail. After confirming that the verification result is passed, the initial preparation parameter combination is obtained by associating the raw material ratio node, sintering temperature, magnetic field orientation strength and cooling rate. The initial preparation parameter combinations are summarized to obtain an initial preparation parameter combination set. The initial preparation parameter combination set is then screened using a pre-defined orthogonal experimental design method to obtain a final preparation parameter combination set.

3. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 2, characterized in that, The process of performing layered cutting on the initial magnet sample to obtain a set of layered magnet sheets includes: Obtain the sample thickness parameters of the initial magnet sample, and calculate the number of layers based on the preset layer thickness threshold and sample thickness parameters; The initial magnet sample is uniformly cut using the number of layers to obtain multiple cut magnet pieces; Perform the following operation on each of the multiple cutting magnets: The surface flatness detection unit is used to detect the cut magnetic steel sheet to obtain the surface flatness value. The surface flatness value is compared with the preset flatness threshold. If the surface flatness value is greater than or equal to the flatness threshold, the cut magnetic steel sheet is ground to obtain a ground magnetic steel sheet. The ground magnetic steel sheet is used as a layered magnetic steel sheet. Otherwise, using the cut magnetic steel sheets as layered magnetic steel sheets, the layered magnetic steel sheets are summarized to obtain a set of layered magnetic steel sheets.

4. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 3, characterized in that, The process of performing eddy current loss detection on each layered magnet in the set of layered magnet sheets yields a set of layered eddy current loss values, including: Obtain a detection frequency sequence for eddy current loss detection, wherein the detection frequency sequence includes multiple detection frequencies and covers the operating frequency range of the high-frequency drive motor. For each layered magnet in the assembly, perform the following operations: For each detection frequency in the detection frequency sequence, perform the following operation: The layered magnetic steel sheet is tested using a detection frequency and a pre-constructed eddy current loss detection device to obtain the frequency eddy current loss value. The frequency eddy current loss value is correlated with the detection frequency to obtain the frequency loss node. The frequency loss nodes are summarized to obtain a frequency loss node set. The weighted eddy current loss value is calculated based on the frequency loss node set. The weighted eddy current loss value is the weighted average of multiple frequency eddy current loss values ​​in the frequency loss node set based on the detection frequency. By summing the weighted eddy current loss values, a layered eddy current loss value set is obtained.

5. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 4, characterized in that, The method of using a detection frequency and a pre-constructed eddy current loss detection device to detect the layered magnetic steel sheet and obtain the frequency eddy current loss value includes: The resistivity and thickness parameters of the layered magnetic steel sheet are obtained, and an alternating magnetic field is applied to the layered magnetic steel sheet using an eddy current loss detection device to obtain the peak value of the magnetic induction intensity. Based on the resistivity parameter, sheet thickness parameter, detection frequency, and peak magnetic induction intensity, the frequency eddy current loss value is calculated using the following formula: in, This represents the frequency eddy current loss value. This indicates the thickness parameter of the sheet. This indicates the detection frequency. This indicates the peak value of the magnetic flux density. This represents the resistivity parameter. The preset value of pi. The preset permeability constant, This indicates the eddy current identifier. Indicates the peak value identifier. This represents the hyperbolic tangent function.

6. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 5, characterized in that, The process of identifying the selected magnet samples and sample loss characteristic nodes based on the hierarchical eddy current loss value set and the target eddy current loss threshold includes: Calculate the mean of the layered eddy current loss values ​​in the set of layered eddy current loss values ​​to obtain the mean of the sample eddy current loss. Calculate the variance of the layered eddy current loss values ​​in the set of layered eddy current loss values ​​to obtain the variance of the sample eddy current loss. The mean value of the sample eddy current loss is compared with the target eddy current loss threshold. If the mean value of the sample eddy current loss is less than or equal to the target eddy current loss threshold, the variance of the sample eddy current loss is compared with a preset variance threshold. If the variance of the sample eddy current loss is less than or equal to the variance threshold, the initial magnet sample is confirmed as the screening magnet sample. The mean of the sample eddy current loss and the variance of the sample eddy current loss are correlated to obtain the sample loss feature node. Otherwise, abnormal eddy current loss values ​​are identified in the layered eddy current loss value set, wherein the absolute difference between the abnormal eddy current loss value and the average eddy current loss value of the sample is greater than a preset deviation threshold. The abnormal eddy current loss value is removed from the layered eddy current loss value set to obtain an updated eddy current loss value set. The updated eddy current loss value set is used as the layered eddy current loss value set, and the step of calculating the average value of the layered eddy current loss values ​​in the layered eddy current loss value set is returned until the screened magnetic steel sample and the sample loss feature node are obtained.

7. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 6, characterized in that, The process of performing grain boundary optimization on the screened magnet samples to obtain optimized magnet samples includes: The microstructure images of the grains of the screened magnetic steel samples are obtained. The microstructure images of the grains are analyzed using a pre-built image analysis model to obtain the grain boundary feature nodes. The grain boundary feature nodes include: average grain size, boundary phase thickness and boundary phase distribution uniformity. Based on the grain boundary feature nodes and the preset boundary optimization parameter range, the boundary optimization processing parameter nodes are identified, including: diffusion annealing temperature, diffusion annealing time, and diffusion element concentration. The grain boundary diffusion process is performed on the screened magnet samples using the boundary optimization processing parameter nodes to obtain diffusion-treated magnet samples. The diffusion-treated magnet samples are then subjected to aging treatment to obtain optimized magnet samples.

8. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 7, characterized in that, The process of identifying boundary optimization processing parameter nodes based on the grain boundary feature nodes and the preset boundary optimization parameter range includes: Obtain the historical optimization dataset, which includes multiple historical optimization nodes, and each historical optimization node includes: historical grain boundary feature nodes, historical boundary optimization processing parameters, and historical optimization effect values; The average grain size, boundary phase thickness, and boundary phase distribution uniformity are extracted from the grain boundary feature nodes. A feature vector is constructed using the average grain size, boundary phase thickness, and boundary phase distribution uniformity. The feature distance value between the feature vector and the feature vector corresponding to each historical grain boundary feature node in the historical optimization dataset is calculated using a pre-constructed Euclidean distance calculation method, thus obtaining a feature distance value set. The feature distance values ​​in the feature distance value set are sorted in ascending order to obtain a feature distance value sequence. The target feature distance value set is then extracted from the feature distance value sequence using a preset extraction quantity. Based on the historical boundary optimization parameters and historical optimization effect values ​​corresponding to the target feature distance value set, the recommended boundary optimization parameters are calculated using the weighted average method. The weighting weights of the weighted average method are positively correlated with the historical optimization effect values ​​and negatively correlated with the feature distance values. After confirming that the recommended boundary optimization parameters are within the range of boundary optimization parameters, the recommended boundary optimization parameters are used as boundary optimization parameter nodes.

9. The method for preparing a low-eddy current loss magnet for a high-frequency drive motor as described in claim 8, characterized in that, The process of identifying the target magnet sample based on the magnetic property detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold includes: The magnetic energy product and coercivity are extracted from the magnetic performance detection node, and the sample eddy current loss mean and sample eddy current loss variance are extracted from the sample loss characteristic node. Using the measured magnetic energy product, coercivity, mean eddy current loss, variance of eddy current loss, target magnetic energy product threshold, and target coercivity threshold, the comprehensive magnetic performance evaluation value is calculated. The calculation formula is as follows: in, This represents the comprehensive magnetic performance evaluation value. , , All are preset weighting coefficients, and , This represents the detected value of the magnetic energy product. This represents the target magnetic energy product threshold. This represents the coercivity test value. This represents the target coercivity threshold. This represents the average eddy current loss of the sample. This represents the target eddy current loss threshold. The preset uniformity adjustment coefficient, This represents the variance of the eddy current loss in the sample. Represents the largest identifier. Indicates the threshold identifier. Indicates coercivity identifier, This represents the variance identifier. Indicates the mean identifier. Indicates the preset base; Once the comprehensive magnetic performance evaluation value is confirmed to be greater than or equal to the preset evaluation threshold, the optimized magnet sample is confirmed as the target magnet sample.

10. A system for manufacturing low-eddy current loss magnets for high-frequency drive motors, characterized in that, The system includes: The parameter node confirmation module is used to receive the magnet preparation instruction and confirm the preparation environment parameter node and the target magnet performance parameter node based on the magnet preparation instruction. The preparation environment parameter node includes: sintering temperature, magnetic field orientation intensity and cooling rate, and the target magnet performance parameter node includes: target eddy current loss threshold, target magnetic energy product threshold and target coercivity threshold. The sample preparation and cutting module is used to obtain a magnetic steel raw material ratio dataset. Based on the magnetic steel raw material ratio dataset, sintering temperature, magnetic field orientation intensity and cooling rate, a set of preparation parameter combinations is constructed. For each preparation parameter combination in the preparation parameter combination set, the following operations are performed: magnetic steel sample is prepared using the preparation parameter combination to obtain an initial magnetic steel sample. The initial magnetic steel sample is then subjected to layer cutting to obtain a set of layered magnetic steel sheets. The eddy current loss screening module is used to perform eddy current loss detection on each layered magnet sheet in the layered magnet sheet set to obtain a layered eddy current loss value set. Based on the layered eddy current loss value set and the target eddy current loss threshold, the screening magnet sample and the sample loss characteristic node are identified. The optimization evaluation and confirmation module is used to perform grain boundary optimization processing on the screened magnetic steel samples to obtain optimized magnetic steel samples. The optimized magnetic steel samples are then tested using a pre-constructed magnetic performance detection unit to obtain magnetic performance detection nodes. The magnetic performance detection nodes include: magnetic energy product detection value and coercivity detection value. Based on the magnetic performance detection node, sample loss characteristic node, target magnetic energy product threshold, and target coercivity threshold, the target magnet sample is identified, thus realizing the preparation of low eddy current loss magnets.

Images