A method and system for intelligent lamination of rotor and stator cores for electric motors

Abstract

This invention provides a method and system for intelligent lamination of rotor and stator cores for motors, relating to the field of rotor and stator processing technology. The method includes: cutting non-oriented silicon steel sheet rolls into rectangular blanks according to preset dimensions; punching rotor and stator slots into the rectangular blanks using a stamping die; dividing the slotted rotor and stator lamination blanks; enlarging the rotor slots of the rotor lamination assembly; pressing the stator laminations axially; assembling; detecting the magnetic properties of the integrated rotor and stator assembly; and performing visual inspection of the integrated rotor and stator assembly based on intelligent image recognition; and classifying and storing the qualified integrated rotor and stator assemblies. This invention achieves dual closed-loop dynamic adjustment through a closed-loop PID tension control system, ensuring that the non-oriented silicon steel sheets are always transported in a flat state, significantly improving raw material utilization and production efficiency; and identifies defect features using an image recognition system based on class activation maps and the Otsu method for adaptive threshold segmentation.

Description

Technical Field

[0001] This invention relates to the field of rotor and stator processing technology, and in particular to a method and system for intelligent lamination of rotor and stator cores for motors. Background Technology

[0002] As a core component of the drive motor in new energy vehicles, the stator assembly's manufacturing technology directly affects the motor's efficiency, power density, and the vehicle's range. The stator, as a key carrier for the conversion of electromagnetic energy in the motor, has undergone multiple generations of technological iterations in its manufacturing process, resulting in the current technological system characterized by high-precision machining, advanced material applications, and intelligent assembly.

[0003] Existing methods for laminating rotor and stator cores for motors often employ open-loop or single feedback control systems to regulate unwinding and rewinding speeds. These systems struggle to respond to tension changes in real time, leading to wrinkles, tensile deformation, or loosening of the silicon steel sheets during transport. This not only affects lamination accuracy but also wastes raw materials, increases equipment downtime for maintenance, and significantly reduces production efficiency. Furthermore, current visual inspection methods rely heavily on manual inspection combined with simple measuring tools, resulting in low inspection efficiency and significant subjective errors. In particular, the identification rate of minute defects is insufficient, potentially allowing defective products to flow into downstream processes and impacting the final motor's performance and reliability.

[0004] To address the shortcomings of the existing technology, this technical solution proposes a method and system for intelligent lamination of rotor and stator cores for motors. Summary of the Invention

[0005] This invention provides a method and system for intelligent lamination of rotor and stator cores for motors, which addresses the deficiencies in the prior art.

[0006] On one hand, the present invention provides a method for intelligent lamination of rotor and stator cores for electric motors, comprising:

[0007] Unroll the non-oriented silicon steel sheet roll, keep the strip flat during conveying through a tension control system, and cut it into rectangular blanks according to the preset size;

[0008] The rectangular blank is punched into the shape of the rotor and stator slots through the stamping die, and the slotted rotor and stator blank is output.

[0009] The slotted stator lamination blank is divided along a preset dividing line to output stator laminations and rotor lamination assemblies;

[0010] The rotor slots of the rotor lamination assembly are enlarged to produce enlarged rotor laminations; the stator laminations are pressed axially to produce a stacked stator core.

[0011] The stacked stator core is assembled with the rotor lamination assembly, and the magnetic circuit skeleton is formed by filling the material through die casting process, resulting in an integrated stator-rotor assembly.

[0012] The magnetic properties of the integrated rotor-stator assembly are tested. If the properties are qualified, the appearance of the integrated rotor-stator assembly is inspected based on image intelligent recognition. If the properties are unqualified, automatic sorting and scrapping are initiated.

[0013] The qualified integrated rotor-stator components are packaged and stored in the warehouse, and classified and stored according to batch labels.

[0014] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes the following steps for maintaining flat conveying of the strip material through a tension control system:

[0015] Real-time monitoring of the conveying tension of non-oriented silicon steel sheet rolls;

[0016] The rotational speeds of the unwinding and take-up rollers are adjusted using a dual closed-loop regulation method based on the dynamics of the conveying tension.

[0017] When the conveying tension is lower than the preset tension value, increase the speed of the unwinding roller and decrease the speed of the winding roller; when the conveying tension is higher than the preset tension value, decrease the speed of the unwinding roller and increase the speed of the winding roller.

[0018] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes the following steps for adjusting the rotational speeds of the unwinding and take-up rollers using a dual closed-loop regulation mechanism:

[0019] Set the preset tension value and inner and outer loop PID parameters, and initialize the drive current and given speed of the unwinding and take-up rollers;

[0020] The inner loop calculates the speed deviation between the unwinding and take-up rollers based on the given speed and the actual speed fed back by the encoder, and adjusts the motor drive current in real time through the speed PID controller to correct the speed deviation and output a stable actual speed.

[0021] The tension deviation is calculated by combining the actual tension fed back by the outer ring tension sensor with the preset tension value; based on the tension deviation, the given speed is updated by the tension PID controller, and the given speed of the outer ring is output.

[0022] The inner ring dynamically adjusts the speed of the unwinding and take-up rollers based on the given speed of the outer ring; the outer ring optimizes the given speed of the outer ring through real-time tension feedback.

[0023] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes updating a given rotational speed via a tension PID controller.

[0024] If the actual tension value is less than the preset tension value, increase the speed of the unwinding roller and decrease the speed of the winding roller.

[0025] If the actual tension value is greater than the preset tension value, reduce the speed of the unwinding roller and increase the speed of the winding roller.

[0026] According to the present invention, a method for intelligent lamination of rotor and stator core for electric motors includes the following steps: forming a magnetic circuit skeleton by filling material through die casting process, and outputting an integrated rotor and stator assembly:

[0027] The stacked stator core and rotor lamination assembly are assembled, and the rotor laminations are precisely embedded into the inner cavity of the stacked stator core and fixed to form a stator-rotor assembly.

[0028] The rotor-stator assembly is placed in a preheated die-casting mold, and molten magnetic material is injected into the mold cavity under high pressure using a die-casting machine; the die-casting mold is kept closed, and the molten magnetic material is cooled and solidified to form a magnetic circuit skeleton;

[0029] The magnetic circuit skeleton is demolded to obtain an integrated rotor-stator assembly.

[0030] According to the present invention, a method for intelligent lamination of rotor and stator core for electric motors includes the following steps for detecting the magnetic properties of an integrated rotor-stator assembly:

[0031] Set the magnetic property detection parameters;

[0032] Based on the magnetic performance detection parameters, measure whether the magnetic field distribution, remanence, and coercivity parameters of the integrated rotor-stator assembly meet the values ​​of the magnetic performance detection parameters;

[0033] Simulate actual working conditions and test the dynamic characteristics of the integrated rotor-stator assembly during rotation;

[0034] Detect whether there are magnetically anomalous regions in dynamic features and output the detection results;

[0035] Compare the test results with the magnetic performance test parameters to determine whether the test is qualified, and output the magnetic performance determination result.

[0036] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes the following steps for visual inspection of the integrated rotor-stator assembly based on image intelligent recognition:

[0037] Based on the magnetic performance determination results, image data of the integrated rotor-stator assembly that passed the determination results were collected;

[0038] The image data is preprocessed, and image features are extracted from the preprocessed image data.

[0039] A defect recognition model is established to identify defects in image features, and the defects are marked and output as defect labels.

[0040] The defect markers are classified, and the pass / fail determination is made based on the classification results. The pass / fail determination results are then output.

[0041] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes the following steps for establishing a defect recognition model to identify defects in image features:

[0042] Defect images from the historical defect database are input into a CNN model for recognition. After passing through a global average pooling layer, the defect feature vector is output and connected to the output layer of the CNN model to generate a class activation map.

[0043] Sampling is performed on the class activation map to obtain the defect heatmap, and the Otsu method is used for adaptive threshold segmentation to output the defect region image;

[0044] Based on morphological operations, the defect area image is denoised and dilated to restore the normal area image of the stator-rotor integrated component, and the morphologically processed image is output.

[0045] Connectivity labeling techniques are used to analyze morphologically processed images and calculate the pixel intensity and related features of each connected region.

[0046] By comparing the relevant feature intensities of different connected regions, the region with the highest pixel intensity is identified as a discriminative region that may have defects or needs special attention.

[0047] According to the present invention, a method for intelligent lamination of rotor and stator cores for motors includes the following steps for denoising and dilating defect area images based on morphological operations:

[0048] The structuring element is used to perform an erosion operation on the defect area image to remove small noise points and abrupt parts, and the eroded image is output.

[0049] Dilation is performed on the eroded image to restore the shape of the normal parts that were eroded away.

[0050] A smart lamination system for rotor and stator cores of an electric motor, provided by the present invention, includes:

[0051] The raw material conveying module is used to unwind non-oriented silicon steel sheet rolls and maintain the flat conveying of the strip through a tension control system;

[0052] The raw material cutting unit is used to cut non-oriented silicon steel sheet rolls into rectangular blanks according to preset dimensions;

[0053] The stamping module is used to punch out the rotor and stator slots on the rectangular blank through the stamping die, and output the slotted rotor and stator lamination blank; and to divide the slotted rotor and stator lamination blank along the preset dividing line to output the stator lamination and rotor lamination assembly.

[0054] The rotor lamination processing module is used to enlarge the rotor slots of the rotor lamination assembly and output expanded rotor laminations.

[0055] The stator core stacking module is used to press stator laminations axially to output stacked stator cores.

[0056] The assembly module is used to assemble the stacked stator core and the rotor lamination assembly, and to fill the material through die casting process to form a magnetic circuit skeleton, and output an integrated stator-rotor assembly.

[0057] The defect detection module is used to detect the magnetic properties of the rotor-stator integrated assembly, determine whether it is qualified, and perform appearance inspection on the qualified rotor-stator integrated assembly based on image intelligent recognition.

[0058] The packaging and storage module is used to package the qualified integrated rotor-stator components and classify and store the packaged integrated rotor-stator components according to batch labels.

[0059] This invention provides a method and system for intelligent lamination of rotor and stator cores for motors. Through a closed-loop PID tension control system, it achieves dual closed-loop dynamic adjustment, ensuring that the non-oriented silicon steel sheets remain flat during transport. This effectively avoids problems such as wrinkles and tensile deformation caused by tension fluctuations in traditional processes, significantly improving raw material utilization and production efficiency. An innovative design using die casting to directly form the magnetic circuit skeleton replaces the traditional lamination and subsequent bonding processes, achieving high-strength structural integration of the rotor and stator assembly, effectively reducing air gap loss and improving the overall energy efficiency of the motor. A multi-layer defect recognition model is constructed using an image recognition system based on class activation maps and the Otsu method for adaptive threshold segmentation to identify defect features. This model can identify minute defects, reducing the false negative and false positive rates. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0061] Figure 1 This is a flowchart of a method for intelligent lamination of rotor and stator cores for motors provided in an embodiment of the present invention;

[0062] Figure 2 This is a schematic diagram of a smart lamination system for rotor and stator cores of an electric motor provided in an embodiment of the present invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0064] Example 1:

[0065] The following is combined Figures 1-2 This invention describes a method and system for intelligent lamination of rotor and stator cores for electric motors.

[0066] like Figure 1 As shown in the figure, an embodiment of the present invention provides a method for intelligent lamination of rotor and stator cores for motors, comprising:

[0067] First, the non-oriented silicon steel sheet roll is unrolled and conveyed flat using a tension control system, then cut into rectangular blanks according to a preset size. During the cutting process, a high-precision cutting device driven by a servo motor cuts along a preset cutting line. The steps for conveying the strip flat using the tension control system include:

[0068] The system monitors the conveying tension of non-oriented silicon steel sheet rolls in real time. It integrates high-precision tension sensors installed at key locations along the material conveying path to detect the tension in the rolls. The sensors transmit the collected tension signals to the controller for real-time processing and analysis.

[0069] The rotational speeds of the unwinding and rewinding rollers are adjusted using a dual closed-loop regulation method based on the dynamics of the conveying tension.

[0070] When the conveying tension is lower than the preset tension value, increase the speed of the unwinding roller and decrease the speed of the take-up roller. When the conveying tension is higher than the preset tension value, decrease the speed of the unwinding roller and increase the speed of the take-up roller. The steps for adjusting the speed of the unwinding roller and take-up roller using a dual closed-loop control method include:

[0071] Set the preset tension value and inner and outer loop PID parameters, and initialize the drive current and given speed of the unwinding and take-up rollers.

[0072] The inner loop calculates the speed deviation between the unwinding and take-up rollers based on the given rotational speed and the actual rotational speed fed back from the encoder. It then uses a speed PID controller to adjust the motor drive current in real time to correct the speed deviation and output a stable actual rotational speed. The formula for PID control is as follows:

[0073]

[0074] In the formula, ΔI is the current regulation amount, e is the speed deviation, and Kp K is the proportional coefficient of the speed PID controller. a K is the integral coefficient of the speed PID controller. d For the speed PID, where is the differential coefficient. For the differential control term of the speed PID controller, This represents the derivative of the rotational speed deviation e with respect to time t.

[0075] The tension deviation is calculated by combining the actual tension fed back by the outer ring tension sensor with the preset tension value. Based on the tension deviation, the given speed is updated by the tension PID controller, and the outer ring given speed is output. The calculation method of tension deviation is as follows:

[0076] ΔQ=Q act -Q pre

[0077] In the formula, ΔQ is the tension deviation, and Q act For the actual tension, Q pre This is the preset tension value.

[0078] Methods for updating the given rotational speed using a tension PID controller include:

[0079] If the actual tension value is less than the preset tension value, increase the unwinding roller speed and decrease the take-up roller speed to increase the material tension. The formula is as follows:

[0080]

[0081] In the formula, ΔN um ΔN is the adjustment amount for the unwinding roll speed. re K p1 K is the proportional coefficient of the tension PID of the unwinding roll. a1 K is the integral coefficient of the PID controller for the tension of the unwinding roll. d1 The differential coefficient of the unwinding roll tension PID is given. The derivative control term of the PID controller for the tension of the unwinding roll. K represents the derivative of tension deviation ΔQ with time t. p2 K is the proportional coefficient of the tension PID of the take-up roll. a2 K is the integral coefficient of the tension PID of the take-up roll. d2 The differential coefficient of the tension PID of the take-up roll is given. This is the derivative control term of the PID controller for the tension of the take-up roller.

[0082] If the actual tension value is greater than the preset tension value, reduce the speed of the unwinding roller and increase the speed of the take-up roller to reduce the material tension. The formula is as follows:

[0083]

[0084] The calculation method is the opposite of the method used when the actual tension value is less than the preset tension value.

[0085] The inner ring dynamically adjusts the speeds of the unwinding and take-up rollers based on the given speed of the outer ring. The outer ring optimizes the given speed through real-time tension feedback. An adaptive control strategy is employed, automatically adjusting control parameters based on material properties, equipment load, and environmental conditions. To ensure system stability, this solution also includes a safety protection mechanism.

[0086] When the tension deviation exceeds the set threshold, the system automatically reduces the equipment speed and issues an alarm. A mechanical limit device is designed to prevent equipment damage caused by material breakage or excessive tension. A dual-motor collaborative control strategy is implemented to ensure the synchronization of the unwinding and rewinding processes. A predictive control algorithm is used to predict tension change trends 30ms in advance, enabling proactive adjustment.

[0087] Next, the rectangular blank is punched with stator slots using a stamping die, resulting in slotted stator lamination blanks. After stamping, the blanks are cleaned with compressed air to remove surface iron filings and oil. Then, a laser micrometer is used to check the slot dimensions to ensure that the dimensional accuracy meets the requirements.

[0088] Furthermore, the slotted stator lamination blanks are divided along preset dividing lines to output stator laminations and rotor lamination assemblies. During the dividing process, a vision inspection system monitors the dividing quality in real time to ensure that the dividing accuracy is within ±0.05mm. After dividing, the system automatically classifies and collects the stator and rotor laminations to avoid mixing.

[0089] Furthermore, the rotor slots of the rotor lamination assembly are enlarged to produce enlarged rotor laminations. The stator laminations are then axially pressed together to produce a laminated stator core. Pressure changes are monitored in real time during the lamination process, and automatic compensation is implemented when the pressure falls below a set value to ensure lamination quality. After completion, magnetic non-destructive testing equipment is used to inspect the laminated stator for internal defects.

[0090] Furthermore, the laminated stator core is assembled with the rotor lamination assembly, and a magnetic circuit frame is formed by filling material through die casting, resulting in an integrated stator-rotor assembly. The steps of forming the magnetic circuit frame through die casting and producing the integrated stator-rotor assembly include:

[0091] The stacked stator core and rotor laminations are assembled, and the rotor laminations are precisely embedded into the inner cavity of the stacked stator core and fixed to form a stator-rotor assembly.

[0092] The stator-rotor assembly is placed in a preheated die-casting mold, and molten magnetic material is injected into the mold cavity under high pressure using a die-casting machine. The die-casting mold is kept closed, and the molten magnetic material cools and solidifies to form the magnetic circuit framework.

[0093] The magnetic circuit skeleton is demolded to obtain an integrated rotor-stator assembly.

[0094] Furthermore, the magnetic properties of the integrated rotor-stator assembly are tested. If they pass, a visual inspection of the integrated rotor-stator assembly is performed based on image intelligent recognition. If they fail, automatic sorting and scrapping are initiated. The steps for testing the magnetic properties of the integrated rotor-stator assembly include:

[0095] The system sets the magnetic performance detection parameters. Based on preset process standards (such as magnetic flux density range and coercivity threshold), the system initializes the detection parameters and simultaneously initiates multi-sensor collaborative calibration, including a gaussmeter and a teslameter, to ensure measurement accuracy reaches ±0.5%. Upon completion of calibration, a calibration report (including error matrix analysis) is automatically generated and stored in the MES system for traceability.

[0096] Static magnetic field characteristic detection is performed. Specifically, this includes: measuring the magnetic field distribution of the integrated rotor-stator assembly based on magnetic performance detection parameters; specifically, using a Hall effect probe array to perform three-dimensional gradient sampling (sampling interval ≤ 1 mm) of the magnetic pole gap and air gap magnetic field of the integrated rotor-stator assembly, generating a BH curve spectrum; measuring whether the remanence and coercivity parameters conform to the values ​​of the magnetic performance detection parameters; acquiring demagnetization curves within a magnetic field strength range of 0-1.5T using the pulsed magnetic field excitation method, and calculating the coercivity error rate (must be ≤ 8% of the nominal value); and identifying abnormal features such as burrs and offsets in the hysteresis loop based on a time-frequency domain analysis algorithm, and setting a dynamic threshold alarm mechanism.

[0097] Simulating actual working conditions, the dynamic characteristics of the integrated rotor-stator assembly during rotation were tested. The closed-loop control system for simulating actual working conditions included: a high-precision servo motor drive (adjustable speed range: 0-3000 rpm); vibration sensors (sampling rate ≥10 kHz) for real-time monitoring of rotor operation; and thermocouples for monitoring temperature field changes (temperature accuracy ±1℃).

[0098] The system detects the presence of magnetic anomaly regions within dynamic features and outputs the detection results. It automatically identifies stress waveform characteristics caused by dynamic hysteresis and expansion. If a local magnetic anomaly region (area > Φ5mm) is detected, the system will output the result. 2 If the depth is greater than 3 layers of coils, the audible and visual alarm device is activated and the detection data is locked. A comprehensive detection report is then generated, including a magnetic flux density distribution cloud map, loss curve, and temperature rise curve.

[0099] Compare the test results with the magnetic performance test parameters to determine whether the test is qualified, and output the magnetic performance determination result.

[0100] The steps for visual inspection of the integrated rotor-stator assembly based on image intelligent recognition include:

[0101] Based on the magnetic performance assessment results, image data of the integrated rotor-stator assembly that has passed the assessment are acquired. Typically, a four-axis robotic arm equipped with a high-resolution industrial camera is used, with a ring-shaped LED light source array to achieve shadowless illumination. It automatically switches between seven inspection angles, and performs camera calibration after each angle switch.

[0102] Image data is preprocessed, and image features are extracted from the preprocessed image data. This includes: median filtering (kernel size 5×5) combined with Wiener filtering for noise reduction; illumination compensation based on Lambert's cosine law; CLAHE algorithm to improve the contrast between defect areas and background; and SIFT feature point matching to achieve multi-scale feature extraction.

[0103] A defect recognition model is established to identify defects in image features, label the defects, and output the defect labels. The steps for establishing a defect recognition model to identify defects in image features include:

[0104] Defect images from a historical defect database are input into a CNN model for identification. After passing through a global average pooling layer, the output defect feature vector is generated and connected to the output layer of the CNN model to generate a class activation map (CAM). The class activation map highlights areas in the stator-rotor integrated component that may contain defects or have important features, such as winding anomalies or core damage. These areas will be the focus of further analysis and processing. CAM is a visualization technique used to highlight regions where the classification network identifies target categories; it is generated by the last convolutional layer of the CNN. A training image is input into the model, passes through a global average pooling (GAP) layer to obtain a feature vector, and is then connected to the model's output layer to generate the CAM. The formula for calculating CAM is as follows:

[0105]

[0106] In the formula, c represents the target category of the defect image o, and fm k ∈R H×W Let S represent the k-th feature map. There are a total of N feature maps. k,c This represents the weights of the fully connected layer.

[0107] Sampling is performed on the class activation map to obtain a defect heatmap, and the Otsu method is used for adaptive thresholding segmentation to output the defect region image. The specific processing procedure and calculation formula of the Otsu method are as follows:

[0108] s1. Calculate the foreground weight W1(T) and background weight W2(T) based on the frequency P(i) of each gray level. The formula is as follows:

[0109]

[0110] W2(T) = 1 - W1(T)

[0111] In the formula, i represents the gray level and T represents the current threshold.

[0112] s2. Calculate the foreground mean, background mean, and between-class variance. The formula is as follows:

[0113]

[0114] Where μ1(T) is the foreground mean, μ2(T) is the background mean, and T* represents the search for the value of T that satisfies the variance between groups within all possible values ​​of T. Find the maximum value of T. This refers to the between-class variance. Between-class variance is a statistic that measures the degree of separation between different classes. A larger between-class variance indicates a more significant difference between the classes, and potentially better segmentation. By adjusting the value of T, the between-class variance can be maximized, thus finding the optimal segmentation threshold.

[0115] s3. Based on morphological operations, denoise and dilate the defect area image to restore the normal area image of the integrated rotor-stator assembly, and output the morphologically processed image. The steps for denoising and dilating the defect area image based on morphological operations include:

[0116] The structuring element is used to perform an erosion operation on the defect area image to remove small noise points and abrupt parts, and the eroded image is output.

[0117] The image after erosion is dilated to restore the shape of the normal parts that were eroded, ensuring the integrity of the outline of the integrated rotor-stator assembly. At the same time, noise interference in the image is effectively smoothed, and the main feature information of the assembly's appearance is preserved.

[0118] s4. Using connectivity labeling techniques, analyze the morphologically processed image to calculate the pixel intensity and related features of each connected region.

[0119] s5. By comparing the relevant feature intensities of different connected regions, the region with the highest pixel intensity is identified as a discriminative region that may have defects or needs special attention. These discriminative regions will be used as the objects for further detailed inspection and evaluation, such as size measurement and defect type classification, in order to accurately determine the appearance quality of the stator-rotor integrated assembly.

[0120] The defect markers are classified, and the pass / fail determination is made based on the classification results. The pass / fail determination results are then output.

[0121] Finally, the qualified integrated rotor-stator components are packaged and stored in the warehouse, and classified and stored according to batch labels.

[0122] In summary, this invention provides an intelligent lamination method for rotor and stator cores in motors. Through a closed-loop PID tension control system, it achieves dual closed-loop dynamic adjustment, ensuring that the non-oriented silicon steel sheets remain flat during transport. This effectively avoids problems such as wrinkles and tensile deformation caused by tension fluctuations in traditional processes, significantly improving raw material utilization and production efficiency. The innovative design of directly forming the magnetic circuit skeleton using die casting replaces the traditional lamination and subsequent bonding processes, achieving high-strength structural integration of the rotor-stator assembly, effectively reducing air gap loss and improving the overall energy efficiency of the motor. A multi-layer defect recognition model is constructed using an image recognition system based on class activation maps and the Otsu method for adaptive threshold segmentation to identify defect features. This can identify minute defects, reducing the false negative and false positive rates.

[0123] The present invention also provides an intelligent lamination system for rotor and stator cores of motors, comprising: a raw material conveying module, a raw material cutting unit, a stamping and forming module, a rotor lamination processing module, a stator core stacking module, an assembly module, a defect detection module, and a packaging and warehousing module.

[0124] The raw material conveying module is used to unwind non-oriented silicon steel sheet rolls and maintain the flat conveying of the strip through a tension control system.

[0125] The raw material cutting unit is used to cut non-oriented silicon steel sheet rolls into rectangular blanks according to preset dimensions.

[0126] The stamping module is used to punch out rotor and stator slots on rectangular blanks using a stamping die, outputting slotted rotor and stator lamination blanks. It then divides the slotted rotor and stator lamination blanks along preset dividing lines, outputting stator laminations and rotor lamination assemblies.

[0127] The rotor lamination processing module is used to enlarge the rotor slots of the rotor lamination assembly and output expanded rotor laminations.

[0128] The stator core stacking module is used to press stator laminations axially and output stacked stator cores.

[0129] The assembly module is used to assemble the stacked stator core and rotor laminations, and to fill the material through die casting to form a magnetic circuit skeleton, outputting an integrated stator-rotor assembly.

[0130] The defect detection module is used to detect the magnetic properties of the rotor-stator integrated assembly, determine whether it is qualified, and perform appearance inspection on the qualified rotor-stator integrated assembly based on image intelligent recognition.

[0131] The packaging and warehousing module is used to package the qualified integrated rotor-stator components and classify and store the packaged integrated rotor-stator components according to batch labels.

[0132] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent lamination of rotor and stator cores for electric motors, characterized in that, include: Unroll the non-oriented silicon steel sheet roll, keep the strip flat during conveying through a tension control system, and cut it into rectangular blanks according to the preset size; The rectangular blank is punched into a rotor-stator slot shape through a stamping die, and a slotted rotor-stator lamination blank is output. The slotted stator lamination blank is divided along a preset dividing line to output stator laminations and rotor lamination assemblies; The rotor slots of the rotor lamination assembly are enlarged to produce enlarged rotor laminations; the stator laminations are pressed axially to produce a stacked stator core. The stacked stator core is assembled with the rotor lamination assembly, and a magnetic circuit skeleton is formed by filling material through die casting process, resulting in an integrated stator-rotor assembly. The magnetic properties of the integrated rotor-stator assembly are tested. If the properties are qualified, the appearance of the integrated rotor-stator assembly is inspected based on image intelligent recognition. If the product fails to meet the requirements, the automatic sorting and scrapping process will be initiated. The qualified integrated rotor-stator components are packaged and stored in the warehouse, and classified and stored according to batch labels.

2. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 1, characterized in that, The steps for maintaining a flat conveyor belt using a tension control system include: Real-time monitoring of the conveying tension of the non-oriented silicon steel sheet roll; Based on the conveying tension, the rotational speeds of the unwinding and take-up rollers are adjusted using a dual closed-loop regulation method. When the conveying tension is lower than the preset tension value, the speed of the unwinding roller is increased and the speed of the take-up roller is decreased; when the conveying tension is higher than the preset tension value, the speed of the unwinding roller is decreased and the speed of the take-up roller is increased.

3. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 2, characterized in that, The steps for adjusting the speed of the unwinding roll and the take-up roll using a dual closed-loop control method include: Set the preset tension value and inner and outer loop PID parameters, and initialize the drive current and given speed of the unwinding and take-up rollers; The inner loop calculates the speed deviation between the unwinding roller and the winding roller based on the given speed and the actual speed fed back by the encoder, and adjusts the motor drive current in real time through the speed PID controller to correct the speed deviation and output a stable actual speed. The tension deviation is calculated by combining the actual tension fed back by the outer ring tension sensor and the preset tension value; based on the tension deviation, the given rotational speed is updated by the tension PID controller, and the outer ring given rotational speed is output. The inner ring dynamically adjusts the rotational speeds of the unwinding and take-up rollers based on the given rotational speed of the outer ring; the outer ring optimizes the given rotational speed of the outer ring through real-time tension feedback.

4. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 3, characterized in that, The methods for updating the given rotational speed using a tension PID controller include: If the actual tension value is less than the preset tension value, increase the speed of the unwinding roller and decrease the speed of the winding roller. If the actual tension value is greater than the preset tension value, reduce the speed of the unwinding roller and increase the speed of the winding roller.

5. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 1, characterized in that, The steps involved in forming a magnetic circuit framework by filling material through die casting and outputting an integrated rotor-stator assembly include: The stacked stator core is assembled with the rotor lamination assembly, and the rotor laminations are precisely embedded into the inner cavity of the stacked stator core and fixed to form a stator-rotor assembly. The rotor-stator assembly is placed in a preheated die-casting mold, and molten magnetic material is injected into the cavity of the die-casting mold under high pressure using a die-casting machine; the die-casting mold is kept closed, and the molten magnetic material is cooled and solidified to form the magnetic circuit skeleton; The magnetic circuit skeleton is demolded to obtain the integrated rotor-stator assembly.

6. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 1, characterized in that, The steps for detecting the magnetic properties of the integrated rotor-stator assembly include: Set the magnetic property detection parameters; Based on the magnetic performance detection parameters, measure whether the magnetic field distribution, remanence, and coercivity parameters of the integrated rotor-stator assembly conform to the values ​​of the magnetic performance detection parameters; Simulate actual working conditions and test the dynamic characteristics of the integrated rotor-stator assembly during rotation; Detect whether there are magnetically anomalous regions in the dynamic features and output the detection results; The test results are compared with the magnetic performance test parameters to determine whether the test is qualified, and the magnetic performance determination result is output.

7. A method for intelligent lamination of rotor and stator cores for motors according to claim 6, characterized in that, The steps for performing appearance inspection on the integrated rotor-stator assembly based on image intelligent recognition include: Based on the magnetic performance determination result, image data of the integrated rotor-stator assembly that meets the determination result is collected; The image data is preprocessed, and image features are extracted from the preprocessed image data; A defect recognition model is established to identify defects in the image features, and the defects are marked and output as defect labels. The defect markers are classified, and a pass / fail determination is made based on the classification results, and the pass / fail determination result is output.

8. The intelligent lamination method for rotor and stator cores of an electric motor according to claim 7, characterized in that, The steps for establishing a defect recognition model to identify defects in the image features include: Defect images from the historical defect database are input into a CNN model for identification. After passing through a global average pooling layer, the defect feature vector is output and connected to the output layer of the CNN model to generate a class activation map. Sampling is performed on the class activation map to obtain a defect heatmap, and the Otsu method is used for adaptive threshold segmentation to output a defect region image; Based on morphological operations, the defective region image is denoised and dilated to restore the normal region image of the integrated rotor-stator component, and the morphologically processed image is output. The morphologically processed image is analyzed using connectivity labeling techniques to calculate the pixel intensity and related features of each connected region. By comparing the intensity of the relevant features in different connected regions, the region with the highest pixel intensity is identified as a discriminative region that may have defects or requires special attention.

9. A method for intelligent lamination of rotor and stator cores for motors according to claim 8, characterized in that, The steps for denoising and dilating the defect region image based on morphological operations include: The image of the defective region is eroded using a structuring element to remove small noise points and abrupt parts, and the eroded image is output. The eroded image is then subjected to a dilation operation to restore the shape of the normal parts that were eroded away.

10. A smart lamination system for rotor and stator cores of electric motors, comprising a smart lamination method for rotor and stator cores of electric motors as described in any one of claims 1 to 9, characterized in that, include: The raw material conveying module is used to unwind non-oriented silicon steel sheet rolls and maintain the flat conveying of the strip through a tension control system; The raw material cutting unit is used to cut the non-oriented silicon steel sheet roll into rectangular blanks according to a preset size; The stamping forming module is used to punch out the rotor-stator slot shape on the rectangular blank through the stamping die, and output the slotted rotor-stator lamination blank; and to divide the slotted rotor-stator lamination blank along the preset dividing line to output stator lamination and rotor lamination assembly; The rotor lamination processing module is used to enlarge the rotor slots of the rotor lamination assembly and output enlarged rotor laminations. The stator core stacking module is used to press the stator laminations axially and output the stacked stator core. The assembly module is used to assemble the stacked stator core with the rotor lamination assembly, and to fill the material through die casting process to form a magnetic circuit skeleton, and output an integrated rotor-stator assembly; The defect detection module is used to detect the magnetic properties of the integrated rotor-stator assembly, determine whether it is qualified, and perform appearance inspection on the qualified integrated rotor-stator assembly based on image intelligent recognition. The packaging and storage module is used to package the qualified integrated rotor-stator components and classify and store the packaged integrated rotor-stator components according to batch labels.

Images