A motor stator coil insulation layer spraying device

Abstract

The application provides a motor stator coil insulation layer spraying device and relates to the field of spraying.The motor stator coil insulation layer spraying device comprises a processing table, a material box, a discharge box, a conveying belt, a cleaning assembly, a blowing assembly, a spraying assembly, a drying assembly, a feeding assembly, a discharging assembly and a control element electrically connected with all the assemblies; the control element is an embedded intelligent controller, internally integrated with a main control module, a multi-sensor fusion sensing module, a process parameter self-adaptive adjustment module and a collaborative control module. The application realizes closed-loop management and control of the whole process of stator coil insulation spraying through multi-source fusion sensing, can self-adaptively optimize the spraying process, accurately collaborates with the actions of each station, effectively solves the industry pain points of uneven spraying of the special-shaped winding of the stator coil and poor film thickness consistency, and greatly improves the insulation spraying qualification rate, production efficiency and the intelligent level of the equipment.

Description

Technical Field

[0001] This invention relates to the field of spraying technology, specifically to a spraying device for the insulation layer of motor stator coils. Background Technology

[0002] Insulation layer spraying, as the core process of stator coil insulation treatment, has become a key link restricting the stable mass production of high-end motors.

[0003] Currently, most mainstream motor stator coil insulation coating equipment in the industry is semi-automatic or simple fully automatic. Their control systems mostly adopt traditional PLC timing control modes, which can only achieve basic start / stop and fixed timing linkage of each station: feeding, cleaning, coating, drying, and unloading. In actual production applications, the following areas need improvement:

[0004] First, the sensing system is too simplistic to achieve precise, closed-loop sensing throughout the entire process. Most existing equipment is only equipped with basic photoelectric sensors for workpiece positioning. Some high-end equipment adds planar vision sensors and single-point film thickness sensors, which can only detect the presence or absence of workpieces, identify the cleanliness of the plane, and detect the overall average film thickness. However, it cannot obtain the complete three-dimensional contour features of the irregular three-dimensional structure at the end of the stator coil winding, making it difficult to identify corners, gaps, and other areas prone to coating blind spots, thus failing to provide accurate spatial data support for coating control.

[0005] Secondly, the process adjustment capability is insufficient, and dynamic adaptive optimization and iteration cannot be achieved. The spraying process parameters of existing equipment all rely on manual trial spraying and debugging in advance and are stored in a static process database. During production, only fixed parameters can be called up, and only simple linear compensation of environmental temperature and humidity can be achieved. It is not possible to dynamically and adaptively adjust the process parameters according to the three-dimensional irregular structure of the stator coil, the real-time viscosity change of the paint, and the individual differences of the workpiece, and it does not have the ability to learn and iterate the process.

[0006] Third, the collaborative control precision is low, and the quality closed-loop handling mechanism is crude. The existing equipment can only achieve basic time-series linkage between each station, and cannot achieve high-precision synchronous control of the workpiece position and the actions of each station. This easily leads to problems such as workpiece misalignment causing spraying position deviation and incomplete cleaning. At the same time, the spraying station and the drying station are independent of each other, and cannot adaptively adjust the drying process parameters according to the thickness distribution of the insulation layer after spraying. This easily leads to defects such as incomplete curing of thick coating areas and over-drying and aging of thin coating areas, affecting the performance of the insulation layer.

[0007] Fourth, the equipment operation and maintenance capabilities are lagging behind, and there is a lack of full life cycle quality traceability capabilities. Existing equipment only has basic fault code display and post-event alarm functions, and cannot monitor the performance degradation trend of core components in real time. It does not have the ability to predict faults and perform predictive maintenance, resulting in a high rate of unplanned equipment downtime. Moreover, after a fault occurs, it can only output fault codes and cannot complete the root cause location and handling guidance, resulting in a long fault repair cycle.

[0008] In summary, the existing control technology of motor stator coil insulation layer spraying equipment can no longer meet the high-quality, high-efficiency, and intelligent production needs of the high-end motor industry. There is an urgent need to develop a dedicated spraying device with full-dimensional closed-loop perception, adaptive process optimization, full-link precise collaboration, intelligent operation and maintenance, and full-process traceability capabilities to solve many industry pain points of the existing technology. Summary of the Invention

[0009] To achieve the above objectives, the present invention provides the following technical solution: a motor stator coil insulation layer spraying device, comprising a processing table, a discharge box on the right side of the processing table, a material box installed at the front end of the processing table, a conveyor belt on the top of the processing table, a cleaning component on the top of the front end of the processing table, a blowing component on the left side of the cleaning component, a spraying component on the left side of the processing table, a control component on the top of the right side of the processing table, a feeding component at one end of the control component, and a feeding component installed on the top of the material box;

[0010] A drying assembly is also provided on the top of the processing table. The drying assembly is located between the spraying assembly and the unloading assembly and is electrically connected to the control component.

[0011] The control component is an embedded intelligent controller, which is electrically connected to the feeding assembly, conveyor belt, cleaning assembly, blowing assembly, spraying assembly, drying assembly and unloading assembly respectively;

[0012] The control unit integrates a main control module, a multi-sensor fusion sensing module, a process parameter adaptive adjustment module, and a collaborative control module.

[0013] This invention provides a device for spraying insulation layer on motor stator coils. It has the following advantages:

[0014] 1. This invention utilizes 3D laser contour scanning and segmented adaptive spraying to stably control the uniformity of the insulation layer film thickness within ±5μm, far exceeding the industry standard. It also incorporates multi-dimensional process compensation based on paint viscosity and ambient temperature and humidity, eliminating film-forming defects such as sagging, pinholes, and insufficient adhesion. Furthermore, it pioneers a spraying-drying linkage control system, resolving issues of incomplete curing in thick coating areas and over-drying and aging in thin coating areas. This improves the single-batch product coating qualification rate, enhances the electrical insulation performance and aging resistance of the insulation layer, and reduces the risk of early motor failure from the source.

[0015] 2. This invention achieves unmanned, precise, and collaborative operation of the entire stator coil insulation spraying process. Based on a high-precision position encoder, it enables synchronous control of workpiece position and all station movements at the ±0.1mm level, improving single-line production efficiency compared to conventional equipment. The built-in digital twin model can generate spraying process packages with one click, shortening the equipment changeover and debugging cycle. It perfectly adapts to the flexible production needs of multiple varieties and small batches, significantly reducing reliance on human experience. The optimized closed-loop handling logic for quality anomalies enables targeted secondary cleaning and precise local respraying, replacing the traditional whole-workpiece rework mode and significantly reducing rework rate and insulation varnish raw material loss. The accompanying process self-learning iteration unit can continuously optimize the process library based on production data, continuously improving batch production stability.

[0016] 3. This invention assigns a unique identifier to each stator coil, binding the entire process data and quality inspection data of a single workpiece to the identifier one by one, generating a complete lifecycle traceability file. This fully meets the quality traceability compliance requirements for core motor components in high-end fields such as new energy vehicles, rail transit, and aerospace. It constructs a cloud-edge collaborative management and control system, supporting the batch distribution of standardized process parameter packages from the cloud, ensuring the consistency of production processes across multiple devices and production lines. At the same time, the optimized process parameters at the edge can be synchronized to the entire production line after review, achieving simultaneous improvement of the process capabilities of the entire production line. It supports remote real-time monitoring of equipment operation data and remote operation and maintenance assistance, significantly reducing the management costs and operation and maintenance response cycles of cross-regional production.

[0017] 4. This invention breaks through the limitations of traditional equipment alarms and passive maintenance. Based on edge computing, the fault prediction and health management module can identify the performance degradation trend of core equipment components in real time and provide early warnings, reducing the rate of unplanned equipment downtime. When equipment malfunctions, it can intelligently locate the root cause of the fault based on multi-source data and simultaneously push standardized handling procedures, shortening the fault repair time. The built-in equipment maintenance ledger management unit automatically records maintenance information and generates maintenance reminders, realizing closed-loop management of the health status of the equipment throughout its entire life cycle and effectively extending the overall service life of the equipment. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall structure of the present invention;

[0019] Figure 2 This is a schematic diagram of the spraying assembly of the present invention;

[0020] Figure 3 This is a schematic diagram of the feeding assembly of the present invention;

[0021] Figure 4 This is a schematic diagram of the overall structure and material flow of the device of the present invention;

[0022] Figure 5 This is a schematic diagram of the workflow of the collaborative control module of the present invention;

[0023] The components include: 1. Processing table; 2. Material box; 3. Discharge box; 4. Control components; 5. Cleaning components; 6. Spraying components; 7. Blowing components; 8. Conveyor belt; 9. Drying components; 10. Feeding components; and 11. Discharging components. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] like Figures 1-4 As shown in the figure, an embodiment of the present invention provides a motor stator coil insulation layer spraying device, including a processing table 1, a discharge box 3 arranged on the right side of the processing table 1, a material box 2 installed at the front end of the processing table 1, a conveyor belt 8 arranged on the top of the processing table 1, a cleaning component 5 arranged on the top front end of the processing table 1, a blowing component 7 arranged on the left side of the cleaning component 5, a spraying component 6 arranged on the left side of the processing table 1, a control component 4 arranged on the top right side of the processing table 1, a feeding component 11 arranged at one end of the control component 4, and a feeding component 10 installed on the top of the material box 2.

[0026] A discharge box 3 is provided on the right side of the processing table 1, which is used to collect the stator coils after the spraying process is completed; a material box 2 is installed at the front end of the processing table 1, which is used to store the stator coils to be processed; a conveyor belt 8 is provided on the top of the processing table 1, which extends along the length of the processing table 1 and is used to transport the stator coils to each processing station in sequence; a cleaning component 5 is provided on the top front end of the processing table 1, which is located above the conveyor belt 8 and is used to clean the surface of the stator coils to remove impurities; a blowing component 7 is provided on the left side of the cleaning component 5, which is arranged adjacent to the cleaning component 5 and is used to blow on the surface of the cleaned stator coils to further remove residual dust; A spraying assembly 6 is provided on the left side of the processing table 1, located above the conveyor belt 8, for uniformly spraying insulating material onto the surface of the stator coil; a control component 4 is provided on the top right side of the processing table 1, electrically connected to the conveyor belt 8, cleaning assembly 5, blowing assembly 7, and spraying assembly 6, for controlling the working status of each component; a feeding assembly 11 is provided at one end of the control component 4, located at the end of the conveyor belt 8, for transferring the sprayed stator coil from the conveyor belt 8 to the discharge box 3; a feeding assembly 10 is installed on the top of the material box 2, located at the beginning of the conveyor belt 8, for transferring the stator coil to be processed from the material box 2 onto the conveyor belt 8.

[0027] A drying component 9 is also provided on the top of the processing table 1. The drying component 9 is located between the spraying component 6 and the unloading component 11, and is electrically connected to the control component 4.

[0028] The control component 4 is an embedded intelligent controller, which is electrically connected to the feeding component 10, the conveyor belt 8, the cleaning component 5, the blowing component 7, the spraying component 6, the drying component 9 and the unloading component 11 respectively.

[0029] The control unit 4 integrates a main control module, a multi-sensor fusion sensing module, a process parameter adaptive adjustment module, and a collaborative control module.

[0030] Furthermore, the overall structure of the device and the material flow of the present invention are as follows: Figure 4 As shown:

[0031] Furthermore, in the above implementation scheme, the multi-sensor fusion sensing module includes a photoelectric sensor, a vision sensor, a film thickness sensor, a 3D laser contour sensor, an online detection sensor for insulating varnish viscosity, a multi-channel infrared temperature sensor, and a high-precision position encoder for the conveyor belt.

[0032] The photoelectric sensor is located above the starting end of the conveyor belt 8 and is used to detect whether the stator coil is placed in place.

[0033] The visual sensor and the 3D laser contour sensor are coaxially disposed between the cleaning component 5 and the spraying component 6. The visual sensor is used to acquire images of the surface of the stator coil after cleaning, and the 3D laser contour sensor is used to scan and acquire three-dimensional point cloud data of the stator coil winding.

[0034] The film thickness sensor is located at the outlet of the spraying assembly 6 and is used to detect the thickness of the insulation layer after spraying in real time.

[0035] The online viscosity detection sensor for insulating varnish is installed at the inlet of the varnish supply pipeline of the spraying assembly 6, and is used to detect the kinematic viscosity of the insulating varnish in real time.

[0036] The multi-channel infrared temperature sensor array is arranged inside the drying component 9 to collect the surface temperature field distribution data of the stator coil insulation layer after spraying in real time.

[0037] The high-precision position encoder of the conveyor belt is coaxially mounted on the drive roller shaft of the conveyor belt 8 to acquire the displacement and speed data of the conveyor belt 8 in real time.

[0038] It should be further noted that, in this embodiment, the multi-sensor fusion sensing module integrates multiple sensors at the hardware level, with the specific configuration as follows:

[0039] Photoelectric sensor: Located on the support above the starting end of conveyor belt 8, its detection beam shines vertically downwards. After the feeding assembly 10 transfers the stator coil from the material box 2 to the conveyor belt 8, the photoelectric sensor detects in real time whether the workpiece is accurately placed in the bearing position of the conveyor belt 8 and transmits the arrival signal to the main control module.

[0040] Visual sensor and 3D laser profile sensor: Both are coaxially mounted on a gantry between cleaning assembly 5 and spraying assembly 6. The visual sensor uses a high-resolution industrial camera to acquire two-dimensional images of the stator coil surface after cleaning, and analyzes the surface cleanliness and impurity distribution through image processing algorithms. The 3D laser profile sensor uses the line laser scanning principle to continuously scan the stator coil winding during the movement of conveyor belt 8, acquire its three-dimensional point cloud data, and accurately reconstruct the geometric features of the irregular structure, corners, and gaps at the winding ends.

[0041] Furthermore, the main control module incorporates a cleanliness detection algorithm, as detailed below:

[0042] After converting the acquired image to grayscale, adaptive threshold segmentation is used to extract the impurity region, and the area ratio of the impurity region is calculated. .when When the cleanliness level falls below a preset threshold (0.5%~2.0%), the cleanliness is determined to be below the preset threshold. Simultaneously, connected component analysis is used to identify impurity distribution areas and generate an impurity mask. This is used for subsequent targeted secondary cleaning.

[0043] Film thickness sensor: Installed on the outlet side of the spraying assembly 6, it adopts the non-contact optical film thickness measurement principle. Immediately after the spraying is completed, it performs multi-point real-time detection of the insulation layer thickness on the surface of the stator coil and uploads the film thickness distribution data in real time.

[0044] Online viscosity detection sensor for insulating varnish: Located at the inlet of the varnish supply pipeline of the spraying assembly 6, it uses a vibration or rotation viscometer to detect the kinematic viscosity value of the insulating varnish before spraying in real time and feeds the data back to the control component 4.

[0045] Multiple infrared temperature sensors are arrayed on the internal cavity of the drying assembly 9 to form a multi-point temperature monitoring network. When the coated stator coil enters the drying assembly 9, the temperature field distribution data of the insulation layer surface is collected in real time at each temperature measuring point for closed-loop control of the drying process.

[0046] High-precision position encoder for conveyor belt: Coaxially mounted at the end of the drive roller shaft of conveyor belt 8, using incremental or absolute encoder, to output displacement pulse signals and speed signals of conveyor belt 8 in real time, providing a position reference for synchronous control of the entire workstation.

[0047] At the data processing level, the main control module has a built-in spatiotemporal synchronized multi-source fusion algorithm, and the specific implementation steps are as follows:

[0048] Step 1: Establish a reference coordinate system

[0049] A global coordinate system O(X,Y,Z) is established with the position of the photoelectric sensor at the starting end of conveyor belt 8 as the origin. The X-axis represents the conveyor belt's movement direction, the Y-axis represents the width direction, and the Z-axis represents the height direction. Each pulse output by the high-precision position encoder of the conveyor belt corresponds to the conveyor belt's movement distance. Where D is the diameter of the drive roller (mm) and N is the number of encoder pulses per revolution.

[0050] Step 2: Sensor coordinate transformation

[0051] Coordinates of each sensor's installation location in the global coordinate system Pre-calibration. When the encoder feedback position is... At that time, the coordinates of the data points collected by the sensor are transformed as follows:

[0052]

[0053] in This represents the offset of the sensor measurement value relative to the origin of the sensor's own coordinate system.

[0054] Step 3: Time Synchronization and Data Interpolation

[0055] Using encoder pulse interruption as the synchronization reference, timestamp This corresponds to the i-th pulse. For sensors with low sampling frequencies (such as vision sensors and film thickness sensors), linear interpolation alignment is used:

[0056]

[0057] in , for Sensor values ​​at any given time.

[0058] Through the above spatiotemporal synchronization processing, all sensor data are precisely bound to the timestamps and position coordinates of the workpiece position encoder, unifying the spatial coordinate system and time reference, and realizing a one-to-one correspondence of the entire process data of a single workpiece.

[0059] Furthermore, the raw point cloud data acquired by the 3D laser contour sensor needs to be processed through the following steps:

[0060] Step 1: Point cloud data preprocessing

[0061] Input the original point cloud Statistical filtering is used to remove outliers, and the average distance within the neighborhood of each point is calculated. and standard deviation Remove points that meet the following conditions:

[0062]

[0063] Where k takes values ​​from 2.0 to 3.0, Let be the distance from point i to the neighborhood center. Then, the moving least squares (MLS) method is used to smoothly reconstruct the point cloud, generating a continuous surface model.

[0064] Step 2: Extraction of features from irregular structures

[0065] For each point By fitting a local plane using neighborhood points, the normal vector is calculated. Curvature Calculate using the following formula:

[0066]

[0067] in These are the eigenvalues ​​of the covariance matrix. When the value is between 0.1 and 0.3, it is marked as a corner or gap region. Clustering is performed on regions with abrupt changes in normal vectors to form a feature region mask. It can accurately identify areas such as irregular structures, corners, and gaps at the ends of the windings that are prone to paint blind spots.

[0068] Furthermore, in the above implementation scheme, the process parameter adaptive adjustment module has a built-in spraying process database, which stores spraying parameters corresponding to different models of stator coils, including spraying pressure, atomization angle, spray gun moving speed and spraying times; the process parameter adaptive adjustment module is also connected to an ambient temperature and humidity sensor, which is used to compensate and correct the spraying parameters according to the environmental data.

[0069] The adaptive adjustment module for process parameters also incorporates a digital twin model of the entire stator coil spraying process and a process self-learning iteration unit. The digital twin model includes a 3D model of the workpiece, a spray gun fluid dynamics model, an insulating varnish film-forming and curing model, and an environmental field model. It is used to map the data collected in real time by the multi-sensor fusion sensing module to the model and simulate the film-forming effect. The process self-learning iteration unit is used to continuously optimize the spraying process database based on the correlation analysis between the workpiece spraying quality data and the corresponding process parameters through reinforcement learning algorithms.

[0070] It should be further explained that the adaptive process parameter adjustment module has built a digital twin-driven adaptive process adjustment and self-learning system within the control unit 4, which is implemented as follows:

[0071] 1. Spraying process database and environmental compensation

[0072] The module has a pre-installed spraying process database that stores basic spraying parameters for different stator coil models, including spraying pressure, atomization angle, spray gun movement speed, spray gun attitude angle, and number of sprays. The module connects to an ambient temperature and humidity sensor to read environmental data in real time and perform linear or nonlinear compensation corrections on the basic parameters.

[0073] 2. Segmented spraying parameter generation

[0074] Based on feature region masking The surface of the stator coil is divided into flat areas. Curved areas Gaps / corner areas And adaptively adjust the parameters according to the following rules:

[0075] Spraying pressure Correction: The values ​​are as follows, depending on the region: , , .

[0076] Spray gun movement speed Correction: ,in ,other =1.

[0077] Atomization angle Correction: ,in ,other =1.

[0078] 3. Calculation of paint viscosity compensation

[0079] Based on real-time detected kinematic viscosity of insulating varnish Compared with reference viscosity Adjust the spray pressure and spray gun speed according to the following formula:

[0080]

[0081] in Take a value of 0.5 to 0.8. Take a value of 0.3 to 0.6.

[0082] 4. Digital twin simulation and film formation effect prediction

[0083] The module has a built-in digital twin model of the entire stator coil spraying process, including a 3D model of the workpiece, a hydrodynamic model of the spray gun, a model of the insulating varnish film formation and curing, and an environmental field model.

[0084] The 3D model of the workpiece is reconstructed based on point cloud data scanned by a 3D laser contour sensor;

[0085] The fluid dynamics model of the spray gun was constructed using experimentally calibrated atomization cone angle and flow-pressure characteristic curves;

[0086] The insulating varnish film-forming and curing model is based on high-speed photography and film thickness calibration experiments, and the parameters of the Gaussian distribution superposition model are obtained by fitting.

[0087] The environmental field model is mapped in real time using multiple temperature and humidity sensors placed inside the equipment.

[0088] Each sub-model has been validated by no less than 200 sets of experimental data, and the film thickness prediction error is ≤ ±3μm.

[0089] The film thickness distribution was simulated using a Gaussian distribution superposition model.

[0090]

[0091] Where T(x,y) is the predicted film thickness (μm). Let σ be the paint flow rate at the j-th spray point (mm³ / s), and σ be the atomization diffusion coefficient (mm). The dwell time (s) of the spray gun at this point is used. Real-time collected data on workpiece contour, paint viscosity, and ambient temperature and humidity are mapped to the model to quickly simulate the film formation effect, uniformity, and curing degree of the insulation layer under the current parameters, and the process parameters are dynamically adjusted accordingly.

[0092] 5. Reinforcement learning self-iterative optimization

[0093] The process self-learning iterative unit performs correlation analysis between the final coating quality data of each workpiece and the corresponding process parameters, and continuously optimizes the process database through reinforcement learning algorithms. Specifically, it adopts the PPO (Proximal Policy Optimization) algorithm, defined as follows:

[0094] State space S: Current workpiece model, ambient temperature and humidity, paint viscosity, 3D contour features

[0095] Action space A: Spraying pressure P, atomization angle ϕ, spray gun speed v, number of sprays n

[0096] reward function ,in For film thickness standard deviation, For the target film thickness, t represents the number of defects. cycle Weighting coefficient for spraying cycle time (s) =0.5, =0.3, =0.2;

[0097] The policy network parameters are updated according to the following formula:

[0098]

[0099] in , ϵ=0.2.

[0100] 6. One-click adaptation of workpiece model

[0101] When it is necessary to change the production model, the operator only needs to input or scan the 3D drawing or model code of the new stator coil on the touch screen. The module will automatically call the digital twin model to perform simulation calculations and generate an initial spraying process parameter package, eliminating the need for repeated manual trial spraying and debugging.

[0102] Furthermore, in the above implementation scheme, the collaborative control module is configured as follows:

[0103] When the photoelectric sensor detects that there is no workpiece at the starting end of the conveyor belt 8, it controls the feeding assembly 10 to transfer the stator coil from the material box 2 to the conveyor belt 8;

[0104] When the stator coil is delivered to the cleaning component 5 station, the conveyor belt 8 is stopped, and the cleaning component 5 and the blowing component 7 are started in sequence.

[0105] When the stator coil is delivered to the spraying assembly 6 station, the conveyor belt 8 is controlled to decelerate and the spraying assembly 6 is started synchronously.

[0106] When the stator coil is conveyed to the end of the conveyor belt 8, the feeding assembly 11 is controlled to transfer the stator coil to the discharge box 3;

[0107] The collaborative control module also uses real-time data from the high-precision position encoder of the conveyor belt to achieve precise synchronous control of the actions and positions of all workstations, with a synchronous control accuracy of no less than ±0.1mm.

[0108] Furthermore, in the above implementation scheme, the collaborative control module is also configured to: when the vision sensor detects that the surface cleanliness of the stator coil does not reach a preset threshold, control the conveyor belt 8 to reverse or delay before entering the spraying station; when the film thickness sensor detects that the spraying thickness exceeds a preset threshold, automatically trigger a re-spraying command to send the stator coil back into the spraying station for secondary spraying; when multiple stator coils have unqualified spraying thicknesses, issue an audible and visual alarm and suspend equipment operation.

[0109] It should be further explained that the targeted secondary cleaning and precise localized touch-up spraying are implemented as follows:

[0110] Targeted secondary cleaning: When the cleanliness test fails, the collaborative control module uses an impurity mask... The real-time coordinates of the high-precision position encoder of the conveyor belt are used to map the coordinates of the impurity area to the execution space of the cleaning component 5 and the blowing component 7. The rotating brush of the cleaning component 5 is controlled to descend and increase its speed only at the corresponding position of the impurity area. At the same time, the nozzle of the blowing component 7 is controlled to adjust the air outlet angle and air pressure according to the position of the impurity area to achieve selective cleaning of the space.

[0111] Precise local replenishment spraying: When the film thickness sensor detects that the film thickness in a local area is lower than the lower limit threshold, the collaborative control module generates a replenishment spraying path and local spraying parameters (spraying pressure reduced by 20%~30%, spray gun moving speed reduced by 40%~60%) based on the film thickness deviation and the coordinates of the abnormal area. The spray gun of the spraying component 6 is controlled to perform local scanning spraying only above the abnormal area. The replenishment spraying path and workpiece movement are controlled in real time by the position encoder in a closed loop.

[0112] Furthermore, in the above implementation scheme, the collaborative control module also constructs a spraying-drying linkage control mechanism and a closed-loop handling logic for quality anomalies, wherein:

[0113] The spraying-drying linkage control mechanism is configured to: adaptively adjust the heating power, airflow distribution and drying time of the drying component 9 based on the insulation layer thickness distribution data collected by the film thickness sensor.

[0114] It should be further explained that the drying component 9 is internally divided into 3 to 6 independent temperature control zones along the conveying direction, each temperature control zone corresponding to an independent heating element and an adjustable damper. (Based on film thickness matrix) Based on the workpiece position, the collaborative control module dynamically adjusts the opening of the dampers in each zone, using the following formula:

[0115]

[0116] in Let k be the opening degree of the damper in zone k. Take a value of 0.3~0.8. The average actual film thickness corresponding to the k-th temperature control zone is used to enhance hot air convection in the thick coating zone and reduce hot air flow in the thin coating zone, thus ensuring uniform curing.

[0117] The targeted closed-loop handling logic for quality anomalies is configured as follows:

[0118] When the surface cleanliness of the stator coil is detected to be below the preset threshold, the impurity distribution area is automatically identified, and the operating parameters of the cleaning component 5 and the blowing component 7 are adaptively adjusted to complete the directional secondary cleaning.

[0119] When the insulation layer thickness is detected to exceed the preset threshold range, the abnormal film thickness area and deviation are automatically identified, directional respraying parameters are generated, and the spraying component 6 is controlled to perform precise local respraying on the abnormal area.

[0120] Furthermore, in the above implementation scheme, the collaborative control module also has a built-in workpiece model error prevention mechanism. When the workpiece model grabbed by the feeding component 10 does not match the currently called process parameter package, the device is automatically prevented from starting and an alarm is issued.

[0121] It should be further explained that the collaborative control module is responsible for the precise timing control and quality closed-loop handling of all workstation actions throughout the entire process. The workflow of the collaborative control module is as follows: Figure 5 As shown, the specific configuration is as follows:

[0122] 1. Basic timing control

[0123] When the photoelectric sensor detects that there is no workpiece at the starting end of the conveyor belt 8, the collaborative control module controls the feeding assembly 10 to grab a stator coil from the material box 2 and place it on the conveyor belt 8.

[0124] When the workpiece is conveyed to the cleaning component 5 station, the control conveyor belt 8 is paused, and the cleaning component 5 and the blowing component 7 are started in sequence.

[0125] When the workpiece is transported to the spraying component 6 station, the control conveyor belt 8 decelerates and the spraying component 6 is started simultaneously.

[0126] When the workpiece is conveyed to the end of the conveyor belt 8, the unloading assembly 11 controls the finished stator coil to be transferred to the discharge box 3.

[0127] 2. High-precision synchronous control

[0128] The collaborative control module uses real-time pulse signals from the high-precision position encoder of the conveyor belt and employs an S-shaped speed planning algorithm to achieve smooth start and stop of the conveyor belt and spray gun movement, avoiding jitter. Maximum acceleration is defined. (mm / s²), maximum jerk (mm / s³), target velocity (mm / s), the segmented velocity curve is planned according to three stages: uniform acceleration, uniform speed, and uniform deceleration.

[0129] Segmented velocity curve:

[0130] Uniform acceleration phase ;

[0131] Uniform velocity stage: ;

[0132] Uniform deceleration phase: ;

[0133] Position feedforward control is corrected according to the following formula:

[0134]

[0135] in For positional deviation, , Ultimately, the synchronization control accuracy between all station movements and workpiece position is no less than ±0.1mm.

[0136] 3. Spraying-drying linkage control mechanism

[0137] Based on the insulation layer thickness distribution data collected by the film thickness sensor, the stator coil surface is divided into M×N grids, and the film thickness matrix is... The heating power, airflow distribution, and drying time of the drying component 9 are adaptively adjusted according to the following formula:

[0138] Zone temperature setting: ,in ;

[0139] Drying time correction: ,in ;

[0140] By controlling the power of the multi-zone heating tubes and the opening of the damper in the drying component 9, the spatial distribution of temperature field and residence time can be regulated, thus solving the problems of incomplete curing in thick coating areas and over-baking and aging in thin coating areas.

[0141] 4. Closed-loop handling logic for quality anomalies

[0142] Handling of Cleanliness Abnormalities: When the vision sensor detects that the cleanliness of the stator coil surface after cleaning does not reach the preset threshold, the collaborative control module controls the conveyor belt 8 to run in reverse or delay to enter the spraying station, automatically identifies the impurity distribution area, and adaptively adjusts the rotation speed of the cleaning component 5 and the air pressure of the blowing component 7 to perform directional secondary cleaning of the impurity area.

[0143] Abnormal Film Thickness Handling: When the film thickness sensor detects that the thickness of the sprayed insulation layer exceeds the preset threshold range, the module automatically identifies the abnormal film thickness area and deviation, generates directional respraying parameters, and controls the spraying component 6 to perform precise local respraying on the abnormal area. If the spraying thickness of multiple stator coils is unqualified, the module issues an audible and visual alarm and suspends equipment operation.

[0144] 5. Workpiece model error prevention mechanism

[0145] The collaborative control module has built-in workpiece model error prevention logic. When the workpiece model picked up by the feeding component 10 is read by visual recognition or RFID, if it does not match the model in the currently invoked process parameter package, the module automatically prevents the equipment from starting and issues an alarm prompt on the touch screen.

[0146] Furthermore, in the above implementation scheme, the control component 4 also includes a fault prediction and health management module based on edge computing. The fault prediction and health management module is used to collect the operating data of each execution component of the entire device in real time, and has a built-in edge fault prediction machine learning model to identify the performance degradation trend of components and issue maintenance warnings in advance. At the same time, when the device malfunctions, it can complete the root cause analysis of the fault based on multi-sensor data and push standardized handling procedures.

[0147] It should be further explained that the fault prediction and health management module builds an edge-intelligent driven device health management system within the control unit 4, and the specific implementation is as follows:

[0148] 1. Data Collection

[0149] The module collects real-time operating data of all execution components of the equipment, including the current and torque of the conveyor belt drive motor, the paint supply pressure fluctuation of the spraying component 6, the motor vibration frequency of the cleaning component 5, and the current and temperature fluctuation of the heating tube of the drying component 9.

[0150] 2. Feature Extraction

[0151] For time-series running data Extract the following features:

[0152] Time-domain characteristics: mean μ, variance Root mean square (RMS), peak value (max), peak-to-peak value (peak-to-peak) ;

[0153] Frequency domain characteristics: Perform FFT transform on the signal to extract the main frequency amplitude. Frequency centroid ;

[0154] Statistical characteristics: skewness , cliff ;

[0155] 3. Fault prediction model

[0156] An LSTM (Long Short-Term Memory) model is used for fault prediction. The network structure is: input layer (d-dimensional) → LSTM layer (64 units) → Dropout (0.2) → fully connected layer (32 units) → Sigmoid output. The loss function is binary cross-entropy.

[0157]

[0158] When predicting probability When the time is right, a maintenance warning is triggered, enabling predictive maintenance.

[0159] 4. Root cause analysis of the failure

[0160] When the device malfunctions, the SHAP value is used to calculate each feature. Contribution to the prediction results:

[0161]

[0162] The components corresponding to the top 3 features with the highest SHAP values ​​are marked as root causes of the failure, and a standardized handling process is pushed on the touch screen.

[0163] 5. Equipment maintenance log management

[0164] The module automatically records the time, content, and replaced parts of each maintenance, generates an equipment maintenance log, and automatically generates regular maintenance reminders based on equipment runtime or processing volume.

[0165] Furthermore, in the above implementation scheme, the control component (4) also includes a wireless communication module, which is used to upload equipment operation data, fault records, process parameters and spraying quality statistics to the remote monitoring platform, and receive remote parameter adjustment instructions;

[0166] The wireless communication module also supports cloud-edge collaboration and full lifecycle quality traceability. It can bind the full process data and quality inspection data of a single stator coil with the unique identifier of the workpiece to generate a single workpiece traceability file and upload it to the cloud platform. At the same time, it supports the batch distribution of standardized process parameter packages in the cloud and the cloud synchronization of optimized process parameters at the edge.

[0167] It should be further explained that, based on the original data upload function, the wireless communication module has been upgraded to build a full-process traceability and cloud-edge collaboration system:

[0168] 1. Full lifecycle traceability for a single workpiece

[0169] During workpiece loading, controller 4 generates a 32-bit UUID for each stator coil captured by loading component 10, in the format {timestamp}-{device ID}-{batch number}-{serial number}. This UUID is physically bound to the workpiece via an RFID reader or QR code printer. A time-series database (such as InfluxDB) is used for storage, with each data record containing {uuid, timestamp, workstation ID, sensor type, sensor value, process parameters, quality inspection result}, forming a complete traceability file for each workpiece throughout the entire process.

[0170] 2. Cloud-edge collaboration and data upload

[0171] Using a wireless communication module (4G / 5G or industrial Wi-Fi) and the MQTT protocol, equipment operation data, fault records, quality statistics, and single-workpiece traceability files are encrypted and uploaded to a remote industrial internet platform (cloud platform). This supports remote real-time monitoring, quality data statistical analysis, and full lifecycle traceability queries.

[0172] 3. Cloud-edge collaborative management of process parameters

[0173] The cloud platform distributes standardized process parameter packages in batches to specified devices or groups of devices via REST API. The optimized process parameters at the edge are managed using a version number mechanism: they are first uploaded for review, and after approval, the version number is updated in the cloud and then synchronized to other devices across the entire production line, achieving synchronized improvement of process capabilities across the entire production line.

[0174] 4. Remote Operation and Maintenance

[0175] When equipment malfunctions, remote maintenance experts can view the equipment's full-dimensional operating data and historical fault records in real time through a cloud platform, assisting on-site personnel in troubleshooting and remote parameter adjustment.

[0176] Furthermore, in the above implementation scheme, the surface of the control component 4 is provided with a touch screen for displaying equipment operating status, process parameters and fault information, and for allowing users to input control commands, switch workpiece model process packages, and view equipment maintenance logs and workpiece traceability files.

[0177] It should be further explained that the touchscreen on the surface of the control component 4 is a human-computer interaction interface, and its displayed content includes:

[0178] Equipment operating status: Real-time display of conveyor belt speed (8 stations), working status of each station, current number of workpieces processed, pass rate, etc.

[0179] Process parameters: Displays key parameters such as the currently used process parameter package model, spraying pressure, atomization angle, and drying temperature, and supports online fine-tuning by operators.

[0180] Fault Information: When a device malfunctions or is predicted to malfunction, the fault information, root cause of the fault, and troubleshooting instructions will be displayed proactively.

[0181] User interface: Allows operators to input control commands (such as start, pause, emergency stop), switch workpiece model process packages, view equipment maintenance logs, and retrieve the full lifecycle traceability file of any processed workpiece.

[0182] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A device for spraying insulation layer of motor stator coil, comprising a processing table (1), characterized in that: A discharge box (3) is provided on the right side of the processing table (1), a material box (2) is installed at the front end of the processing table (1), a conveyor belt (8) is provided on the top of the processing table (1), a cleaning component (5) is provided on the top of the front end of the processing table (1), a blowing component (7) is provided on the left side of the cleaning component (5), a spraying component (6) is provided on the left side of the processing table (1), a control component (4) is provided on the top right side of the processing table (1), a feeding component (11) is provided at one end of the control component (4), and a feeding component (10) is installed on the top of the material box (2). A drying assembly (9) is also provided on the top of the processing table (1). The drying assembly (9) is located between the spraying assembly (6) and the unloading assembly (11) and is electrically connected to the control component (4). The control component (4) is an embedded intelligent controller, which is electrically connected to the feeding component (10), conveyor belt (8), cleaning component (5), blowing component (7), spraying component (6), drying component (9) and unloading component (11), respectively. The control unit (4) integrates a main control module, a multi-sensor fusion sensing module, a process parameter adaptive adjustment module, and a collaborative control module.

2. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The multi-sensor fusion sensing module includes photoelectric sensors, vision sensors, film thickness sensors, 3D laser contour sensors, online detection sensors for insulating varnish viscosity, multi-channel infrared temperature sensors, and a high-precision position encoder for the conveyor belt. The photoelectric sensor is located above the starting end of the conveyor belt (8) and is used to detect whether the stator coil is placed in place; The visual sensor and the 3D laser contour sensor are coaxially arranged between the cleaning component (5) and the spraying component (6). The visual sensor is used to collect images of the surface of the stator coil after cleaning, and the 3D laser contour sensor is used to scan and obtain three-dimensional point cloud data of the stator coil winding. The film thickness sensor is located at the outlet of the spraying assembly (6) and is used to detect the thickness of the insulation layer after spraying in real time. The online viscosity detection sensor for insulating varnish is installed at the inlet of the varnish supply pipeline of the spraying assembly (6) and is used to detect the kinematic viscosity of the insulating varnish in real time. The multi-channel infrared temperature sensor array is arranged inside the drying component (9) to collect the surface temperature field distribution data of the stator coil insulation layer after spraying in real time. The high-precision position encoder of the conveyor belt is coaxially mounted on the drive roller shaft of the conveyor belt (8) to acquire the displacement and speed data of the conveyor belt (8) in real time.

3. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The adaptive adjustment module for process parameters has a built-in spraying process database, which stores spraying parameters corresponding to different models of stator coils, including spraying pressure, atomization angle, spray gun moving speed, and number of sprays. The adaptive adjustment module for process parameters is also connected to an ambient temperature and humidity sensor, which is used to compensate and correct the spraying parameters based on environmental data. The adaptive adjustment module for process parameters also incorporates a digital twin model of the entire stator coil spraying process and a process self-learning iteration unit. The digital twin model includes a 3D model of the workpiece, a spray gun fluid dynamics model, an insulating varnish film-forming and curing model, and an environmental field model. It is used to map the data collected in real time by the multi-sensor fusion sensing module to the model and simulate the film-forming effect. The process self-learning iteration unit is used to continuously optimize the spraying process database based on the correlation analysis between the workpiece spraying quality data and the corresponding process parameters through reinforcement learning algorithms.

4. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The collaborative control module is configured as follows: When the photoelectric sensor detects that there is no workpiece at the starting end of the conveyor belt (8), it controls the feeding assembly (10) to transfer the stator coil from the material box (2) to the conveyor belt (8); When the stator coil is delivered to the cleaning component (5) station, the conveyor belt (8) is stopped and the cleaning component (5) and the blowing component (7) are started in sequence. When the stator coil is delivered to the spraying assembly (6) station, the conveyor belt (8) is controlled to decelerate and the spraying assembly (6) is started synchronously. When the stator coil is conveyed to the end of the conveyor belt (8), the feeding assembly (11) is controlled to transfer the stator coil to the discharge box (3). The collaborative control module also uses real-time data from the high-precision position encoder of the conveyor belt to achieve precise synchronous control of the actions and positions of all workstations, with a synchronous control accuracy of no less than ±0.1mm.

5. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The collaborative control module is also configured to: when the vision sensor detects that the surface cleanliness of the stator coil does not reach the preset threshold, control the conveyor belt (8) to reverse or delay before entering the spraying station; when the film thickness sensor detects that the spraying thickness exceeds the preset threshold, automatically trigger the re-spraying command and send the stator coil back into the spraying station for secondary spraying; when multiple stator coils have unqualified spraying thickness, issue an audible and visual alarm and suspend the operation of the equipment.

6. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The collaborative control module also includes a spraying-drying linkage control mechanism and a closed-loop logic for handling quality anomalies, wherein: The spraying-drying linkage control mechanism is configured to: adaptively adjust the heating power, air field distribution and drying time of the drying component (9) based on the insulation layer thickness distribution data collected by the film thickness sensor; The targeted closed-loop handling logic for quality anomalies is configured as follows: When the surface cleanliness of the stator coil is detected to be below the preset threshold, the impurity distribution area is automatically identified, and the operating parameters of the cleaning component (5) and the blowing component (7) are adaptively adjusted to complete the directional secondary cleaning. When the insulation layer thickness is detected to exceed the preset threshold range, the abnormal film thickness area and deviation are automatically identified, directional spraying parameters are generated, and the spraying component (6) is controlled to perform precise local spraying on the abnormal area.

7. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The control unit (4) also includes a fault prediction and health management module based on edge computing. The fault prediction and health management module is used to collect the operating data of each execution component of the whole equipment in real time, and has a built-in edge fault prediction machine learning model to identify the performance degradation trend of the components and issue maintenance warnings in advance. Simultaneously, when equipment malfunctions, it can perform root cause analysis based on multi-sensor data and push standardized handling procedures.

8. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The control unit (4) also includes a wireless communication module, which is used to upload equipment operation data, fault records, process parameters and spraying quality statistics to the remote monitoring platform, and receive remote parameter adjustment instructions; The wireless communication module also supports cloud-edge collaboration and full lifecycle quality traceability. It can bind the full process data and quality inspection data of a single stator coil with the unique identifier of the workpiece to generate a single workpiece traceability file and upload it to the cloud platform. At the same time, it supports the batch distribution of standardized process parameter packages in the cloud and the cloud synchronization of optimized process parameters at the edge.

9. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The surface of the control component (4) is equipped with a touch screen, which is used to display the equipment operating status, process parameters and fault information, and to allow users to input control commands, switch workpiece model process packages, and view equipment maintenance logs and workpiece traceability files.

10. The motor stator coil insulation layer spraying device according to claim 1, characterized in that: The collaborative control module also has a built-in workpiece model error prevention mechanism. When the workpiece model captured by the feeding component (10) does not match the currently called process parameter package, the device will be automatically prevented from starting and an alarm will be issued.

Images