Valve element intelligent detection method and system based on gear rotation positioning and electronic equipment

Abstract

The invention discloses an intelligent valve element detection method and system based on gear rotation positioning and electronic equipment, and relates to the technical field of valve element detection.The method comprises the steps that a material disc provided with a to-be-detected valve element is placed on a top plate; a to-be-detected valve core is placed on a workpiece placing seat by utilizing a clamping jaw, the workpiece placing seat is moved to a detection area through a guide rail transmission device, and the to-be-detected valve core is driven by a gear to rotate and be positioned; performing multi-dimensional detection on the to-be-detected valve core through the sensor group, and performing defect identification on the multi-dimensional detection data of the valve core; and performing qualified product judgment on the to-be-detected valve element based on the valve element defect detection information, and performing classified placement treatment on the to-be-detected valve element according to a judgment result. The technical problems of detection omission, low efficiency and high misjudgment rate caused by insufficient accuracy of rotation positioning precision, multi-dimensional data synchronous acquisition and defect identification in the prior art are solved, and the technical effects of improving the coverage rate, precision, efficiency and reliability of valve element detection are achieved.

Description

Technical Field

[0001] This invention relates to the field of valve core detection technology, and in particular to a valve core intelligent detection method and system, and electronic equipment based on gear rotation positioning. Background Technology

[0002] The valve core is a core functional component of fluid control equipment such as hydraulic control systems, gas transmission valves, and water supply and drainage pipeline valves. Its structural precision, surface quality, internal integrity, and sealing performance directly determine the valve's opening and closing reliability, fluid control accuracy, and safe operating life. Currently, valve core testing technology in the industry generally suffers from low levels of automation throughout the entire process, with most testing procedures relying on manual pallet handling and numerous manual intervention steps. Furthermore, the rotational positioning accuracy is insufficient; existing valve core rotation methods often employ simple motors and manual calibration, failing to accurately match the rotation angle according to the valve core specifications, leading to missed detections in some circumferential testing areas. Additionally, the data correlation between multiple sensors is poor. These shortcomings make it difficult for existing testing solutions to meet the demands of high-precision, high-efficiency, and high-reliability batch testing of valve cores.

[0003] In summary, existing technologies suffer from insufficient rotational positioning accuracy, multi-dimensional data synchronous acquisition, and defect identification accuracy, leading to detection omissions, low efficiency, and high false positive rates. Summary of the Invention

[0004] This application provides a valve core intelligent detection method, system, and electronic equipment based on gear rotation positioning, which solves the technical problems in the prior art such as insufficient rotation positioning accuracy, multi-dimensional data synchronous acquisition, and defect identification accuracy, leading to detection omissions, low efficiency, and high false judgment rate.

[0005] The first aspect of this application provides a valve core intelligent detection method based on gear rotation positioning, the method comprising:

[0006] A tray containing the valve core to be tested is placed on the top plate. The top plate is raised by a bottom cylinder, and the tray is moved horizontally to the loading area via a screw drive. The valve core to be tested is picked up from the tray by grippers and placed on a workpiece placement seat. The motor is started, and the workpiece placement seat is moved to the testing area via a guide rail drive. In the testing area, the valve core to be tested is rotated and positioned by gears. The rotated and positioned valve core to be tested is subjected to multi-dimensional detection by a sensor group to obtain multi-dimensional detection data of the valve core. Defect identification is performed on the multi-dimensional detection data of the valve core to determine the valve core defect detection information. Based on the valve core defect detection information, the valve core to be tested is judged as qualified, and the valve core to be tested is classified and placed according to the judgment result.

[0007] A second aspect of this application provides a valve core intelligent detection system based on gear rotation positioning, the system comprising:

[0008] The feeding module places a tray containing the valve core to be tested on a top plate, raises the top plate using a bottom cylinder, and moves the tray to the loading area via a screw drive. The positioning module uses grippers to remove the valve core from the tray and place it on a workpiece holder. A motor is started, and the workpiece holder is moved to the testing area via a guide rail drive. Within the testing area, gears rotate the valve core for positioning. The testing module uses a sensor array to perform multi-dimensional testing on the rotated valve core, obtaining multi-dimensional testing data. Defects are identified from this data to determine valve core defect information. The classification module determines the valve core's conformity based on the defect information and classifies it according to the results.

[0009] A third aspect of this application provides an electronic device comprising: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to implement a valve core intelligent detection method based on gear rotation positioning.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] This application places a tray containing the valve core to be tested on a top plate. A bottom cylinder pushes the top plate upwards, and a screw drive moves the tray horizontally to the loading area. A gripper picks up the valve core from the tray and places it on a workpiece holder. A motor is started, and a guide rail drive moves the workpiece holder to the testing area. In the testing area, gears rotate and position the valve core. A sensor array performs multi-dimensional detection on the rotated valve core, obtaining multi-dimensional detection data. Defects are identified from this data to determine valve core defect information. Based on the defect information, the valve core is judged as a qualified product, and then classified and placed according to the judgment results. Through fully automated gear rotation positioning, synchronous multi-dimensional detection, and intelligent defect recognition, the technical effect of improving the coverage, accuracy, efficiency, and reliability of valve core detection is achieved. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic flowchart of the intelligent valve core detection method based on gear rotation positioning provided in the embodiments of this application.

[0014] Figure 2 This is a schematic diagram of the valve core intelligent detection system based on gear rotation positioning provided in the embodiments of this application.

[0015] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0016] Explanation of reference numerals in the attached drawings: feeding module 11, positioning module 12, detection module 13, sorting module 14, input device 301, processor 302, memory 303, output device 304. Detailed Implementation

[0017] This application provides a valve core intelligent detection method, system, and electronic equipment based on gear rotation positioning, which solves the technical problems in the prior art such as insufficient rotation positioning accuracy, multi-dimensional data synchronous acquisition, and defect identification accuracy, leading to detection omissions, low efficiency, and high false judgment rate.

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0020] Example 1, as Figure 1 As shown, a valve core intelligent detection method based on gear rotation positioning is described, wherein the method includes:

[0021] S100: Place the tray containing the valve core to be tested on the top plate, push the top plate up through the bottom cylinder, and move the tray to the feeding area through the screw drive device.

[0022] Specifically, firstly, the automated feeding equipment smoothly places the tray containing the valve cores to be tested onto the top plate. The tray is a standard chemical component that supports the valve cores, and its interior has an arc-shaped positioning groove adapted to the valve core's shape, ensuring a neat arrangement of multiple valve cores and preventing displacement during movement. The top plate serves as the direct support for the tray, and a cylinder is located at its bottom. This bottom cylinder is a linear power element based on pneumatic transmission. When compressed air enters the rodless chamber of the cylinder, the piston rod extends and pushes the top plate, along with the tray, smoothly upwards along the guide columns on both sides of the equipment until the top plate reaches the preset height. Simultaneously, the screw drive device is activated. This device consists of a high-precision ball screw, a matching nut slider, and double parallel guide rails. The transmission system consists of a servo motor and a planetary reduction gear. The servo motor is connected to one end of a ball screw via a coupling. The nut slider is fixed to a slot on the side of the top plate. When the servo motor receives a translation command, the nut slider moves linearly along the guide rail under the drive of the screw thread, thereby driving the top plate and the material tray to translate upwards to the material area. The material area is equipped with a photoelectric sensor and a mechanical stop. When the material tray is translated to this area, the photoelectric sensor detects the positioning line on the edge of the material tray and immediately sends a position signal to the control system. The servo motor brakes urgently, and at the same time, the mechanical stop pops out and abuts the side of the material tray, realizing secondary positioning of the material tray and ensuring that the final position of the material tray is completely matched with the subsequent gripping trajectory of the gripper. This completes the step.

[0023] S200: The valve core to be tested is clamped and removed from the material tray by the gripper and placed on the workpiece placement seat. The motor is started and the workpiece placement seat is moved to the testing area by the guide rail transmission device. In the testing area, the valve core to be tested is rotated and positioned by the gear.

[0024] Specifically, after the material tray is positioned in the loading area, an electric gripper descends to a preset height above the valve core, based on the preset arrangement coordinates of the valve core within the material tray, determined by prior equipment calibration, and aligns with the positioning slot of the material tray. The gripper then closes to stably hold the valve core. After holding, the gripper moves along a preset trajectory to the workpiece placement seat. The top of the workpiece placement seat has a tapered positioning hole that matches the bottom of the valve core, and the bottom of the placement seat integrates a T-shaped connecting block for fixing to the slider of the guide rail transmission device. After the gripper smoothly places the valve core into the positioning hole, the pressure sensor provides feedback of the placement in place signal, the gripper releases and resets to the loading area for standby, completing the transfer of the valve core from the material tray to the workpiece placement seat. Next, the motor is started, and the control system outputs power to the drive motor, which is transmitted to the guide rail transmission device through the planetary reduction mechanism. The guide rail transmission device consists of a high-precision linear guide rail, a ball screw, and a nut slider. The nut slider is rigidly locked to the T-shaped connecting block at the bottom of the workpiece placement seat. Under the drive of the motor, the ball screw rotates, causing the nut slider to move linearly along the linear guide rail, thereby pulling the workpiece placement seat and the valve core to be tested to move towards the detection area. When the workpiece placement seat triggers the photoelectric calibration signal in the detection area, the servo motor immediately brakes and locks its position, and the mechanical limit block pops out simultaneously to perform secondary positioning of the placement seat, ensuring that the workpiece placement seat is in the detection area. Subsequently, based on the specifications and testing requirements of the valve core to be tested, such as scratches on the side circumference of the valve core, the depth of the sealing groove, and the coaxiality of the holes, the control system presets a rotation angle. For example, for a valve core with three evenly distributed sealing grooves, the preset rotation angles are 0°, 120°, and 240° respectively, to cover the testing of all sealing grooves. At this time, another independent servo motor starts, and its output shaft is directly connected to a gear through a coupling. This gear meshes with a rack on the base fixed to the bottom side of the workpiece placement seat next to the guide rail in the testing area. The gear rolls along the rack under the drive of the motor, and at the same time, it drives the workpiece placement seat to rotate around the central axis of the valve core through meshing transmission. During the rotation, the angle encoder installed in the testing area collects the valve core rotation angle data in real time and feeds it back to the control system. When the angle value reaches the preset rotation angle, the motor stops, completing a single rotation positioning. If multiple surfaces need to be tested, the above gear transmission rotation action is repeated until all preset angle positioning is completed.

[0025] S300: The valve core to be tested is rotated and positioned by a sensor group to perform multi-dimensional detection, obtain multi-dimensional detection data of the valve core, identify defects in the multi-dimensional detection data of the valve core, and determine the valve core defect detection information.

[0026] Specifically, the sensor group first performs multi-dimensional detection on the valve core under test, which is in a rotating positioning state. The sensor group is a set of multiple types of sensors pre-set in the scheme, including laser displacement sensor, vision sensor, eddy current sensor and pressure flow sensor. The valve core under test is stably placed in the conical positioning hole of the workpiece placement seat without displacement deviation. All sensors operate according to the pre-set synchronous acquisition strategy of the scheme, and finally generate multi-dimensional detection data of the valve core. This data covers the size value sequence output by the laser displacement sensor, the high-definition image data output by the vision sensor, the electromagnetic signal waveform data output by the eddy current sensor, and the pressure and flow rate change curve data output by the pressure flow sensor. All data are associated with the corresponding rotation angle information of the valve core to form a structured dataset.

[0027] Subsequently, defect identification is performed. Based on the characteristics of the valve core multidimensional detection data, such as the fluctuation range of dimensional data, the pixel distribution of image data, and the frequency characteristics of electromagnetic signals, data filters are initialized. For example, Kalman filtering is used for dimensional data, and Gaussian filtering is used for image data to filter out noise such as environmental vibration and light interference, obtaining usable valve core multidimensional detection data. Next, the usable data is multimodally aligned according to spatiotemporal information, and key features are extracted from each dimension of data to form a valve core multidimensional key feature set. Then, weighting factors are determined according to the output accuracy of each sensor, and the multidimensional key feature set is weighted accordingly. Multimodal fusion is used to obtain the valve core fusion key feature set. Finally, the fusion key feature set is analyzed by a pre-built valve core defect identifier to determine the valve core defect detection information. This information clearly includes whether the valve core has defects, the type of defects such as surface scratches, dimensional deviations, internal cracks, and sealing leaks, the degree of defects such as slight, moderate, and severe, and is divided according to preset thresholds, such as scratch depth <0.01mm as slight, 0.01-0.03mm as moderate, and >0.03mm as severe, and the location of defects such as the sealing groove at 120° on the side circumference of the valve core and the edge of the center hole on the end face of the valve core.

[0028] S400: Based on the valve core defect detection information, the valve core to be tested is judged as qualified, and the valve core to be tested is classified and placed according to the judgment result.

[0029] Specifically, the process of entering the qualified product judgment stage involves an automated comparison and analysis of valve core defect detection information based on preset valve core qualification judgment standards. These qualification judgment standards require a pre-built standard library tailored to the valve core's application scenario (e.g., hydraulic system valve cores, gas valve cores), industry specifications, and customer customization requirements. The library clearly defines the allowable thresholds for different defect types, such as surface scratch depth ≤0.01mm, dimensional deviation ≤±0.02mm, sealing leakage ≤5mL / min, and no internal cracks. Furthermore, it supports automatically calling the corresponding standard based on the specifications and model of the valve core to be tested, such as diameter and structural type. The control system first reads the model information of the valve core to be tested, matches and loads the qualified standard for that model from the standard library. The system first determines the threshold, then compares each indicator in the valve core defect detection information, such as defect type (surface scratch), defect degree (depth 0.008mm), and defect location (60° on the side circumference), with the threshold one by one. If all defect indicators are within the allowable range, the product is judged as qualified. If one or more indicators exceed the threshold but the deviation is small and does not affect the core function, the product is judged as awaiting re-inspection. If any indicator seriously exceeds the threshold or there is a critical functional defect, the product is judged as unqualified. This completes the qualified product judgment and generates a clear judgment result, namely, qualified product, product awaiting re-inspection, and unqualified product. At the same time, the system automatically records the judgment process data, such as comparison items, thresholds, and actual values, forming a traceable judgment log.

[0030] Subsequently, the valve cores to be tested are classified and placed according to the judgment results. Specifically, based on the judgment results generated from the qualified product determination, an automated actuator moves the valve cores to be tested to the corresponding storage areas. The equipment frame is divided into three independent storage areas according to the judgment results: a qualified product bin, a bin for products awaiting re-inspection, and a non-qualified product bin. Each bin is equipped with a photoelectric positioning sensor and an identification tag to ensure that the gripper can accurately identify the bin's position. The control system sends action commands to the gripper based on the judgment results. The gripper first moves to above the workpiece placement seat in the testing area and smoothly clamps the completed product. The system detects the valve core and then plans its movement trajectory based on the judgment result. When the gripper moves above the target hopper, the photoelectric positioning sensor of the hopper provides feedback that the hopper is in position. The gripper descends to the preset height and then releases, placing the valve core into the positioning slot inside the hopper to prevent collision damage. It also receives feedback from the pressure sensor at the bottom of the hopper that the valve core is in place. After confirming successful placement, the gripper resets to the standby position. If the hopper is full, the liquid level sensor inside the hopper will detect this and the system will issue an alarm signal and suspend the corresponding type of sorting and placement until an empty hopper is replaced, ensuring that there is no mixing or omission during the sorting and placement process.

[0031] Furthermore, the method of using gears to drive the valve core to be tested to rotate and position includes:

[0032] Based on the specifications and testing requirements of the valve core to be tested, a preset rotation angle is determined; the gear is driven to rotate by a motor, wherein the output shaft of the motor is connected to the gear, and the gear meshes with the rack at the bottom of the workpiece placement seat. The gear rolls along the rack, causing the valve core to be tested on the workpiece placement seat to rotate and be positioned to the preset rotation angle.

[0033] Specifically, the testing requirements depend on the specifications and model of the valve core to be tested. This refers to the physical structural parameters of the valve core, such as its diameter, the number of circumferential sealing grooves, the spacing of the holes, and end face features. Different specifications and models of valve cores have different spatial positions for the parts to be tested. For example, a valve core with a diameter of 20mm and three evenly distributed sealing grooves has a completely different circumferential spacing between its testing surfaces compared to a valve core with a diameter of 15mm and two sealing grooves. The testing requirements also need to be considered, specifying the types of valve core defects that require focused inspection. For example, when inspecting scratches on the valve core's side circumference, the inspection should cover the entire circumference; when inspecting the depth of the sealing grooves, the inspection should be aligned with the center of each sealing groove; and when inspecting the coaxiality of the holes, the inspection should be aligned with the hole axis. In the linear direction, the preset rotation angle is determined by the parameter matching module of the equipment control system. For example, if the valve core to be tested is a model with three circumferentially evenly distributed sealing grooves, and the testing requirement is to fully test the depth and surface flatness of each sealing groove, then the preset rotation angle needs to be set to 0°, 120°, and 240° to ensure that each sealing groove can be accurately aligned with the detection axis of the sensor group. If the testing requirement is to check for overall scratches on the side circumference of the valve core, the preset rotation angle can be set in 24 increments at 15° intervals to achieve no dead angle coverage of the circumference. After the parameters are determined, the preset rotation angle data will be synchronously transmitted to the motor control unit.

[0034] Subsequently, the motor output shaft is rigidly connected to the gear via a coupling. This gear forms a meshing transmission structure with the rack at the bottom of the workpiece placement seat. When the motor drives the gear to rotate, the gear rolls along the rack tooth surface under the constraint of the rack. At the same time, due to the indirect fixed relationship between the gear and the bottom of the workpiece placement seat, the rolling process will drive the workpiece placement seat to rotate synchronously around the central axis of the valve core to be tested. During the rotation, the angle encoder installed in the detection area will feed back the angle signal to the control system in real time. When the deviation between the feedback angle and the preset rotation angle is less than 0.01°, the control system immediately sends a braking command to the motor, and the motor stops rotating. At this time, the valve core to be tested stays at the preset rotation angle position, completing a single rotation positioning. If multi-position rotation positioning is required, the above process is repeated until all preset rotation angles have been executed.

[0035] Furthermore, the obtained multidimensional detection data of the valve core includes:

[0036] Configure a sensor group, which includes a laser displacement sensor, a vision sensor, an eddy current sensor, and a pressure flow sensor; construct a synchronous acquisition strategy, in which all sensors simultaneously detect the valve core according to a preset acquisition frequency; use the synchronous acquisition strategy to perform synchronous multidimensional detection on the rotated and positioned valve core under test through the sensor group to obtain multidimensional detection data of the valve core.

[0037] Specifically, the first step is to configure a sensor group, a high-precision sensor suite integrated to meet the multi-dimensional detection needs of the valve core. This group includes four types of sensors: a laser displacement sensor, which is installed on the side and top of the valve core, aligning with the detection areas for the valve core's outer diameter, sealing groove depth, and end-face flatness, respectively; a vision sensor, installed directly above and to the side of the valve core, with adjustable light source brightness to eliminate glare interference from the valve core surface; and an eddy current sensor, installed close to the non-magnetic metal parts of the valve core, such as the valve body. It emits an alternating electromagnetic field into the valve core and determines hidden defects such as internal cracks and inclusions based on changes in the phase and amplitude of the eddy current signal. Finally, a pressure flow sensor connects to the valve core's inlet and outlet via a customized interface, allowing the introduction of compressed air or hydraulic oil at a preset pressure, selected according to the valve core type, to monitor fluid pressure loss and flow rate changes in real time to assess sealing performance. During the configuration process, the installation angle and distance of each sensor need to be adjusted using calibration fixtures to ensure that the detection area completely covers the valve core to be detected, and that the sensor output signals are all connected to the same data acquisition card to achieve synchronous acquisition.

[0038] Subsequently, a synchronous acquisition strategy is constructed. This strategy ensures that all sensors collect data from the same detection area of ​​the valve core at the same time dimension. Its core is the unification of timestamps and frequency matching. First, a preset acquisition frequency needs to be determined. This frequency should be set based on the dwell time of the valve core after rotation and positioning, as well as the required detection accuracy. For example, setting it to 100Hz means acquiring 100 sets of data per second. This ensures sufficient data to support analysis at a single angle while avoiding data redundancy due to excessive frequency. Second, the acquisition trigger signals of each sensor need to be synchronously calibrated through the control system, ensuring that all sensors start acquisition at the same millisecond time point. This ensures that the data collected by different sensors corresponds to the same state of the valve core at the current rotation angle. For example, while the laser displacement sensor collects dimensional data, the vision sensor captures the surface image at that angle, avoiding misalignment of data and detection area due to time differences. Simultaneously, the strategy includes a preset data caching mechanism to temporarily store the real-time data collected by each sensor in the format of rotation angle-timestamp-sensor type, preventing data loss due to data transmission delays.

[0039] Finally, synchronous multidimensional detection is performed to obtain multidimensional detection data of the valve core. When the control system receives a signal that the valve core has rotated to the set angle and locked, it immediately sends a synchronous acquisition command to all sensors. Each sensor starts detection simultaneously according to a preset acquisition frequency. The laser displacement sensor continuously outputs a sequence of dimensional values ​​such as the outer diameter and sealing groove depth of the valve core at the current angle; the vision sensor captures images of the valve core surface at a set frequency and converts them into digital image data; the eddy current sensor outputs electromagnetic signal waveform data in real time and records the electromagnetic response of the internal structure of the valve core; the pressure flow sensor monitors and records the pressure change curve and real-time flow data in the flow channel. All data collected by the sensors are synchronously transmitted to the system database through the data acquisition card, and each set of data is associated with a corresponding rotation angle identifier and timestamp identifier to ensure that the data can be traced back to the specific detection part of the valve core. After the acquisition at the preset angle is completed, the valve core continues to rotate to the next preset angle, and the above synchronous acquisition process is repeated until the detection of all preset rotation angles is completed. Finally, a structured multidimensional detection data set of the valve core is formed, covering the valve core size, appearance, internal structure, sealing performance, and associated angle information.

[0040] Furthermore, determining the valve core defect detection information includes:

[0041] Based on the characteristic information of the valve core multidimensional detection data, a data filter is initialized; the data filter is used to preprocess the valve core multidimensional detection data to obtain usable valve core multidimensional detection data; the usable valve core multidimensional detection data is then subjected to multimodal alignment and defect identification to determine valve core defect detection information.

[0042] Specifically, the first step is to initialize the data filter based on the characteristic information of the valve core's multidimensional detection data. This characteristic information includes the inherent properties and noise characteristics of different dimensions of data, such as the dimensional value sequence output by the laser displacement sensor (susceptible to environmental vibration interference and exhibiting linear fluctuations), the image data output by the vision sensor (susceptible to changes in lighting and containing isolated noise points), the electromagnetic signal waveform output by the eddy current sensor (susceptible to electromagnetic interference and containing high-frequency noise), and the pressure-flow curve output by the pressure-flow sensor (susceptible to fluid pulsation and exhibiting periodic fluctuations). The data filter is designed to address these different noise types. The signal processing tool is designed to retain valid detection information and filter out irrelevant interference. The initialization process requires matching the corresponding filtering algorithm with the data characteristics. For example, a Kalman filter is selected for linear fluctuations in size data, and deviations are predicted and corrected in real time through state equations. A Gaussian filter is selected for isolated noise in image data, and smoothing is achieved by weighted averaging of neighboring pixel gray levels. A wavelet filter is selected for high-frequency noise in electromagnetic signals, and signals and noise are separated synchronously in the time and frequency domains. A moving average filter is selected for pulsations in pressure-flow curves, and fluctuations are weakened by averaging continuous data points, ensuring that there is a suitable filtering scheme for each dimension of data.

[0043] Subsequently, the initialized filters are used to perform targeted preprocessing on the valve core multidimensional detection data. The numerical sequences of laser displacement diameter and sealing groove depth are input into a Kalman filter, and vibration deviations are corrected in real time through preset noise covariance to output stable dimensional data. The visual images are input frame by frame into a Gaussian filter to remove light noise and retain the contours of defects such as scratches and burrs. The eddy current electromagnetic waveform is input into a wavelet filter to decompose and eliminate high-frequency interference and retain the abnormal response of internal cracks. The pressure and flow data are input into a moving average filter to reduce pulsation and highlight the pressure and flow anomalies of sealing leakage. After preprocessing, usable valve core multidimensional detection data is obtained, which is a structured dataset with interference filtered out, core features retained, and uniform format.

[0044] Next, multimodal alignment and defect identification are performed. Multimodal alignment refers to unifying the spatiotemporal association of data from different sensors. Due to differences in sensor installation positions and response speeds, the original data may have timestamp misalignments or mismatched detection locations. For example, when a laser measures a certain angle, the vision system may not have captured that location yet. Therefore, the preset rotation angle of the valve core rotation positioning is used as a benchmark. That is, each angle corresponds to a specific circumferential detection position. Combined with the data timestamps, the size, image, electromagnetic, and pressure data at the same angle are associated and matched to ensure that all data correspond to the same part of the valve core, forming standard data that is spatiotemporally synchronized. Subsequently, key features of each dimension are extracted from the standard data, including the deviation value of size from the qualified threshold, the area / contour of the image defect region, the electromagnetic abnormal peak value, and the pressure leakage. The features are integrated through a weighted fusion algorithm based on sensor accuracy and then input into a pre-trained valve core defect identifyer. After model classification and judgment, the valve core defect detection information is finally determined. This information clearly includes whether the valve core has a defect; the defect type, including surface scratches, dimensional deviations, internal cracks, and sealing leaks; the defect severity, including minor, moderate, and severe, and the rotation angle corresponding to the defect location.

[0045] Furthermore, the step of performing multimodal alignment and defect identification on the available valve core multidimensional detection data to determine valve core defect detection information includes:

[0046] The available valve core multidimensional detection data is aligned in a multimodal manner according to spatiotemporal information to obtain standard valve core multidimensional detection data; key features are extracted from each dimension of the standard valve core multidimensional detection data to obtain a valve core multidimensional key feature set; multimodal fusion and defect identification are performed based on the valve core multidimensional key feature set to determine valve core defect detection information.

[0047] Specifically, the available multidimensional detection data of the valve core is multimodally aligned according to spatiotemporal information. This spatiotemporal information includes: spatial information, namely the rotational positioning angle of the valve core under test in the detection area, such as preset rotation angles like 0°, 120°, and 240°, each angle corresponding to a specific circumferential detection part of the valve core; and temporal information, namely the unified timestamp when each sensor collects data, synchronized and calibrated by the system clock to ensure millisecond-level time accuracy. During the alignment process, the control system uses the rotational positioning angle as a spatial reference to categorize the data collected by all sensors at the same angle; then, using the timestamp as a time reference, it eliminates time-displaced data caused by sensor response delays. Finally, the size, image, electromagnetic, pressure, and flow data under the same spatiotemporal dimension are integrated into a set of structured data. The integrated data from all angles is summarized to form standard multidimensional detection data of the valve core. This data eliminates the spatiotemporal deviation of multi-sensor data, ensuring that each set of data accurately corresponds to the same detection time state of a specific part of the valve core.

[0048] Subsequently, based on the characteristics of different dimensions of the standard valve core multidimensional detection data, key features were extracted to form a valve core multidimensional key feature set. Specifically, for the dimensional data output by the laser displacement sensor, dimensional accuracy features were extracted by calculating the deviation between the actual measured value and the qualified standard value, and the fluctuation amplitude of multiple measurements at the same location; for the image data output by the vision sensor, defect contours were extracted using edge detection algorithms, such as the Canny algorithm, and the defect area, perimeter, and grayscale contrast were calculated by combining pixel statistics to obtain appearance defect features; for the electromagnetic signal waveform output by the eddy current sensor, the time domain signal was converted into a frequency domain signal using Fourier transform, and abnormal peaks and phase shifts in the signal were extracted to capture hidden defect features such as internal cracks and inclusions; for the pressure flow curve output by the pressure flow sensor, sealing performance features were extracted by calculating the difference between the actual pressure loss value and the standard value, and the leakage amount during the flow stabilization stage. The key features extracted from all dimensions were organized in the format of sensor type-rotation angle-feature category to form a valve core multidimensional key feature set representing all dimensions of valve core defects.

[0049] Finally, based on preset sensor weighting factors, a weighted summation algorithm is used to fuse similar features in the multidimensional key feature set of the valve core, forming a valve core fusion key feature set that comprehensively reflects the overall state of the valve core, avoiding the limitations of single sensor data. Subsequently, defect identification is performed. The valve core fusion key feature set is input into a pre-built valve core defect identifyr. The identifyr, through feature matching and model inference, outputs complete information including the presence, type, severity, and location of defects, forming valve core defect detection information.

[0050] Furthermore, the step of performing multimodal fusion and defect identification based on the multidimensional key feature set of the valve core to determine valve core defect detection information includes:

[0051] Based on the output accuracy of each sensor in the sensor group, a sensor weighting factor is determined; multimodal fusion is performed on the valve core multidimensional key feature set according to the sensor weighting factor to obtain a valve core fusion key feature set; a valve core defect identifyr is constructed, and defect identification is performed on the valve core fusion key feature set based on the valve core defect identifyr to determine the valve core defect detection information.

[0052] Specifically, sensor weighting factors are determined based on the output accuracy of each sensor in the sensor group. When determining the weighting factors, the principle of higher accuracy corresponds to a larger weight percentage is adopted. Weights are allocated through normalization calculations, converting the base weights of each sensor into coefficients that sum to 1, ensuring that the weight allocation conforms to a total percentage of 100%. For example, the weighting factor for the laser displacement sensor is set to 0.3, the vision sensor to 0.25, the eddy current sensor to 0.25, and the pressure flow sensor to 0.2, with the sum of all sensor weighting factors being 1.

[0053] Subsequently, multimodal fusion is performed on the valve core's multidimensional key feature set according to sensor weighting factors. The valve core's multidimensional key feature set is composed of core defect characterization data extracted from standard valve core multidimensional detection data, including dimensional deviation values ​​corresponding to laser displacement sensors, appearance defect parameters corresponding to vision sensors, internal defect signal values ​​corresponding to eddy current sensors, and sealing performance parameters corresponding to pressure flow sensors. The fusion process employs a weighted summation algorithm, multiplying the quantified value of each dimension's key feature by the corresponding sensor's weighting factor, and then summing all weighted values ​​to form a fused key feature set that comprehensively reflects the valve core's full-dimensional defect state. For example, the weighted value for a valve core's dimensional deviation is 0.005 × 0.3 = 0.0015, the weighted value for scratch length is 0.8 × 0.25 = 0.2, the weighted value for electromagnetic anomaly peak value is 0.3 × 0.25 = 0.075, and the weighted value for leakage flow is 3 × 0.2 = 0.6. After summing, the fused feature value is 0.8765.

[0054] Finally, a valve core defect identifyer is constructed. This identifyer is based on a deep neural network such as a CNN-LSTM hybrid model. First, a large dataset of historical valve core defect detection data is collected, covering multi-dimensional key feature sets of valve cores of different specifications and models, along with corresponding actual defect situations. The dataset is labeled to clarify the defect type, severity, and location of each sample, forming a labeled valve core defect detection sample set. Then, the sample set is divided into a training set and a validation set in a 7:3 ratio. The deep neural network is trained using the training set, and the model parameters are continuously adjusted through backpropagation to reduce prediction errors. Simultaneously, cross-validation optimization is performed using the validation set to avoid model overfitting and ensure accurate identification of unseen samples until the model's recognition accuracy stabilizes above 99%, completing the construction of the valve core defect identifyer. The obtained fused key feature set of the valve core is then input into the identifyer. The identifyer performs feature matching (comparing the fused features with defect feature templates in the training set) and model inference (calculating the defect probability using the trained parameters). The output is a complete judgment result including the presence, type, severity, and location of the defect, ultimately forming the valve core defect detection information.

[0055] Furthermore, the construction of the valve core defect identifier includes:

[0056] Collect a historical defect detection dataset of valve cores, identify the defect type and degree of the historical defect detection dataset to obtain a valve core defect detection sample set; use a deep neural network structure to train defect recognition and optimize the valve core defect detection sample set to construct the valve core defect recognizer.

[0057] Specifically, to build a valve core defect identifier, the first step is to collect and label the historical defect detection dataset of valve cores. The historical defect detection dataset of valve cores refers to the original data set accumulated from past valve core testing operations, covering valve core testing information in multiple scenarios. During collection, it is necessary to ensure the comprehensiveness of the data coverage, including valve cores of different specifications and models to be tested, as well as various typical defect types. At the same time, it is necessary to simultaneously record the multi-dimensional testing data of the corresponding valve cores and the actual defect situation after manual disassembly or verification by professional equipment. After data collection, the dataset is labeled with defect type and defect severity. Defect type labeling follows a pre-defined classification standard, such as four main categories: surface defects, dimensional defects, internal structural defects, and sealing performance defects. Each main category is further subdivided into specific subtypes, such as surface defects including scratches, burrs, and dents, clearly labeling each data point with its defect category. Defect severity labeling uses quantitative thresholds set according to industry standards and practical application requirements. For example, surface scratches with a depth <0.01mm are labeled as minor, 0.01-0.03mm as moderate, and >0.03mm as severe; dimensional deviations within ±0.01mm are considered minor, ±0.01-0.02mm as moderate, and above ±0.02mm as severe, ensuring consistency in severity classification. After labeling, the original dataset is transformed into a structured valve core defect detection sample set with multi-dimensional detection data, defect type labels, and defect severity labels. Each sample has clearly defined input features and output labels.

[0058] Subsequently, a deep neural network structure was used to train and optimize the valve core defect detection sample set for defect identification. The obtained valve core defect detection sample set, i.e., a structured dataset labeled with defect types and degrees, includes dimensional data from laser displacement sensors, image data from vision sensors, electromagnetic signal data from eddy current sensors, and pressure and flow data from pressure flow sensors, as well as labels for surface scratches, dimensional deviations, internal cracks, and sealing leaks, and severity labels (minor, moderate, and severe). The image data was grayscaled and pixel resolution was unified. The time-series data such as size, electromagnetic, pressure, and flow were normalized using Min-Max and mapped to the 0-1 range to eliminate the interference of data magnitude differences on training. Then, the sample set was randomly divided into a training set (7:3 ratio) for model parameter learning and an initial validation set for preliminary monitoring of training effects.

[0059] The preprocessed training data is then input into the deep neural network in batches. The error is calculated using the cross-entropy loss function. The Adam optimizer is used, with the initial learning rate set to 0.001 and adjusted every 10 rounds by a decay factor of 0.9 to balance training speed and convergence stability. The network weights and bias parameters are iteratively updated. During training, the initial validation set is input into the model in real time, and the changes in the validation set accuracy and loss value are monitored. If the validation set loss value does not decrease for 10 consecutive rounds, the model is considered to have converged, and training is paused. At this point, the model initially has the ability to identify defects.

[0060] Finally, cross-validation optimization was performed using a 5-fold cross-validation method. The complete valve core defect detection sample set was divided into 5 non-overlapping subsets. Each time, 4 subsets were selected as temporary training sets and 1 subset as test sets, and the training and testing process was repeated 5 times. After each test, the model's defect type recognition accuracy and severity judgment accuracy on the test set were recorded, and the average accuracy of the 5 tests was calculated. If the initial average accuracy was lower than 98%, the network hyperparameters were adjusted accordingly, such as increasing the number of CNN convolutional layers to 4 layers to improve image feature extraction capabilities, or adjusting the number of LSTM hidden units to 256 to optimize temporal feature capture effects. This process continued until the average accuracy of the 5-fold cross-validation stabilized above 99%, and the difference in test accuracy among the subsets was less than 2%, ensuring that the model had good generalization ability. At this point, the training and optimization of the deep neural network were completed, and the final network structure and parameters were saved, thus completing the construction of a valve core defect identifyer that can automatically identify valve core defects.

[0061] In summary, the intelligent valve core detection method based on gear rotation positioning provided in this application has the following technical effects:

[0062] This application places a tray containing the valve core to be tested on a top plate. A bottom cylinder pushes the top plate upwards, and a screw drive moves the tray horizontally to the loading area. A gripper picks up the valve core from the tray and places it on a workpiece holder. A motor is started, and a guide rail drive moves the workpiece holder to the testing area. In the testing area, gears rotate and position the valve core. A sensor array performs multi-dimensional detection on the rotated valve core, obtaining multi-dimensional detection data. Defects are identified from this data to determine valve core defect information. Based on the defect information, the valve core is judged as a qualified product, and then classified and placed according to the judgment results. Through fully automated gear rotation positioning, synchronous multi-dimensional detection, and intelligent defect recognition, the technical effect of improving the coverage, accuracy, efficiency, and reliability of valve core detection is achieved.

[0063] Example 2, as Figure 2As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a valve core intelligent detection system based on gear rotation positioning, the system comprising:

[0064] The feeding module 11 places a tray containing the valve core to be tested on a top plate, raises the top plate using a bottom cylinder, and moves the tray to the loading area via a screw drive. The positioning module 12 uses grippers to remove the valve core to be tested from the tray and places it on a workpiece placement seat. A motor is started, and the workpiece placement seat is moved to the testing area via a guide rail drive. Within the testing area, gears rotate the valve core to be tested for positioning. The testing module 13 uses a sensor array to perform multi-dimensional detection on the rotated and positioned valve core to obtain multi-dimensional detection data. Defects are identified from this data to determine valve core defect detection information. The classification module 14 determines the valve core to be tested as qualified based on the defect detection information and classifies and places the valve cores according to the determination results.

[0065] Furthermore, the positioning module 12 in the intelligent valve core detection system based on gear rotation positioning is used for:

[0066] Based on the specifications and testing requirements of the valve core to be tested, a preset rotation angle is determined; the gear is driven to rotate by a motor, wherein the output shaft of the motor is connected to the gear, and the gear meshes with the rack at the bottom of the workpiece placement seat. The gear rolls along the rack, causing the valve core to be tested on the workpiece placement seat to rotate and be positioned to the preset rotation angle.

[0067] Furthermore, the detection module 13 in the intelligent valve core detection system based on gear rotation positioning is used for:

[0068] Configure a sensor group, which includes a laser displacement sensor, a vision sensor, an eddy current sensor, and a pressure flow sensor; construct a synchronous acquisition strategy, in which all sensors simultaneously detect the valve core according to a preset acquisition frequency; use the synchronous acquisition strategy to perform synchronous multidimensional detection on the rotated and positioned valve core under test through the sensor group to obtain multidimensional detection data of the valve core.

[0069] Furthermore, the detection module 13 in the intelligent valve core detection system based on gear rotation positioning is also used for:

[0070] Based on the characteristic information of the valve core multidimensional detection data, a data filter is initialized; the data filter is used to preprocess the valve core multidimensional detection data to obtain usable valve core multidimensional detection data; the usable valve core multidimensional detection data is then subjected to multimodal alignment and defect identification to determine valve core defect detection information.

[0071] Furthermore, the detection module 13 in the intelligent valve core detection system based on gear rotation positioning is also used for:

[0072] The available valve core multidimensional detection data is aligned in a multimodal manner according to spatiotemporal information to obtain standard valve core multidimensional detection data; key features are extracted from each dimension of the standard valve core multidimensional detection data to obtain a valve core multidimensional key feature set; multimodal fusion and defect identification are performed based on the valve core multidimensional key feature set to determine valve core defect detection information.

[0073] Furthermore, the detection module 13 in the intelligent valve core detection system based on gear rotation positioning is also used for:

[0074] Based on the output accuracy of each sensor in the sensor group, a sensor weighting factor is determined; multimodal fusion is performed on the valve core multidimensional key feature set according to the sensor weighting factor to obtain a valve core fusion key feature set; a valve core defect identifyr is constructed, and defect identification is performed on the valve core fusion key feature set based on the valve core defect identifyr to determine the valve core defect detection information.

[0075] Furthermore, the detection module 13 in the intelligent valve core detection system based on gear rotation positioning is also used for:

[0076] Collect a historical defect detection dataset of valve cores, identify the defect type and degree of the historical defect detection dataset to obtain a valve core defect detection sample set; use a deep neural network structure to train defect recognition and optimize the valve core defect detection sample set to construct the valve core defect recognizer.

[0077] Example 3, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides an electronic device, the electronic device comprising:

[0078] The memory 303 is used to store executable instructions; the processor 302 is used to execute the executable instructions stored in the memory 303 to implement the intelligent valve core detection method based on gear rotation positioning.

[0079] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing the embodiments of the present invention. Figure 3The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention. This electronic device is in the form of a general-purpose computing device, and its components may include, but are not limited to, an input device 301, a processor 302, a memory 303, and an output device 304. The processor 302 may be one or more; the memory 303 may include a computer-readable medium and at least one program product having a set (at least one) of program modules configured to perform the functions of the embodiments of this application.

[0080] The memory 303 shown in this embodiment of the invention can be any combination of one or more computer-readable media. The computer-readable storage media can be, but is not limited to, infrared, semiconductor systems, devices or components, or any combination thereof, used to store software programs, computer-executable programs and modules, such as the program instructions / modules corresponding to an optimized manufacturing method for a composite copper-aluminum block in this embodiment of the invention. The processor 302 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 303, that is, to realize the above-mentioned intelligent valve core detection method based on gear rotation positioning.

[0081] The intelligent valve core detection system based on gear rotation positioning provided in this embodiment of the invention can execute the intelligent valve core detection method based on gear rotation positioning provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0082] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0083] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A valve core intelligent detection method based on gear rotation positioning, characterized in that, The method includes: Place the tray containing the valve core to be tested on the top plate, push the top plate up through the bottom cylinder, and move the tray to the feeding area through the screw drive device; The valve core to be tested is picked up from the material tray by the gripper and placed on the workpiece placement seat. The motor is started and the workpiece placement seat is moved to the testing area through the guide rail transmission device. In the testing area, the valve core to be tested is rotated and positioned by the gear. The valve core to be tested is subjected to multi-dimensional detection by a sensor group in a rotating position to obtain multi-dimensional detection data of the valve core. Defect identification is performed on the multi-dimensional detection data of the valve core to determine the defect detection information of the valve core. Based on the valve core defect detection information, the valve core to be tested is judged to be qualified, and the valve core to be tested is classified and placed according to the judgment result.

2. The intelligent valve core detection method based on gear rotation positioning as described in claim 1, characterized in that, The method of using gears to drive the valve core to be tested to rotate and position includes: Determine the preset rotation angle based on the specifications and testing requirements of the valve core to be tested; The gear is driven to rotate by a motor, wherein the output shaft of the motor is connected to the gear, and the gear meshes with the rack at the bottom of the workpiece placement seat. The gear rolls along the rack, causing the valve core to be tested on the workpiece placement seat to rotate and be positioned to the preset rotation angle.

3. The intelligent valve core detection method based on gear rotation positioning as described in claim 1, characterized in that, The obtained multidimensional detection data of the valve core includes: Configure a sensor group, which includes a laser displacement sensor, a vision sensor, an eddy current sensor, and a pressure flow sensor; A synchronous acquisition strategy is constructed, wherein all sensors simultaneously detect the valve core according to a preset acquisition frequency; The synchronous acquisition strategy is used to perform synchronous multidimensional detection on the rotating and positioned valve core under test through the sensor group to obtain the multidimensional detection data of the valve core.

4. The intelligent valve core detection method based on gear rotation positioning as described in claim 1, characterized in that, The determination of valve core defect detection information includes: Initialize the data filter based on the characteristic information of the valve core multidimensional detection data; The data filter is used to preprocess the valve core multidimensional detection data to obtain usable valve core multidimensional detection data; The available valve core multidimensional detection data are subjected to multimodal alignment and defect identification to determine valve core defect detection information.

5. The intelligent valve core detection method based on gear rotation positioning as described in claim 4, characterized in that, The step of performing multimodal alignment and defect identification on the available valve core multidimensional detection data to determine valve core defect detection information includes: The available valve core multidimensional detection data is multimodally aligned according to spatiotemporal information to obtain standard valve core multidimensional detection data. Key features are extracted from each dimension of the standard valve core multidimensional detection data to obtain the valve core multidimensional key feature set. Based on the multi-dimensional key feature set of the valve core, multi-modal fusion and defect identification are performed to determine the valve core defect detection information.

6. The intelligent valve core detection method based on gear rotation positioning as described in claim 5, characterized in that, The process of performing multimodal fusion and defect identification based on the multidimensional key feature set of the valve core to determine valve core defect detection information includes: The sensor weighting factor is determined based on the output accuracy of each sensor in the sensor group; The valve core multidimensional key feature set is fused using the sensor weighting factors to obtain the valve core fused key feature set. A valve core defect identifier is constructed, and the valve core defect identifier is used to identify defects in the valve core fused key feature set to determine the valve core defect detection information.

7. The intelligent valve core detection method based on gear rotation positioning as described in claim 6, characterized in that, The valve core defect identifier includes: Collect a historical defect detection dataset of the valve core, and identify the defect type and degree of the historical defect detection dataset of the valve core to obtain a valve core defect detection sample set; The valve core defect identifyer is constructed by using a deep neural network structure to train and optimize the defect identification of the valve core defect detection sample set through defect identification and cross-validation.

8. A valve core intelligent detection system based on gear rotation positioning, characterized in that, The system is used to implement the intelligent valve core detection method based on gear rotation positioning as described in any one of claims 1-7, the system comprising: The feeding module places a tray containing the valve core to be tested on the top plate, pushes the top plate up through the bottom cylinder, and moves the tray to the loading area through the screw drive device. The positioning module uses grippers to pick up the valve core to be tested from the material tray and place it on the workpiece placement seat. The motor is started, and the workpiece placement seat is moved to the testing area through the guide rail transmission device. In the testing area, the valve core to be tested is rotated and positioned by the gear. The detection module performs multi-dimensional detection on the rotating and positioned valve core to be detected through a sensor group, obtains multi-dimensional detection data of the valve core, identifies defects in the multi-dimensional detection data of the valve core, and determines the valve core defect detection information. The classification module determines the quality of the valve core to be tested based on the valve core defect detection information, and classifies and places the valve core to be tested according to the determination result.

9. An electronic device, characterized in that, The electronic device includes: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the intelligent valve core detection method based on gear rotation positioning as described in any one of claims 1 to 7.

Images