A multi-sensor fusion-based coil geometry online detection and closed-loop control system and method
Patent Information
- Application Number
- CN202610528502.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-21
- Publication Date
- 2026-09-04
AI Technical Summary
传统盘管生产依赖人工抽检和离线测量,存在显著技术缺陷:一是检测滞后性,离线测量无法实时反馈生产过程的质量波动,当发现尺寸超差时,已产生大批量不合格品,造成材料与产能的大量浪费;二是抽样局限性,人工抽检无法实现100%全检,存在漏检风险,部分隐性尺寸缺陷流入下游环节后会引发设备故障等严重后果;三是效率与精度短板,人工测量劳动强度大、速度慢,难以适配高速连续化生产节奏,且人工判断易受主观因素影响,测量数据缺乏统一客观标准;四是无闭环调控能力,现有生产设备多依靠人工经验设定工艺参数,不具备检测-反馈-调整的联动机制,无法从源头预防不合格品产生
该一种基于多传感融合的盘管几何尺寸在线检测与闭环控制系统及方法,通过三组激光扫描传感器重构盘管截面完整轮廓,可精准获取微米级的外径、圆度数据;搭配双工业相机及专用光源,同步实现节距测量、盘形规整度监测和表面缺陷识别,覆盖了盘管核心质量指标,实现100%在线非接触全检,相比传统人工抽检和单一传感检测,既消除了漏检风险,又突破了检测维度的限制,同时非接触式测量避免了对盘管表面的损伤,保障了产品表面质量,为高端盘管产品的质量管控提供了可靠技术支撑;
Smart Images

Figure CN122682948A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of automated manufacturing and precision measurement technology, specifically relating to an online detection and closed-loop control system and method for the geometric dimensions of coils based on multi-sensor fusion. Background Technology
[0002] As a key basic component in the industrial field, the consistency of geometric dimensions such as outer diameter, roundness, and pitch of coils directly determines the performance and reliability of the product. Traditional coil production relies on manual sampling and offline measurement, which has significant technical defects: First, there is a lag in detection; offline measurement cannot provide real-time feedback on quality fluctuations in the production process. By the time dimensional deviations are detected, a large number of defective products have already been generated, resulting in a significant waste of materials and production capacity. Second, there are limitations in sampling; manual sampling cannot achieve 100% inspection, and there is a risk of missed inspections. Some hidden dimensional defects can flow into downstream processes and cause serious consequences such as equipment failure. Third, there are shortcomings in efficiency and accuracy; manual measurement is labor-intensive and slow, making it difficult to adapt to the pace of high-speed continuous production. Moreover, human judgment is easily affected by subjective factors, and measurement data lacks unified objective standards. Fourth, there is no closed-loop control capability; existing production equipment mostly relies on manual experience to set process parameters and lacks a linkage mechanism of detection-feedback-adjustment, making it impossible to prevent the generation of defective products from the source. Summary of the Invention
[0003] The purpose of this invention is to provide an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion, so as to solve the problems mentioned in the background art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion, comprising a drive production unit, a first adjustment unit, a second adjustment unit, a third adjustment unit, a fourth adjustment unit, a high-precision laser scanning measurement unit, a machine vision recognition unit, and a central processing control unit; The drive production unit includes a coiling machine and a coiling bracket disposed on one side of the coiling machine. A first fixing frame is fixedly installed on the top of the coiling machine, and a crossbeam is fixedly installed on one side of the first fixing frame. The crossbeam is located directly above the coiling bracket, and a second fixing frame is fixedly installed at one end of the crossbeam. The coiling machine is used to produce coil bodies, and the coiling bracket can support the coil bodies. The coiling machine is equipped with drive components such as a feeding servo motor, a bending die hydraulic valve, and a wire guide stepper motor. The first adjustment unit includes a first fixed frame, a first motor is fixedly installed on one side of the first fixed frame, the output end of the first motor passes through the inside of the first fixed frame and is fixedly connected to a bidirectional lead screw, one end of the bidirectional lead screw is rotatably connected to the inner wall of the first fixed frame, and two first adjustment blocks are threaded to the two ends of the outer surface of the bidirectional lead screw respectively, and a first sliding component is provided at the bottom of each of the two first adjustment blocks; The second, third, and fourth adjustment units each include a second fixed frame. A second motor is fixedly installed on the inner wall of the second fixed frame. A threaded rod is fixedly installed on the output end of the second motor. One end of the threaded rod is rotatably connected to the inner wall of the second fixed frame. A second adjustment block is threadedly connected to the outer surface of the threaded rod. A second sliding component is provided on one side of the second adjustment block. The first fixing frame is horizontally installed on the top of the coiling machine. The second fixing frame included in the second adjustment unit is vertically installed inside the first fixing frame and is perpendicular to the first fixing frame. The second fixing frame included in the third adjustment unit is horizontally installed at the bottom of the crossbeam. The second fixing frame included in the fourth adjustment unit is vertically installed inside the second fixing frame.
[0005] Through the above technical solution, the first adjustment unit can synchronously adjust the distance between the first adjustment blocks on both sides through bidirectional screw drive, adapting to the horizontal detection requirements of coils with different diameters. The second to fourth adjustment units can realize the vertical or horizontal position fine adjustment of the detection components through threaded rod drive, so that each detection unit can accurately align with the detection area of the coil. This solves the drawbacks of poor adaptability and fixed detection position of traditional detection devices, and provides a structural basis for multi-sensor precision detection.
[0006] In a preferred embodiment, the high-precision laser scanning measurement unit includes a first laser scanning sensor, a second laser scanning sensor, and a third laser scanning sensor. The first and second laser scanning sensors are respectively connected to two first adjustment blocks included in the first adjustment unit. The third laser scanning sensor is connected to a second adjustment block included in the second adjustment unit, and is used to reconstruct the complete outline of the coil body cross-section. The machine vision recognition unit includes a first industrial camera and a second industrial camera. The first industrial camera is connected to a second adjustment block included in the third adjustment unit, and the second industrial camera is connected to a second adjustment block included in the fourth adjustment unit. The first industrial camera is equipped with a backlight for measuring the pitch of the coil body, and the second industrial camera is equipped with a ring coaxial light source for monitoring the coil body's disc shape regularity and surface defects such as scratches and pits.
[0007] Through the above technical solution, the complementarity and full coverage of multi-sensor detection data are achieved. The three sets of laser scanning sensors are arranged in multiple directions, which can reconstruct the complete outline of the coil cross-section through multi-view data and accurately obtain core dimensional parameters such as outer diameter and roundness. The dual industrial cameras, with dedicated light sources, respectively realize high-precision measurement of pitch and visual monitoring of disc shape and surface defects. The precise dimensional detection of laser scanning and the macroscopic morphological detection of machine vision complement each other, solving the problem of limited dimensions of single-sensor detection and achieving comprehensive coverage of coil quality indicators.
[0008] In a preferred embodiment, the central processing control unit consists of an industrial computer (IPC) and a programmable logic controller (PLC). The first laser scanning sensor, the second laser scanning sensor, the third laser scanning sensor, the first industrial camera, and the second industrial camera are all electrically connected to the central processing control unit. The central processing control unit is electrically connected to the internal drive components of the coiling machine (such as a feeding servo motor, a bending die hydraulic valve, and a wire guide stepper motor) and is used to send the processed control commands to the corresponding drive components.
[0009] Through the above technical solution, a linkage hub for detection and execution is established. The industrial computer is responsible for complex data fusion and algorithm calculation, while the programmable logic controller is responsible for high-speed instruction issuance and logic control. This realizes the centralized reception and processing of multi-sensor data and the precise control of drive components, opening up the "detection-analysis-control" link and providing a core control carrier for closed-loop control. Compared with traditional independent detection systems, it has the ability to respond quickly to instructions and link production.
[0010] In a preferred embodiment, the first sliding assembly includes a first slider, one end of which is fixedly connected to a first adjusting block, and the inner wall of the first fixing frame is provided with a first groove that is adapted to the first slider.
[0011] Through the above technical solution, the sliding cooperation between the first slider and the first groove provides a guide limit for the first adjusting block, avoiding the adjustment block from shifting or jamming during the bidirectional screw drive process, ensuring that the laser scanning sensors on both sides can always maintain a symmetrical and horizontal spacing, improving the calibration accuracy of the horizontal detection position, and providing structural protection for the accurate measurement of outer diameter and roundness.
[0012] In a preferred embodiment, the second sliding component includes a second slider, one end of which is fixedly connected to a second adjusting block, and the inner wall of the second fixing frame is provided with a second sliding groove that is adapted to the second slider.
[0013] Through the above technical solution, the matching structure between the second slider and the second slide groove can restrict the rotational freedom of the second adjusting block, so that the adjusting block can only move in a straight line along the slide groove direction. This ensures the stability of the laser scanning sensor and industrial camera during the position fine adjustment process, avoids the deviation of detection data caused by component shaking, and improves the measurement stability of the entire detection system.
[0014] In a preferred embodiment, the central processing control unit executes the following data fusion algorithm: a. Perform Kalman filtering on the outer diameter and roundness measurement data acquired by the high-precision laser scanning measurement unit to eliminate random noise; b. Perform sub-pixel edge extraction on the pitch and disk-related visual images acquired by the machine vision recognition unit to improve the accuracy of dimensional measurement; c. Establish a comprehensive quality evaluation model based on fuzzy logic or neural network, and integrate multiple parameters such as outer diameter, roundness, pitch and surface defect level into a comprehensive quality index by weighting and fusing them together. d. Built-in sampling point judgment rules, which can preset the threshold monitoring conditions for N consecutive sampling points. When a certain size parameter exceeds the warning range for N consecutive sampling points, the corresponding deviation analysis and compensation logic is triggered (the value of N can be preset according to the production accuracy requirements, coil material characteristics and equipment response speed. For example, setting it to 5 sampling points ensures that the system is sensitive to continuous quality fluctuations and does not respond to occasional fluctuations, thus improving the control accuracy).
[0015] Through the above technical solutions, Kalman filtering can effectively filter environmental interference noise in laser measurement data, ensuring the authenticity of dimensional data. Subpixel edge extraction technology breaks through the accuracy bottleneck of traditional pixel-level recognition, greatly improving the accuracy of visual measurement parameters such as pitch. Fuzzy logic or neural network models realize the weighted fusion of multi-dimensional quality parameters, which can output intuitive comprehensive quality indicators, solving the problem of independent evaluation of multi-source data and the inability to quantify comprehensive quality, and providing a scientific basis for quality decision-making.
[0016] In a preferred embodiment, after receiving the control signal from the central processing control unit, the internal drive assembly of the coiling machine adjusts at least one of the following parameters in real time to compensate for coil geometric deviations: a. The speed of the feeding servo motor; b. Hydraulic or servo pressure of the bending die; c. The ratio of the cable guide's moving speed to the spindle speed.
[0017] By using the above technical solutions, corresponding control parameters can be matched for different types of dimensional deviations. For example, when the outer diameter is out of tolerance, the speed of the feeding servo motor can be adjusted; when the roundness is not up to standard, the pressure of the bending die can be adjusted; and when the pitch is abnormal, the speed ratio between the wire guide and the spindle can be optimized. Production deviations can be quickly corrected, and quality problems can be eliminated in the production process. Compared with the traditional method of manually stopping the machine for adjustment, the control efficiency and accuracy are greatly improved.
[0018] A method for online detection and closed-loop control of coil geometry based on multi-sensor fusion, utilizing the aforementioned online detection and closed-loop control system for coil geometry based on multi-sensor fusion, includes the following steps: S1: The online detection and closed-loop control system for coil geometry based on multi-sensor fusion is initialized by loading preset coil body specification parameters and tolerance thresholds through the parameter import function of the central processing control unit. S2: Start the first motor to drive the bidirectional lead screw to rotate, which in turn moves the two first adjustment blocks to calibrate the horizontal distance between the first laser scanning sensor and the second laser scanning sensor; start each second motor to drive the corresponding threaded rod to rotate, which in turn moves the second adjustment block to adjust the vertical height of the third laser scanning sensor, the position of the first industrial camera, and the vertical height of the second industrial camera, so that each detection unit can accurately adapt to the detection requirements of the coil body. S3: A high-precision laser scanning measurement unit and a machine vision recognition unit perform real-time data acquisition; S4: The central processing and control unit integrates multi-source data to calculate key dimensions such as the outer diameter, roundness, and pitch of the current coil body (when the laser sensor detects that the roundness value of the coil body exceeds the upper limit warning value for 5 consecutive sampling points, the system determines that the bending pressure may be too high). S5: Compare the calculated size with the preset threshold (including the warning range and the out-of-tolerance range). If it is within the acceptable range, return to S3 to continue monitoring; if it enters the warning range, only record and prompt; if it enters the out-of-tolerance range, proceed to S6. S6: Based on the direction and magnitude of the deviation, calculate the compensation amount for the corresponding actuator using the prediction model; S7: Sends instructions to the internal drive components of the coiling machine to complete the fine-tuning of process parameters; S8: The system continuously monitors the effect of the adjustment until the size returns to the acceptable range, and records the intervention event.
[0019] In a preferred embodiment, in S4, the system sets a warning interval and an out-of-tolerance interval. The threshold range of the warning interval and the out-of-tolerance interval can be preset according to the coil specifications, production accuracy requirements and equipment response characteristics. The hierarchical control ensures production stability and quality reliability.
[0020] Through the above technical solutions, the early warning range can identify the trend of quality fluctuations in advance, making it easier for staff to predict the production status. The out-of-tolerance range triggers automatic compensation, realizing rapid closed-loop handling of quality problems. The hierarchical control not only avoids production fluctuations caused by excessive regulation, but also prevents the expansion of quality defects, thus taking into account both production stability and quality reliability.
[0021] Compared with the prior art, the beneficial effects of the present invention are: This invention relates to an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion. By reconstructing the complete outline of the coil cross-section using three sets of laser scanning sensors, it can accurately acquire micron-level outer diameter and roundness data. Combined with dual industrial cameras and a dedicated light source, it simultaneously realizes pitch measurement, coil regularity monitoring, and surface defect identification, covering the core quality indicators of the coil and achieving 100% online non-contact full inspection. Compared with traditional manual sampling and single-sensor detection, it eliminates the risk of missed detection and breaks through the limitations of detection dimensions. At the same time, non-contact measurement avoids damage to the coil surface, ensuring product surface quality and providing reliable technical support for the quality control of high-end coil products. This invention relates to an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion. The system relies on a multi-source data fusion algorithm in a central processing unit to reduce noise, refine, and comprehensively evaluate laser and visual inspection data, quickly identifying dimensional deviations. It can also precisely adjust process parameters such as feeding speed, bending pressure, and wire winding speed ratio according to the type of deviation, forming a closed-loop control of "real-time detection - intelligent analysis - dynamic compensation." This mechanism transforms the traditional "post-inspection and rejection of defective products" model into "pre-inspection and in-process control." When the detection data enters the out-of-tolerance range, parameter fine-tuning can be completed within seconds, controlling the dimensional deviation within the tolerance range, significantly reducing the defect rate, and avoiding the production capacity loss caused by manual downtime adjustments, thus improving production continuity and stability.
[0022] This paper presents an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion. This invention relates to an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion. The motors on the multiple adjustment units can achieve multi-directional position adjustment of the detection components through corresponding bidirectional lead screws or threaded rods. It can quickly adapt to the detection needs of coils with different pipe diameters and coil diameters, without the need to change special detection fixtures, thus reducing the time and cost of switching between multiple product specifications. Attached Figure Description
[0023] Figure 1 This is a schematic diagram of the overall structure of the present invention; Figure 2 This is a schematic cross-sectional view of the first fixing frame of the present invention; Figure 3This is a schematic cross-sectional view of the second fixing frame of the present invention.
[0024] In the diagram: 1. Coiling machine; 2. Coiling bracket; 3. First fixed frame; 4. Crossbeam; 5. Second fixed frame; 6. First fixed bracket; 7. First motor; 8. Bidirectional lead screw; 9. First adjusting block; 10. First sliding assembly; 101. First slider; 102. First slide groove; 11. Second fixed bracket; 12. Second motor; 13. Threaded rod; 14. Second adjusting block; 15. Second sliding assembly; 151. Second slider; 152. Second slide groove; 161. First laser scanning sensor; 162. Second laser scanning sensor; 163. Third laser scanning sensor; 171. First industrial camera; 172. Second industrial camera; 18. Central processing control unit; 19. Coiling body. Detailed Implementation
[0025] The present invention will be further described below with reference to embodiments.
[0026] The following embodiments are used to illustrate the present invention, but should not be used to limit the scope of protection of the present invention. The conditions in the embodiments can be further adjusted according to specific conditions, and simple improvements to the method of the present invention under the premise of the concept of the present invention are all within the scope of protection claimed by the present invention.
[0027] Please see Figure 1-3 The present invention provides an online detection and closed-loop control system and method for coil geometry based on multi-sensor fusion, including a drive production unit, a first adjustment unit, a second adjustment unit, a third adjustment unit, a fourth adjustment unit, a high-precision laser scanning measurement unit, a machine vision recognition unit, and a central processing control unit 18.
[0028] Furthermore, the driving production unit includes a coiling machine 1 (a commercially available product) and a coiling bracket 2 disposed on one side of the coiling machine 1. A first fixing frame 3 is fixedly installed on the top of the coiling machine 1, and a crossbeam 4 is fixedly installed on one side of the first fixing frame 3. The crossbeam 4 is located directly above the coiling bracket 2, and a second fixing frame 5 is fixedly installed at one end of the crossbeam 4. The coiling machine 1 is equipped with drive components such as a feeding servo motor, a bending die hydraulic valve, and a wire guide stepper motor for producing the coil body 19. The coiling bracket 2 can support the coil body 19.
[0029] Furthermore, the first adjustment unit includes a first fixed frame 6. A first motor 7 is fixedly installed on one side of the first fixed frame 6. The output end of the first motor 7 passes through the interior of the first fixed frame 6 and is fixedly connected to a bidirectional lead screw 8. One end of the bidirectional lead screw 8 is rotatably connected to the inner wall of the first fixed frame 6. Two first adjustment blocks 9 are threadedly connected to both ends of the outer surface of the bidirectional lead screw 8. A first sliding assembly 10 is provided at the bottom of each of the two first adjustment blocks 9. The first sliding assembly 10 includes a first slider 101. One end of the first slider 101 is fixedly connected to the first adjustment block 9. A first groove 102 adapted to the first slider 101 is opened on the inner wall of the first fixed frame 6. The sliding cooperation between the first slider 101 and the first groove 102 provides a guide limit for the first adjustment block 9, avoiding the adjustment block from shifting or jamming during the transmission of the bidirectional lead screw 8. This ensures that the laser scanning sensors on both sides can always maintain a symmetrical and horizontal spacing, improving the calibration accuracy of the horizontal detection position and providing structural protection for the accurate measurement of outer diameter and roundness.
[0030] Furthermore, the second, third, and fourth adjustment units all include a second fixed frame 11. A second motor 12 is fixedly installed on the inner wall of the second fixed frame 11. A threaded rod 13 is fixedly installed on the output end of the second motor 12. One end of the threaded rod 13 is rotatably connected to the inner wall of the second fixed frame 11. A second adjustment block 14 is threadedly connected to the outer surface of the threaded rod 13. A second sliding assembly 15 is provided on one side of the second adjustment block 14. The second sliding assembly 15 includes a second slider 151. One end of the second slider 151 is fixedly connected to the second adjustment block 14. A second sliding groove 152 adapted to the second slider 151 is opened on the inner wall of the second fixed frame 11. The matching structure between the second slider 151 and the second sliding groove 152 can restrict the rotational freedom of the second adjustment block 14, so that the adjustment block can only move linearly along the direction of the sliding groove. This ensures the stability of the laser scanning sensor and the industrial camera during the position fine adjustment process, avoids the deviation of detection data caused by component shaking, and improves the measurement stability of the entire detection system.
[0031] Furthermore, the first fixing frame 6 is horizontally installed on the top of the coiling machine 1, the second fixing frame 11 included in the second adjustment unit is vertically installed inside the first fixing frame 3 and is perpendicular to the first fixing frame 6, the second fixing frame 11 included in the third adjustment unit is horizontally installed at the bottom of the crossbeam 4 and is arranged along the length of the crossbeam 4, and the second fixing frame 11 included in the fourth adjustment unit is vertically installed inside the second fixing frame 5.
[0032] Furthermore, the high-precision laser scanning measurement unit includes a first laser scanning sensor 161, a second laser scanning sensor 162, and a third laser scanning sensor 163. These sensors utilize the Mi-Iridium ScanCONTROL series. The first and second laser scanning sensors 161 and 162 are respectively connected to the two first adjustment blocks 9 included in the first adjustment unit, and the third laser scanning sensor 163 is connected to the second adjustment block 14 included in the second adjustment unit. This is used to reconstruct the complete profile of the coil body 19 cross-section. The three sets of laser scanning sensors are arranged in a multi-directional manner, allowing for the reconstruction of the complete profile of the coil cross-section through multi-view data, accurately acquiring core dimensional parameters such as outer diameter and roundness. This unit is installed after the bending and forming point of the coil body 19, aiming to capture the forming quality at the earliest possible time.
[0033] Furthermore, the machine vision recognition unit includes a first industrial camera 171 and a second industrial camera 172, both of which are Keyence CV-X series cameras. The first industrial camera 171 is connected to the second adjustment block 14 included in the third adjustment unit, and the second industrial camera 172 is connected to the second adjustment block 14 included in the fourth adjustment unit. The first industrial camera 171 is equipped with a backlight for measuring the pitch of the coil body 19, and the second industrial camera 172 is equipped with a ring coaxial light source. Through a carefully designed lighting scheme (such as using backlight to measure the contour and using coaxial light to detect the surface), combined with image processing algorithms (such as edge detection, blob analysis, and template matching), this unit is responsible for extracting information such as pitch, coil outer diameter, coil shape regularity, and macroscopic surface defects.
[0034] Furthermore, the central processing control unit 18 consists of an industrial computer (IPC) and a programmable logic controller (PLC). The central processing control unit 18 uses a Siemens S7-1500 series PLC and an industrial computer. The PLC is responsible for high-speed data acquisition and logic control, while the industrial computer runs a Windows system and has built-in custom-developed SCADA software responsible for running data fusion algorithms and providing a visual interface. The first laser scanning sensor 161, the second laser scanning sensor 162, the third laser scanning sensor 163, the first industrial camera 171, and the second industrial camera 172 are all electrically connected to the central processing control unit 18. The central processing control unit 18 is also electrically connected to the internal drive components of the coiling machine 1, such as the feeding servo motor, the bending die hydraulic valve, and the wire guide stepper motor. The digital commands issued by the central processing control unit 18 are converted into physical control of the internal drive components of the coiling machine 1 (such as the servo motor, hydraulic valve, and stepper motor), thereby adjusting key process parameters such as feeding speed, bending pressure, and wire guide speed in real time, forming a closed-loop control. In operation, a linkage hub for detection and execution was established. An industrial computer handles complex data fusion and algorithm computation, while a programmable logic controller (PLC) issues high-speed commands and performs logic control. This enables centralized reception and processing of multi-sensor data, as well as precise control of drive components, establishing a seamless "detection-analysis-control" link. It provides a core control platform for closed-loop regulation and, compared to traditional independent detection systems, offers rapid command response and production linkage capabilities. This unit goes beyond simply comparing measured values with thresholds; it analyzes data trends, predicts quality changes, and calculates the optimal compensation strategy using advanced algorithms (such as Kalman filtering, fuzzy control, or simple PID control).
[0035] Furthermore, the central processing unit 18 executes the following data fusion algorithm: a. Perform Kalman filtering on the outer diameter and roundness measurement data acquired by the high-precision laser scanning measurement unit to eliminate random noise (Kalman filtering can effectively filter environmental interference noise in laser measurement data and ensure the authenticity of dimensional data); b. Perform subpixel edge extraction on the pitch and disk-related visual images acquired by the machine vision recognition unit to improve the accuracy of dimensional measurement (subpixel edge extraction technology breaks through the accuracy bottleneck of traditional pixel-level recognition, and greatly improves the accuracy of visual measurement parameters such as pitch). c. Establish a comprehensive quality evaluation model based on fuzzy logic or neural networks, which weights and integrates multiple parameters such as outer diameter, roundness, pitch, and surface defect level into a comprehensive quality index (the fuzzy logic or neural network model realizes the weighted integration of multi-dimensional quality parameters, can output an intuitive comprehensive quality index, solves the problem of independent evaluation of multi-source data and inability to quantify comprehensive quality, and provides a scientific basis for quality decision-making). d. Built-in sampling point judgment rules, which can preset the threshold monitoring conditions for N consecutive sampling points. When a certain size parameter exceeds the warning range for N consecutive sampling points, the corresponding deviation analysis and compensation logic is triggered (the value of N can be preset according to the production accuracy requirements, coil material characteristics and equipment response speed. For example, setting it to 5 sampling points ensures that the system is sensitive to continuous quality fluctuations and does not respond to occasional fluctuations, thus improving the control accuracy).
[0036] Furthermore, after receiving the control signal from the central processing control unit 18, the internal drive component of the coiling machine 1 adjusts at least one of the following parameters in real time to compensate for coil geometric deviations: a. Speed of the feeding servo motor (adjust the speed of the feeding servo motor when the outer diameter is out of tolerance); b. Hydraulic or servo pressure of the bending die (adjust the bending die pressure when the roundness is not up to standard); c. The ratio of the cable guide's moving speed to the spindle speed (optimize the ratio of cable guide to spindle speed when the pitch is abnormal).
[0037] Working principle of the invention: S1: Initialize the online detection and closed-loop control system for coil geometry based on multi-sensor fusion, and load the preset specifications and tolerance thresholds of the coil body 19 through the parameter import function of the central processing control unit 18. S2: Start the first motor 7 to drive the bidirectional lead screw 8 to rotate, which in turn moves the two first adjusting blocks 9 to calibrate the horizontal distance between the first laser scanning sensor 161 and the second laser scanning sensor 162; start each second motor 12 to drive the corresponding threaded rod 13 to rotate, which in turn moves the second adjusting block 14 to adjust the vertical height of the third laser scanning sensor 163, the position of the first industrial camera 171, and the vertical height of the second industrial camera 172, so that each detection unit can accurately adapt to the detection requirements of the coil body 19; S3: A high-precision laser scanning measurement unit and a machine vision recognition unit perform real-time data acquisition; S4: The central processing control unit 18 integrates multi-source data to calculate key dimensions such as the outer diameter, roundness, and pitch of the current coil body 19 (when the laser sensor detects that the roundness value of the coil body 19 exceeds the upper limit warning value for 5 consecutive sampling points, the system determines that the bending pressure may be too high). S5: Compare the calculated size with the preset threshold (including the warning range and the out-of-tolerance range). If it is within the acceptable range, return to S3 to continue monitoring; if it enters the warning range, only record and prompt; if it enters the out-of-tolerance range, proceed to S6. S6: Based on the direction and magnitude of the deviation, calculate the compensation amount for the corresponding actuator using the prediction model; S7: Send instructions to the internal drive components of the coiling machine 1 to complete the fine-tuning of process parameters; S8: The system continuously monitors the effect of the adjustment until the size returns to the acceptable range, and records the intervention event.
[0038] In S4, the system sets early warning intervals and out-of-tolerance intervals. The threshold ranges of the early warning intervals and out-of-tolerance intervals can be preset according to the coil specifications, production accuracy requirements and equipment response characteristics. This hierarchical control ensures production stability and quality reliability.
[0039] 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 multi-sensor fusion-based online detection and closed-loop control system for coil geometry, characterized in that: It includes a drive production unit, a first adjustment unit, a second adjustment unit, a third adjustment unit, a fourth adjustment unit, a high-precision laser scanning measurement unit, a machine vision recognition unit, and a central processing control unit (18). The driving production unit includes a coiling machine (1) and a coiling bracket (2) set on one side of the coiling machine (1). A first fixing frame (3) is fixedly installed on the top of the coiling machine (1). A crossbeam (4) is fixedly installed on one side of the first fixing frame (3). The crossbeam (4) is located directly above the coiling bracket (2). A second fixing frame (5) is fixedly installed at one end of the crossbeam (4). The coiling machine (1) is used to produce coil bodies (19). The coiling bracket (2) can support the coil bodies (19). The coiling machine (1) is equipped with drive components such as a feeding servo motor, a bending mold hydraulic valve, and a wire guide stepper motor. The first adjustment unit includes a first fixed frame (6), a first motor (7) is fixedly installed on one side of the first fixed frame (6), the output end of the first motor (7) passes through the inside of the first fixed frame (6) and is fixedly connected to a bidirectional lead screw (8), one end of the bidirectional lead screw (8) is rotatably connected to the inner wall of the first fixed frame (6), and two first adjustment blocks (9) are threaded to the two ends of the outer surface of the bidirectional lead screw (8), and a first sliding component (10) is provided at the bottom of each of the two first adjustment blocks (9). The second adjustment unit, the third adjustment unit and the fourth adjustment unit all include a second fixed frame (11). A second motor (12) is fixedly installed on the inner wall of the second fixed frame (11). A threaded rod (13) is fixedly installed on the output end of the second motor (12). One end of the threaded rod (13) is rotatably connected to the inner wall of the second fixed frame (11). A second adjustment block (14) is threadedly connected to the outer surface of the threaded rod (13). A second sliding component (15) is provided on one side of the second adjustment block (14). The first fixing frame (6) is horizontally installed on the top of the coiling machine (1). The second fixing frame (11) included in the second adjustment unit is vertically installed inside the first fixing frame (3) and is perpendicular to the first fixing frame (6). The second fixing frame (11) included in the third adjustment unit is horizontally installed at the bottom of the crossbeam (4). The second fixing frame (11) included in the fourth adjustment unit is vertically installed inside the second fixing frame (5).
2. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 1, characterized in that: The high-precision laser scanning measurement unit includes a first laser scanning sensor (161), a second laser scanning sensor (162), and a third laser scanning sensor (163). The first laser scanning sensor (161) and the second laser scanning sensor (162) are respectively connected to the two first adjustment blocks (9) included in the first adjustment unit. The third laser scanning sensor (163) is connected to the second adjustment block (14) included in the second adjustment unit and is used to reconstruct the complete outline of the cross-section of the coil body (19). The machine vision recognition unit includes a first industrial camera (171) and a second industrial camera (172). The first industrial camera (171) is connected to the second adjustment block (14) included in the third adjustment unit. The second industrial camera (172) is connected to the second adjustment block (14) included in the fourth adjustment unit. The first industrial camera (171) is equipped with a backlight source for measuring the pitch of the coil body (19). The second industrial camera (172) is equipped with a ring coaxial light source for monitoring the regularity of the coil body (19) and defects such as surface scratches and pits.
3. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 2, characterized in that: The central processing control unit (18) is composed of an industrial computer (IPC) and a programmable logic controller (PLC). The first laser scanning sensor (161), the second laser scanning sensor (162), the third laser scanning sensor (163), the first industrial camera (171), and the second industrial camera (172) are all electrically connected to the central processing control unit (18). The central processing control unit (18) is electrically connected to the internal drive components of the coiling machine (1) (such as the feeding servo motor, the bending mold hydraulic valve, and the wire guide stepper motor) and is used to send the processed control commands to the corresponding drive components.
4. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 1, characterized in that: The first sliding component (10) includes a first slider (101), one end of which is fixedly connected to the first adjusting block (9), and the inner wall of the first fixing frame (6) is provided with a first groove (102) that is adapted to the first slider (101).
5. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 1, characterized in that: The second sliding assembly (15) includes a second slider (151), one end of which is fixedly connected to the second adjusting block (14), and the inner wall of the second fixing frame (11) is provided with a second sliding groove (152) that is adapted to the second slider (151).
6. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 1, characterized in that: The central processing control unit (18) executes the following data fusion algorithm: a. Perform Kalman filtering on the outer diameter and roundness measurement data acquired by the high-precision laser scanning measurement unit to eliminate random noise; b. Perform sub-pixel edge extraction on the pitch and disk-related visual images acquired by the machine vision recognition unit to improve the accuracy of dimensional measurement; c. Establish a comprehensive quality evaluation model based on fuzzy logic or neural network, and integrate multiple parameters such as outer diameter, roundness, pitch and surface defect level into a comprehensive quality index by weighting and fusion. d. Built-in sampling point judgment rules, which can preset the threshold monitoring conditions for N consecutive sampling points. When a certain size parameter exceeds the warning range for N consecutive sampling points, the corresponding deviation analysis and compensation logic is triggered (the value of N can be preset according to the production accuracy requirements, coil material characteristics and equipment response speed. For example, setting it to 5 sampling points ensures that the system is sensitive to continuous quality fluctuations and does not respond to occasional fluctuations, thus improving the control accuracy).
7. The online detection and closed-loop control system for coil geometry based on multi-sensor fusion according to claim 6, characterized in that: After receiving the control signal from the central processing control unit (18), the internal drive component of the coiling machine (1) adjusts at least one of the following parameters in real time to compensate for the geometric deviation of the coil: a. The speed of the feeding servo motor; b. Hydraulic or servo pressure of the bending die; c. The ratio of the cable guide's moving speed to the spindle speed.
8. The method for online detection and closed-loop control of coil geometry based on multi-sensor fusion according to claim 1, characterized in that: The online detection and closed-loop control system for coil geometry based on multi-sensor fusion as described in any one of claims 1 to 7 includes the following steps: S1: The online detection and closed-loop control system for coil geometry based on multi-sensor fusion is initialized by loading the preset specifications and tolerance thresholds of the coil body (19) through the parameter import function of the central processing control unit (18). S2: Start the first motor (7) to drive the bidirectional lead screw (8) to rotate, thereby moving the two first adjustment blocks (9) to calibrate the horizontal distance between the first laser scanning sensor (161) and the second laser scanning sensor (162); start each second motor (12) to drive the corresponding threaded rod (13) to rotate, thereby moving the second adjustment block (14) to adjust the vertical height of the third laser scanning sensor (163), the position of the first industrial camera (171), and the vertical height of the second industrial camera (172) respectively, so that each detection unit can accurately adapt to the detection requirements of the coil body (19); S3: A high-precision laser scanning measurement unit and a machine vision recognition unit perform real-time data acquisition; S4: The central processing control unit (18) integrates multi-source data and calculates key dimensions such as the outer diameter, roundness, and pitch of the current coil body (19). (When the laser sensor detects that the roundness value of the coil body (19) exceeds the upper limit warning value for 5 consecutive sampling points, the system determines that the bending pressure may be too high.) S5: Compare the calculated size with the preset threshold (including the warning range and the out-of-tolerance range). If it is within the acceptable range, return to S3 to continue monitoring; if it enters the warning range, only record and prompt; if it enters the out-of-tolerance range, proceed to S6. S6: Based on the direction and magnitude of the deviation, calculate the compensation amount for the corresponding actuator using the prediction model; S7: Send instructions to the internal drive components of the coiling machine (1) to complete the fine-tuning of process parameters; S8: The system continuously monitors the effect of the adjustment until the size returns to the acceptable range, and records the intervention event.
9. The method for online detection and closed-loop control of coil geometry based on multi-sensor fusion according to claim 8, characterized in that: In S4, the system sets early warning intervals and out-of-tolerance intervals. The threshold ranges of the early warning intervals and out-of-tolerance intervals can be preset according to the coil specifications, production accuracy requirements and equipment response characteristics. This hierarchical control ensures production stability and quality reliability.
