Intelligent Assembly System and Assembly Method for Gas Detectors

Abstract

This invention belongs to the field of automation equipment technology, specifically an intelligent assembly system and method for gas detectors. The intelligent assembly system for gas detectors includes a production line and an integrated control system. The production line is equipped with assembly, tightening, and back cover assembly stations according to the material flow direction, with each station configured with an execution robot and a conveyor line. The integrated control system comprises a perception layer, a control layer, and an execution layer, forming a perception network through vision sensors, torque sensors, etc. The control layer integrates vision processing, central control, and data management units, realizing dynamic compensation of robot trajectories and adaptive assembly based on visual feedback, and binding all process data with product identity for traceability. The assembly method includes flexible production changeover, vision-guided assembly, online inspection, data traceability, and closed-loop control steps. This invention solves the problems of poor flexibility, low precision, and lack of traceability in traditional assembly methods, achieving efficient, high-quality, and intelligent production of gas detectors.

Description

Technical Field

[0001] This invention belongs to the field of automation equipment technology, specifically relating to precision instruments, and in particular to an intelligent assembly system and assembly method for a gas detector. Background Technology

[0002] As the manufacturing industry transforms towards intelligent and flexible production, the market demand for diverse varieties and small batches poses a severe challenge to traditional assembly production methods. In the assembly of precision instruments such as gas detectors, the internal components typically include external components like alarm indicator lights, gas sensors, and double-nut aviation connectors, as well as internal components such as displays, buzzers, and circuit boards, requiring high assembly precision. Currently, assembly in this field mainly exists in the following modes, each with its inherent drawbacks:

[0003] (1) Manual assembly mode

[0004] This model heavily relies on the skill and focus of operators, resulting in three major drawbacks: First, it is inefficient and inconsistent; manual assembly is slow and it is difficult to ensure the consistency of assembly quality for numerous components, especially in areas heavily influenced by human factors, such as the installation of seals and the tightening torque of nuts. Second, it cannot achieve precise force control and data traceability; for screw and nut tightening, manual operation cannot quantify and record precise torque-angle curves, leading to missing key process parameters. In the event of quality problems, traceability lacks data support. Third, it is costly and unsustainable; with rising labor costs, this model has poor long-term economic benefits and hinders the digital upgrade of production management.

[0005] (2) Rigid automated assembly line

[0006] To improve efficiency, dedicated automated equipment has emerged for single-model gas detectors. However, such rigid production lines have the following drawbacks: First, they lack flexibility and are difficult to change products. When the product model changes, even with only minor structural adjustments, significant time and costs are required for mechanical adjustments, fixture replacements, and program rewriting, making them unsuitable for rapid changeovers. Second, they have low fault tolerance, easily leading to equipment damage and product scrap. The production line lacks "sensing" and "adaptive" capabilities, unable to compensate for minor deviations in incoming material tolerances, fixture wear, or material positioning in real time. This can easily cause serious consequences such as assembly failures, part damage, or even equipment collisions—the so-called "blind assembly" problem.

[0007] (3) Automated equipment with basic visual assistance

[0008] While some equipment incorporates vision technology, it doesn't fundamentally solve the problems: First, the systems are isolated, forming "information silos." The vision system, robot control system, and upper-level management software are often independent, resulting in poor data interoperability. Visual guidance and online quality inspection functions are not effectively integrated, failing to form a closed-loop control system of "guidance-execution-verification," thus limiting the level of intelligence. Second, there is a lack of a complete quality data chain. Existing vision applications are mostly limited to positioning or simple inspection of single processes, failing to bind visual data from each process (such as hole positioning images before component installation and post-assembly positioning inspection images) to the unique identity of the product, thus hindering the construction of a complete, end-to-end visualized quality traceability system.

[0009] In summary, existing technologies either rely on manual labor, leading to low efficiency and quality; or while highly automated, they are rigid and inflexible, unable to adapt to the demands of flexible production; or, although advanced sensing technologies are introduced, system-level integration and data fusion are not achieved. Therefore, there is an urgent need in this field for an intelligent assembly solution that deeply integrates visual perception, robotic flexible execution, end-to-end quality monitoring, and data traceability for products like gas detectors, in order to achieve high-quality, efficient, and flexible production. Summary of the Invention

[0010] The technical problem this invention aims to solve is: to address the issues of low manual efficiency, difficulty in rigidly switching to automation, and the isolation of initial automation systems leading to information silos and a lack of full-process data traceability in existing technologies, as mentioned in the background, this invention provides an intelligent assembly system for gas detectors. By integrating a perception layer, a control layer, and an execution layer, it utilizes visual guidance and dynamic trajectory compensation to achieve adaptive robot assembly and online quality inspection. Simultaneously, it binds all process data to product identity, forming a traceable electronic file, thereby solving the industry problems of poor flexibility, difficulty in quality control, and lack of traceability.

[0011] The technical solution adopted by this invention to solve its technical problem is: an intelligent assembly system for a gas detector, comprising:

[0012] The production line includes an assembly station, a tightening station, and a back cover assembly station arranged sequentially along the material flow direction, as well as a conveyor line connecting each station; the assembly station includes a first group of robots for transporting the housing and first type of components, the tightening station includes a second group of robots for transporting second type of components, and the back cover assembly station includes a third robot for assembling the back cover.

[0013] An integrated control system, which communicates with the production line, includes:

[0014] The perception layer includes 2D / 3D vision sensors installed at each workstation, torque sensors for acquiring tightening torque data, and RFID readers or QR code scanners for product identification binding.

[0015] The execution layer includes the first group of robots, the second group of robots, the third robot, an automatic tightening device, and an automatic feeding mechanism;

[0016] The control layer includes:

[0017] The vision processing unit is used to process image data uploaded from the perception layer and execute visual-guided localization algorithms and online quality detection algorithms.

[0018] The central control unit has pre-stored assembly process recipes for various gas detector models. It communicates with the vision processing unit, each robot in the execution layer, and the automatic tightening device. It is used to dynamically adjust the robot's working trajectory based on the output of the vision processing unit and coordinate the orderly operation of each execution mechanism.

[0019] The data management unit is used to establish and store the association between the product's unique identifier and the entire production process data, which includes at least visual guidance images, online inspection results, and tightening torque curves.

[0020] By integrating a three-layer control system, the fundamental defects of the background technology, such as system isolation, lack of flexibility, and lack of traceability, are solved, realizing the transformation from rigid automated or manual assembly to flexible intelligent assembly.

[0021] Furthermore, the central control unit receives the recognition results of the actual pose of the workpiece from the vision processing unit, and calculates the target grasping or assembly pose of the robot in combination with the hand-eye calibration matrix, so as to dynamically compensate the pre-stored robot standard trajectory and realize adaptive assembly.

[0022] By using visual feedback and dynamic trajectory compensation, the cumulative errors in the material supply, fixtures and other processes are actively offset, transforming traditional blind assembly into intelligent assembly. This increases the first-time success rate of assembly and reduces reliance on high-precision, high-cost special tooling, thus addressing the shortcomings of rigid automation in the background technology, which has low fault tolerance.

[0023] Furthermore, the active end effector of the first group of robots and / or the second group of robots is equipped with a quick-change mechanism and at least two types of execution grippers, the execution grippers including grippers for holding housings and large components, pneumatic suction nozzles or grippers for picking up nuts, and flaring tools for fitting sealing rings.

[0024] By combining a quick-change mechanism with a multi-functional fixture, a single robot can alternately perform different tasks such as picking up components, picking up nuts, and putting on seals, reducing robot waiting time and the number of production line devices, achieving a compact layout and efficient collaboration, and optimizing production costs and cycle time.

[0025] Furthermore, it also includes a digital twin monitoring interface connected to the central control unit, which is used to display the status of production line equipment, robot movement trajectory and production data in real time in a three-dimensional visualization format.

[0026] The digital twin interface presents the real-time status of the physical production line in a three-dimensional visualization, enabling operators to intuitively and comprehensively control the entire production process and supporting remote intervention, thereby reducing the operating threshold and maintenance difficulty of advanced automation systems.

[0027] Furthermore, the assembly station is also equipped with an automatic nut feeding mechanism. Under the scheduling of the central control unit, the first group of robots alternately performs the operations of picking up nuts from the feeding mechanism and assembling components onto the housing using visual guidance.

[0028] The first group of robots can autonomously complete the continuous action from picking up nuts to assembly. In the collaboration of multiple robots, responsibilities are clearly defined, and the refined task scheduling optimizes the assembly cycle, reduces non-value-adding waiting time, and improves overall work efficiency.

[0029] A method for assembling the intelligent gas detector assembly system described above is also provided, comprising the following steps:

[0030] S1. Production preparation and flexible production changeover: The central control unit receives production work orders, calls up the assembly process formula corresponding to the product model, and completes the automatic loading and switching of equipment parameters.

[0031] S2, Visual Guidance and Adaptive Assembly:

[0032] S2.1. Obtain the position and orientation of the workpiece to be assembled using a vision sensor;

[0033] S2.2 The vision processing unit identifies the actual pose of the workpiece, and the central control unit dynamically calculates the target pose of the robot based on the actual pose, and drives the robot to complete the gripping and assembly of the workpiece.

[0034] S3. Online intelligent inspection and quality judgment: After the key assembly process is completed, the assembly results are captured and analyzed by vision sensors, and the OK / NG judgment result is output.

[0035] S4. Data-driven quality traceability: During the assembly process, the real-time torque-angle curve data of screws, visual guidance images, online inspection images and results are bound and stored with the unique identifier of the current product.

[0036] S5. Closed-loop feedback and process control: The central control unit determines whether the product should flow into the next process based on the online detection results.

[0037] S6. Finished Product Judgment and Warehousing: After the appearance, quantity and specifications are checked, the central control unit decides whether to release the product. If released, it enters the waiting area for warehousing and awaits allocation of storage location.

[0038] Furthermore, during the assembly process of S2, a force / torque hybrid control strategy is adopted by combining the force / torque information fed back by the torque sensor to achieve compliant insertion and assembly of the robot.

[0039] Through force-position hybrid control, the robot can sense and respond to contact forces during operations such as inserting connectors, preventing damage to parts or products caused by hard collisions and improving the reliability and yield of precision assembly.

[0040] Furthermore, the online intelligent inspection in S3 includes at least the following: integrity inspection of component mounting holes on the tester housing at the assembly station; inspection of the placement of installed displays, buzzers, and circuit boards at the tightening station; and inspection of the tightening status of back cover screws at the back cover assembly station.

[0041] Real-time visual inspection is carried out at key stages such as assembly, tightening, and final assembly, and multiple inspections are conducted to achieve a shift from post-production sampling to full-process inspection, thus preventing the circulation of defective products.

[0042] A method for assembling a gas detector is also provided, applied to the intelligent assembly system for gas detectors described above, characterized by comprising the following sequentially executed process steps:

[0043] At the assembly station, the alarm indicator light, gas sensor, double-nut aviation plug and its corresponding nut are assembled onto the detector housing through visual guidance and robot operation.

[0044] At the tightening station, the display, buzzer, and circuit board are assembled into the housing through visual guidance and robotic operation, and then secured with screws.

[0045] At the rear cover assembly station, the rear cover is laser-marked, then assembled onto the housing and secured with screws with insulating washers.

[0046] After at least one workstation is completed, online visual inspection and / or electrical performance testing are performed, and the products are diverted to the OK finished product conveyor line or the NG finished product conveyor line according to the test results.

[0047] For the first time, the complex, traditionally manual assembly process of gas detectors has been solidified into a standardized, automated precision assembly process, ensuring consistency and high efficiency in product assembly.

[0048] Furthermore, at the assembly station, the alarm indicator light is assembled onto the detector housing, specifically including:

[0049] First, the loading robot of the first group of robots uses the first gripper to grab the alarm indicator light and place it into the housing mounting hole that has been visually positioned for detection. Then, it replaces the second gripper to put the sealing ring on it. At the same time, the cooperating robot of the first group of robots takes the corresponding nut from the nut feeding mechanism and locks it.

[0050] By using one robot specifically for component handling and final assembly, and another robot specifically for retrieving nuts, a high degree of parallel operation and precise coordination can be achieved, improving the efficiency and reliability of this specific process.

[0051] Furthermore, at the assembly station, the double-nut aviation connector is assembled onto the tester housing, specifically including:

[0052] The mounting holes for the double-nut aviation plug on the housing are located and detected using a visual sensor.

[0053] After passing the inspection, the first group of robots' loading robot grabs the double-nut aviation plug, while the first group of robots' collaborating robot takes two nuts from the feeding mechanism.

[0054] The first group of robots, working collaboratively, placed the double-nut aviation plug into the positioned mounting hole and locked it in place.

[0055] By using two robots working together, the plug body and two nuts are handled separately, ensuring the stability of the assembly process and the quality of tightening, thus solving the problems of low efficiency and inconsistent torque in manual assembly of this component.

[0056] Further, at the tightening station, the circuit board is assembled onto the housing of the testing instrument, specifically including:

[0057] After the second group of robots picks up the circuit board, it is first placed on an inclined plane for secondary positioning.

[0058] After the secondary positioning is completed, the second group of robots' loading robot grabs the circuit board that has completed the pose correction again and places it into the predetermined position inside the detector housing;

[0059] After the circuit board is clamped and fixed by the upper and lower positioning fixtures set on the tightening station, the screws are installed by the collaborative robot of the second group of robots.

[0060] By using inclined plane secondary positioning to actively correct the deviation of the circuit board grasping posture, it is ensured that the circuit board can be installed into the housing without damage and with precision, thereby improving the assembly success rate and quality of this station.

[0061] Further, at the tightening station, the buzzer is assembled onto the housing of the testing instrument, specifically including:

[0062] The second group of robots uses a loading robot to pick up the buzzer; a vision sensor is used to locate the mounting position of the buzzer inside the housing; the loading robot of the second group of robots places the buzzer into the located mounting position, and the cooperating robot of the second group of robots installs screws to fix it.

[0063] Furthermore, at the rear cover assembly station, the rear cover is assembled onto the testing instrument housing, specifically including:

[0064] First, the third robot uses the first end effector to grab the back cover and place it into the laser marking machine for marking;

[0065] After marking is completed, the third robot is replaced by the second end effector. The second end effector performs the following operations in sequence: picks up screws from the screw feeding mechanism, puts on insulating washers from the insulating pad feeding mechanism, moves to the top of the detector housing, and locks the back cover to the detector housing with screws.

[0066] By integrating multiple sub-processes such as laser marking, screw removal, insulating pad placement, and final tightening into a single workstation and having them smoothly completed by a single robot, not only is material handling and the number of workstations reduced, but the accuracy of the marking information binding to the product and the quality of the final assembly are also ensured.

[0067] The beneficial effects of this invention are:

[0068] Compared with existing technologies, the intelligent assembly system and method for gas detectors provided by this invention, by constructing an intelligent architecture integrating perception, decision-making, execution, and traceability, achieves a fundamental leap from traditional rigid automation to flexible intelligence, specifically bringing the following outstanding beneficial effects:

[0069] (1) This invention pre-stores assembly process formulas for multiple products in the central control unit and software-izes core process data such as visual parameters of different models of gas detectors, robot motion trajectories, and tightening torque standards. When a change of production is required, the system can automatically call and load the corresponding formula to achieve instantaneous switching of all equipment parameters without the need for hardware adjustment. It breaks free from the constraints of physical reconstruction, enabling the production line to quickly respond to market orders for multiple varieties and small batches, greatly improving production flexibility and overall efficiency.

[0070] (2) By deploying 2D / 3D vision sensors and deeply integrating them with the control system, the present invention constructs a real-time closed loop of perception-decision-execution; the system can accurately acquire the actual pose of the workpiece, and the central control unit dynamically calculates and compensates the robot's motion trajectory to achieve high-precision adaptive assembly; this greatly reduces the system's dependence on high-precision, high-cost special tooling, allowing the use of lower-cost standardized fixtures, which can reduce the initial investment and maintenance costs of the production line while ensuring or even improving the assembly success rate.

[0071] Third, this invention deeply embeds online intelligent inspection into the assembly process. After the completion of key workstations, the system automatically uses visual sensors to collect images of the assembly results and performs real-time analysis and OK / NG judgment through algorithms. This enables the immediate detection and interception of defective products, preventing them from flowing into subsequent processes and causing greater losses. It is equivalent to transforming the traditional inefficient post-inspection or final inspection into an efficient and reliable full-process inspection, eliminating the flow and amplification of defects from the source and ensuring the high quality and consistency of the products leaving the factory.

[0072] (3) This invention innovatively binds the unique identifier of each product with its entire production data—including visual guidance images, online inspection results, tightening torque curves, etc.—through a data management unit, forming a complete and tamper-proof electronic file. When a quality problem occurs, all relevant process data and visual evidence can be quickly retrieved through the product serial number, enabling accurate and rapid location of the problem from the phenomenon to the root cause of production. This improves the efficiency and accuracy of quality analysis, provides a solid data foundation for process optimization and quality traceability, and promotes the digitalization and refinement of production management.

[0073] (4) The present invention coordinates and schedules all subsystems in a unified manner through a central control unit, and can be equipped with a digital twin monitoring interface to provide a single control entry and intuitive three-dimensional visualization interaction; operators no longer need to face multiple independent units such as vision system and robot controller, but can complete one-click production change, process monitoring and intervention in a unified interface; simplifying the operation process, reducing the composite requirements of multiple professional skills of operators, making such an advanced automation system easier to master and use, improving human-machine collaboration efficiency, and reducing the later training and maintenance costs. Attached Figure Description

[0074] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0075] Figure 1 This is a schematic diagram of the gas detector that needs to be assembled according to the present invention.

[0076] Figure 2This is a block diagram of the integrated control system architecture of the intelligent assembly system for the gas detector of the present invention.

[0077] Figure 3 This is a schematic diagram of the gas detector production line that needs to be assembled according to the present invention.

[0078] Figure 4 yes Figure 3 Top view.

[0079] Figure 5 yes Figure 3 A schematic diagram of the structure of the nut assembly unit.

[0080] Figure 6 yes Figure 3 A schematic diagram of the tightening station.

[0081] Figure 7 yes Figure 6 A schematic diagram of the structure of the screw assembly unit.

[0082] Figure 8 yes Figure 3 A schematic diagram of the structure of the middle and rear cover assembly unit.

[0083] Figure 9 This is a flowchart of the intelligent assembly control method for the gas detector of the present invention.

[0084] In the diagram: 100. Detector housing; 101. First alarm indicator light; 102. Gas sensor; 103. Second alarm indicator light; 104. Double-nut aviation connector; 105. Display; 106. Buzzer; 107. Circuit board; 108. Back cover; 1. First host computer; 2. Second host computer; 3. Machine vision sensor; 4. Assembly station collaborative robot; 5. Assembly station loading robot; 6. Peripheral loading buffer line; 7. Assembly station three-axis loading module; 8. Tightening station loading robot; 9. Tightening station collaborative robot; 10. Back cover collaborative robot; 11. Tightening station three-axis loading module; 12. Tightening station loading conveyor line; 13. Laser marking machine; 14. Performance testing box; 15. Manual assistance station;

[0085] 50. Fixed nut assembly table; 51. Plug hole position detection vision sensor; 52. Hole position detection vision sensor; 53. First housing positioning cylinder; 54. Material gripping fixture; 55. Automatic nut locking mechanism; 56. Double nut aviation plug nut feeding mechanism; 57. Gas detector nut feeding mechanism; 58. Alarm indicator nut feeding mechanism; 60. Movable housing assembly base; 61. Second housing positioning cylinder; 62. Positioning fixture on the display; 63. Positioning fixture on the circuit board; 64. Positioning fixture below the display; 65. 111. Circuit board under-positioning fixture; 112. Automatic screw feeder I; 113. First automatic screw tightening machine; 114. Automatic screw feeder II; 115. Vision sensor; 70. Vision light source; 71. Rear cover clamping cylinder; 72. Third housing positioning cylinder; 73. Screw insulating pad feeder; 74. Automatic screw feeder; 75. Rear cover vision sensor; 76. Second automatic screw tightening machine; 77. Rear cover gripper; 78. Finished product rotary positioning cylinder; 120. OK finished product conveyor line; 121. NG finished product conveyor line. Detailed Implementation

[0086] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0087] like Figure 1 As shown, the gas detector in this embodiment includes a detector housing 100. A first alarm indicator light 101 and a second alarm indicator light 103 with the same structure need to be installed on the front end face of the detector housing 100. A gas sensor 102 needs to be installed between the two alarm indicator lights. A double-nut aviation plug 104 needs to be installed on one side face of the detector housing 100. Inside the detector housing 100, a display 105, a buzzer 106, and a circuit board 107 need to be installed. The circuit board 107 needs to be installed above the display 105. Finally, the back cover 108 needs to be installed.

[0088] Example 1

[0089] like Figure 2As shown, an intelligent assembly system for a gas detector includes a production line and an integrated control system. The production line includes an assembly station, a tightening station, and a back cover assembly station arranged sequentially along the material flow direction, as well as a conveyor line connecting each station. The assembly station includes a first group of robots for handling the housing and first-type components; the tightening station includes a second group of robots for handling second-type components; and the back cover assembly station includes a third robot for assembling the back cover. The moving ends of the first group of robots and / or the second group of robots are equipped with quick-change mechanisms and at least two types of execution grippers. The execution grippers include grippers for holding the housing and large components, pneumatic suction nozzles or grippers for picking up nuts, and flaring tools for fitting sealing rings. The assembly station also has an automatic nut feeding mechanism. Under the scheduling of the central control unit, the first group of robots alternately performs the operations of picking up nuts from the feeding mechanism and assembling components onto the housing using vision guidance.

[0090] The integrated control system communicates with the production line and includes a perception layer, a control layer, and an execution layer. The perception layer includes 2D / 3D vision sensors at each workstation, torque sensors for acquiring tightening torque data, and RFID readers or QR code scanners for product identification. The execution layer includes a first group of robots, a second group of robots, a third group of robots, an automatic tightening device, and an automatic feeding mechanism. The control layer includes a vision processing unit, a central control unit, and a data management unit. The vision processing unit processes the image data uploaded from the perception layer and executes vision-guided positioning algorithms and online quality inspection algorithms. The central control unit pre-stores assembly process recipes for various gas detector models and communicates with the vision processing unit, each robot in the execution layer, and the automatic tightening device. It dynamically adjusts the robot's work trajectory based on the output of the vision processing unit and coordinates the orderly operation of each execution mechanism. The data management unit establishes and stores the association between the unique product identifier and the entire production process data, which includes at least vision-guided images, online inspection results, and tightening torque curves.

[0091] The central control unit receives the recognition results of the actual workpiece pose from the vision processing unit and, in conjunction with the hand-eye calibration matrix, calculates the robot's target grasping or assembly pose to dynamically compensate for the pre-stored robot standard trajectory, achieving adaptive assembly. The hand-eye calibration matrix is ​​used for the transformation between the camera and robot coordinate systems.

[0092] In a preferred embodiment, the system also includes a digital twin monitoring interface connected to the central control unit, used to display the status of production line equipment, robot motion trajectory and production data in real time in a three-dimensional visualization format. Specific implementation examples:

[0094] like Figure 3 and Figure 4As shown, the assembly station on the production line is equipped with a nut assembly unit A, which requires tightening the first alarm indicator light 101, the second alarm indicator light 103, the gas sensor 102, and the double-nut aviation plug 104 to their corresponding positions on the detector housing 100 using nuts; the tightening station is equipped with a screw assembly unit B, which requires tightening the display 105, the buzzer 106, and the circuit board 107 to their corresponding positions inside the detector housing 100 using screws; the rear cover assembly station is equipped with a rear cover assembly unit C, which requires assembling the rear cover 108 onto the detector housing 100. The first group of robots includes a collaborative robot 4 at the assembly station and a loading robot 5 at the assembly station. The second group of robots includes a loading robot 8 at the tightening station and a collaborative robot 9 at the tightening station. The third robot is the collaborative robot 10 for the back cover. It also includes an external loading buffer line 6, a three-axis loading module 7 at the assembly station, a three-axis loading module 11 at the tightening station, a loading conveyor line 12 at the tightening station, a unloading conveyor line 59 at the assembly station, a laser marking machine 13, a performance testing box 14, and a manual assistance station 15.

[0095] like Figure 5 As shown, the nut assembly unit A is equipped with a fixed nut assembly table 50, a plug hole position detection vision sensor 51, a hole position detection vision sensor 52, a first housing positioning cylinder 53, a material gripping fixture 54, an automatic nut locking mechanism 55, a double nut aviation plug nut feeding mechanism 56, a gas detector nut feeding mechanism 57, and an alarm indicator nut feeding mechanism 58. The fixed nut assembly table 50 is located in the middle of the nut assembly unit A. The first housing positioning cylinder 53 is set next to one side of the fixed nut assembly table 50. The hole position detection vision sensor 52 is set to the left of the fixed nut assembly table 50. The material gripping fixture 54 is set near the hole position detection vision sensor 52. The plug hole position detection vision sensor 51 is set directly behind the first housing positioning cylinder 53. The alarm indicator nut feeding mechanism 58 is located to the right of the plug hole position detection vision sensor 51. The double nut aviation plug nut feeding mechanism 56 and the gas detector nut feeding mechanism 57 are set side by side to the right of the alarm indicator nut feeding mechanism 58. The automatic nut locking mechanism 55 is located between the gas detector nut feeding mechanism 57 and the alarm indicator nut feeding mechanism 58.

[0096] like Figure 6 and Figure 7 As shown, the screw assembly unit B is equipped with an automatic screw feeder 111, a first automatic screw tightening machine 112, an automatic screw feeder 113, a vision sensor 114, a movable housing assembly base 60, a second housing positioning cylinder 61, a positioning fixture on the display 62, a positioning fixture on the circuit board 63, a positioning fixture on the lower part of the display 64, and a positioning fixture on the lower part of the circuit board 65.

[0097] like Figure 8As shown, the rear cover assembly unit C includes a vision light source 70, a rear cover clamping cylinder 71, a third housing positioning cylinder 72, a screw insulating pad feeder 73, an automatic screw feeder 74, a rear cover vision sensor 75, a second automatic screw tightening machine 76, a rear cover gripping fixture 77, and a finished product rotation positioning cylinder 78. The vision light source 70 is located at the far right front of the rear cover assembly unit C, the rear cover clamping cylinder 71 is located to the left of the vision light source 70, and the third housing positioning cylinder 72 is located to the right rear of the rear cover clamping cylinder 71. The screw insulating pad feeder 73 is located in the middle right of the rear cover assembly unit C. The second automatic screw tightening machine 76 and the rear cover gripping fixture 77 are sequentially arranged to the left of the screw insulating pad feeder 73. The automatic screw feeder 74 is arranged behind the second automatic screw tightening machine 76, and the rear cover vision sensor 75 is arranged to the right of the automatic screw feeder 74.

[0098] The integrated control system includes a first host computer 1, a second host computer 2, and machine vision sensors 3. The first host computer 1, acting as the main control server, primarily undertakes the core functions of the central control unit and data management unit, responsible for production scheduling, recipe management, data storage, and traceability. The second host computer 2, acting as a dedicated vision processing server, primarily handles the computational needs of the vision processing unit, processing image data uploaded from all workstation vision sensors and executing vision-guided positioning and online quality inspection algorithms. The two host computers are interconnected via industrial Ethernet, working collaboratively to ensure the real-time performance and stability of system control and vision processing. Machine vision sensors 3 is a collective term for all vision sensors in the system, uniformly labeled in the attached diagram. Specific examples include, but are not limited to: plug hole position detection vision sensor 51 and hole position detection vision sensor 52 in nut assembly unit A, vision sensor 114 in screw assembly unit B, and rear cover vision sensor 75 in rear cover assembly unit C, etc. These sensors collectively constitute the visual perception network of the perception layer.

[0099] The detector housing 100 is placed in a transfer box with dimensions of 600mm×400mm×230mm and four layers. It is fed by an AGV trolley and placed on the outer feeding buffer line 6. The outer feeding buffer line 6 rotates 90° and transports the material to the designated position in a width direction of 600mm.

[0100] Example 2

[0101] like Figure 9 As shown, the assembly method of the intelligent assembly system for gas detectors includes the following steps:

[0102] Step 1: Production Preparation and Flexible Changeover: The central control unit receives the production work order, calls up the assembly process formula corresponding to the product model, and completes the automatic loading and switching of equipment parameters.

[0103] Step 2, Visual Guidance and Adaptive Assembly:

[0104] (1) Obtain the position and orientation of the workpiece to be assembled using a vision sensor;

[0105] (2) The vision processing unit identifies the actual position of the workpiece, and the central control unit dynamically calculates the target position of the robot based on the actual position, and drives the robot to complete the gripping and assembly of the workpiece. During the assembly process, the force / torque information fed back by the torque sensor is combined with the force-position hybrid control strategy to realize the robot's compliant insertion assembly.

[0106] Step 3, Online Intelligent Inspection and Quality Judgment: After the key assembly process is completed, the assembly results are image-acquired and analyzed by vision sensors, and OK / NG judgment results are output. This includes integrity inspection of component mounting holes on the housing at the assembly station; inspection of the position of the installed display 105 and circuit board 107 at the tightening station; and inspection of the tightening status of the screws on the back cover 108 at the back cover assembly station.

[0107] Step 4, Data-driven quality traceability: During the assembly process, the real-time torque-angle curve data of the screws, visual guidance images, online inspection images and results are bound and stored with the unique identifier of the current product.

[0108] Step 5, Closed-loop feedback and process control: The central control unit determines whether the product should flow into the next process based on the online detection results.

[0109] Step Six: Finished Product Judgment and Warehousing: After the appearance, quantity and specifications are checked, the central control unit decides whether to release the product. If released, it enters the waiting area for warehousing and awaits allocation of storage space. Specific implementation examples:

[0111] An assembly method for a gas detector includes the following sequentially performed process steps:

[0112] At the assembly station, the alarm indicator light, gas sensor 102, double nut aviation plug 104 and its corresponding nuts are assembled onto the detector housing 100 through visual guidance and robot operation.

[0113] At the tightening station, the display 105, buzzer 106, and circuit board 107 are assembled into the detector housing 100 through visual guidance and robot operation, and then locked in place with screws.

[0114] At the rear cover assembly station, the rear cover 108 is laser-marked, and then the rear cover 108 is assembled onto the detector housing 100 and secured with screws with insulating washers.

[0115] After at least one workstation is completed, online visual inspection and / or electrical performance testing are performed, and the products are diverted to OK finished product conveyor line 120 or NG finished product conveyor line 121 according to the test results.

[0116] The specific operating steps are as follows:

[0117] In nut assembly unit A, the assembly station loading robot 5 picks up the detector housing 100 from the material bin and places it on the fixed nut assembly table 50. Then, the first housing positioning cylinder 53 pushes and clamps the detector housing 100. The hole position detection vision sensor 52 takes pictures of the hole positions of the alarm indicator lights (first alarm indicator light 101 and second alarm indicator light 103) and the gas sensor 102 on the housing. After taking pictures, the hole position detection vision sensor 52 acquires images of the hole positions on the housing. The vision processing unit uses the built-in image processing algorithm or deep learning model to analyze the integrity of the assembly and the component positioning in real time, and outputs the OK / NG judgment result.

[0118] It should be noted that the vision processing unit uses built-in image processing algorithms or deep learning models to perform real-time analysis on the integrity of the assembly and the position of components. For those skilled in the art, it is a conventional technical means to implement these functions based on publicly available computer vision libraries (such as OpenCV) or mature deep learning frameworks (such as TensorFlow, PyTorch).

[0119] The hole position detection vision sensor 52 outputs an OK result. The assembly station loading robot 5 changes the material gripper 54 to pick up the alarm indicator light and then puts on the rubber ring. At the same time, the assembly station collaborative robot 4 goes to the alarm indicator light nut feeding mechanism 58 to pick up the nut of the alarm indicator light. After picking up the nut, it tightens the nut to the alarm indicator light housing.

[0120] Similarly, the assembly station loading robot 5 uses the material gripper 54 to pick up the gas sensor 102 and then puts on the rubber ring. At the same time, the assembly station collaborative robot 4 goes to the gas detector nut feeding mechanism 57 to pick up the nut of the gas sensor 102. After picking up the nut, it tightens the nut to the gas sensor 102 into the housing.

[0121] Similarly, the plug hole detection vision sensor 51 takes pictures of the hole positions of the double nut aviation plug 104 on the housing. After taking pictures, the plug hole detection vision sensor 51 acquires images of the housing hole positions. The vision processing unit uses built-in image processing algorithms or deep learning models to analyze the integrity of the assembly and the component placement in real time, and outputs OK / NG judgment results.

[0122] When the plug hole position detection vision sensor 51 outputs an OK result, the assembly station loading robot 5 uses the material gripper 54 to pick up the double-nut aviation plug 104. At the same time, the assembly station collaborative robot 4 uses the automatic nut locking mechanism 55 to pick up the nut of the double-nut aviation plug 104 from the plug hole position detection vision sensor 51. After picking up the nut, the nut is tightened to the double-nut aviation plug 104 into the housing. If there is no abnormality after tightening, the assembly station loading robot 5 picks up the housing and places it on the normal product conveyor line and moves it to the manual position. If there is an abnormality after tightening, the robot picks up the housing and places it on the NG product conveyor line.

[0123] At the manual auxiliary station 15, the operator sorts the semi-finished products on the assembly station unloading conveyor line 59 into tooling, manually scans the detector housing 100 and gas sensor 102, uploads the data to the first host computer 1, and binds these components and housings.

[0124] The assembled product is manually placed on the tightening station's feeding conveyor line 12 and transferred to the screw assembly unit B. After the semi-finished product reaches the designated position, the second housing positioning cylinder 61 extends to position the semi-finished product. Simultaneously, the tightening station's feeding robot 8 picks up the display 105, and the tightening station's collaborative robot 9 picks up the screws from the automatic screw feeder 111. The screws then wait in the designated position. The vision sensor 114 takes pictures of the screw hole positions on the display 105 and the circuit board 107. After taking pictures, the vision sensor 114 performs image acquisition of the screw hole positions. The vision processing unit uses built-in image processing algorithms or deep learning models to analyze the integrity of the assembly, the size of the screw holes, and the spacing between the holes in real time, and outputs an OK / NG judgment result.

[0125] If the vision sensor 114 outputs an OK result, then the tightening station loading robot 8 places the display 105 into the housing. The positioning fixture 62 on the display and the positioning fixture 64 on the lower part of the display are fixed diagonally. After detecting the positioning success signal, the tightening station loading robot 8 is removed, and the tightening station collaborative robot 9 performs screw-driving assembly.

[0126] Similarly, the installation position of the buzzer 106 inside the housing is located by the vision sensor 114, the tightening station loading robot 8 places the buzzer 106 into the housing, the tightening station collaborative robot 9 removes the screws and the automatic nail feeder 113 takes the screws and assembles the buzzer 106.

[0127] Similarly, the tightening station loading robot 8 grabs the circuit board 107, then places it on an inclined plane for secondary positioning. After secondary positioning, the circuit board 107 is placed into the housing. After the circuit board 107 is positioned by the upper positioning fixture 63 and the lower positioning fixture 65, the tightening station loading robot 8 is removed, and the tightening station collaborative robot 9 performs screw-driving assembly. After successful assembly, the housing flows out through the tightening station loading conveyor line 12 to the manual auxiliary station 15.

[0128] The back cover collaborative robot 10 picks up the back cover 108 from the material bin and places it on the laser marking machine 13 for laser marking. Then, the housing flows into the back cover assembly unit C via a conveyor belt. After the laser marking is completed, the back cover collaborative robot 10 picks up the housing and places it on the back cover assembly unit C. After picking up the marked back cover 108 and placing it on the housing, another back cover 108 is picked up and placed on the laser marking machine 13 for later use. The third housing positioning cylinder 72 first positions the housing, and the back cover clamping cylinder 71 presses down to clamp the back cover 108. The back cover collaborative robot 10 replaces the second automatic screw tightening machine 76, then goes to the automatic screw feeder 74 to pick up screws, and then goes to the screw insulating washer feeder 73 to put on nylon insulating washers. After completion, the back cover collaborative robot 10 moves to the top of the housing to tighten the screws. After the screws are tightened, the back cover collaborative robot 10 changes the back cover gripper 77 to pick up the shell from the assembly table and place it at the finished product rotary positioning cylinder 78. The back cover vision sensor 75 takes pictures of the screws on the finished back cover. After taking pictures, the back cover vision sensor 75 acquires images of the screw hole positions. The vision processing unit uses built-in image processing algorithms or deep learning models to analyze the integrity of the assembly and the integrity of the screws in real time and outputs the OK / NG judgment result.

[0129] If the screws are tightened to OK as shown in the photo, the back cover collaborative robot 10 picks up the finished product and places it in the performance test box 14 for pressure resistance and insulation performance testing. After the pressure resistance and insulation tests are qualified, the back cover collaborative robot 10 automatically puts the product into the right fixture for other comprehensive performance tests. After the comprehensive performance tests are OK, the back cover collaborative robot 10 places the finished product on the OK finished product conveyor line 120.

[0130] Similarly, if the screw tightening is not acceptable (NG) during the photo inspection, the back cover collaborative robot 10 picks up the finished product and places it on the NG finished product conveyor line 121. If the overall performance test is not acceptable (NG), the back cover collaborative robot 10 places the finished product on the NG finished product conveyor line 121.

[0131] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. An intelligent assembly system for a gas detector, characterized in that, include: The production line includes an assembly station, a tightening station, and a back cover assembly station arranged sequentially along the material flow direction, as well as a conveyor line connecting each station; the assembly station includes a first group of robots for transporting the housing and first type of components, the tightening station includes a second group of robots for transporting second type of components, and the back cover assembly station includes a third robot for assembling the back cover. An integrated control system, which communicates with the production line, includes: The perception layer includes 2D / 3D vision sensors installed at each workstation, torque sensors for acquiring tightening torque data, and RFID readers or QR code scanners for product identification binding. The execution layer includes the first group of robots, the second group of robots, the third robot, an automatic tightening device, and an automatic feeding mechanism; The control layer includes: The vision processing unit is used to process image data uploaded from the perception layer and execute visual-guided localization algorithms and online quality detection algorithms. The central control unit has pre-stored assembly process recipes for various gas detector models. It communicates with the vision processing unit, each robot in the execution layer, and the automatic tightening device. It is used to dynamically adjust the robot's working trajectory based on the output of the vision processing unit and coordinate the orderly operation of each execution mechanism. The data management unit is used to establish and store the association between the product's unique identifier and the entire production process data, which includes at least visual guidance images, online inspection results, and tightening torque curves. The central control unit receives the recognition results of the actual pose of the workpiece from the vision processing unit, and calculates the target grasping or assembly pose of the robot in combination with the hand-eye calibration matrix, so as to dynamically compensate the pre-stored robot standard trajectory and realize adaptive assembly. The first group of robots and / or the second group of robots are equipped with quick-change mechanisms at their active ends and at least two types of execution grippers, including grippers for holding housings and large components, pneumatic nozzles or grippers for picking up nuts, and flaring tools for fitting sealing rings.

2. The intelligent assembly system for gas detectors according to claim 1, characterized in that: It also includes a digital twin monitoring interface connected to the central control unit, which is used to display the status of production line equipment, robot movement trajectory and production data in real time in a three-dimensional visualization.

3. The intelligent assembly system for gas detectors according to claim 1, characterized in that: The assembly station is also equipped with an automatic nut feeding mechanism. Under the scheduling of the central control unit, the first group of robots alternately performs the operations of picking up nuts from the feeding mechanism and assembling components onto the housing using visual guidance.

4. An assembly method for an intelligent gas detector assembly system according to any one of claims 1 to 3, characterized in that, Includes the following steps: S1. Production preparation and flexible production changeover: The central control unit receives production work orders, calls up the assembly process formula corresponding to the product model, and completes the automatic loading and switching of equipment parameters. S2, Visual Guidance and Adaptive Assembly: S2.

1. Obtain the position and orientation of the workpiece to be assembled using a vision sensor; S2.2 The vision processing unit identifies the actual pose of the workpiece, and the central control unit dynamically calculates the target pose of the robot based on the actual pose, and drives the robot to complete the gripping and assembly of the workpiece. S3. Online intelligent inspection and quality judgment: After the key assembly process is completed, the assembly results are captured and analyzed by vision sensors, and the OK / NG judgment result is output. S4. Data-driven quality traceability: During the assembly process, the real-time torque-angle curve data of screws, visual guidance images, online inspection images and results are bound and stored with the unique identifier of the current product. S5. Closed-loop feedback and process control: The central control unit determines whether the product should flow into the next process based on the online detection results. S6. Finished Product Judgment and Warehousing: After the appearance, quantity and specifications are checked, the central control unit decides whether to release the product. If released, it enters the waiting area for warehousing and awaits allocation of storage location.

5. The assembly method of the intelligent assembly system for gas detectors according to claim 4, characterized in that, During the assembly process of S2, a force / torque hybrid control strategy is adopted by combining the force / torque information fed back by the torque sensor to achieve compliant insertion and assembly of the robot.

6. The assembly method of the intelligent assembly system for gas detectors according to claim 4, characterized in that, The online intelligent inspection in S3 includes at least the following: integrity inspection of the component mounting holes on the housing (100) of the inspection instrument at the assembly station; inspection of the position of the installed display (105), buzzer (106) and circuit board (107) at the tightening station; and inspection of the tightening status of the screws on the back cover (108) at the back cover assembly station.

7. A method for assembling a gas detector, applied to an intelligent assembly system for a gas detector as described in any one of claims 1 to 5, characterized in that, The following process steps are performed sequentially: At the assembly station, the alarm indicator light, gas sensor (102), double nut aviation plug (104) and its corresponding nuts are assembled onto the detector housing through visual guidance and robot operation; At the tightening station, the display (105), buzzer (106), and circuit board (107) are assembled into the detector housing (100) through visual guidance and robot operation, and then secured with screws; At the rear cover assembly station, the rear cover (108) is laser-marked, and then the rear cover (108) is assembled onto the detector housing (100) and secured with screws with insulating washers. After at least one workstation is completed, online visual inspection and / or electrical performance testing are performed, and the products are diverted to the OK finished product conveyor line (120) or the NG finished product conveyor line (121) according to the test results.

8. The assembly method of the gas detector according to claim 7, characterized in that, At the assembly station, the alarm indicator light is assembled onto the detector housing (100), specifically including: First, the loading robot of the first group of robots uses the first gripper to grab the alarm indicator light and place it into the housing mounting hole that has been visually positioned for detection. Then, it replaces the second gripper to put the sealing ring on it. At the same time, the cooperating robot of the first group of robots takes the corresponding nut from the nut feeding mechanism and locks it.

9. The assembly method of the gas detector according to claim 7, characterized in that, At the assembly station, the double-nut aviation plug (104) is assembled onto the tester housing (100), specifically including: The mounting holes for the double-nut aviation plug on the housing are located and detected using a visual sensor. After the inspection is passed, the first group of robots' loading robot grabs the double-nut aviation plug (104), and at the same time, the first group of robots' collaborating robot takes two nuts from the feeding mechanism. The collaborative robot of the first group of robots places the double-nut aviation plug (104) into the positioned mounting hole and locks the double-nut aviation plug (104) in place.

10. The assembly method of the gas detector according to claim 7, characterized in that, At the tightening station, the circuit board (107) is assembled onto the detector housing (100), specifically including: After the second group of robots picks up the circuit board (107), it first places it on an inclined plane for secondary positioning. After the secondary positioning is completed, the second group of robots' loading robot grabs the circuit board (107) that has completed the pose correction again and places it in the predetermined position inside the detector housing (100); After the circuit board (107) is clamped and fixed by the upper and lower positioning fixtures set on the tightening station, the screws are installed by the collaborative robot of the second group of robots.

11. The assembly method of the gas detector according to claim 7, characterized in that, At the tightening station, the buzzer (106) is assembled onto the housing (100) of the testing instrument, specifically including: The second group of robots' loading robot grabs the buzzer (106); the installation position of the buzzer (106) inside the housing is located by the vision sensor (114); the second group of robots' loading robot places the buzzer (106) into the positioned installation position, and the second group of robots' collaborative robot installs screws to fix it.

12. The assembly method of the gas detector according to claim 7, characterized in that, At the rear cover assembly station, the rear cover (108) is assembled onto the detector housing (100), specifically including: First, the third robot uses the first end effector to grab the back cover (108) and place it into the laser marking machine for marking; After marking is completed, the third robot is replaced by the second end effector. The second end effector performs the following operations in sequence: picks up screws from the screw feeding mechanism, puts on the insulating washer from the insulating pad feeding mechanism, moves to the top of the detector housing and locks the back cover (108) to the detector housing (100) with screws.

Images