Intelligent cleaning workshop linkage system and control method

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent cleaning workshop linkage system and a control method, and an execution device comprises a European style double-beam crane, an ultrasonic cleaning machine, a purging mechanical arm and a circulating system; the control device generates a cleaning path instruction through the dynamic distribution module, the feedback control module decomposes the instruction to adjust parameters of a water pressure valve, the decision closed-loop module calculates a quality parameter alpha based on visual data and outputs a spiral coverage control instruction, and the load balancing module distributes a calculation task to the multi-node controller. The method comprises the steps of pollution level judgment path generation, cavitation intensity linkage water pressure adjustment, driving of a three-dimensional cleaning track by a quality parameter alpha, task fragmentation load balancing and real-time feedback termination control. The cleaning intensity is adjusted through pollution atlas driven automatic scheduling, cavitation intensity linkage equipment control and quality parameter alpha feedback, so that process collaborative optimization, resource utilization rate improvement and cleaning quality stability are realized.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent cleaning workshop linkage system and control method. Background Technology

[0002] Existing intelligent cleaning workshop linkage systems adopt a multi-device collaborative architecture. Sensors monitor the status of the objects being cleaned and the environmental impact in real time, and transmit the data to the control center for analysis and processing. The control center outputs commands after making logical judgments based on preset rules, driving robotic arms or nozzles and other execution units to perform cleaning actions, while coordinating the position adjustment of conveyor belts or other auxiliary equipment. The system achieves efficient interoperability between components through standardized communication protocols, effectively reducing human intervention and process interruptions, and improving the level of automation and repeatability of the cleaning process in industrial environments.

[0003] Existing intelligent cleaning workshop linkage systems suffer from the following technical pain points: The lack of integrated automatic control systems and real-time data processing mechanisms in existing cleaning workshops leads to reliance on individual manual operations, resulting in low cleaning efficiency, cumbersome procedures, and unstable cleaning quality. Individual manual operations cannot achieve coordinated equipment scheduling and dynamic parameter adjustment, causing interruptions in process connections and resource waste. For example, in the cleaning of marine turbocharger parts, workers must manually disassemble and load the parts, performing handling, ultrasonic treatment, and drying in stages. However, the lack of sensor feedback and intelligent decision support makes it difficult to thoroughly remove oil stains from complex structural crevices, and repeated manual intervention increases operational complexity and the risk of quality fluctuations. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent cleaning workshop linkage system and control method. This invention solves the technical problems caused by the reliance on manual individual operation mode in existing cleaning workshops, which leads to discrete equipment scheduling, lagging process parameter control, and lack of quality feedback, resulting in insufficient coordination of cleaning processes, low resource utilization, and large fluctuations in cleaning quality.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows: In a first aspect, the present invention provides an intelligent cleaning workshop linkage system, comprising: The sensing devices include a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor; The actuator includes a European-style double-girder crane 4, an ultrasonic cleaner 7, a first blowing and cleaning robotic arm 9, a blowing and cleaning positioning mechanism 10, a second blowing and cleaning robotic arm 11, and a cleaning fluid circulation system. The control device, which connects the sensing device and the actuating device via a network, includes: The dynamic allocation module receives workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt map, and water pressure data from the sensing device. It converts the surface dirt map into contamination level parameters, which are divided according to the percentage of oil stain coverage area to the total surface area of ​​the workpiece: a percentage higher than 30% is defined as high contamination level, a percentage between 10% and 30% is defined as medium contamination level, and a percentage lower than 10% is defined as low contamination level. Based on the workpiece type data and contamination level parameters obtained from the vision sensor, it generates cleaning path instructions. High contamination level triggers a preset collaborative path of ultrasonic cleaner 7, first blow-wash robotic arm 9, and second blow-wash robotic arm 11; medium contamination level triggers a preset independent cleaning path of ultrasonic cleaner 7; and low contamination level triggers a preset direct ventilation drying station path. The feedback control module receives the cleaning path instruction, decomposes the cleaning path instruction into equipment control parameters, and adjusts the water level valve opening, heater power, robotic arm water pressure value, and air drying trigger conditions. The decision-making closed-loop module receives the drying completion signal from the second blow-washing robotic arm 11 and the secondary data collected by the vision sensor, and outputs the quality judgment command. The load balancing module receives computing tasks from the dynamic allocation module, feedback control module, and decision-making closed-loop module, and distributes them to a multi-node programmable controller. The cleaning path command output by the dynamic allocation module triggers the feedback control module to generate equipment control parameters; The air drying trigger condition output by the feedback control module starts the air drying gun of the second blowing robot arm 11. When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, the air drying completion signal is sent to the decision closed loop module. The quality judgment command output by the decision-making closed-loop module is distributed to the controller node for execution via the load balancing module. The execution result data generated by the controller after executing the quality judgment instruction is fed back to the dynamic allocation module to update the pollution level judgment rules.

[0006] Furthermore, in the intelligent cleaning workshop linkage system of the present invention, the dynamic allocation module includes: The dynamic allocation module includes a task topology mapping unit, which is used to generate corresponding path planning according to the needs of the cleaning task. After generating the cleaning path instruction, the cleaning path instruction is transmitted to the control unit of the European double girder crane 4. After receiving the cleaning path command, the control unit of the European-style double girder crane 4 analyzes it and outputs displacement control signals to the drum drive motor 4b, the upper pulley bottom motor 4g, and the lower pulley bottom motor 4h of the European-style double girder crane 4. The drum drive motor 4b drives the drum to rotate according to the displacement control signal, thereby controlling the hook 4d to complete the vertical lifting and lowering movement; At the same time, the upper motor 4g at the bottom of the pulley and the lower motor 4h at the bottom of the pulley work together to drive the gantry 4c to move horizontally along the track. By coordinating vertical lifting and horizontal movement, the workpiece is precisely transferred and positioned in the basket 5.

[0007] Furthermore, in the intelligent cleaning workshop linkage system of the present invention, the feedback control module includes: The ultrasonic parameter adaptive unit receives the water level threshold and temperature setting value in the cleaning path instruction. When the water level sensor detects a value lower than the water level threshold, it opens the output valve 16 of the water storage tank 15 in the cleaning fluid circulation system. When the temperature sensor detects a value more than ±2℃ away from the temperature setting value, the ultrasonic parameter adaptive unit adjusts the power of the first heater 20 and the second heater 21 to bring the temperature detection value back to the set value range. The robotic arm coordination unit receives the cavitation intensity threshold and residual moisture threshold in the cleaning path instruction. When the vibration data of the ultrasonic cleaner 7 exceeds the cavitation intensity threshold, the water pressure of the first high-pressure water gun 9a of the first blowing and washing robotic arm 9 and the second high-pressure water gun 11a of the second blowing and washing robotic arm 11 is reduced to 70%. When the residual moisture coverage rate detected by the visual sensor received by the dynamic allocation module exceeds 15%, the action of the second air drying gun 11b of the second blowing and washing robotic arm 11 is triggered. The cleaning path instructions include the following working modes: a preset collaborative path of ultrasonic cleaner 7 with first blow-wash robotic arm 9 and second blow-wash robotic arm 11, a preset cleaning path of ultrasonic cleaner 7 alone, and a preset direct ventilation drying station path. The collaborative path of the ultrasonic cleaner 7 with the first blow-washing robotic arm 9 and the second blow-washing robotic arm 11 is configured as follows: the European double-beam crane 4 is instructed to carry the workpiece on the basket 5 and transfer it to the ultrasonic cleaner 7 for ultrasonic cleaning. After completion, it is transferred to the blow-washing positioning mechanism 10 where the first blow-washing robotic arm 9 is located for high-pressure water gun rinsing, and to the area where the second blow-washing robotic arm 11 is located for air drying. The ultrasonic cleaner 7 is configured with a separate cleaning path as follows: the European double-beam crane 4 is instructed to transport the workpiece on the basket 5 to the ultrasonic cleaner 7 for ultrasonic cleaning, and after completion, it is directly transported to the area where the second blow-washing robotic arm 11 is located for air drying. The direct ventilation drying station path is configured such that the European-style double-girder crane 4 will directly transfer the workpiece loaded with basket 5 to the area where the second blowing and washing robot arm 11 is located for air drying. The water pressure reduction action triggers the first heater 20 and the second heater 21 of the main liquid circuit 23 of the cleaning fluid circulation system to heat up, and starts the loop filter device 25 of the liquid circuit 22 of the cleaning fluid circulation system.

[0008] Furthermore, in the intelligent cleaning workshop linkage system of the present invention, the decision-making closed-loop module includes: The multi-threshold grading unit receives secondary acquisition data from the vision sensor, extracts the oil residue area ratio β, rust coverage γ, and reflectivity δ, calculates the quality parameter α=0.6β+0.3γ+0.1δ, and outputs it to the quality judgment unit. The quality judgment unit receives the quality parameter α. When α≤0.05, it triggers the European double-girder crane 4 to return to the initial position. When α>0.1, it triggers the screw cover control command. The spiral coverage control command adjusts the first rotary motor 9f of the base of the first blow-washing robotic arm 9 and the second rotary motor 11f of the base of the second blow-washing robotic arm 11 to generate a three-dimensional spiral trajectory. At the same time, it triggers the first drive motor 9c of the execution device of the first blow-washing robotic arm 9 and the second drive motor 11c of the execution device of the second blow-washing robotic arm 11 to increase the tilt angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a by 20°, and adjusts the heating power of the ultrasonic cleaner 7 to increase the water temperature by 5°C.

[0009] Furthermore, in the intelligent cleaning workshop linkage system of the present invention, the load balancing module includes: The task sharding unit receives the computation tasks from the dynamic allocation module, the feedback control module, and the decision closed-loop module, and decomposes the cleaning path instruction task from the dynamic allocation module, the equipment control parameter task from the feedback control module, and the quality judgment instruction task from the decision closed-loop module to three independent controller nodes. The queue optimization unit receives load rate feedback data from the controller nodes. When the load rate of any of the three independent controller nodes is >80%, it transfers 20% of the computing tasks of the controller node to the idle node with the lowest load rate. It also receives obstacle avoidance control commands generated by the drum drive motor 4b, the right side motor 4e, the left side motor 4f, the upper bottom motor 4g, and the lower bottom motor 4h of the European double girder crane 4 during operation, and inserts the obstacle avoidance control commands at the front of the task queue for priority execution.

[0010] Furthermore, the intelligent cleaning workshop linkage system of the present invention also includes: The cleaning path instruction generated by the task topology mapping unit in the dynamic allocation module is transmitted to the control unit of the European double-girder crane 4. The control unit parses the cleaning path instructions and outputs displacement control signals to the upper pulley bottom motor 4g, the lower pulley bottom motor 4h, and the drum drive motor 4b of the pulley assembly 4a of the European double girder crane 4. The upper motor 4g at the bottom of the pulley and the lower motor 4h at the bottom of the pulley drive the frame 4c of the European double girder crane 4 to move horizontally, and the drum drive motor 4b controls the vertical lifting of the hook 4d of the European double girder crane 4. The workpiece is transferred and positioned by the coordinated control of horizontal movement and vertical lifting.

[0011] Furthermore, the intelligent cleaning workshop linkage system of the present invention also includes: The water pressure reduction action triggers the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency; The water pressure reduction action of the robotic arm coordination unit triggers the main drive pump 18 of the main liquid circuit 23 of the cleaning fluid circulation system to operate at a reduced frequency. The water pressure reduction action simultaneously triggers the start time of the loop drive pump 24 of the liquid circuit 22 of the cleaning fluid circulation system to be brought forward to before the heating operation of the main liquid circuit 23; After the loop drive pump 24 starts, it draws wastewater into the loop filter device 25 of the liquid circuit 22. The filtered wastewater flows back to the water storage tank 15 of the cleaning fluid circulation system through the loop back pressure valve 26 of the liquid circuit 22.

[0012] Furthermore, the intelligent cleaning workshop linkage system of the present invention also includes: The spiral coverage control command is input to the first rotating motor 9f and the second rotating motor 11f of the base, which controls the first blowing and washing robotic arm 9 and the second blowing and washing robotic arm 11 to rotate and generate a three-dimensional spiral trajectory. The first drive motor 9c of the first blow-washing robotic arm 9 and the second drive motor 11c of the second blow-washing robotic arm 11 are synchronously input, and the tilt angle of the first high-pressure water gun 9a of the first blow-washing robotic arm 9 and the second high-pressure water gun 11a of the second blow-washing robotic arm 11 is increased by 20°. During the execution of the spiral cover control command, the vision sensor collects the surface reflectivity δ in real time and transmits it to the decision closed-loop module. When the recalculated quality parameter α ≤ 0.05, the decision closed-loop module stops the spiral cover control command.

[0013] Furthermore, the intelligent cleaning workshop linkage system of the present invention also includes: The task sharding unit of the load balancing module distributes the decomposed computing tasks to three independent controller nodes; One of the controller nodes receives the device control parameter task from the feedback control module, which is used to adjust the opening of the water tank output valve 16 of the ultrasonic cleaner 7 and the power of the first heater 20 and the second heater 21. Another controller node receives the cleaning path instruction task from the dynamic allocation module, which is used to generate the motion trajectory of the first blow-wash robotic arm 9 and the second blow-wash robotic arm 11. The third controller node receives the quality judgment instruction task from the decision-making closed-loop module and is used to calculate the quality parameter α of the decision-making closed-loop module. The queue optimization unit of the load balancing module collects the load rate feedback data of the three independent controller nodes and transmits it to the system control computer 12 of the system control center 3 via the local area network.

[0014] Secondly, the present invention provides an intelligent cleaning workshop linkage control method, applied to the aforementioned intelligent cleaning workshop linkage system, comprising: Step 1: Receive workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt pattern, and water pressure data through a sensing device, wherein the sensing device includes a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor. Step 2: Convert the surface dirt map into contamination level parameters. The contamination level parameters are divided according to the percentage of oil stain coverage area to the total surface area of ​​the workpiece: the percentage of oil stain coverage area to the total surface area of ​​the workpiece is higher than 30% and is defined as high contamination level; the percentage of oil stain coverage area to the total surface area of ​​the workpiece is between 10% and 30% and is defined as medium contamination level; the percentage of oil stain coverage area to the total surface area of ​​the workpiece is lower than 10% and is defined as low contamination level. Based on the workpiece type data and contamination level parameters obtained by the vision sensor, a cleaning path instruction is generated. The high contamination level triggers the ultrasonic cleaner 7 and the first blowing robot arm 9 and the second blowing robot arm 11 to work together; the low contamination level triggers the direct ventilation drying station path. Step 3: Decompose the cleaning path instruction into equipment control parameters, adjust the water level valve opening, heater power, robotic arm water pressure value, and air drying trigger condition. The water level valve opening corresponds to the water tank output valve 16 of the cleaning fluid circulation system, the heater power corresponds to the first heater 20 and the second heater 21, the robotic arm water pressure value corresponds to the water pressure of the first high-pressure water gun 9a and the second high-pressure water gun 11a, and the air drying trigger condition starts the second air drying air gun 11b of the second blowing robotic arm 11. When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, send an air drying completion signal. Step 4: Receive the drying completion signal from the second blow-washing robotic arm 11 and the secondary data collected by the vision sensor, and output the quality judgment command; Step 5: Assign computing tasks to the multi-node programmable controller, wherein the computing tasks include generating cleaning path instructions, equipment control parameters and quality judgment instructions; Step 6: After the controller executes the quality judgment instruction, the execution result data generated is fed back to the dynamic allocation module to update the pollution level judgment rules.

[0015] In step 2, the cleaning path instruction is transmitted to the control unit of the European double-girder crane 4. After parsing, the control unit outputs displacement control signals to the drum drive motor 4b, the upper motor 4g at the bottom of the pulley, and the lower motor 4h at the bottom of the pulley, driving the hook 4d to complete vertical lifting and the gantry 4c to move horizontally, so as to transfer and position the workpiece on the basket 5.

[0016] Beneficial effects of this invention: The dynamic allocation module of this invention collects surface dirt maps using a visual sensor and converts them into pollution level parameters. This drives the task topology mapping unit to generate differentiated cleaning path instructions. The instructions trigger the European double-girder crane 4 to transfer the workpiece onto the basket 5, eliminating process interruptions caused by manual scheduling and improving space utilization. The feedback control module decomposes the cleaning path instructions into equipment control parameters, linking the vibration data of the ultrasonic cleaner 7 with the water pressure adjustment of the first high-pressure water gun 9a and the second high-pressure water gun 11a. This synchronously triggers the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency and the loop drive pump 24 of the liquid circuit 22 to start in advance, achieving cross-system collaborative control between sensor data and the execution device, overcoming the lag defect of process parameters. The decision-making closed-loop module is based on the oil residue surface. The quality parameter α is calculated by fusing the area ratio, rust coverage rate, and reflectivity. The output spiral coverage control command drives the first rotating motor 9f and the second rotating motor 11f of the base to generate a three-dimensional trajectory and adjust the tilt angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a. The visual sensor iteratively feeds back the α value in real time to control the cleaning intensity, thus solving the quality fluctuations caused by human experience misjudgment. The load balancing module decomposes the calculation task to three controller nodes. The queue optimization unit monitors the load rate, migrates tasks, and prioritizes obstacle avoidance commands. The system's central control computer 12 synchronizes node data to ensure reduced response latency. The execution result data is fed back to dynamically update the pollution level judgment rules, forming a complete technical system of scheduling discretization elimination, parameter control optimization, and quality feedback closed loop. Attached Figure Description

[0017] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0018] Figure 1 This is a system architecture diagram of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention.

[0019] Figure 2 This is a top view of the overall structure of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention.

[0020] Figure 3 This is a front three-dimensional structural diagram of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention. Figure 4 This is a rear three-dimensional structural diagram of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention. Figure 5 The diagram shows the structure of a European-style double-girder crane for the intelligent cleaning workshop linkage system provided in this embodiment of the invention. Figure 6 This is a first partial schematic diagram of the European-style double-girder crane structure of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention. Figure 7 This is a second partial schematic diagram of the European-style double-beam crane structure of the intelligent cleaning workshop linkage system provided in an embodiment of the present invention. Figure 8 The diagram shows the structure of the blow-washing robotic arm in the intelligent cleaning workshop linkage system provided in this embodiment of the invention. Explanation of reference numerals in the attached diagrams: 1. Intelligent cleaning area; 2. Manual cleaning area; 3. System control center; 4. European-style double-girder crane; 4a. Pulley assembly; 4b. Drum drive motor; 4c. Overhead frame; 4d. Hook; 4e. Right side motor of the overhead frame; 4f. Left side motor of the overhead frame; 4g. Upper motor at the bottom of the pulley; 4h. Lower motor at the bottom of the pulley; 5. Workpiece loading basket; 6. Ultrasonic cleaning machine control cabinet; 7. Ultrasonic cleaning machine; 8. Oil-water filtration device; 9. First blowing and cleaning robotic arm; 9a. First high-pressure water gun; 9b. First air drying gun; 9c. First drive motor of the actuator; 9d. First drive motor of the forearm; 9e. First drive motor of the boom; 9f. First rotation of the base. 10. Electric motor; 11. Blowing and washing positioning mechanism; 11. Second blowing and washing robotic arm; 11a. Second high-pressure water gun; 11b. Second air drying gun; 11c. Second drive motor of the actuator; 11d. Second drive motor of the forearm; 11e. Second drive motor of the upper arm; 11f. Second rotary motor of the base; 12. System central control computer; 13. Floor drain mechanism; 14. Cleaning workshop foundation; 15. Water storage tank; 16. Water tank output valve; 17. Safety valve; 18. Main drive pump; 19. Main filter device; 20. First heater; 21. Second heater; 22. Liquid circuit; 23. Main liquid circuit; 24. Circuit drive pump; 25. Circuit filter device; 26. Circuit back pressure valve; 27. Main circuit back pressure valve. Detailed Implementation

[0021] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0022] Firstly, please refer to Figures 1 to 8 The present invention provides an intelligent cleaning workshop linkage system, comprising: The sensing devices include a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor; The actuator includes a European-style double-girder crane 4, an ultrasonic cleaner 7, a first blowing and cleaning robotic arm 9, a blowing and cleaning positioning mechanism 10, a second blowing and cleaning robotic arm 11, and a cleaning fluid circulation system. The control device, which connects the sensing device and the actuating device via a network, includes: The dynamic allocation module receives workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt map, and water pressure data from the sensing device. It converts the surface dirt map into contamination level parameters, which are divided according to the percentage of oil stain coverage area to the total surface area of ​​the workpiece: a percentage higher than 30% is defined as high contamination level, a percentage between 10% and 30% is defined as medium contamination level, and a percentage lower than 10% is defined as low contamination level. Based on the workpiece type data and contamination level parameters obtained from the vision sensor, it generates cleaning path instructions. High contamination level triggers a preset collaborative path of ultrasonic cleaner 7, first blow-wash robotic arm 9, and second blow-wash robotic arm 11; medium contamination level triggers a preset independent cleaning path of ultrasonic cleaner 7; and low contamination level triggers a preset direct ventilation drying station path. The feedback control module receives the cleaning path instruction, decomposes the cleaning path instruction into equipment control parameters, and adjusts the water level valve opening, heater power, robotic arm water pressure value, and air drying trigger conditions. The decision-making closed-loop module receives the drying completion signal from the second blow-washing robotic arm 11 and the secondary data collected by the vision sensor, and outputs the quality judgment command. The load balancing module receives computing tasks from the dynamic allocation module, feedback control module, and decision-making closed-loop module, and distributes them to a multi-node programmable controller. The cleaning path command output by the dynamic allocation module triggers the feedback control module to generate equipment control parameters; The air drying trigger condition output by the feedback control module starts the air drying gun of the second blowing robot arm 11. When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, the air drying completion signal is sent to the decision closed loop module. The quality judgment command output by the decision-making closed-loop module is distributed to the controller node for execution via the load balancing module. The execution result data generated by the controller after executing the quality judgment instruction is fed back to the dynamic allocation module to update the pollution level judgment rules.

[0023] The sensing device includes a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor, used to collect real-time data on workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt patterns, and water pressure. The vision sensor acquires images of the workpiece surface through an image acquisition unit, and the image processing unit analyzes the dirt distribution characteristics to provide input data for determining the contamination level.

[0024] The dynamic allocation module receives workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt map, and water pressure data transmitted from the sensing device. The dirt map analysis unit converts the surface dirt map into contamination level parameters. These contamination level parameters are categorized into high, medium, and low contamination levels based on the percentage of oil-covered area relative to the total workpiece surface area. The dynamic allocation module generates cleaning path instructions based on workpiece type data acquired from the vision sensor and the contamination level parameters, using the task topology mapping unit. The cleaning path instructions trigger preset paths according to the contamination level: high contamination triggers a collaborative path between the ultrasonic cleaner 7, the first blowing robot arm 9, and the second blowing robot arm 11; medium contamination triggers a solo cleaning path for the ultrasonic cleaner 7; and low contamination triggers a direct ventilation drying station path.

[0025] The feedback control module receives the cleaning path instructions output by the dynamic allocation module and decomposes these instructions into equipment control parameters by the ultrasonic parameter adaptive unit and the robotic arm coordination unit. The ultrasonic parameter adaptive unit adjusts the water level valve opening and heater power. The water level valve opening corresponds to the output valve 16 of the water storage tank 15 in the cleaning fluid circulation system, and the heater power corresponds to the first heater 20 and the second heater 21. The robotic arm coordination unit adjusts the robotic arm water pressure values ​​of the first high-pressure water gun 9a and the second high-pressure water gun 11a, and sets the air-drying trigger conditions. The air-drying trigger conditions activate the second air-drying air gun 11b of the second blowing robotic arm 11. When the air-drying operation reaches the preset time or the sensor determines that the workpiece surface dryness meets the standard, the feedback control module sends an air-drying completion signal to the decision-making closed-loop module.

[0026] The decision-making closed-loop module receives the drying completion signal from the feedback control module and secondary data acquired by the vision sensor. It then extracts features such as the oil residue area ratio, rust coverage, and reflectivity through a multi-threshold grading unit to calculate quality parameters. Based on these parameters, the quality judgment unit outputs quality judgment commands to drive subsequent cleaning actions or terminate the process. These commands are distributed to controller nodes for execution via the load balancing module.

[0027] The load balancing module receives computational tasks from the dynamic allocation module, feedback control module, and decision-making closed-loop module. Through the task sharding unit, it breaks down and distributes cleaning path instruction tasks, equipment control parameter tasks, and quality judgment instruction tasks to multi-node programmable controllers. The queue optimization unit monitors the load rate of the controller nodes and dynamically adjusts task allocation to achieve computational efficiency. After the controller nodes execute the quality judgment instructions, the resulting execution data flows back to the dynamic allocation module, updating the contamination level judgment rules in the contamination map analysis unit, thus achieving system adaptive optimization.

[0028] The entire system forms a collaborative mechanism through data acquisition by sensors, path planning by the dynamic allocation module, parameter adjustment by the feedback control module, quality assessment by the decision-making closed-loop module, and task allocation by the load balancing module. Each module exchanges data and transmits commands via network communication, improving the automation level and stability of the cleaning process.

[0029] When the intelligent cleaning workshop linkage system is activated, the hook position sensor, water level sensor, temperature sensor, vision sensor, and pressure sensor of the sensing device collect the spatial coordinates of the workpiece mounted on the basket 5, the water level of the water tank 15 in the cleaning fluid circulation system, the workpiece surface temperature, the surface dirt pattern, and the water pressure data of the main fluid path 23, respectively. The vision sensor analyzes the surface dirt pattern through the image processing unit, calculates the percentage of oil stain coverage area to the total surface area of ​​the workpiece, and classifies it into high-contamination level, medium-contamination level, or low-contamination level according to a preset threshold. The dynamic allocation module generates cleaning path instructions based on the contamination level parameters and workpiece type data. High-contamination level triggers the ultrasonic cleaner 7 and the first blowing robot arm 9 and the second blowing robot arm 11 to work together on a path; medium-contamination level triggers the ultrasonic cleaner 7 to work alone on a path; and low-contamination level triggers the direct ventilation drying station path.

[0030] The cleaning path command is transmitted to the control unit of the European-style double-girder crane 4. After parsing the command, the control unit outputs displacement control signals to the drum drive motor 4b, the upper pulley bottom motor 4g, and the lower pulley bottom motor 4h. The drum drive motor 4b drives the hook 4d to complete vertical lifting. The upper pulley bottom motor 4g and the lower pulley bottom motor 4h work together to drive the gantry 4c to move horizontally, transferring the workpiece onto the basket 5 to the target workstation. After receiving the cleaning path command, the feedback control module decomposes it into equipment control parameters. The ultrasonic parameter adaptive unit adjusts the opening of the water tank output valve 16 and the power of the first heater 20 and the second heater 21 according to the water level threshold and temperature setpoint. The robotic arm coordination unit lowers the water pressure of the first high-pressure water gun 9a and the second high-pressure water gun 11a or triggers the action of the second air drying gun 11b according to the cavitation intensity threshold and the residual moisture threshold.

[0031] The decision-making closed-loop module receives the drying completion signal from the second blowing and washing robotic arm 11 and secondary acquisition data from the vision sensor. The multi-threshold grading unit extracts the oil residue area ratio, rust coverage rate, and reflectivity features to calculate the quality parameter α. Based on the α value, the quality judgment unit triggers the European double-girder crane 4 to return to the initial position or outputs a spiral coverage control command. The spiral coverage control command adjusts the first rotating motor 9f and the second rotating motor 11f of the base to generate a three-dimensional spiral trajectory. At the same time, it adjusts the first drive motor 9c and the second drive motor 11c of the actuator to increase the tilt angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a, and increases the heating power of the ultrasonic cleaner 7. The load balancing module distributes the calculation tasks of the dynamic allocation module, the feedback control module, and the decision-making closed-loop module to three independent controller nodes. The queue optimization unit monitors the node load rate and prioritizes the processing of obstacle avoidance control commands. The execution result data is fed back to the dynamic allocation module to update the pollution level judgment rules.

[0032] The system achieves process optimization, improved resource utilization, and stable cleaning quality by using pollution map-driven automatic scheduling, cavitation intensity-linked equipment control, and quality parameter α feedback to adjust cleaning intensity, thus solving the problem of incomplete removal of oil stains from complex crevices of ship parts.

[0033] Specifically, the intelligent cleaning workshop linkage system of the present invention includes a dynamic allocation module comprising: The dynamic allocation module includes a task topology mapping unit, which is used to generate corresponding path planning according to the needs of the cleaning task. After generating the cleaning path instruction, the cleaning path instruction is transmitted to the control unit of the European double girder crane 4. After receiving the cleaning path command, the control unit of the European-style double girder crane 4 analyzes it and outputs displacement control signals to the drum drive motor 4b, the upper pulley bottom motor 4g, and the lower pulley bottom motor 4h of the European-style double girder crane 4. The drum drive motor 4b drives the drum to rotate according to the displacement control signal, thereby controlling the hook 4d to complete the vertical lifting and lowering movement; At the same time, the upper motor 4g at the bottom of the pulley and the lower motor 4h at the bottom of the pulley work together to drive the gantry 4c to move horizontally along the track. By coordinating vertical lifting and horizontal movement, the workpiece is precisely transferred and positioned in the basket 5.

[0034] The dynamic allocation module includes a task topology mapping unit. Based on the requirements of the cleaning task, the task topology mapping unit generates an optimized path planning scheme by analyzing the workpiece type, cleaning sequence, and workshop layout. The path planning scheme is converted into specific cleaning path instructions, which include the target coordinate point sequence and motion trajectory parameters. The task topology mapping unit transmits the cleaning path instructions to the control unit of the European double girder crane 4.

[0035] After receiving the cleaning path instruction, the control unit of the European-style double girder crane 4 processes the instruction content through the built-in parsing algorithm and extracts the displacement control parameters. The displacement control parameters include the vertical displacement and the horizontal displacement. The control unit generates the corresponding displacement control signal according to the displacement control parameters, and outputs the displacement control signal to the drum drive motor 4b, the upper motor 4g at the bottom of the pulley, and the lower motor 4h at the bottom of the pulley.

[0036] After receiving the displacement control signal, the drum drive motor 4b drives the drum to rotate. The rotation of the drum causes the wire rope to wind or release, thereby controlling the hook 4d to complete the vertical lifting motion. The vertical lifting motion allows the workpiece mounted on the basket 5 to adjust its position in the height direction, realizing the lifting or lowering operation of the workpiece.

[0037] At the same time, the upper motor 4g and the lower motor 4h at the bottom of the pulley work together. After receiving the displacement control signal, they drive the frame 4c to move horizontally along the track. The horizontal movement is achieved by synchronous control of the motors to realize the translation of the frame 4c in the X and Y axes, so as to realize the precise positioning of the workpiece mounted on the basket 5 in the plane.

[0038] Through the coordination of vertical lifting and horizontal movement, the European-style double-girder crane 4 transports the workpiece on the basket 5 to the cleaning station or storage location; during the movement, the control unit monitors the position feedback in real time and dynamically adjusts the motor operation status to ultimately complete the precise transfer and positioning of the workpiece on the basket 5.

[0039] Specifically, the feedback control module of the intelligent cleaning workshop linkage system of the present invention includes: The ultrasonic parameter adaptive unit receives the water level threshold and temperature setting value in the cleaning path instruction. When the water level sensor detects a value lower than the water level threshold, it opens the output valve 16 of the water storage tank 15 in the cleaning fluid circulation system. When the temperature sensor detects a value more than ±2℃ away from the temperature setting value, the ultrasonic parameter adaptive unit adjusts the power of the first heater 20 and the second heater 21 to bring the temperature detection value back to the set value range. The robotic arm coordination unit receives the cavitation intensity threshold and residual moisture threshold in the cleaning path instruction. When the vibration data of the ultrasonic cleaner 7 exceeds the cavitation intensity threshold, the water pressure of the first high-pressure water gun 9a of the first blowing and washing robotic arm 9 and the second high-pressure water gun 11a of the second blowing and washing robotic arm 11 is reduced to 70%. When the residual moisture coverage rate detected by the visual sensor received by the dynamic allocation module exceeds 15%, the action of the second air drying gun 11b of the second blowing and washing robotic arm 11 is triggered. The cleaning path instructions include the following working modes: a preset collaborative path of ultrasonic cleaner 7 with first blow-wash robotic arm 9 and second blow-wash robotic arm 11, a preset cleaning path of ultrasonic cleaner 7 alone, and a preset direct ventilation drying station path. The collaborative path of the ultrasonic cleaner 7 with the first blow-washing robotic arm 9 and the second blow-washing robotic arm 11 is configured as follows: the European double-beam crane 4 is instructed to carry the workpiece on the basket 5 and transfer it to the ultrasonic cleaner 7 for ultrasonic cleaning. After completion, it is transferred to the blow-washing positioning mechanism 10 where the first blow-washing robotic arm 9 is located for high-pressure water gun rinsing, and to the area where the second blow-washing robotic arm 11 is located for air drying. The ultrasonic cleaner 7 is configured with a separate cleaning path as follows: the European double-beam crane 4 is instructed to transport the workpiece on the basket 5 to the ultrasonic cleaner 7 for ultrasonic cleaning, and after completion, it is directly transported to the area where the second blow-washing robotic arm 11 is located for air drying. The direct ventilation drying station path is configured such that the European-style double-girder crane 4 will directly transfer the workpiece loaded with basket 5 to the area where the second blowing and washing robot arm 11 is located for air drying. The water pressure reduction action triggers the first heater 20 and the second heater 21 of the main liquid circuit 23 of the cleaning fluid circulation system to heat up, and starts the loop filter device 25 of the liquid circuit 22 of the cleaning fluid circulation system.

[0040] The feedback control module of the intelligent cleaning workshop linkage system integrates an ultrasonic parameter adaptive unit and a robotic arm coordination unit to achieve precise control and adjustment of the cleaning process. The ultrasonic parameter adaptive unit receives the water level threshold and temperature setpoint from the cleaning path command. When the water level sensor detects a value lower than the water level threshold, the ultrasonic parameter adaptive unit opens the output valve 16 of the water storage tank 15 in the cleaning fluid circulation system to replenish the cleaning fluid and maintain the water level. When the temperature sensor detects a value that deviates from the temperature setpoint beyond the allowable range, the ultrasonic parameter adaptive unit adjusts the power of the first heater 20 and the second heater 21 to bring the temperature detection value back to the setpoint range, thereby ensuring the stable cleaning effect of the ultrasonic cleaner 7.

[0041] The robotic arm coordination unit receives the cavitation intensity threshold and residual moisture threshold from the cleaning path instruction. When the vibration data of the ultrasonic cleaner 7 exceeds the cavitation intensity threshold, the robotic arm coordination unit lowers the water pressure of the first high-pressure water gun 9a of the first blowing robotic arm 9 and the second high-pressure water gun 11a of the second blowing robotic arm 11 to a specified ratio to reduce the risk of over-cleaning. When the residual moisture coverage detected by the visual sensor received by the dynamic allocation module exceeds the set value, the robotic arm coordination unit triggers the action of the second air drying gun 11b of the second blowing robotic arm 11 to accelerate the workpiece drying process.

[0042] The cleaning path instructions include three preset working modes to guide the transfer operations of the European-style double-girder crane 4. The ultrasonic cleaner 7, in conjunction with the first and second cleaning robotic arms 9 and 11, guides the European-style double-girder crane 4 to transfer the workpiece, mounted on a basket 5, to the ultrasonic cleaner 7 for ultrasonic cleaning. After cleaning, the workpiece is sequentially transferred to the cleaning positioning mechanism 10 (where the first cleaning robotic arm 9 is located) for high-pressure water rinsing, and then to the area where the second cleaning robotic arm 11 is located for air drying, achieving multi-step collaborative cleaning. The ultrasonic cleaner 7's independent cleaning path instructions guide the European-style double-girder crane 4 to transfer the workpiece, mounted on a basket 5, to the ultrasonic cleaner 7 for ultrasonic cleaning. After cleaning, the workpiece is directly transferred to the area where the second cleaning robotic arm 11 is located for air drying, simplifying the cleaning process. The direct air-drying station path instructions guide the European-style double-girder crane 4 to directly transfer the workpiece, mounted on a basket 5, to the area where the second cleaning robotic arm 11 is located for air drying, suitable for workpieces that do not require cleaning.

[0043] The water pressure reduction action simultaneously triggers the first heater 20 and the second heater 21 of the main liquid circuit 23 of the cleaning fluid circulation system to heat the fluid and maintain a stable cleaning fluid temperature. It also activates the loop filter 25 of the liquid circuit 22 of the cleaning fluid circulation system to enhance filtration and achieve higher cleaning fluid cleanliness. This coordinated response allows the system to optimize the overall cleaning environment while adjusting the robotic arm's operation.

[0044] Specifically, the intelligent cleaning workshop linkage system of the present invention includes a decision-making closed-loop module comprising: The multi-threshold grading unit receives secondary acquisition data from the vision sensor, extracts the oil residue area ratio β, rust coverage γ, and reflectivity δ, calculates the quality parameter α=0.6β+0.3γ+0.1δ, and outputs it to the quality judgment unit. The quality judgment unit receives the quality parameter α. When α≤0.05, it triggers the European double-girder crane 4 to return to the initial position. When α>0.1, it triggers the screw cover control command. The spiral coverage control command adjusts the first rotary motor 9f of the base of the first blow-washing robotic arm 9 and the second rotary motor 11f of the base of the second blow-washing robotic arm 11 to generate a three-dimensional spiral trajectory. At the same time, it triggers the first drive motor 9c of the execution device of the first blow-washing robotic arm 9 and the second drive motor 11c of the execution device of the second blow-washing robotic arm 11 to increase the tilt angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a by 20°, and adjusts the heating power of the ultrasonic cleaner 7 to increase the water temperature by 5°C.

[0045] The multi-threshold grading unit in the decision-making closed-loop module receives secondary acquisition data from the vision sensor. This unit extracts three key indicators from the acquisition data: oil residue area ratio β, rust coverage rate γ, and reflectivity δ. Based on the preset weight relationship, it calculates the comprehensive quality parameter α. The calculated quality parameter α is transmitted to the quality judgment unit for subsequent judgment.

[0046] The quality assessment unit makes a graded decision based on the range of the received quality parameter α. When the quality parameter α does not exceed the low threshold, the cleaning quality is deemed to meet the requirements, and the command to return the European double-girder crane 4 to the initial position is triggered. When the quality parameter α exceeds the high threshold, the workpiece surface is deemed to require further processing, and a spiral cover control command is generated to start the enhanced cleaning process.

[0047] The spiral coverage control command first coordinates the control of the first rotary motor 9f on the base of the first cleaning robot arm 9 and the second rotary motor 11f on the base of the second cleaning robot arm 11, driving the first high-pressure water gun 9a and the second high-pressure water gun 11a to move along a three-dimensional spiral trajectory to expand the cleaning coverage area. At the same time, the command triggers the first drive motor 9c and the second drive motor 11c of the actuator to adjust the spray angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a, increasing the contact intensity between the jet and the workpiece surface. In addition, the command synchronously adjusts the heating power of the ultrasonic cleaner 7 to appropriately increase the temperature of the cleaning solution to enhance the dissolution and cleaning effect.

[0048] Through the collaboration of the multi-threshold grading unit and the quality judgment unit, the system can dynamically adjust the cleaning strategy based on real-time detection data; the spiral cover control command controls the linkage of the first blow-washing robotic arm 9, the second blow-washing robotic arm 11 and the ultrasonic cleaner 7, forming an adaptive cleaning response mechanism for different levels of contamination, which improves the intelligence level and processing effect of the cleaning process.

[0049] Specifically, the intelligent cleaning workshop linkage system of the present invention includes a load balancing module comprising: The task sharding unit receives the computation tasks from the dynamic allocation module, the feedback control module, and the decision closed-loop module, and decomposes the cleaning path instruction task from the dynamic allocation module, the equipment control parameter task from the feedback control module, and the quality judgment instruction task from the decision closed-loop module to three independent controller nodes. The queue optimization unit receives load rate feedback data from the controller nodes. When the load rate of any of the three independent controller nodes is >80%, it transfers 20% of the computing tasks of the controller node to the idle node with the lowest load rate. It also receives obstacle avoidance control commands generated by the drum drive motor 4b, the right side motor 4e, the left side motor 4f, the upper bottom motor 4g, and the lower bottom motor 4h of the European double girder crane 4 during operation, and inserts the obstacle avoidance control commands at the front of the task queue for priority execution.

[0050] The task sharding unit in the load balancing module receives computing tasks from the dynamic allocation module, feedback control module, and decision-making closed-loop module. The task sharding unit identifies the task type and granularity of the cleaning path instruction tasks generated by the dynamic allocation module, the equipment control parameter tasks generated by the feedback control module, and the quality judgment instruction tasks generated by the decision-making closed-loop module. Based on the task characteristics, it allocates them to the three independent controller nodes for execution, thereby achieving initial division of computing resources and task isolation.

[0051] The queue optimization unit continuously collects real-time load rate data from the three independent controller nodes. When the load rate of a controller node exceeds a set threshold, the queue optimization unit initiates a dynamic task scheduling mechanism to migrate some of the computational tasks from that node to the idle controller node with the lowest current load rate, thereby maintaining cluster load balance. In addition, the queue optimization unit also receives obstacle avoidance control commands generated during the operation of the European-style double-girder crane 4 by its drum drive motor 4b, right-side gantry motor 4e, left-side gantry motor 4f, upper pulley bottom motor 4g, and lower pulley bottom motor 4h. These commands have high real-time performance and high priority; the queue optimization unit inserts them at the front of the task queue for immediate execution, thus ensuring the safety and timely response of the European-style double-girder crane 4 during operation.

[0052] Through the coordinated operation of the task sharding unit and the queue optimization unit, the load balancing module not only achieves reasonable distribution of computing tasks and optimized resource utilization, but also realizes real-time response to key control commands, thereby enhancing the reliability and efficiency of the overall control of the intelligent cleaning workshop linkage system.

[0053] Specifically, the intelligent cleaning workshop linkage system of the present invention further includes: The cleaning path instruction generated by the task topology mapping unit in the dynamic allocation module is transmitted to the control unit of the European double-girder crane 4. The control unit parses the cleaning path instructions and outputs displacement control signals to the upper pulley bottom motor 4g, the lower pulley bottom motor 4h, and the drum drive motor 4b of the pulley assembly 4a of the European double girder crane 4. The upper motor 4g at the bottom of the pulley and the lower motor 4h at the bottom of the pulley drive the frame 4c of the European double girder crane 4 to move horizontally, and the drum drive motor 4b controls the vertical lifting of the hook 4d of the European double girder crane 4. The workpiece is transferred and positioned by the coordinated control of horizontal movement and vertical lifting.

[0054] The task topology mapping unit in the dynamic allocation module generates cleaning path instructions based on the topology of the cleaning task and the workshop layout. These instructions include a sequence of target location coordinates and a movement order. The cleaning path instructions are transmitted to the control unit of the European-style double-girder crane 4 via an industrial Ethernet communication protocol to support real-time data transmission and integrity.

[0055] After receiving the cleaning path instruction, the control unit of the European-style double girder crane 4 uses its built-in microprocessor to parse the instruction and extract the horizontal and vertical displacement parameters. The control unit converts the displacement parameters into pulse signals or analog voltage signals and outputs displacement control signals to the upper pulley bottom motor 4g, the lower pulley bottom motor 4h, and the drum drive motor 4b of the pulley assembly 4a. The displacement control signals include speed, direction, and distance information.

[0056] The upper motor 4g and lower motor 4h at the bottom of the pulley drive the gantry 4c to move horizontally along the track via a gear transmission mechanism, realizing the lateral and longitudinal positioning of the European-style double-girder crane 4 within the workshop. The drum drive motor 4b drives the drum to rotate via a reducer, controlling the winding and unwinding of the wire rope, thereby adjusting the vertical lifting height of the hook 4d.

[0057] The control unit coordinates horizontal movement and vertical lifting actions through a closed-loop control system. Utilizing encoder feedback of the real-time position information of the gantry 4c and hook 4d, it dynamically adjusts the outputs of the upper pulley motor 4g, the lower pulley motor 4h, and the drum drive motor 4b to achieve precise transfer and positioning of the workpiece mounted in the basket 5. This collaborative control ensures that the workpiece mounted in the basket 5 accurately reaches the working area of ​​the ultrasonic cleaner 7 or the blowing robot arm, completing the cleaning process.

[0058] Specifically, the intelligent cleaning workshop linkage system of the present invention further includes: The water pressure reduction action triggers the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency; The water pressure reduction action of the robotic arm coordination unit triggers the main drive pump 18 of the main liquid circuit 23 of the cleaning fluid circulation system to operate at a reduced frequency. The water pressure reduction action simultaneously triggers the start time of the loop drive pump 24 of the liquid circuit 22 of the cleaning fluid circulation system to be brought forward to before the heating operation of the main liquid circuit 23; After the loop drive pump 24 starts, it draws wastewater into the loop filter device 25 of the liquid circuit 22. The filtered wastewater flows back to the water storage tank 15 of the cleaning fluid circulation system through the loop back pressure valve 26 of the liquid circuit 22.

[0059] The water pressure reduction action receives a pressure sensor signal through the system control center 3, triggering the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency. The reduced frequency operation of the main drive pump 18 reduces the output frequency and the fluid delivery of the main liquid circuit 23, thereby adjusting the system water pressure to the set range to meet the cleaning operation requirements.

[0060] The water pressure reduction action of the robotic arm coordination unit is sent to the system control center 3 by the actuator drive motor of the first blowing robotic arm 9 or the second blowing robotic arm 11. The system control center 3 controls the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency. The reduced frequency operation of the main drive pump 18 coordinates with the operation of the robotic arm to maintain stable water pressure during the cleaning process and improve the cleaning effect.

[0061] The water pressure reduction action simultaneously triggers the system control center 3 to adjust the timing logic, so that the start time of the loop drive pump 24 of the liquid circuit 22 is advanced to before the heating operation of the first heater 20 or the second heater 21 of the main liquid circuit 23; the loop drive pump 24 starts in advance to realize wastewater pretreatment, prepare for the heating operation, and optimize the system energy efficiency.

[0062] After the loop drive pump 24 starts, it collects wastewater from the intelligent cleaning zone 1 or the manual cleaning zone 2 through the extraction pipeline and inputs it into the loop filter device 25 of the liquid circuit 22. The loop filter device 25 removes impurities, and the filtered wastewater is returned to the water storage tank 15 of the cleaning fluid circulation system after the pressure is regulated by the loop back pressure valve 26 of the liquid circuit 22, thus completing the wastewater recycling.

[0063] Specifically, the intelligent cleaning workshop linkage system of the present invention further includes: The spiral coverage control command is input to the first rotating motor 9f and the second rotating motor 11f of the base, which controls the first blowing and washing robotic arm 9 and the second blowing and washing robotic arm 11 to rotate and generate a three-dimensional spiral trajectory. The first drive motor 9c of the first blow-washing robotic arm 9 and the second drive motor 11c of the second blow-washing robotic arm 11 are synchronously input, and the tilt angle of the first high-pressure water gun 9a of the first blow-washing robotic arm 9 and the second high-pressure water gun 11a of the second blow-washing robotic arm 11 is increased by 20°. During the execution of the spiral cover control command, the vision sensor collects the surface reflectivity δ in real time and transmits it to the decision closed-loop module. When the recalculated quality parameter α ≤ 0.05, the decision closed-loop module stops the spiral cover control command.

[0064] When the intelligent cleaning workshop linkage system performs cleaning operations, the system control center sends spiral coverage control commands to the first rotary motor 9f and the second rotary motor 11f on the base, driving the first blowing and cleaning robotic arm 9 and the second blowing and cleaning robotic arm 11 to rotate around their mounting bases respectively. The speed and direction of the two motors are coordinated and controlled so that the movement trajectories of the first high-pressure water gun 9a and the second high-pressure water gun 11a are combined in three-dimensional space into a spiral path covering the surface of the workpiece, thereby achieving cleaning without dead angles.

[0065] Simultaneously with the generation of the spiral trajectory, the system control center sends synchronous commands to the first drive motor 9c and the second drive motor 11c of the actuator, respectively driving the end actuators of the first rinsing robotic arm 9 and the second rinsing robotic arm 11 to perform pitching motion. Through a precise angle control mechanism, the spray angles of the first high-pressure water gun 9a and the second high-pressure water gun 11a are adjusted to a preset range, thereby enhancing the adaptability of the cleaning medium to complex curved surfaces and improving rinsing efficiency.

[0066] During the spiral cleaning process, the vision sensor integrated on the blowing and positioning mechanism 10 continuously collects optical information of the workpiece surface and calculates the real-time reflectivity δ accordingly. This data is transmitted to the decision-making closed-loop module, which is processed and analyzed in real time by the system's central control computer 12. The system infers the degree of cleanliness and updates the quality parameter α based on the change in reflectivity. When the system determines that the quality parameter α has reached the preset threshold, the decision-making closed-loop module immediately terminates the output of the spiral coverage control command, and the first blowing and cleaning robotic arm 9 and the second blowing and cleaning robotic arm 11 stop moving, thereby ensuring the cleaning quality while avoiding resource waste.

[0067] Specifically, the intelligent cleaning workshop linkage system of the present invention further includes: The task sharding unit of the load balancing module distributes the decomposed computing tasks to three independent controller nodes; One of the controller nodes receives the device control parameter task from the feedback control module, which is used to adjust the opening of the water tank output valve 16 of the ultrasonic cleaner 7 and the power of the first heater 20 and the second heater 21. Another controller node receives the cleaning path instruction task from the dynamic allocation module, which is used to generate the motion trajectory of the first blow-wash robotic arm 9 and the second blow-wash robotic arm 11. The third controller node receives the quality judgment instruction task from the decision-making closed-loop module and is used to calculate the quality parameter α of the decision-making closed-loop module. The queue optimization unit of the load balancing module collects the load rate feedback data of the three independent controller nodes and transmits it to the system control computer 12 of the system control center 3 via the local area network.

[0068] The task sharding unit of the load balancing module logically decomposes the comprehensive instructions issued by the system's central control computer 12 into discrete subsets of computing tasks based on the type and resource requirements of the computing tasks. The task sharding unit then assigns these subsets of computing tasks to three independent controller nodes based on a pre-defined allocation strategy, thereby achieving parallel processing and improving system response efficiency.

[0069] The first controller node is dedicated to handling equipment control parameter tasks from the feedback control module. Based on real-time sensor data and set process parameters, this node generates adjustment commands through a control algorithm to drive the actuators. Specifically, this node outputs control signals to the water tank output valve 16 of the ultrasonic cleaner 7 to precisely adjust its opening degree to control the cleaning fluid flow rate; simultaneously, it outputs power adjustment signals to the first heater 20 and the second heater 21 to dynamically manage the cleaning fluid temperature.

[0070] The second controller node receives the cleaning path instruction from the dynamic allocation module. Based on the spatial position information of the workpiece mounted on the basket 5 and the cleaning process requirements, this node performs kinematic calculations and trajectory planning. The planning results are converted into servo drive instructions, controlling the movement of each axis of the first and second cleaning robotic arms 9 and 11, respectively. Specifically, this includes driving the first rotary motor 9f of the base, the first drive motor 9e of the upper arm, the first drive motor 9d of the lower arm, and the first drive motor 9c of the actuator, as well as driving the second rotary motor 11f of the base, the second drive motor 11e of the upper arm, the second drive motor 11d of the lower arm, and the second drive motor 11c of the actuator, causing the first high-pressure water gun 9a, the first air-drying gun 9b, the second high-pressure water gun 11a, and the second air-drying gun 11b to move along predetermined paths.

[0071] The third controller node processes the quality assessment instructions from the decision-making closed-loop module. This node invokes a built-in quality evaluation algorithm to extract features and analyze patterns from the collected cleaning process data and terminal detection data, thereby calculating a comprehensive quality parameter α that characterizes the cleaning quality level. The calculation result of quality parameter α provides a basis for system process optimization and closed-loop control decisions.

[0072] The queue optimization unit of the load balancing module continuously monitors the operating status of the three independent controller nodes, collecting real-time load rate data from each node as feedback signals. The queue optimization unit transmits the encapsulated load rate feedback data packets to the system control center's central computer 12 via the local area network. Based on this data, the system control center's central computer 12 dynamically adjusts the task sharding strategy and resource allocation scheme, thereby achieving load optimization and stable operation of the entire interconnected system.

[0073] Secondly, the present invention provides an intelligent cleaning workshop linkage control method, applied to the aforementioned intelligent cleaning workshop linkage system, comprising: Step 1: Receive workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt pattern, and water pressure data through a sensing device, wherein the sensing device includes a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor. Step 2: Convert the surface dirt map into contamination level parameters. The contamination level parameters are divided according to the percentage of oil stain coverage area to the total surface area of ​​the workpiece: the percentage of oil stain coverage area to the total surface area of ​​the workpiece is higher than 30% and is defined as high contamination level; the percentage of oil stain coverage area to the total surface area of ​​the workpiece is between 10% and 30% and is defined as medium contamination level; the percentage of oil stain coverage area to the total surface area of ​​the workpiece is lower than 10% and is defined as low contamination level. Based on the workpiece type data and contamination level parameters obtained by the vision sensor, a cleaning path instruction is generated. The high contamination level triggers the ultrasonic cleaner 7 and the first blowing robot arm 9 and the second blowing robot arm 11 to work together; the low contamination level triggers the direct ventilation drying station path. Step 3: Decompose the cleaning path instruction into equipment control parameters, adjust the water level valve opening, heater power, robotic arm water pressure value, and air drying trigger condition. The water level valve opening corresponds to the water tank output valve 16 of the cleaning fluid circulation system, the heater power corresponds to the first heater 20 and the second heater 21, the robotic arm water pressure value corresponds to the water pressure of the first high-pressure water gun 9a and the second high-pressure water gun 11a, and the air drying trigger condition starts the second air drying air gun 11b of the second blowing robotic arm 11. When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, send an air drying completion signal. Step 4: Receive the drying completion signal from the second blow-washing robotic arm 11 and the secondary data collected by the vision sensor, and output the quality judgment command; Step 5: Assign computing tasks to the multi-node programmable controller, wherein the computing tasks include generating cleaning path instructions, equipment control parameters and quality judgment instructions; Step 6: After the controller executes the quality judgment instruction, the execution result data generated is fed back to the dynamic allocation module to update the pollution level judgment rules.

[0074] In step 2, the cleaning path instruction is transmitted to the control unit of the European double-girder crane 4. After parsing, the control unit outputs displacement control signals to the drum drive motor 4b, the upper motor 4g at the bottom of the pulley, and the lower motor 4h at the bottom of the pulley, driving the hook 4d to complete vertical lifting and the gantry 4c to move horizontally, so as to transfer and position the workpiece on the basket 5.

[0075] Step 1 involves real-time collection of initial operation data using various sensors deployed within the intelligent cleaning zone 1. These sensors include a hook position sensor that continuously monitors the spatial coordinates of the workpiece mounted on the basket 5, a water level sensor that monitors the liquid level in the water tank 15, a temperature sensor that collects the surface temperature data of the workpiece, a vision sensor that captures the surface image of the workpiece and generates a surface dirt map, and a pressure sensor that detects the water pressure data in the main liquid path 23. All of this data is transmitted to the system control center 3 for centralized processing via wired or wireless communication.

[0076] Step 2 involves performing image analysis on the surface dirt map transmitted by the vision sensor within the system control center 3. A contamination level parameter is generated by calculating the ratio of the oil stain coverage area to the total surface area of ​​the workpiece. Based on preset thresholds, high, medium, and low contamination levels are defined. A cleaning path instruction is generated by combining workpiece type data with the contamination level parameter. The high contamination level instruction triggers the ultrasonic cleaner 7 to form a collaborative operation path with the first blow-wash robotic arm 9 and the second blow-wash robotic arm 11. The low contamination level instruction generates a direct ventilation drying station path. This instruction is transmitted to the control unit of the European-style double-girder crane 4 via the industrial bus.

[0077] Step 3 parses the cleaning path instructions into specific equipment control parameters: the cleaning fluid flow rate is controlled by adjusting the opening of the water tank output valve 16, the power output of the first heater 20 and the second heater 21 is adjusted according to the temperature feedback, the spray pressure values ​​of the first high-pressure water gun 9a and the second high-pressure water gun 11a are set, and the triggering conditions of the second air drying gun 11b are configured; when the air drying time reaches the preset value or the humidity sensor detects that the dryness of the workpiece surface meets the standard, the second blowing and washing robotic arm 11 sends an air drying completion signal to the system control center 3.

[0078] Step 4: The system control center 3 receives the drying completion signal transmitted by the second blowing and washing robotic arm 11. At the same time, it starts the vision sensor to perform secondary image acquisition on the cleaned workpiece. By comparing the differences in surface dirt patterns before and after cleaning, a quality judgment instruction is generated. The judgment results are divided into two categories: qualified products and rework products.

[0079] Step 5: The system control center 3 dynamically allocates the calculation tasks of generating cleaning path instructions, equipment control parameters, and quality judgment instructions to the multi-node programmable controller. The distributed computing architecture is used to improve the instruction processing efficiency. The control unit of the European double-girder crane 4 is responsible for parsing the path instructions and outputting displacement control signals to the drum drive motor 4b, the right side motor 4e of the gantry, the left side motor 4f of the gantry, the upper bottom motor 4g of the pulley, and the lower bottom motor 4h of the pulley, which work together to drive the hook 4d to complete the three-dimensional spatial positioning.

[0080] Step 6: After the multi-node programmable controller executes the quality judgment instruction, it feeds back the execution result data, including workpiece type, contamination level, cleaning parameters, and quality assessment results, to the dynamic allocation module. The contamination level judgment rules are optimized through machine learning algorithms to achieve adaptive iterative updates of the cleaning strategy.

[0081] This invention uses a dynamic allocation module and a dirt map analysis unit to convert surface dirt maps collected by a visual sensor into dirt level parameters. Oil stain coverage exceeding a set threshold is defined as a high dirt level, while coverage below the threshold is defined as a low dirt level. The task topology mapping unit generates cleaning path instructions based on the dirt level parameters: a high dirt level triggers a collaborative path between the ultrasonic cleaner 7, the first blowing robot arm 9, and the second blowing robot arm 11; a low dirt level triggers a direct ventilation drying station path. This instruction triggers a European-style double-girder crane 4 to transfer the workpiece onto a basket 5. The upper motor 4g and lower motor 4h at the bottom of the pulleys drive the gantry 4c to move horizontally, while the drum drive motor 4b controls the vertical lifting of the hook 4d, achieving automatic transfer from the disassembly area to the cleaning station. This closed-loop path decision-making and mechanical execution eliminates process interruptions caused by manual handling, improving space utilization.

[0082] The feedback control module decomposes the cleaning path instructions into equipment control parameters: the ultrasonic parameter adaptive unit dynamically adjusts the opening of the water tank output valve 16 and the power of the first heater 20 and the second heater 21 based on the water level threshold and temperature setpoint; the robotic arm coordination unit monitors the vibration data of the ultrasonic cleaner 7, and when the cavitation intensity exceeds the threshold, it synchronously reduces the water pressure of the first high-pressure water gun 9a and the second high-pressure water gun 11a to the set ratio, and links the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency; when the visual sensor detects that the residual moisture coverage exceeds the set threshold, it immediately triggers the action of the second air drying gun 11b of the second blowing robotic arm 11. The water pressure reduction action simultaneously starts the loop drive pump 24 of the liquid circuit 22 in advance, so that wastewater filtration takes priority over heating operation. Real-time feedback of sensor data and cross-system coordination of equipment actions eliminate the response delay of manual adjustment.

[0083] The decision-making closed-loop module extracts the oil residue area ratio, rust coverage, and reflectivity from secondary visual acquisition through a multi-threshold grading unit, and calculates the quality parameter α. When α exceeds the set threshold, the quality judgment unit outputs a spiral coverage control command, driving the first rotating motor 9f and the second rotating motor 11f of the base to generate a three-dimensional spiral trajectory, and synchronously adjusting the first drive motor 9c and the second drive motor 11c of the execution device to increase the tilt angle of the first high-pressure water gun 9a and the second high-pressure water gun 11a. During execution, the visual sensor provides real-time feedback on the surface reflectivity and iteratively calculates the α value. When α is lower than the set threshold, the decision-making closed-loop module immediately stops the spiral coverage control command and triggers the transfer of the European double-girder crane 4. The quality parameter α dynamically guides the adjustment of cleaning intensity, avoiding quality fluctuations caused by human experience misjudgment.

[0084] The load balancing module breaks down and distributes the computational tasks to three independent controller nodes: the first node processes ultrasonic control parameters, the second node generates the motion trajectories of the first and second cleaning robotic arms 9 and 11, and the third node calculates the mass parameter α. The queue optimization unit monitors the node load rate, migrating tasks and prioritizing obstacle avoidance commands when the load exceeds a set threshold. All load data is synchronized to the system's central control computer 12 via a local area network, achieving low-latency response to control commands. Execution result data is fed back to update the pollution level determination rules, forming a dynamic optimization closed loop.

[0085] Please see Figures 2 to 8 The specific embodiments of this invention are as follows: The system is activated in a marine turbocharger parts cleaning scenario. The visual sensor of the sensing device collects images of the workpiece surface and transmits them to the dynamic allocation module. The dirt map analysis unit identifies the contours of the oil stain area through grayscale analysis. When the oil stain coverage area exceeds 30%, a high pollution level parameter is output. The task topology mapping unit generates cleaning path instructions based on the pollution level parameters. The high pollution level triggers the ultrasonic cleaner 7 to work collaboratively with the first blowing robot arm 9 and the second blowing robot arm 11. The path instructions are transmitted to the control unit of the European double-girder crane 4 via the local area network. After parsing the instructions, the control unit outputs a displacement signal: the upper motor 4g and the lower motor 4h at the bottom of the pulley drive the gantry 4c to move horizontally along the top beam slide rail to the ultrasonic cleaning station. The drum drive motor 4b controls the hook 4d to descend and lift the workpiece onto the basket 5 to complete the transfer.

[0086] The feedback control module decomposes the cleaning path instructions into equipment control parameters: the ultrasonic parameter adaptive unit opens the water tank output valve 16 of the water storage tank 15 to inject water into the ultrasonic cleaner 7 according to the water level threshold; when the temperature sensor detects a deviation from the set value, it adjusts the power output of the first heater 20. The robotic arm coordination unit monitors the vibration data of the ultrasonic cleaner 7; when the cavitation intensity exceeds the threshold, it synchronously reduces the water pressure of the first high-pressure water gun 9a and the second high-pressure water gun 11a to 70%. The water pressure reduction action triggers the main drive pump 18 of the main liquid circuit 23 to operate at a reduced frequency, and the loop drive pump 24 of the liquid circuit loop 22 is started in advance. When the visual sensor detects that the residual moisture coverage exceeds 15%, it triggers the second air drying gun 11b to spray in a directional manner.

[0087] After receiving the drying completion signal, the decision-making closed-loop module extracts the oil residue area ratio, rust coverage, and reflectivity from the secondary visual acquisition, and calculates the quality parameter α. When α is greater than 0.1, the quality judgment unit outputs a spiral coverage control command: the first rotating motor 9f and the second rotating motor 11f drive the robotic arm to generate a three-dimensional spiral trajectory, and the first drive motor 9c synchronously increases the tilt angle of the first high-pressure water gun 9a by 20 degrees. During execution, the visual sensor collects the surface reflectivity in real time and iteratively calculates the α value. When α is less than or equal to 0.05, the decision-making closed-loop module stops the spiral command and triggers the European double-girder crane 4 to return to the initial position.

[0088] The load balancing module breaks down and distributes quality judgment instructions: the task sharding unit assigns equipment control parameter tasks to the first node to adjust the opening of the water tank output valve 16, cleaning path instructions to the second node to generate the robotic arm's motion trajectory, and quality parameter α calculation tasks to the third node. When the queue optimization unit detects that the load rate of the second node exceeds 80%, it transfers 20% of the trajectory planning tasks to idle nodes and inserts the obstacle avoidance instructions triggered by the infrared sensors of the European double-girder crane 4 at the front of the queue. The load rate data of each node is synchronized to the central control computer 12 via the local area network.

[0089] The cleaning results data are fed back to the dynamic allocation module: the dirt map analysis unit dynamically updates the pollution level judgment threshold based on the actual cleaning effect, for example, increasing the sensitivity of oil stain coverage area judgment by 5% for complex structural components such as turbocharger blades. The system solves the problem of incomplete removal of oil stains from complex gaps in ship parts through a closed-loop mechanism of automatic scheduling driven by pollution map, equipment control linked to cavitation intensity, and adjustment of cleaning intensity by feedback of quality parameter α, thereby reducing the interruption time between processes.

[0090] The technical features of this invention are explained below: The dirt map analysis unit is a dedicated processing unit within the dynamic allocation module. Its core function is to receive raw image data from the vision sensor and run image processing algorithms. These algorithms analyze the grayscale, texture, and color features of the image to identify and segment oily areas on the workpiece surface, ultimately calculating the precise percentage of oil-covered area relative to the total surface area, providing quantitative data input for subsequent contamination level determination.

[0091] Task Topology Mapping Unit: The Task Topology Mapping Unit is the decision-making core of the dynamic allocation module. It receives contamination level parameters and workpiece type data from the Fouling Map Analysis Unit. This unit has a built-in cleaning strategy rule base and can perform path planning calculations based on the process requirements of the cleaning task (such as cleaning intensity and step sequence) and the physical layout of the workshop (such as equipment location), generating a detailed cleaning path plan that includes a target coordinate sequence and movement instructions.

[0092] Ultrasonic parameter adaptive unit: The ultrasonic parameter adaptive unit is a sub-unit in the feedback control module that is specifically responsible for regulating the ultrasonic cleaning environment. It receives water level and temperature setpoints and compares them with sensor feedback in real time to automatically adjust the power of the water tank valves and heaters, so that the ultrasonic cleaner can operate under the optimal process parameters. Robotic Arm Coordination Unit: The robotic arm coordination unit is the sub-unit in this module responsible for controlling the blowing and washing action. It monitors the dynamic data of the cleaning process (such as cavitation intensity) and coordinates the control of the high-pressure water guns and water pressure on the two robotic arms accordingly to achieve adaptive adjustment of the cleaning intensity.

[0093] Multi-threshold grading unit: The multi-threshold grading unit is the data preprocessing part of this module. It is responsible for extracting multiple key quality features (such as β, γ, δ) from the secondary visual detection data and fusing them according to a preset formula to calculate a comprehensive quality score (α).

[0094] Quality Judgment Unit: The quality judgment unit is the decision-making part. It receives the comprehensive quality score (α) and makes a final judgment based on the preset quality pass and fail threshold range, and outputs the corresponding control instructions (such as returning to the initial work position or triggering enhanced cleaning).

[0095] Task Sharding Unit: The task sharding unit is responsible for receiving various computing tasks generated by the system, breaking them down and classifying them according to task type and computational load, and then distributing them reasonably to three different dedicated controller nodes for execution, thereby realizing the parallelization of computation and the initial distribution of load.

[0096] Queue Optimization Unit: The queue optimization unit is responsible for monitoring the running status of the entire computing cluster. It collects the load status of each control node in real time, and when it finds that a node is overloaded, it dynamically reallocates some of the tasks on that node to idle nodes. At the same time, it also manages a priority queue to ensure that high-priority emergency tasks (such as obstacle avoidance commands) can be responded to and processed immediately.

[0097] The dirt map analysis algorithm analyzes the workpiece surface image acquired by the vision sensor using image processing technology, identifies the outline of the oil stain area using grayscale analysis and edge detection methods, calculates the ratio of the oil stain coverage area to the total surface area of ​​the workpiece, and classifies the pollution level according to the preset percentage threshold.

[0098] The task topology mapping algorithm constructs a path planning model for cleaning tasks based on graph theory principles. It combines the layout of workshop equipment and the characteristics of workpiece type to generate the optimal cleaning path instruction sequence, realizing conflict-free transfer planning from the initial workstation to the target cleaning workstation.

[0099] The multi-threshold grading algorithm uses a multi-feature fusion calculation method to extract three features from secondary visual acquisition data: oil residue area ratio, rust coverage rate, and surface reflectivity. These features are then combined using a weighted summation formula to synthesize comprehensive quality parameters, providing a quantitative basis for quality judgment.

[0100] The quality judgment algorithm sets a dual threshold judgment rule for quality parameters based on the principle of fuzzy logic. When the quality parameter is lower than the qualified threshold, a process termination instruction is triggered. When the quality parameter is higher than the rework threshold, an enhanced cleaning program is started to achieve adaptive decision-making.

[0101] The task sharding algorithm adopts a load-aware task decomposition strategy. Based on the type characteristics and resource requirements of the computing tasks, the comprehensive instructions are split into discrete sub-task sets, and dynamically allocated according to the processing capabilities of each control node.

[0102] The queue optimization algorithm monitors the load status of distributed computing nodes in real time. When the node load exceeds the set threshold, the task migration mechanism is activated to redistribute some computing tasks to idle nodes. At the same time, a priority queue model is used to achieve immediate response to high real-time instructions.

[0103] The dirt map analysis model analyzes the workpiece surface image collected by the vision sensor through image processing algorithms, uses grayscale segmentation and contour extraction technology to identify the boundaries of oil stain areas, calculates the ratio of the oil stain coverage area to the total surface area of ​​the workpiece, and classifies the pollution level according to the preset percentage threshold.

[0104] The task topology mapping model constructs a cleaning path planning network based on graph theory algorithms. Combining the layout of workshop equipment and the characteristics of workpiece type, it generates the optimal motion trajectory sequence to achieve a conflict-free transfer scheme from the initial position to the target cleaning station.

[0105] The multi-threshold grading model uses a multi-feature fusion calculation method to extract three features from the secondary visual acquisition data: oil residue area ratio, rust coverage rate, and surface reflectivity. The comprehensive quality parameters are then synthesized through a weighted summation formula.

[0106] The quality assessment model uses fuzzy logic to set dual-threshold decision rules. When the quality parameter is lower than the qualified threshold, a process termination instruction is triggered. When the quality parameter is higher than the rework threshold, an enhanced cleaning procedure is started to achieve adaptive quality assessment.

[0107] The ultrasonic parameter adaptive model adjusts the water level valve opening and heater power through a closed-loop control algorithm, and dynamically adjusts the output parameters based on the deviation between the sensor feedback data and the set value to maintain stable cleaning liquid level and temperature.

[0108] The robotic arm collaborative model monitors the vibration data of the ultrasonic cleaner. When the cavitation intensity exceeds the threshold, it synchronously reduces the water pressure of the high-pressure water gun and triggers the action of the air drying gun based on the residual moisture coverage detected by the visual sensor.

[0109] The spiral coverage control model generates a three-dimensional spiral trajectory through kinematic algorithms, coordinates the base rotation motor and the actuator drive motor, and adjusts the spray angle and movement path of the high-pressure water gun.

[0110] Based on the type of computing task and the characteristics of resource requirements, the task sharding model uses a load-aware strategy to split the comprehensive instructions into discrete sub-task sets and dynamically allocate them to multi-node controllers.

[0111] The queue optimization model monitors the load status of distributed nodes in real time. When node overload is detected, a task migration mechanism is initiated, and a priority queue is used to process high real-time obstacle avoidance commands.

[0112] The pollution level determination rule update model analyzes historical cleaning result data through machine learning algorithms, dynamically optimizes the threshold parameters of the dirt map analysis unit, and achieves adaptive iterative improvement of the system.

[0113] α represents a comprehensive quality parameter, derived by weighting three indicators: oil residue area ratio, rust coverage, and reflectivity. It is used to comprehensively assess the cleaning quality level. This parameter uses weight ratios of 0.6, 0.3, and 0.1 to reflect the degree of influence of different contaminants on quality. The final value is used to determine whether the cleaning meets the standards or requires intensive treatment.

[0114] β represents the oil residue area ratio, which is the ratio of the area of ​​residual oil on the workpiece surface to the total surface area after secondary cleaning. This parameter is calculated by image processing algorithms after images are acquired by a vision sensor, reflecting the oil removal effect after ultrasonic cleaning and high-pressure water jet rinsing.

[0115] γ represents rust coverage, which refers to the percentage of the workpiece surface area covered by rust. This parameter is used to assess the degree of corrosion and the effectiveness of the cleaning process in removing rust by detecting rust features and calculating the coverage ratio using a visual sensor.

[0116] δ represents surface reflectivity, which is an indicator of the ability of a workpiece surface to reflect light after cleaning. This parameter is obtained by measuring the intensity of reflected light from the surface using an optical sensor, reflecting the cleanliness and smoothness of the surface after cleaning. A higher value indicates less surface residue.

[0117] The percentage thresholds (30%, 10%) in the pollution level parameter classification are preset fixed values. The system allows these thresholds to be dynamically adjusted through the system's central control computer according to the workpiece type and process requirements.

[0118] The collaborative path of the ultrasonic cleaner and the dual robotic arms includes a clear execution sequence: the ultrasonic cleaner first completes the basic cleaning, then the first blow-wash robotic arm performs high-pressure water rinsing, and finally the second blow-wash robotic arm performs the air-drying operation. This sequence cannot be reversed and the steps are sequential.

[0119] The trigger condition for the drying completion signal is an "OR" logic relationship, that is, the signal is sent immediately when either the preset time is reached or the sensor determines that the dryness meets the standard. The system prioritizes sensor data, and the timeout mechanism is only used as a backup.

[0120] The three-dimensional spiral trajectory in the spiral coverage control command is generated by the base rotary motor according to a preset algorithm. This algorithm controls the end effector of the robotic arm to move in an involute motion along the workpiece surface at a constant linear velocity, achieving coverage without blind spots.

[0121] The quality parameter α is calculated using a weighted summation model (0.6β+0.3γ+0.1δ). The weight coefficients are set based on the degree of influence of oil stains, rust, and reflection on the cleaning quality. The system allows for dynamic optimization of the weight allocation through a machine learning module.

[0122] The water pressure is reduced to 70% as a relative value. The baseline value is the current standard water pressure value of the process. The reduction action is achieved through a proportional valve to achieve linear pressure regulation, rather than a step-like sudden change, to avoid water hammer effect.

[0123] The "early start" of the liquid circuit drive pump is defined as starting at least 5 seconds before the main liquid circuit heater is energized, so that the filtration system runs before the heating system. This time difference can be adjusted by system parameters.

[0124] The criteria for determining an "idle node" in the load balancing module are that the load rate is below 40% for 30 consecutive seconds and the memory availability is above 60%, to avoid misjudging an idle state due to a momentary low load.

[0125] The calculation of "residual moisture coverage" in the secondary data acquired by the visual sensor adopts image binarization processing. The water film area and the dry area are distinguished by grayscale threshold segmentation, and the grayscale value of moisture reflection is set to 180-255.

[0126] The pollution level determination rule is updated using an incremental learning algorithm. Each time the data is fed back, only the decision boundary parameters of the determination model are adjusted without changing the basic classifier structure, thus achieving system stability.

Claims

1. A smart cleaning workshop linkage system, characterized in that, include: The sensing devices include a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor; The actuators include a European-style double-girder crane (4), an ultrasonic cleaner (7), a first blow-wash robotic arm (9), a blow-wash positioning mechanism (10), a second blow-wash robotic arm (11), and a cleaning fluid circulation system; The control device, which connects the sensing device and the actuating device via a network, includes: The dynamic allocation module receives the workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt map, and water pressure data from the sensing device. It converts the surface dirt map into pollution level parameters. The pollution level parameters are divided according to the percentage of oil stain coverage area to the total surface area of ​​the workpiece: a percentage of oil stain coverage area to the total surface area of ​​the workpiece is defined as high pollution level, a percentage of oil stain coverage area to the total surface area of ​​the workpiece is defined as medium pollution level, and a percentage of oil stain coverage area to the total surface area of ​​the workpiece is defined as low pollution level. Based on the workpiece type data and pollution level parameters obtained by the vision sensor, a cleaning path instruction is generated. The high pollution level triggers the preset ultrasonic cleaner (7) and the first blow-washing robot arm (9) and the second blow-washing robot arm (11) collaborative path. The medium pollution level triggers the preset ultrasonic cleaner (7) individual cleaning path. The low pollution level triggers the preset direct ventilation dry station path. The feedback control module receives the cleaning path instruction, decomposes the cleaning path instruction into equipment control parameters, and adjusts the water level valve opening, heater power, robotic arm water pressure value, and air drying trigger conditions. The decision-making closed-loop module receives the drying completion signal from the second blow-washing robotic arm (11) and the secondary data collected by the vision sensor, and outputs the quality judgment command. The load balancing module receives computing tasks from the dynamic allocation module, feedback control module, and decision-making closed-loop module, and distributes them to a multi-node programmable controller. The cleaning path command output by the dynamic allocation module triggers the feedback control module to generate equipment control parameters; The air drying trigger condition output by the feedback control module starts the air drying gun of the second blowing robot arm (11). When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, the air drying completion signal is sent to the decision closed loop module. The quality judgment command output by the decision-making closed-loop module is distributed to the controller node for execution via the load balancing module. The execution result data generated by the controller after executing the quality judgment instruction is fed back to the dynamic allocation module to update the pollution level judgment rules.

2. The intelligent cleaning workshop linkage system according to claim 1, characterized in that, The dynamic allocation module includes: The dynamic allocation module includes a task topology mapping unit, which is used to generate corresponding path planning according to the needs of the cleaning task. After generating the cleaning path instruction, the cleaning path instruction is transmitted to the control unit of the European double girder crane (4). After receiving the cleaning path instruction, the control unit of the European double girder crane (4) analyzes it and outputs displacement control signals to the drum drive motor (4b), the upper pulley bottom motor (4g), and the lower pulley bottom motor (4h) of the European double girder crane (4). The drum drive motor (4b) drives the drum to rotate according to the displacement control signal, thereby controlling the hook (4d) to complete the vertical lifting motion; At the same time, the upper motor (4g) at the bottom of the pulley and the lower motor (4h) at the bottom of the pulley work together to drive the gantry (4c) to move horizontally along the track; By combining vertical lifting and horizontal movement, the workpiece is accurately transported and positioned by the basket (5).

3. The intelligent cleaning workshop linkage system according to claim 2, characterized in that, The feedback control module includes: The ultrasonic parameter adaptive unit receives the water level threshold and temperature setting value in the cleaning path instruction. When the water level sensor detects a value lower than the water level threshold, it opens the output valve (16) of the water storage tank (15) in the cleaning fluid circulation system. When the temperature sensor detects a value more than ±2℃ away from the temperature setting value, the ultrasonic parameter adaptive unit adjusts the power of the first heater (20) and the second heater (21) to bring the temperature detection value back to the set value range. The robotic arm coordination unit receives the cavitation intensity threshold and residual moisture threshold in the cleaning path instruction. When the vibration data of the ultrasonic cleaner (7) exceeds the cavitation intensity threshold, the water pressure of the first high-pressure water gun (9a) of the first blowing robotic arm (9) and the second high-pressure water gun (11a) of the second blowing robotic arm (11) is reduced to 70%. When the residual moisture coverage rate detected by the visual sensor received by the dynamic allocation module exceeds 15%, the action of the second air drying gun (11b) of the second blowing robotic arm (11) is triggered. The cleaning path instructions include the following working modes: a preset working path of ultrasonic cleaner (7) and first blow-washing robotic arm (9) and second blow-washing robotic arm (11), a preset cleaning path of ultrasonic cleaner (7) alone, and a preset direct ventilation drying station path. The ultrasonic cleaner (7) and the first blow-washing robotic arm (9) and the second blow-washing robotic arm (11) are configured to work together as follows: the European double-girder crane (4) is instructed to transport the workpiece in a basket (5) to the ultrasonic cleaner (7) for ultrasonic cleaning, and after completion, it is transferred to the blow-washing positioning mechanism (10) where the first blow-washing robotic arm (9) is located for high-pressure water gun rinsing, and to the area where the second blow-washing robotic arm (11) is located for air drying; The ultrasonic cleaner (7) is configured with a separate cleaning path as follows: the European double-beam crane (4) is instructed to transport the workpiece on a basket (5) to the ultrasonic cleaner (7) for ultrasonic cleaning, and after completion, it is directly transported to the area where the second blow-washing robot arm (11) is located for air drying. The direct ventilation drying station path is configured such that the European double-girder crane (4) is instructed to directly transfer the workpiece loaded with basket (5) to the area where the second blowing and washing robot arm (11) is located for air drying; The water pressure reduction action triggers the first heater (20) and the second heater (21) of the main liquid circuit (23) of the cleaning fluid circulation system to heat up, and starts the loop filter device (25) of the liquid circuit (22) of the cleaning fluid circulation system.

4. The intelligent cleaning workshop linkage system according to claim 3, characterized in that, The decision-making closed-loop module includes: The multi-threshold grading unit receives secondary acquisition data from the vision sensor, extracts the oil residue area ratio β, rust coverage γ, and reflectivity δ, calculates the quality parameter α=0.6β+0.3γ+0.1δ, and outputs it to the quality judgment unit. The quality judgment unit receives the quality parameter α. When α≤0.05, it triggers the European double-girder crane (4) to return to the initial position. When α>0.1, it triggers the spiral cover control command. The spiral coverage control command adjusts the first rotary motor (9f) of the base of the first blow-washing robotic arm (9) and the second rotary motor (11f) of the base of the second blow-washing robotic arm (11) to generate a three-dimensional spiral trajectory. At the same time, it triggers the first drive motor (9c) of the execution device of the first blow-washing robotic arm (9) and the second drive motor (11c) of the execution device of the second blow-washing robotic arm (11) to increase the tilt angle of the first high-pressure water gun (9a) and the second high-pressure water gun (11a) by 20°, and adjusts the heating power of the ultrasonic cleaner (7) to increase the water temperature by 5°C.

5. The intelligent cleaning workshop linkage system according to claim 4, characterized in that, The load balancing module includes: The task sharding unit receives the computation tasks from the dynamic allocation module, the feedback control module, and the decision closed-loop module, and decomposes the cleaning path instruction task from the dynamic allocation module, the equipment control parameter task from the feedback control module, and the quality judgment instruction task from the decision closed-loop module to three independent controller nodes. The queue optimization unit receives the load rate feedback data of the controller node. When the load rate of any controller node among the three independent controller nodes is > 80%, it transfers 20% of the computing tasks of the controller node to the idle node with the lowest load rate. It receives the obstacle avoidance control commands generated by the drum drive motor (4b), the right side motor (4e), the left side motor (4f), the upper bottom motor (4g), and the lower bottom motor (4h) of the European double girder crane (4) during operation, and inserts the obstacle avoidance control commands into the front of the task queue for priority execution.

6. The intelligent cleaning workshop linkage system according to claim 5, characterized in that, Also includes: The cleaning path instruction generated by the task topology mapping unit in the dynamic allocation module is transmitted to the control unit of the European double-girder crane (4). The control unit parses the cleaning path command and outputs displacement control signals to the upper pulley bottom motor (4g), lower pulley bottom motor (4h), and drum drive motor (4b) of the pulley assembly (4a) of the European double girder crane (4). The upper motor (4g) at the bottom of the pulley and the lower motor (4h) at the bottom of the pulley drive the frame (4c) of the European double girder crane (4) to move horizontally, and the drum drive motor (4b) controls the hook (4d) of the European double girder crane (4) to lift vertically. The workpiece is transferred and positioned by the basket (5) through coordinated control of horizontal movement and vertical lifting.

7. The intelligent cleaning workshop linkage system according to claim 6, characterized in that, Also includes: The water pressure reduction action triggers the main drive pump (18) of the main liquid circuit (23) to operate at a reduced frequency; The water pressure reduction action of the robotic arm coordination unit triggers the main drive pump (18) of the main liquid circuit (23) of the cleaning fluid circulation system to operate at a reduced frequency. The water pressure reduction action simultaneously triggers the start time of the loop drive pump (24) of the liquid circuit (22) of the cleaning fluid circulation system to be brought forward to before the heating operation of the main liquid circuit (23); After the loop drive pump (24) starts, it draws wastewater into the loop filter device (25) of the liquid circuit (22). The filtered wastewater flows back to the water storage tank (15) of the cleaning fluid circulation system through the loop back pressure valve (26) of the liquid circuit (22).

8. The intelligent cleaning workshop linkage system according to claim 7, characterized in that, Also includes: The spiral coverage control command is input to the first rotary motor (9f) and the second rotary motor (11f) of the base, controlling the first blow-washing robotic arm (9) and the second blow-washing robotic arm (11) to rotate and generate a three-dimensional spiral trajectory; The first drive motor (9c) of the first blow-washing robotic arm (9) and the second drive motor (11c) of the second blow-washing robotic arm (11) are synchronously input, and the tilt angle of the first high-pressure water gun (9a) of the first blow-washing robotic arm (9) and the second high-pressure water gun (11a) of the second blow-washing robotic arm (11) is increased by 20°. During the execution of the spiral cover control command, the vision sensor collects the surface reflectivity δ in real time and transmits it to the decision closed-loop module. When the recalculated quality parameter α ≤ 0.05, the decision closed-loop module stops the spiral cover control command.

9. The intelligent cleaning workshop linkage system according to claim 8, characterized in that, Also includes: The task sharding unit of the load balancing module distributes the decomposed computing tasks to three independent controller nodes; One of the controller nodes receives the equipment control parameter task from the feedback control module, which is used to adjust the opening of the water tank output valve (16) of the ultrasonic cleaner (7) and the power of the first heater (20) and the second heater (21); Another controller node receives the cleaning path instruction task from the dynamic allocation module and is used to generate the motion trajectory of the first blow-washing robot arm (9) and the second blow-washing robot arm (11). The third controller node receives the quality judgment instruction task from the decision-making closed-loop module and is used to calculate the quality parameter α of the decision-making closed-loop module. The queue optimization unit of the load balancing module collects the load rate feedback data of the three independent controller nodes and transmits it to the system control computer (12) of the system control center (3) via the local area network.

10. A method for intelligent cleaning workshop linkage control, applied to the intelligent cleaning workshop linkage system according to any one of claims 1 to 9, characterized in that, include: Step 1: Receive workpiece coordinates, water level in the reservoir, workpiece surface temperature, surface dirt pattern, and water pressure data through a sensing device, wherein the sensing device includes a hook position sensor, a water level sensor, a temperature sensor, a vision sensor, and a pressure sensor. Step 2: Convert the surface dirt map into a contamination level parameter. The contamination level parameter is divided according to the percentage of the oil stain coverage area to the total surface area of ​​the workpiece: the percentage of the oil stain coverage area to the total surface area of ​​the workpiece is higher than 30% and is defined as high contamination level; the percentage of the oil stain coverage area to the total surface area of ​​the workpiece is between 10% and 30% and is defined as medium contamination level; the percentage of the oil stain coverage area to the total surface area of ​​the workpiece is lower than 10% and is defined as low contamination level. Based on the workpiece type data and contamination level parameter obtained by the vision sensor, a cleaning path instruction is generated. The high contamination level triggers the ultrasonic cleaner (7) and the first blow-washing robot arm (9) and the second blow-washing robot arm (11) to work together. The low contamination level triggers the direct ventilation drying station path. Step 3: Decompose the cleaning path instruction into equipment control parameters, adjust the water level valve opening, heater power, robotic arm water pressure value and air drying trigger condition, wherein the water level valve opening corresponds to the water tank output valve (16) of the cleaning fluid circulation system, the heater power corresponds to the first heater (20) and the second heater (21), the robotic arm water pressure value corresponds to the water pressure of the first high-pressure water gun (9a) and the second high-pressure water gun (11a), and the air drying trigger condition starts the second air drying air gun (11b) of the second blowing robotic arm (11). When the air drying reaches the preset time or the sensor determines that the surface dryness of the workpiece meets the standard, send the air drying completion signal. Step 4: Receive the drying completion signal and secondary data collected by the vision sensor from the second blow-washing robotic arm (11), and output the quality judgment command; Step 5: Assign computing tasks to the multi-node programmable controller, wherein the computing tasks include generating cleaning path instructions, equipment control parameters and quality judgment instructions; Step 6: After the controller executes the quality judgment instruction, the execution result data generated is fed back to the dynamic allocation module to update the pollution level judgment rules; In this process, the cleaning path instruction generated in step 2 is transmitted to the control unit of the European double girder crane (4). After parsing, the control unit outputs displacement control signals to the drum drive motor (4b), the upper motor at the bottom of the pulley (4g), and the lower motor at the bottom of the pulley (4h), driving the hook (4d) to complete vertical lifting and horizontal movement of the gantry (4c) in order to transfer and position the workpiece on the basket (5).

Images