Multi-sensor fusion and hydraulic cooperative control deep water base bed leveling equipment and operation method

Abstract

The present application belongs to the technical field of deep water foundation construction, and specifically relates to a multi-sensor fusion and hydraulic coordinated control deep water foundation leveling equipment and operation method, which comprises: early preparation and positioning, frame adjustment, feeding and leveling operation, excess material control and displacement, data feedback and optimization. Compared with the prior art, the present application not only breaks through the scene and functional limitations of the prior art which is only suitable for ≤10m shallow water, single sensor positioning and hydraulic independent control, but also specifically solves the problems of principle deviation in measurement, lack of collaborative logic in driving and imbalance between equipment and project demand in small and medium-sized projects. Through multi-system deep integration, efficient adaptation to complex sea conditions of 0-20m water depth, wind force ≤6, wave height ≤1.0m and flow rate ≤1.0m / s is achieved, filling the technical gap of integrated intelligent construction of "perception-decision-execution-protection" of small and medium-sized deep water foundations in the industry.

Description

Technical Field

[0001] This invention belongs to the field of deep water foundation construction technology, specifically involving multi-sensor fusion and hydraulic collaborative control deep water foundation leveling equipment and operation method. It is applicable to the intelligent leveling construction of small and medium-sized deep water foundation projects with ≤50 caissons and operating water depths of 0-20m in port terminals, cross-sea channels, breakwaters and other projects. Background Technology

[0002] In port engineering construction, deep-water subgrade leveling is a crucial preliminary process to ensure the accuracy of subsequent caisson installation, structural stability, and engineering durability. While current mainstream leveling technologies can meet the needs of foundation construction, they have significant limitations in terms of efficiency, accuracy, cost, and adaptability to different scenarios, as detailed below:

[0003] 1. Limitations of traditional manual underwater leveling

[0004] Manual underwater leveling relies on direct underwater operations by divers and was the main method for early deep-water foundation construction, but it has insurmountable drawbacks:

[0005] Extremely low efficiency: Due to the limitations of diving depth usually ≤10m, single operation time ≤1h and sea state with wind force ≤4 and wave height ≤0.5m, the leveling operation time for a single container bed can be as long as 2.5-3 days, which cannot meet the needs of tight project schedule;

[0006] Poor accuracy and instability: Relying on the experience of divers, the leveling accuracy is poor and is affected by water flow and waves. The flatness deviation of different areas in the same caisson position can reach 15cm, and additional leveling costs are required for subsequent caisson installation.

[0007] High safety risks: Divers face safety hazards such as hypoxia, decompression sickness, and swell impact. Underwater visibility is limited, and they are prone to work-related injuries due to falling rocks or equipment collisions.

[0008] High labor intensity: The underwater working environment is harsh, divers' physical strength is quickly depleted, the effective working time per day is less than 4 hours, and labor costs account for more than 30% of the total construction cost of the project.

[0009] 2. Limitations of specialized leveling vessel operations

[0010] While professional leveling vessels offer improved efficiency compared to manual leveling, their large size and high cost make them poorly suited for small to medium-sized projects.

[0011] The equipment costs are high: the cost of a single professional leveling vessel exceeds 100 million yuan, with annual depreciation costs of approximately 8-10 million yuan. In addition, the cost of a single dispatch, including channel dredging and vessel towing, is estimated at 500,000 to 1 million yuan, which is difficult for small and medium-sized projects to afford.

[0012] Strict scene restrictions: Due to the requirement that the water depth of the channel must be ≥8m and the working space must be ≥50m×30m, it is not possible to enter shallow water harbor basins, narrow channels or small near-shore project sites.

[0013] Low resource utilization: Small and medium-sized projects typically have ≤50 caissons. The single operation cycle of a professional leveling vessel is short, generally ≤15 days, but the round-trip dispatch time is as long as 7-10 days, resulting in an equipment idle rate of over 60% and a waste of resources.

[0014] 3. Limitations of underwater foundation leveling machines

[0015] Chinese patent CN110485492A discloses an underwater bed leveling machine consisting of a "frame + single hydraulic drive + foundation positioning" system. While it achieves partial mechanization, it still suffers from multiple technical shortcomings and cannot meet the demands of deep-water, high-precision operations.

[0016] The sensing system is simple and lacks accuracy: it only relies on GPS positioning and prism measurement to achieve basic position monitoring, without attitude monitoring and force feedback functions. The leveling accuracy can only reach ±5cm, and the attitude deviation cannot be corrected in real time when disturbed by water flow, making it difficult to adapt to complex water environments with a depth of >10m.

[0017] The hydraulic control is independent and the anti-interference ability is weak: the outrigger cylinder and the leveling head drive are independently controlled and there is no coordinated adjustment logic. When encountering waves with a wave height of >0.5m and water flow with a velocity of >0.5m / s, the outrigger extension and retraction are not synchronized and the leveling head material is offset, resulting in local leveling deviation.

[0018] Suitable for shallow water, but lacking in deep water performance: It is not designed with deep water sealing structures such as IP68-level sealing box and watertight joints, and can only operate in water depths of ≤10m. Moreover, it has no anti-tipping device, and the frame is at high risk of tipping over under the impact of deep water flow.

[0019] Low modularity and difficult relocation: The frame is an integral structure without standardized disassembly interfaces. Each module weighs over 50t and requires large lifting equipment such as cranes with a capacity of ≥100t for transportation. The relocation cost is high and the efficiency is low, which is not suitable for the construction characteristics of small and medium-sized projects with "multiple construction sites and short cycles".

[0020] 4. Limitations of the combined positioning method for deep-water gravel subgrade

[0021] Chinese patent CN118549966A, with "GNSS-RTK + ultra-short baseline + wire surveying" as its core, has overcome the challenge of positioning in deep water up to 100 meters deep. However, it has unavoidable shortcomings in small and medium-sized deep water projects, directly resulting in its inability to meet the requirements of "high precision, high efficiency, and low cost."

[0022] (1) The measurement system has a fundamental deviation, and the accuracy cannot be consistently met.

[0023] The shortcomings of ultra-short baseline for shallow water depth adaptation: Its planar positioning relies on ultra-short baselines. The accuracy of this technology depends on the "acoustic signal propagation time difference calculation". In shallow water depth scenarios of 0-20m: ① The signal propagation distance is short, and the proportion of time difference measurement error increases significantly. The actual planar positioning error can reach ±8~12cm, far exceeding the accuracy requirement of ±5cm for small and medium-sized projects; ② The acoustic signals of underwater beacons and transmitting and receiving units are easily affected by near-shore sediment scattering interference, resulting in a high signal loss rate. Repeated retesting is required, and the single positioning time exceeds 30 minutes, which is inefficient.

[0024] Mechanical deformation error of wire measurement: The elevation measurement uses a steel wire and relies on an absolute encoder to calculate the length, but does not consider: ① the elastic deformation of the steel wire under deep water pressure; ② the "false elongation" caused by pulley friction; ③ the lateral displacement of the wire under the impact of water flow. The combination of these three factors makes the actual elevation error reach ±12~18cm, which is completely unable to meet the centimeter-level accuracy required for caisson installation.

[0025] (2) There is no coordinated logic between drive and control, and the stability of operation depends on manual operation.

[0026] The outriggers and leveling head operate independently, resulting in weak anti-interference capabilities: No hydraulic collaborative drive system is designed; outrigger extension and retraction rely on manual valve control, and leveling head movement requires separate winch adjustment. There is no linkage logic between the two: ① When encountering 0.5m high swells, the frame tilts ≥0.5°. If the outriggers are not adjusted in time, the leveling head fabric will exhibit the problem of "one side being too thick and the other too thin"; ② There is no matching algorithm between the feeding speed and the leveling head movement speed, leading to lag in manual judgment and easily causing "excessive material accumulation" or "insufficient material and missed areas," requiring rework.

[0027] The lack of automated path planning and low standardization mean that path leveling requires manual placement of marker stakes underwater, which are prone to shifting due to water flow, resulting in path deviation. Furthermore, there is no "missed point detection" function, requiring manual re-measurement with a depth sounder at each workstation. If a missed point is found, the equipment needs to be moved again, resulting in a total work time of over 6 hours per workstation, which is far higher than the "4 hours / workstation" schedule requirement for small and medium-sized projects.

[0028] (3) The equipment structure is mismatched with the needs of small and medium-sized projects, resulting in poor economic efficiency.

[0029] The relocation cost of the integrated structure is high: the leveling device including the measuring tower is an integrated design, weighing more than 120t, requiring: ① hoisting with a crawler crane of ≥150t; ② transfer by a special transport ship; ③ a hoisting site of ≥50m×30m on the project site. However, the total cost of a single project is low for small and medium-sized projects, and the cost of equipment relocation and site costs account for a high proportion, resulting in an economic imbalance.

[0030] The fixed design lacks adaptability: the fixed height of the measuring tower makes it prone to collisions with passing ships when operating in narrow waterways near the shore, requiring additional anti-collision piers and further increasing project costs.

[0031] 5. Limitations of underwater crushed stone subgrade laying devices and methods

[0032] Chinese patent CN110004933A discloses a hull-type riprap laying device with folding pontoon rails. By alternately positioning the hull and pontoon rails, the chute slides along the length of the hull to achieve "Z"-shaped riprap laying. This solution is suitable for large-scale construction, but it has the following shortcomings in small to medium-sized deep-water bed leveling scenarios:

[0033] (1) The equipment is large and its applicable scenarios are limited: The device is based on a large ship hull, with many supporting structures and a large overall size. It is suitable for open waters and large-scale projects, but not for near-shore, narrow waterways and small and medium-sized projects. The relocation cost is high.

[0034] (2) The main method is paving, and the leveling accuracy is limited: This scheme mainly realizes the stone throwing and initial paving, without setting up a dedicated fine leveling mechanism. The leveling relies on the stone throwing process control, which is difficult to meet the high precision requirements required for caisson installation. There are few monitoring points and a lack of closed-loop monitoring of the overall posture, stress and actuator of the equipment. Under external interference such as water flow, the uniformity of material distribution and the stability of elevation accuracy are poor.

[0035] (3) There are deficiencies in construction path and efficiency: its “Z” shaped construction is a segmented straight paving and hull lateral movement operation mode. The paving path is fixed for each time, and the single work station needs to be moved and repositioned multiple times, and the auxiliary operation time is long.

[0036] (4) Independent control and low degree of automation: Each actuator is controlled independently, lacking collaborative control logic between outriggers, walking and feeding, and has no automatic path planning and accuracy compensation function, making it less adaptable to external interference.

[0037] (5) Inadequate deep-water protection and safety structure: The device does not have a dedicated sealing protection, anti-overturning and overload protection structure for deep-water operations. The protection level of underwater components is insufficient, and the reliability and safety of deep-water operations need to be improved.

[0038] In summary, existing technologies cannot simultaneously meet the requirements of "high precision, high efficiency, low cost, and strong adaptability" for small and medium-sized deep-water subgrade projects. Developing a set of intelligent leveling equipment and operation methods that integrates "multi-sensor fusion perception, hydraulic collaborative drive, deep-water safety assurance, and modular transfer" is the key to breaking through the industry's technical bottlenecks and promoting the intelligent transformation of port construction. Summary of the Invention

[0039] The core objective of this invention is to address the technical pain points of traditional manual leveling, which suffers from low efficiency, poor accuracy, and high risk; the high cost and weak adaptability of specialized leveling vessels; and the limitations of single-sensor positioning, independent hydraulic control, suitability only for shallow water ≤10m, measurement biases, lack of collaborative drive logic, and mismatch between equipment and the needs of small and medium-sized projects. By deeply integrating five major systems—modular leveling framework, multi-sensor fusion perception, hydraulic collaborative drive, automated control, and safety assurance—a fully automated operation system encompassing "perception-decision-execution-assurance" is constructed. Simultaneously, standardized operating methods are implemented to achieve efficient, accurate, and safe construction of small and medium-sized deep-water foundation beds in water depths of 0-20m.

[0040] To achieve the above objectives, the present invention adopts the following technical solution:

[0041] This deep-water bed leveling equipment utilizes multi-sensor fusion and hydraulic collaborative control. Each system in this equipment features innovative designs to address current technological limitations. The specific structure and functions are as follows:

[0042] 1. Modular leveling frame – solves the problems of “monolithic structure, difficult relocation, and poor adaptability”.

[0043] The design employs a "segmented design + standardized interface" approach, balancing structural stability with flexibility for site transitions, as detailed below:

[0044] The main frame is constructed from three sections of 8m-long H800×280×12×20 steel profiles made of Q355B material with a yield strength ≥345MPa, connected by M30 high-strength bolts of performance grade 8.8 and preload torque 800-850N・m. The ends are reinforced with welded φ630×16mm round tube beams made of Q355B material, forming a closed load-bearing structure of "long beam-short beam-round tube beam," with an overall bending stiffness ≥5×10⁻⁶. 5 The pressure rating is N·m², which can withstand water flow impact of ≤1.0m / s. The frame is equipped with detachable DN800 flanges on both sides, with a pressure rating of 1.6MPa. The sealing surface uses raised face + oil-resistant rubber gasket, which facilitates later maintenance and component replacement. The effective internal leveling area is 20.5×9.5m, which can cover the leveling needs of half of a conventional caisson. Compared with the "small area single workstation" design, it reduces the number of workstation switching times by 50%.

[0045] Stone-filling chute: Addressing the issues of "no segmented design and difficult disassembly and assembly," it adopts a segmented structure: 800mm diameter, 12mm wall thickness, Q355B material, service life ≥5000㎡, total length 23m, consisting of a 3.5m base section (including the material inlet), a 1.5m long flared 2000mm diameter feeding inlet, two 6m standard sections, and two 3m adjusting sections. Each section is connected by a DN800 flange and equipped with 24 M24 bolts for quick disassembly and assembly. A single person can complete the disassembly and assembly of a single section within 30 minutes, solving the problem of transporting the entire chute. The lower leveling head of the chute is equipped with a 30° flared opening to expand the material distribution range to 1.5m wide and a 1.5×1.5m scraper plate. A high-level alarm is installed 8m away from the feed inlet. When the height of the stone in the chute is greater than 8m, the alarm is triggered and the feeding is stopped. A low-level alarm is installed 2m away from the feed inlet. When the height of the stone in the chute is less than 2m, the alarm is triggered and the material distribution stops, thus avoiding the problem of "no feeding control and stone overflow".

[0046] Transportation adaptability: The maximum weight of a single module is ≤30t, the main frame is 28t, and a single section of the chute is 2.5t. It is compatible with a 25t flatbed truck with a wheelbase of 6m and a bearing surface size of 12×2.5m. No large hoisting equipment is required for transportation. When relocating, it only needs to be disassembled into several modules. Compared with "an integral frame weighing more than 50t, which requires a crane of ≥100t for transportation", the relocation cost is reduced by 80% and the relocation efficiency is increased by 3 times.

[0047] 2. Multi-sensor fusion perception system – solving the problems of “single GPS positioning, no attitude or force feedback, and low accuracy”.

[0048] By integrating four types of high-precision sensors, a three-dimensional monitoring network is constructed to provide millisecond-level accurate data support for automated control, as detailed below:

[0049] Dual-antenna Beidou RTK positioning module: Addressing the issue of "poor horizontal and vertical accuracy with single GPS positioning," this module employs a dual-antenna design, installed on both sides of the frame receiving port, achieving a horizontal accuracy of ±(10+1×10⁻⁻⁴). 6 ×D)mm, elevation accuracy ±(20+1×10⁻ 6 ×D)mm, where D is the baseline length in mm, the data update rate is 1Hz, and the positioning deviation is ≤±2mm, providing a benchmark for frame spatial positioning and elevation calibration.

[0050] Dual-axis tilt sensor: Addressing the issue of "no attitude monitoring and inability to correct frame tilt", this sensor employs MEMS technology, with a measurement range of ±5°, accuracy of ±0.1°, and zero drift ≤0.01° / h. It is installed at the bottom of the central crossbeam of the frame, 0.5m from the bottom surface of the frame, and collects the lateral and longitudinal tilt angles of the frame in real time. The data is transmitted via RS485 bus with a delay of ≤10ms, providing a basis for leveling adjustment and avoiding leveling deviations caused by surges.

[0051] Magnetostrictive displacement sensor: To address the issues of "no outrigger extension and retraction monitoring and poor synchronization", it is installed on the outside of the cylinder of the four outrigger cylinders. It has a range of 0-1.2m, an accuracy of ±0.05%FS, a resolution of 0.01mm, and collects the outrigger extension and retraction in real time. The data transmission frequency is 1kHz, ensuring that the synchronous extension and retraction error of the four outriggers is ≤±2mm.

[0052] Pressure transmitter: To address the issue of "no outrigger force monitoring and high overload risk", it is installed at the rodless chamber oil port of the outrigger cylinder. The measurement range is 0-30MPa, the accuracy is ±0.25%FS, and it collects outrigger force data in real time. When the force exceeds 50t, which is 1.2 times the rated force, an overload alarm is triggered, and the cylinder is controlled to stop moving to avoid damage to the frame structure.

[0053] 3. Hydraulic Cooperative Drive System – Solves the problems of “independent hydraulic control, weak disturbance rejection, and asynchronous action”.

[0054] With "dual pumps, one in use and one on standby + diversion and collection valve assembly" as the core, the outriggers and leveling head are driven in tandem to counteract the interference of surge waves and water flow, as detailed below:

[0055] Power source: A dual-plunger hydraulic pump with a displacement of 10.56L / min, rated pressure of 25MPa, and maximum speed of 2000r / min is used. Installed in the mother ship's hydraulic station, it employs a one-in-one-outstand design. When the main pump pressure fluctuates by more than ±0.5MPa or the flow rate drops by 10%, the standby pump automatically switches via an electromagnetic directional valve, with a switching time of ≤1s, ensuring continuous operation. Compared to the "single-pump drive, immediate stop upon failure" design of CN110485492A, operational reliability is improved by 100%. It is equipped with a 120L hydraulic oil tank, including an oil temperature sensor, oil level gauge, and air filter, maintaining the working oil temperature between 30-60℃.

[0056] Outrigger drive unit: The four outrigger cylinders have an inner diameter of φ194mm, a piston rod diameter of φ160mm, and a stroke of 1m. Through a flow divider and collector valve assembly, the synchronization accuracy is ≤3%, enabling synchronous extension and retraction of the four outriggers at a speed of 0.4m / min.

[0057] The leveling head drive unit adopts a "gear and rack transmission + hydraulic motor reducer" linkage structure. Addressing the issues of poor leveling head movement accuracy and uneven material distribution, the specific design is as follows: the gear and rack have a module of 8 and 23 teeth, installed inside the sliding beam; the hydraulic motor reducer outputs a torque of 200 N·m and a speed of 15 r / min, driving the gear rotation via a coupling, which in turn moves the leveling head along the sliding beam. The material distribution speed is 1.05 m / min, and the path deviation is ≤ ±1 cm. Load-bearing wheels are installed at the connection between the sliding beam and the frame to reduce frictional resistance and ensure smooth movement.

[0058] 4. Automated control system – solves the problem of “reliance on manual operation and large fluctuations in accuracy”.

[0059] Using a PLC as the core, a closed-loop control system of "data acquisition - algorithm processing - action execution" is achieved, reducing manual intervention, as detailed below:

[0060] Hardware composition: The core controller adopts Siemens PLC, integrating a 4G or Bluetooth wireless transmission module with a transmission distance ≤500m and a data rate ≥1Mbps, as well as two button-type remote controls; equipped with a 7-inch MCGS touch screen, which displays the cylinder extension length, outrigger force, leveling head position and equipment operating status (such as pump start / stop, alarm information) in real time, which is convenient for operators to monitor. The cylinder extension length accuracy is ±0.1mm, the outrigger force accuracy is ±0.1MPa, and the leveling head position accuracy is ±1cm.

[0061] Control logic:

[0062] Positioning logic: Receives dual-antenna BeiDou RTK data, calculates the deviation between the actual position of the frame and the preset position through a "coordinate comparison algorithm", with an X / Y axis deviation of ≤ ±2cm, and outputs commands to control the crane for fine-tuning. Compared with "manual observation positioning", the positioning accuracy is improved by 25 times.

[0063] The levelness adjustment logic is as follows: It receives data from the bidirectional inclinometer and uses a PID algorithm to calculate the adjustment amount of the outriggers. The proportional coefficient is 1.2, the integral time is 0.5s, and the derivative time is 0.1s. It prioritizes controlling the extension of the cylinder on the side with lower elevation at a speed of 0.2m / min. If the cylinder extension length reaches 90% and the level is still not achieved, it controls the cylinder on the side with higher elevation to descend at a speed of 0.1m / min until the levelness of the frame is ≤0.1°.

[0064] Path planning logic: Automatically generates a serpentine leveling path. The overlap length can be set to 5-10cm via the touch screen, with a default of 8cm. The single leveling width is 1.38m. After each lateral movement in the longitudinal direction, the path moves 1.3m. The single leveling width minus the overlap length ensures no areas are missed. Compared to "manual path planning", the fabric laying efficiency is increased by 2 times.

[0065] Accuracy compensation logic: When the leveling accuracy deviation exceeds ±3cm, the system automatically triggers adjustment by comparing the Beidou RTK and total station. If the deviation is caused by the tilt angle >0.1° due to the frame tilt, the corresponding support leg cylinder is adjusted, with a single support leg adjustment amount ≤5cm. If the deviation is caused by uneven material distribution, the leveling head path is corrected, with an offset amount ≤2cm and a compensation response time ≤2s. Compared with "manual re-measurement and adjustment", the accuracy correction efficiency is improved by 10 times.

[0066] Operating modes: Supports both "control room touch screen operation" and "on-site remote control operation" modes. The two modes can be seamlessly switched via an emergency stop button with a switching time of ≤0.5s, making it suitable for mother ship control room monitoring or on-site emergency adjustment scenarios.

[0067] 5. Safety Protection System – Solves the problem of “shallow water compatibility, lack of deep water sealing / anti-tipping design”

[0068] To address the risks associated with operations in water depths of 0-20m, a triple protection system of "sealing, anti-tipping, and residual material control" is designed, as detailed below:

[0069] Deep-water sealing unit: Addressing the issue of "lack of deep-water sealing and easy water ingress damage to components," the hydraulic valve assembly and electrical components, including the PLC and sensor junction boxes, are installed in a 316L stainless steel underwater sealing enclosure. This enclosure has a wall thickness of 10mm, dimensions of 800×600×500mm, and an IP68 protection rating. A pressure balancing valve is installed on the top of the enclosure to accommodate pressure changes from 0-2.5MPa, preventing deformation of the enclosure due to deep-water pressure. Cables and oil pipes use M20×1.5 type watertight joints with an IP68 sealing rating and a water pressure resistance of ≥3MPa. Waterproof tape is wrapped around the joints to resist seawater corrosion and ensure no leakage at a depth of 20m.

[0070] Anti-tipping unit: To address the problem of "no anti-tipping device and easy to tip over in deep water", four GCr15 anti-tipping hooks are symmetrically arranged at the bottom of the gearbox. They are 200mm in diameter, 50mm in width, and have a hardness of HRC60-62. The bottom of the hook is 10cm from the bottom of the frame and can resist the impact of water flow with a velocity of ≤1.0m / s.

[0071] The operation method of the deep-water bed leveling equipment with multi-sensor fusion and hydraulic collaborative control is based on the above equipment and equipped with a standardized operation process. Based on the three-stage process of "positioning-laying-inspection", it adds the steps of "land pre-test", "residual material control" and "data feedback optimization". The specific steps are as follows:

[0072] S1: Preliminary Preparations and Positioning

[0073] Land-based pre-test: Before operation, a 25×10m simulated base bed was constructed on land. The foundation was C30 concrete, and the surface was covered with crushed stone of three specifications: 30-60mm, 60-90mm, and 60-120mm, to simulate the actual base bed stone gradation. After assembling the equipment modules, performance tests were conducted: ① Multi-sensor system consistency test: Beidou RTK and total station, accuracy ±1.5+2×10⁻ 6×D mm comparison, positioning deviation ≤±2mm; ② Hydraulic system synchronization test: the four leg cylinders extend synchronously by 1m, and the extension error ≤±2mm; ③ Leveling accuracy test: the leveling head walks along a serpentine path, and the total station is used to check according to a 2m×2m grid. The leveling accuracy is ≤±1cm. Ensure that the equipment performance meets the design requirements before transporting it to the site to avoid the problem of "directly launching into the water and difficulty in troubleshooting".

[0074] Marine positioning: ① Parameter input: Import the coordinates of the first and last ends of the leveling frame in the engineering design into the automated control system with an accuracy of ±1mm for the X / Y / Z axes; ② Crane transfer: Use a 150-ton crawler crane with an operating radius of 15-20m and a rated lifting capacity of ≥50t to lift the frame. Attach a 20m long, 20mm diameter nylon sway rope to the crane's wire rope with a breaking strength of ≥50kN. Adjust the horizontal rotation angle of the frame by the sway rope, with a deviation of ≤0.3°; ③ Suspension stabilization: Lower the frame to 50cm from the bed and suspend it for ≥30s. After the frame's sway amplitude due to water flow is ≤±5cm, prepare for the next operation. Positioning stability is improved by 80%.

[0075] S2: Framework Adjustment

[0076] Outrigger bottoming out: The hydraulic system is activated, and the four outrigger cylinders are controlled to extend synchronously at a speed of 0.4 m / min. The load data of the crane torque limiter is received in real time. When the load is <10t, it indicates that the outrigger has bottomed out and is under force. The cylinder action is stopped, and the crane hook is lowered to a state where it is ≥10cm away from the top of the frame and is not under force, so as to avoid the crane vibration being transmitted to the frame.

[0077] Levelness adjustment: Read the data from the bidirectional inclinometer. If the frame tilts laterally or longitudinally by more than 0.1°, prioritize extending and lifting the outrigger cylinders on the lower side at a speed of 0.2 m / min. If the cylinders are extended to 90% of their length and still not level, lower the outrigger cylinders on the higher side at a speed of 0.1 m / min until the inclinometer displays a levelness of ≤0.1°.

[0078] Elevation calibration: The real-time coordinates of the four corners of the frame are collected by the dual-antenna Beidou RTK positioning module, the elevation of the four corners is calculated, and compared with the design leveling elevation. If the deviation is > ±0.8cm, the extension and retraction of the corresponding outrigger cylinders are adjusted in linkage. The adjustment amount of a single outrigger is ≤5cm until the elevation of the four corners meets the design requirements. The elevation calibration efficiency is improved by 5 times.

[0079] S3: Feeding and Leveling Operations

[0080] Feeding preparation: Control the leveling head to move via remote control, align the stone chute with one corner of the frame end; the crawler crane lifts a 2.5m³ Q355B material self-discharge hopper to the top of the stone chute feeding port, with the hopper outlet and feeding port aligned with a deviation of ≤±10cm to avoid stone spillage.

[0081] Quantitative feeding: Lower the hopper and open the hopper outlet to feed material into the stone chute. When the high material level alarm is triggered, stop feeding. When the low material level alarm is triggered, stop moving and leveling to ensure that there is always material in the chute and avoid "uneven material distribution caused by material interruption".

[0082] Serpentine leveling: The automated control system starts the hydraulic motor, driving the leveling head to move along a preset serpentine path: ① Lateral movement: The leveling head moves along the sliding beam at a speed of 1.05m / min, with a single leveling width of 1.38m and an 8cm overlap between adjacent paths to ensure no areas are missed; ② Longitudinal movement: After each lateral movement, the leveling head moves longitudinally along the frame by 1.3m, which is the single leveling width minus the overlap length. A total of 8 serpentine movements are made to complete the 20.5×9.5m coverage of a single workstation. During the process, the leveling accuracy is checked every 2m with a total station. If the deviation is > ±2cm, the height of the leveling head is adjusted in real time, with an adjustment amount ≤ 3cm. This improves leveling efficiency by 2 times and accuracy stability by 40%.

[0083] S4: Residual Material Control and Displacement

[0084] Leftover material estimation and handling: When the rock-throwing chute is 5m away from the end position of the work station, the system automatically calculates the leftover material amount through the "feeding amount - walking distance" algorithm. Leftover material amount = total feeding amount - already laid amount, already laid amount = walking distance × laying width × average stone thickness, with the average stone thickness taken as 0.3m; if the leftover material amount > 0.1m³, the rock-throwing chute is controlled to continue moving forward to 1m beyond the work station, and the leftover material is discharged into the collection trough. Then, a small crane is used to lift the collection trough to the mother ship to avoid the leftover material contaminating the completed surface.

[0085] Workstation relocation: After leveling a single workstation, the outrigger cylinders are retracted at a speed of 0.4 m / min, and the crawler crane lifts the frame onto the mother ship's deck. The bottom of the frame is then secured with sleepers. The mother ship moves to the next workstation via GPS positioning, with a positioning deviation of ≤ ±50 cm. Steps S1-S3 are repeated. Leveling time for a single workstation is 3-4 hours, hoisting and positioning time is 1 hour, and the total operation time for a single container is 4-5 hours, resulting in a 30% increase in efficiency.

[0086] S5: Data Feedback and Optimization

[0087] Real-time data acquisition: During operation, the multi-sensor fusion sensing system acquires data at a frequency of 1Hz, including: ① Leveling accuracy, comparison value between Beidou RTK and total station, accuracy ±0.1cm; ② Equipment attitude, lateral / longitudinal tilt angle of bidirectional inclinometer, accuracy ±0.01°; ③ Outrigger force, pressure transmitter data, accuracy ±0.01MPa; ④ Leveling head position, magnetostrictive displacement sensor data, accuracy ±0.01mm; All data is stored in a local database of ≥100GB, backed up using an SD card with a capacity of 128GB, and uploaded to the cloud platform via a 4G module, supporting web-based viewing for easy later traceability.

[0088] Dynamic parameter optimization: After daily operations, control parameters are optimized through data analysis: ① If the outrigger synchronization error is >3%, adjust the opening of the diversion and collection valve group by ±5%; ② If the leveling accuracy fluctuates greatly, with the proportion of points with deviation >±2cm exceeding 5%, optimize the PID algorithm parameters, with a proportional coefficient of ±0.1 and an integral time of ±0.1s; ③ If the material laying speed and feeding speed are mismatched, and the high material level alarm interval is <5min or >15min, adjust the single feeding amount by ±0.2m³; Through continuous optimization, ensure that the flatness qualification rate is ≥95% and the accuracy stability is improved by 10%.

[0089] Compared with the prior art, the present invention has the following advantages:

[0090] I. Strong adaptability to various scenarios and excellent construction economy

[0091] This invention abandons the use of specialized leveling vessels and large ship-borne carriers, instead adopting a fully modular frame-type leveling structure. This fundamentally solves the core pain points of large shipborne equipment: high cost, high relocation costs, inability to access narrow or near-shore waters, and low equipment utilization in small and medium-sized projects. It also avoids the shortcomings of monolithic structures, such as low modularity and high requirements for relocation and hoisting. Each module of this invention has a maximum weight of ≤30t, allowing for transport via conventional flatbed trucks. Relocation costs are reduced by more than 80% compared to large shipborne equipment, and the overall cost of the equipment is less than 5% of that of specialized leveling vessels. It is perfectly suited for small and medium-sized deep-water foundation projects, breaking the long-standing industry dilemma of "large equipment being unaffordable and small equipment being unusable." It also significantly reduces labor costs compared to traditional manual diving leveling.

[0092] II. High leveling accuracy and good construction stability

[0093] This invention utilizes a multi-sensor fusion sensing system to construct a three-dimensional closed-loop monitoring network encompassing spatial positioning, attitude monitoring, and force feedback. This completely avoids the inherent errors of acoustic and wire-guided measurements, as well as the insufficient accuracy and weak anti-interference capabilities caused by single-sensor positioning and the lack of closed-loop monitoring. Coupled with a dedicated leveling mechanism and real-time accuracy compensation mechanism, it overcomes the inherent shortcomings of heavy-duty paving without precise leveling control. The leveling accuracy of this invention can be stably controlled within ±3cm, with a flatness qualification rate exceeding 95%, representing a 70% improvement over traditional manual leveling. It also solves the problems of uneven material distribution and thickness deviations easily caused by water flow and wave interference in existing technologies, reliably meeting the high-precision construction requirements of caisson installation.

[0094] 3. High work efficiency and high degree of automation

[0095] This invention utilizes a hydraulic collaborative drive system to achieve synchronized extension and retraction of outriggers, coordinated movement of the leveling head, and feeding operations, solving the common problems of independent control of actuators, lack of collaborative logic, and high dependence on manual operation. Coupled with automated serpentine leveling path planning, it can complete full-area coverage of a single workstation within a fixed frame in one go, eliminating the need for repeated vessel relocation and repositioning, significantly reducing auxiliary operation time. The leveling operation time for a single workstation is only 3-4 hours, and the total operation time for a single container position can be controlled within 4-5 hours, improving efficiency by more than 200% compared to traditional manual leveling and more than 60% compared to existing shipborne paving solutions. Simultaneously, it significantly reduces omissions and rework caused by human operation, effectively extending the annual construction window.

[0096] IV. Strong adaptability to deep water, safe and reliable operation.

[0097] This invention adopts a fully mechanized underwater operation mode, eliminating the need for underwater divers and completely eliminating the high safety risks of traditional manual diving operations. It is equipped with IP68-level deep-water sealing protection, anti-tipping hook wheel, and outrigger overload protection design, which solves the shortcomings of only being suitable for shallow water, lacking deep-water protection structure, and lacking a deep-water operation safety guarantee system. It can stably adapt to water depths of 0-20m and complex sea conditions, greatly improving the equipment reliability and construction safety redundancy in deep-water conditions.

[0098] V. Excellent environmental performance, meeting the requirements of green construction.

[0099] This invention achieves unmanned underwater operation mode by "mechanizing to replace manpower", completely eliminating safety hazards such as lack of oxygen and swell impact in diving operations, reducing safety risks by 80%; in terms of environmental performance, by using "feeding-walking" linkage control and precise material placement design, it reduces stone waste by 15% and reduces the disturbance of marine ecology to marine construction, which fully meets the requirements of green construction in the current engineering construction field.

[0100] In summary, compared with existing technologies, this invention not only breaks through the limitations of existing technologies, which are only suitable for shallow water of ≤10m, single sensor positioning, and independent hydraulic control, but also specifically solves the problems of measurement principle deviation, lack of collaborative logic in drive, and mismatch between equipment and project requirements in small and medium-sized projects. Through deep integration of multiple systems, it achieves efficient adaptation to complex sea conditions such as water depth of 0-20m, wind force ≤6, wave height ≤1.0m, and current velocity ≤1.0m / s, filling the technological gap in the industry for integrated intelligent construction of small and medium-sized deep-water foundation beds with "perception-decision-execution-guarantee". Attached Figure Description

[0101] Figure 1 This is a front view of the overall structure of the deep-water bed leveling equipment in this invention;

[0102] Figure 2 This is a top view of the overall structure of the deep-water bed leveling equipment in this invention;

[0103] Figure 3 This is a schematic diagram of the hydraulic collaborative drive system lifting mechanism in this invention;

[0104] Figure 4 This is a schematic diagram of the hydraulic cooperative drive system descending in this invention;

[0105] Figure 5 This is a schematic diagram of the gear and rack transmission mechanism and bearing housing in this invention;

[0106] Figure 6 This is a structural diagram of the self-pouring hopper in this invention (before material feeding);

[0107] Figure 7 This is a structural diagram of the self-discharging hopper in this invention (during material feeding);

[0108] Figure 8 This is a diagram of the serpentine walking path in this invention;

[0109] Figure 9 This is a diagram of the multi-sensor fusion sensing system in this invention;

[0110] Figure 10 This is a layout diagram of the measurement and control system in this invention;

[0111] In the diagram: 1-Feeding hopper; 2-Stone chute; 3-Traveling mechanism; 4-Feeding port; 5-Long beam; 6-Outrigger hydraulic mechanism; 7-Short beam; 8-Drive shaft one; 9-Circular tube beam; 10-Upper box of outrigger hydraulic mechanism; 11-Lower box of outrigger hydraulic mechanism; 12-Pointed leg of outrigger hydraulic mechanism base plate; 13-Drive shaft two; 14-Bearing wheel; 15-Traveling gear; 16-Hook wheel; 17-Self-dumping hopper; 18-Dual-antenna Beidou RTK positioning module; 19-Dual-axis tilt sensor; 20-Altimeter; 21-Single-axis tilt sensor; 22-Magnetostrictive displacement sensor; 23-Pressure transmitter. Detailed Implementation

[0112] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the present invention is not limited to the specific embodiments.

[0113] Example 1

[0114] like Figure 1 , Figure 2 As shown, a deep-water bed leveling equipment based on multi-sensor fusion and hydraulic collaborative control includes:

[0115] The modular leveling frame adopts a standardized interface design, including a frame body, a chute 2, and DN800 connecting flanges. The frame body is composed of three 8m long H800×280×12×20 steel sections connected by M30 high-strength bolts. The ends are equipped with φ630×16mm round pipe beams 9 made of Q355B material, and both sides have detachable DN800 flanges. The pressure rating is 1.6MPa, and the effective internal leveling area is 20.5×9.5m. The chute 2 has a diameter of 800mm, a wall thickness of 12mm, is made of Q355B material, and has a total length of 23m. It consists of a 3.5m base section, two 6m standard sections, two 3m adjusting sections, and a 1.5m feeding port 4. It is quickly assembled and disassembled via DN800 connecting flanges. The maximum weight of a single module is ≤30t to accommodate 25t flatbed truck transportation.

[0116] The multi-sensor fusion sensing system uses a PLC controller inside an underwater sealed enclosure as its core control unit. All sensors are mechanically fixed and electrically connected via waterproof shielded cables, IP68 watertight connectors, and an RS485 bus, constructing a three-dimensional monitoring closed loop of "spatial positioning - attitude monitoring - elevation feedback - displacement acquisition - force monitoring". The dual-antenna Beidou RTK positioning module 18 is symmetrically bolted to the top of the leveling frame's pillars using stainless steel brackets and connected to the PLC via waterproof cables, providing a planar accuracy of ±(10+1×10⁻⁻⁶). 6 ×D)mm, elevation accuracy ±(20+1×10⁻ 6A global positioning reference with a baseline length of ×D)mm (D is the baseline length) and an update rate of 1Hz; a dual-axis tilt sensor 19 is embedded and bolted to the center of gravity of the central column of the frame, and connected to the PLC via RS485 bus (transmission delay ≤10ms, accuracy ±0.1°, measurement range ±5°) to monitor the horizontal / longitudinal tilt angle of the frame in real time; a height gauge 20 is bolted to the bottom of the column near the discharge port and connected to the PLC via a waterproof cable to detect the elevation of the base bed before and after the material is laid in real time; single-axis tilt sensors 21 are fixed at the four corners of the frame and connected via shielded cables. The system is connected to a PLC to assist in monitoring the local unidirectional tilt angle of the leveling head and the rock-throwing chute. A magnetostrictive displacement sensor 22 is built into the cylinder of each of the four outrigger cylinders and connected to the PLC via a signal cable (range 0-1.2m, accuracy ±0.05%FS, resolution 0.01mm) to accurately collect the outrigger extension and retraction displacement. A pressure transmitter 23 is connected in series in the hydraulic inlet pipeline of the outrigger cylinders and connected to the PLC via an analog cable (measurement range 0-30MPa, accuracy ±0.25%FS, rated force ≤50t) to monitor the working load and system pressure in real time. All sensor data is uniformly integrated into the PLC and processed by a multi-sensor fusion algorithm, providing core data support for positioning correction, frame leveling, hydraulic collaborative control, overload protection, and accuracy control in deep-water bed leveling operations, comprehensively ensuring operational accuracy and stability.

[0117] like Figure 3 , Figure 4 As shown, the hydraulic collaborative drive system uses a dual-plunger hydraulic pump as the power source, with a displacement of 10.56 L / min and a rated pressure of 25 MPa. It adopts a one-in-one-outstand design with a fault switching time ≤1 s. It is equipped with a flow divider / combiner valve group with a synchronization accuracy ≤3%, and four-leg cylinders with a cylinder inner diameter of φ194 mm and a piston rod diameter of φ160 mm. Each four-leg cylinder is equipped with an upper support hydraulic mechanism box 10, a lower support hydraulic mechanism box 11, and a base plate pointed leg 12. The cylinder stroke is 1 m, and the extension / retraction speed is 0.4 m / min. It is connected to the system via a transmission shaft 13. Figure 5 The gear and rack transmission mechanism shown is linked to achieve synchronous extension and retraction of the outriggers and coordinated movement of the flattening head. The flattening head's fabric walking speed is 1.05m / min, and the path deviation is ≤±1cm.

[0118] The automated control system includes a PLC controller, a 4G or Bluetooth wireless transmission module with a transmission distance of ≤500m, and two button-type remote controls with a battery life of ≥8h. It receives data from a multi-sensor fusion perception system and automatically plans a serpentine leveling path using a preset PID algorithm with a proportional coefficient of 1.2, an integral time of 0.5s, and a derivative time of 0.1s. The overlap length is 5-10cm, and the single leveling width is 1.38m. It controls the hydraulic system to perform outrigger extension and retraction, leveling head movement, and feeding start and stop actions. It is equipped with a 7-inch MCGS touch screen with a resolution of 800×480, which displays the cylinder extension and retraction length, outrigger force, and leveling head position in real time.

[0119] The safety protection system includes a 316L stainless steel underwater sealing box, an M20×1.5 watertight connector with an IP68 sealing rating, and an anti-tipping hook wheel with a diameter of 200mm installed at the bottom of the gearbox. The underwater sealing box has a wall thickness of 10mm, an IP68 protection rating, and contains a hydraulic valve group and electrical components. It is suitable for operating environments with water depths of 0-20m, wind force ≤6, wave height ≤1.0m, and current velocity ≤1.0m / s.

[0120] The lower leveling head of the stone chute 2 is equipped with a 30° flared mouth and a 1.5×1.5m wear-resistant steel scraper plate with a service life of ≥5000㎡. An ultrasonic high-level alarm is installed in the middle of the stone chute, with a detection distance of 0.3-3m and an accuracy of ±1%. The alarm is triggered when the height of the stone in the stone chute reaches 2m, and the single replenishment amount is 2.5m³.

[0121] The magnetostrictive displacement sensor 22 is installed in the cylinder of the four-leg hydraulic cylinder to collect data on the extension and retraction of the outriggers in real time; the pressure transmitter 23 is installed in the rodless chamber of the outrigger cylinder to collect data on the force on the outriggers; both have a data transmission frequency of milliseconds to provide real-time feedback to the PLC controller to ensure that the synchronization accuracy of the outriggers is ≤±2mm.

[0122] The automated control system supports both touchscreen operation in the control room and remote control operation on-site, with a switching time of ≤0.5s between the two modes. When the leveling accuracy deviation exceeds ±3cm, the system automatically triggers an adjustment command to achieve accuracy compensation through extension and retraction of the outrigger cylinders or correction of the leveling head path. The compensation response time is ≤2s, the adjustment amount of a single outrigger is ≤5cm, and the offset of the leveling head path correction is ≤2cm.

[0123] Example 2

[0124] This embodiment applies the equipment described in Embodiment 1 to the Changhai Test Base Pier 1 and 2 project. The project comprises 31 caisson foundations, with a maximum bottom dimension of 18×17m and a maximum depth of 18m. The riprap foundations use 10-100mm boulders. The operating environment is characterized by water depth ≤20m, wind force ≤6, wave height ≤1.0m, and current velocity ≤1.0m / s. The final leveling accuracy reaches ±3cm, with a flatness pass rate of 96%. Compared to traditional manual leveling, this represents a 250% increase in efficiency, a 50% reduction in leveling time, a 15.4% cost saving per caisson, a 0% safety accident rate, and a 15% reduction in stone waste, achieving significant economic, safety, and social benefits.

[0125] Specifically, the steps include the following:

[0126] S1: Preliminary Preparations and Positioning

[0127] Land-based pre-test: Before operation, a 25×10m simulated base bed was constructed on land. The foundation was C30 concrete, and the surface was covered with crushed stone of three specifications: 30-60mm, 60-90mm, and 60-120mm, to simulate the actual base bed stone gradation. After assembling the equipment modules, performance tests were conducted: ① Multi-sensor system consistency test: Beidou RTK and total station, accuracy ±1.5+2×10⁻ 6 ×D mm comparison, positioning deviation ≤±2mm; ② Hydraulic system synchronization test: the four leg cylinders extend synchronously by 1m, and the extension error ≤±2mm; ③ Leveling accuracy test: the leveling head walks along a serpentine path, and the total station is tested according to a 2m×2m grid. The leveling accuracy ≤±1cm, and the sensor data fluctuation for 1 hour is ≤±0.5%, which meets the design requirements. It is then transported to the site to avoid the problem of "direct watering and difficulty in troubleshooting".

[0128] Marine Positioning: ① Parameter Input: Import the coordinates of the first caisson foundation into the dual-antenna Beidou RTK positioning module 18 of the automated control system, with an X / Y / Z axis accuracy of ±1mm; ② Crane Transfer: Use a 150-ton crawler crane with an operating radius of 15-20m and a rated lifting capacity of ≥50t to lift the frame. Attach a 20m long, 20mm diameter nylon sway rope to the crane's wire rope, with a breaking strength of ≥50kN. Adjust the horizontal rotation angle of the frame by the sway rope, with a deviation of ≤0.3°; ③ Suspension Stability: Lower the frame to 50cm from the foundation and suspend it for ≥30s. After the frame's sway amplitude due to water flow is ≤±5cm, prepare for the next operation. Positioning stability is improved by 80%.

[0129] S2: Framework Adjustment

[0130] Outrigger bottoming: Activate the hydraulic system to control the four outrigger cylinders to extend synchronously at a speed of 0.4m / min with a load of 9t. Stop the cylinder action to allow the crane hook to fall to a state of no force with a distance of ≥10cm from the top of the frame, thus preventing the crane vibration from being transmitted to the frame.

[0131] Levelness adjustment: Read the data from the bidirectional inclinometer, which shows a lateral tilt of 0.3°. Extend the two outrigger cylinders on the south side by 3cm at a speed of 0.2m / min to correct the levelness to 0.08°.

[0132] Elevation calibration: The real-time coordinates of the four corners of the frame are collected by the dual-antenna Beidou RTK positioning module 18, the elevation of the four corners is calculated, and the outriggers are adjusted to make the level head reach the design elevation (-11.5m), with a deviation of ±0.8cm.

[0133] S3: Feeding and Leveling Operations

[0134] Feeding preparation: Control the leveling head to move using the remote control, align the stone chute 2 with one corner of the frame end; the crawler crane lifts a 2.5m³ Q355B material self-discharge hopper 17 to above the stone chute feeding port 4, with the hopper outlet and feeding port 4 having an alignment deviation of ≤±10cm to avoid stone spillage. Figures 6-7 The diagram shown is a structural diagram of the self-discharging hopper 17.

[0135] Quantitative feeding: Lower the hopper and open the hopper outlet to feed material into the stone chute 2 until the high material level alarm is triggered and feeding stops.

[0136] Snake-shaped leveling: such as Figure 8 As shown, the automated control system starts the hydraulic motor, driving the leveling head to move along a preset serpentine path: ① Lateral movement: The leveling head moves along the sliding beam at a speed of 1.05m / min, with a single leveling width of 1.38m and an 8cm overlap between adjacent paths to ensure no areas are missed; ② Longitudinal movement: After each lateral movement, the leveling head moves longitudinally along the frame by 1.3m, which is the single leveling width minus the overlap length. A total of 8 lateral movements are made to complete the coverage of a single workstation of 20.5×9.5m. During the process, the leveling accuracy is checked every 2m with a total station, with a deviation of ≤±2cm. Each material replenishment is 2.5m.

[0137] S4: Residual Material Control and Displacement

[0138] Leftover material estimation and handling: When the rock chute 2 is 5m away from the end position of the work station, the system automatically calculates the leftover material amount through the "feed amount - travel distance" algorithm. Leftover material amount = total feed amount - already laid amount, already laid amount = travel distance × laying width × average stone thickness, and the average stone thickness is taken as 0.3m; if the leftover material amount > 0.1m³, control the rock chute 2 to continue moving forward to 1m beyond the work station, discharge the leftover material into the collection trough, and then use a small crane to lift the collection trough to the mother ship to avoid the leftover material from contaminating the completed surface.

[0139] Workstation relocation: After leveling a single workstation, the outrigger cylinders are retracted at a speed of 0.4 m / min, and the crawler crane lifts the frame onto the mother ship's deck. The bottom of the frame is then secured with sleepers. The mother ship moves to the next workstation via GPS positioning, with a positioning deviation of ≤ ±50 cm. Steps S1-S3 are repeated. Leveling time for a single workstation is 4 hours, hoisting and positioning time is 1 hour, and the total operation time for a single caisson is 5 hours. All 31 caisson foundation beds were leveled as required.

[0140] S5: Data Feedback and Optimization

[0141] Real-time data acquisition: During operation, the multi-sensor fusion sensing system acquires data at a frequency of 1Hz, including: ① Leveling accuracy, comparison value between Beidou RTK and total station, accuracy ±0.1cm; ② Equipment attitude, lateral / longitudinal tilt angle of dual-axis tilt sensor 19, accuracy ±0.01°; ③ Outrigger force, data from pressure transmitter 23, accuracy ±0.01MPa; ④ Leveling head position, data from magnetostrictive displacement sensor 22, accuracy ±0.01mm; All data is stored in a local database of ≥100GB, backed up using an SD card with a capacity of 128GB, and uploaded to the cloud platform via a 4G module, supporting web-based viewing for easy later traceability.

[0142] Dynamic parameter optimization: After daily operations, control parameters are optimized through data analysis: ① If the outrigger synchronization error is >3%, adjust the opening of the diversion and collection valve group by ±5%; ② If the leveling accuracy fluctuates greatly, with the proportion of points with deviation >±2cm exceeding 5%, optimize the PID algorithm parameters, with a proportional coefficient of ±0.1 and an integral time of ±0.1s; ③ If the material laying speed and feeding speed are mismatched, and the high material level alarm interval is <5min or >15min, adjust the single feeding amount by ±0.2m³; Through continuous optimization, ensure that the flatness qualification rate is ≥95% and the accuracy stability is improved by 10%.

[0143] This invention is not limited to the above embodiments. Without departing from the core technology of this invention, the structure and parameters of each component can be appropriately adjusted. All technical improvements based on this invention fall within the protection scope of this invention. For example, replacing the dual-antenna Beidou RTK positioning module with a GPS+GLONASS dual-mode positioning module, or adding an AI visual recognition module to assist in monitoring the flatness of the subgrade, all fall within the protection scope of the claims of this invention.

Claims

1. A deep-water bed leveling equipment with multi-sensor fusion and hydraulic collaborative control, characterized in that, include: The modular leveling frame adopts a standardized interface design, including a frame body, a chute, and DN800 connecting flanges. The frame body is composed of three 8m long H800×280×12×20 steel sections connected by M30 high-strength bolts. The ends are equipped with φ630×16mm round pipe beams made of Q355B material, and both sides have detachable DN800 flanges. The pressure rating is 1.6MPa, and the effective internal leveling area is 20.5×9.5m. The chute has a diameter of 800mm, a wall thickness of 12mm, is made of Q355B material, and has a total length of 23m. It consists of a 3.5m base section, two 6m standard sections, two 3m adjusting sections, and a 1.5m feeding port. It is quickly assembled and disassembled via DN800 connecting flanges. The maximum weight of a single module is ≤30t to accommodate 25t flatbed truck transportation. The multi-sensor fusion sensing system uses a PLC controller inside the underwater sealed enclosure as the core control unit. Each sensor is electrically and data-wise connected via waterproof shielded cables, IP68 watertight connectors, and an RS485 bus. Dual-antenna Beidou RTK positioning modules are symmetrically installed at the top of the leveling frame's columns, providing a global positioning reference for the entire machine. Dual-axis tilt sensors are fixed at the center of gravity of the frame's central column, monitoring the machine's lateral and longitudinal tilt angles in real time. Single-axis tilt sensors are positioned at the four corners of the frame to assist in monitoring the local attitude of the leveling head and the rock-throwing chute. The level gauge is installed near the discharge port to detect the elevation of the base bed before and after material placement in real time; the magnetostrictive displacement sensor is built into the cylinder of the four outrigger cylinders to accurately collect the outrigger extension and retraction displacement data; the pressure transmitter is connected in series in the hydraulic oil inlet pipeline of the outrigger cylinder to monitor the working load and system pressure in real time; all sensor data are uniformly integrated into the PLC controller, and after being processed by the multi-sensor fusion algorithm, it provides data support for positioning correction, frame leveling, hydraulic collaborative control, overload protection and accuracy control of deep water base bed leveling operations, and comprehensively ensures the accuracy of operation and the stability of operation; The hydraulic collaborative drive system uses a dual-plunger hydraulic pump as the power source, with a displacement of 10.56L / min and a rated pressure of 25MPa. It adopts a one-in-one standby design with a fault switching time of ≤1s. It is equipped with a flow divider and combiner valve group with a synchronization accuracy of ≤3%, four-leg cylinders with a cylinder inner diameter of φ194mm and a piston rod diameter of φ160mm, a cylinder stroke of 1m, an extension speed of 0.4m / min, and a gear and rack transmission mechanism to achieve synchronous extension and retraction of the outriggers and coordinated movement of the leveling head. The leveling head's material laying speed is 1.05m / min, and the path deviation is ≤±1cm. The automated control system includes a PLC controller, a 4G or Bluetooth wireless transmission module with a transmission distance of ≤500m, and two button-type remote controls with a battery life of ≥8h. It receives data from a multi-sensor fusion perception system and automatically plans a serpentine leveling path using a preset PID algorithm with a proportional coefficient of 1.2, an integral time of 0.5s, and a derivative time of 0.1s. The overlap length is 5-10cm, and the single leveling width is 1.38m. It controls the hydraulic system to perform outrigger extension and retraction, leveling head movement, and feeding start and stop actions. It is equipped with a 7-inch MCGS touch screen with a resolution of 800×480, which displays the cylinder extension and retraction length, outrigger force, and leveling head position in real time. The safety protection system includes a 316L stainless steel underwater sealing box, an M20×1.5 watertight connector with an IP68 sealing rating, and an anti-tipping hook wheel with a diameter of 200mm installed at the bottom of the gearbox. The underwater sealing box has a wall thickness of 10mm, an IP68 protection rating, and contains a hydraulic valve group and electrical components. It is suitable for operating environments with water depths of 0-20m, wind force ≤6, wave height ≤1.0m, and current velocity ≤1.0m / s.

2. The deep-water bed leveling equipment with multi-sensor fusion and hydraulic collaborative control according to claim 1, characterized in that, The lower leveling head of the stone chute is equipped with a 30° flared mouth and a 1.5×1.5m wear-resistant steel scraper plate with a service life of ≥5000㎡. An ultrasonic high-level alarm is installed in the middle of the stone chute, with a detection distance of 0.3-3m and an accuracy of ±1%. The alarm is triggered when the height of the stone in the stone chute reaches 2m, and the single replenishment amount is 2.5m³.

3. The deep-water bed leveling equipment with multi-sensor fusion and hydraulic collaborative control according to claim 1, characterized in that, The magnetostrictive displacement sensor is installed in the cylinder of the four outrigger cylinders to collect outrigger extension and retraction data in real time; the pressure transmitter is installed in the rodless chamber of the outrigger cylinder to collect outrigger force data; both have a data transmission frequency of millisecond level to provide real-time feedback to the PLC controller to ensure outrigger synchronization accuracy ≤ ±2mm.

4. The deep-water bed leveling equipment with multi-sensor fusion and hydraulic collaborative control according to claim 1, characterized in that, The automated control system supports both touchscreen operation in the control room and remote control operation on-site, with a switching time of ≤0.5s between the two modes. When the leveling accuracy deviation exceeds ±3cm, the system automatically triggers an adjustment command to achieve accuracy compensation through extension and retraction of the outrigger cylinders or correction of the leveling head path. The compensation response time is ≤2s, the adjustment amount of a single outrigger is ≤5cm, and the offset of the leveling head path correction is ≤2cm.

5. A method for deep-water bed leveling operations using multi-sensor fusion and hydraulic collaborative control of the equipment described in claim 1, characterized in that, Includes the following steps: S1. Preliminary Preparation and Positioning: Before the operation, a 25×10m simulated base bed is built on land to test the consistency of data from multiple sensor systems and the synchronization of the hydraulic system. The deviation between Beidou RTK and total station is ≤±2mm, and the extension error of the outriggers is ≤±2mm. During the operation at sea, the coordinates of the first and last ends of the leveling frame are input through the dual-antenna Beidou RTK positioning module to guide the 150-ton crawler crane to move the frame to the predetermined plane position. The horizontal rotation deviation of the frame is adjusted to ≤0.3° by using a 20m long and 20mm diameter swaying rope. The frame is lowered to 50cm from the base bed and hovered for ≥30s until the attitude is stable. S2. Frame Adjustment: The outrigger cylinders extend synchronously, and the crane load is monitored in real time. When the load is <10t, the operation stops, allowing the crane hook to fall to a state of no force with a distance ≥10cm from the top of the frame. Based on the data from the bidirectional inclinometer, the cylinder on the side with the lower elevation is adjusted first to extend and lift at a speed of 0.2m / min. If the cylinder extension length reaches 90% and the frame is still not level, the cylinder on the side with the higher elevation is lowered at a speed of 0.1m / min until the frame levelness is ≤0.1°. The elevation of the four corners of the frame is calculated using Beidou RTK, and the outrigger cylinders are adjusted in conjunction to make the leveling head reach the design leveling elevation, with the deviation controlled within ±0.8cm. S3. Feeding and Leveling Operation: Move the chute to one corner of the frame end. The crawler crane lifts the self-discharging hopper to above the chute feeding port, with an alignment deviation ≤ ±10cm. The hopper is then lowered for automatic unloading. Feeding stops after the high-level alarm is triggered. The automated control system starts the hydraulic motor, driving the leveling head to move along a serpentine path: laterally, it moves along the sliding beam at a speed of 1.05m / min, with a single leveling width of 1.38m and an overlap length of 8cm; longitudinally, it moves 1.3m after each lateral movement, for a total of 8 movements to cover a single workstation. During the process, the leveling accuracy is checked every 2m with a total station. If the deviation is > ±2cm, the leveling head height is adjusted in real time. S4. Residual Material Control and Relocation: When the rock dumping chute is 5m away from the end position of the work station, the residual material is estimated by the feed rate minus the travel distance algorithm. After the leveling of a single work station is completed, the outrigger cylinders are retracted at a speed of 0.4m / min. The crawler crane lifts the frame to the mother ship's special H-beam support seat, and the bottom of the frame is padded with sleepers and the outriggers are fixed. The mother ship moves to the next work station with a positioning deviation of ≤±50cm. Steps S1-S3 are repeated. The leveling time for a single work station is 3-4 hours, the hoisting and positioning time is 1 hour, and the total operation time for a single container position is 4-5 hours. S5. Data Feedback and Optimization: The multi-sensor fusion sensing system collects data on leveling accuracy, equipment posture, outrigger force, and leveling head position at a frequency of 1Hz, stores it in a local database of ≥100GB, and uploads it to the cloud platform. After the daily operation, if the outrigger synchronization error is >3%, adjust the opening of the diversion and collection valve group by ±5%. If the leveling accuracy fluctuates greatly, optimize the PID algorithm parameters, with a proportional coefficient of ±0.1 and an integral time of ±0.1s. If the material distribution and feeding speeds are mismatched, adjust the single feeding amount by ±0.2m³ to ensure a flatness qualification rate of ≥95%.

6. The method for deep-water bed leveling operation using multi-sensor fusion and hydraulic collaborative control according to claim 5, characterized in that, The surface of the simulated substrate for the land pre-test in S1 is paved with crushed stone of different particle sizes: 30-60mm, 60-90mm, and 60-120mm. The crawler crane used in the offshore positioning process has an operating radius of 15-20m and a rated lifting capacity of ≥50t.

7. The method for deep-water bed leveling operation using multi-sensor fusion and hydraulic collaborative control according to claim 5, characterized in that, The self-pouring hopper in S3 has a volume of 2.5m³ and is made of Q355B material.

Images