Real-time assembly quality control method for buoyancy tank wharf

Abstract

The invention relates to a buoyancy tank wharf real-time assembly quality control method, and belongs to the technical field of port engineering automatic construction. According to the method, aiming at the problems of large geometric positioning deviation, uneven contact surface stress and difficulty in real-time control of attitude instability in buoyancy tank assembly, laser range finders are mounted at four corners of a buoyancy tank through a geometric positioning layer to monitor horizontal deviation, and pressure sensors are arranged on a butt joint surface through a mechanical sensing layer to generate a pre-tightening force curve; the attitude monitoring layer adopts a bottom inclination angle sensor to detect inclination; according to the horizontal deviation, the curve similarity deviation and the inclination angle deviation, the real-time management and control platform triggers a hydraulic adjusting mechanism, a torque wrench and a balance weight drive to conduct adjustment; meanwhile, data are stored through a buoyancy tank number, a timestamp and a sensor type three-dimensional index, and a rechecking instruction is generated based on the difference between the current adjusting quantity and the first three mean values. According to the invention, real-time control and dynamic adjustment of the assembly process are realized, and the modular assembly precision and safety of the floating wharf are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automated construction of port projects, and more particularly to a method for real-time assembly quality control of a pontoon dock. Background Art

[0002] As a modular water structure, the assembly quality of a pontoon wharf directly affects the overall stability and service life of the wharf. The assembly of traditional pontoon wharves usually adopts the method of segmented lifting and water assembly, in which the pontoon modules are initially fixed by positioning devices pre-installed on a temporary support platform. The temporary support platform is usually composed of a group of piles driven into the seabed, with a steel frame on top to bear the assembly load of the pontoon modules. However, due to factors such as waves, tides and construction machinery disturbances in the marine environment, the pontoon modules are prone to cumulative errors during the assembly process, resulting in problems such as misalignment of the docking surface, uneven distribution of bolt preload or module tilt.

[0003] In the prior art, the assembly quality control of pontoon terminals mostly relies on manual measurement and periodic adjustments. For example, the coordinates of the module corners are regularly detected by a total station, or a hydraulic jack is used to make local height corrections. However, such methods have defects such as poor real-time performance and feedback lag, and are difficult to cope with micro-deformations in dynamic construction environments. Especially when multiple modules are assembled continuously, the error transmission effect will be significantly amplified, resulting in a surge in the cost of subsequent adjustments. Public document CN112095543A proposes a pontoon positioning system based on laser scanning, but it only focuses on geometric positioning accuracy, and does not conduct coordinated monitoring of the mechanical state of the joints and the module posture, resulting in a single dimension of assembly quality assessment.

[0004] However, the assembly quality judgment standards for pontoon terminals generally cover three core indicators: geometric deviation, mechanical distribution, and attitude stability. Geometric deviation requires that the horizontal misalignment of adjacent pontoon modules does not exceed 5mm; mechanical distribution must ensure that the similarity between the warning force curve of the pressure sensor on the docking surface and the theoretical curve is greater than 85% to avoid local stress concentration; attitude stability stipulates that the module tilt angle must be controlled within 0.5° to prevent the risk of capsizing due to center of gravity shift. However, existing technologies lack the ability to fuse and analyze multi-source data for the above indicators. For example, the correlation between pressure sensor data and tilt angle data has not been effectively explored, resulting in insufficient coordination of adjustment strategies.

[0005] Furthermore, existing assembly and adjustment methods mostly use independent actuators, such as hydraulic jacks or torque wrenches that are operated separately, and lack system-level linkage control. For example, when adjusting the horizontal deviation, if the counterweight distribution is not corrected synchronously, new tilt problems may occur. In addition, the utilization rate of historical adjustment data is low, and it is difficult to optimize subsequent adjustment parameters through machine learning or statistical models. Although patent CN114197326A proposes a pontoon docking method based on pressure feedback, its adjustment logic is only for a single sensor type, and a three-dimensional index data storage architecture is not constructed, which makes it impossible to achieve dynamic correlation analysis across timestamps and sensor categories. The above technical bottlenecks have led to problems such as low efficiency and large fluctuations in accuracy in the assembly process of the pontoon terminal. There is an urgent need for a real-time quality control method that integrates multi-level sensor networks and intelligent decision-making. Summary of the Invention

[0006] An object of the present invention is to solve at least the above problems and to provide at least the advantages which will be described hereinafter.

[0007] Another object of the present invention is to provide a real-time assembly quality control method for a pontoon terminal, which realizes multi-dimensional precise control of horizontal deviation, torque and counterweight through real-time collaborative control of a three-layer sensor network of geometric positioning, mechanical perception and posture monitoring, and dynamic response of the assembly adjustment module. Combined with historical data optimization and correlation analysis, it significantly improves the full-process control capability of the assembly quality of the pontoon terminal and reduces error accumulation and capsizing risks.

[0008] In order to achieve these purposes and other advantages according to the present invention, a method for real-time assembly quality control of a pontoon terminal is provided, comprising: A three-layer sensor network consisting of geometric positioning, mechanical sensing, and attitude monitoring is set up on the assembly surface of the pontoon module. The geometric positioning layer includes laser rangefinders installed at the four corners of the pontoon module. The mechanical sensing layer includes pressure sensors arranged at equal intervals on the edges of the docking surfaces of adjacent pontoon modules. The attitude monitoring layer includes an inclination sensor installed at the bottom of the pontoon module. An assembly quality standard module and an assembly adjustment module are deployed within the real-time assembly quality control platform. When the laser rangefinder detects a horizontal deviation exceeding 5 mm, when the similarity between the actual warning force distribution curve generated by the pressure sensor and the standard curve is less than 85%, or when the inclination sensor detects a tilt angle exceeding 0.5°, the assembly adjustment module is triggered to control the underwater hydraulic adjustment mechanism to perform horizontal deviation adjustment, the torque wrench to perform torque adjustment, and the counterweight drive to perform counterweight adjustment. Among them, an assembly adjustment data processing module is established inside the assembly quality control platform. The assembly process data processing module stores data according to the three-dimensional index of the pontoon number, timestamp and sensor type. The index dimensions include the laser rangefinder number, the pressure sensor group number and the inclination sensor orientation code; at the same time, the assembly process data processing module extracts the previous three adjustments of the pontoon module at the same position and calculates the average value of the adjustment. When the correlation of the three layers of sensor data at the same timestamp is greater than 0.7 and any current adjustment value of the horizontal deviation adjustment, torque adjustment and counterweight adjustment of the pontoon module at the same position exceeds 15% of the average adjustment value of the previous three times, a review instruction is generated.

[0009] Preferably, the pontoon module is arranged on a temporary support platform on the dock, and the underwater hydraulic adjustment mechanism includes four hydraulic jack units arranged on the temporary support platform, and the four hydraulic jack units are just opposite the four corners of the bottom of the pontoon module. The piston rod on the top of any hydraulic jack unit is connected to the reinforcing steel plate arranged at the bottom of the pontoon module through a ball joint; the torque wrench is driven by an electric servo, and its shell is fixed to the surface of the pontoon module within 200mm from the center of the connector by a magnetic clamp, and the front end of the torque wrench is provided with a hexagonal socket adapter matching the specifications of the pontoon module connector, the coaxiality deviation of the hexagonal socket adapter and the axis of the connector bolt does not exceed 0.1mm, and the rear end of the hexagonal socket adapter is integrated with a dynamic torque sensor; the counterweight adjustment is achieved by a pair of counterweight boxes symmetrically arranged at the bottom of the pontoon module, and the pair of counterweight boxes are respectively arranged on a pair of parallel slide rails and controlled by corresponding counterweight drives.

[0010] Preferably, when the underwater hydraulic adjustment mechanism performs horizontal deviation adjustment, it is implemented according to the following multi-stage control strategy: When the horizontal deviation value reaches 5mm for the first time, the preload compensation mode is activated: the two diagonally arranged jack units are controlled to perform initial displacement compensation at a synchronous lifting speed of 1mm / s. At the same time, the readings of the two sets of sensors closest to the adjustment point in the pressure sensor group are monitored in real time. When the pressure value of any set drops by more than 10% of its initial value, the lifting is immediately stopped and the system switches to the balanced lifting mode. In the balanced lifting mode, the four jack units are lifted alternately according to a spiral progressive algorithm: each time the jack on one side is lifted by 0.5 mm, the jack on the opposite side is switched to lift. The alternating frequency is synchronized with the frequency of updating the laser rangefinder data. The module geometric center offset is recalculated after each lifting cycle. When the offset calculated value converges to the range of ±1 mm for three consecutive iterative calculations, the adjustment is determined to be complete.

[0011] Preferably, when the torque wrench performs torque adjustment, the following steps are performed: In the three sets of pressure sensors at the interface of adjacent pontoon modules, each set of sensors acquires contact pressure values at a sampling frequency of 10 Hz. The actual preload force distribution curve is generated using the cubic spline interpolation method and discretized into 50 equally spaced data points. The standard preload curve is stored in the assembly quality standard module. During the comparison, the dynamic time warping algorithm is used to calculate the similarity of the two curves, and a torque adjustment instruction is generated when the calculated result is less than 85%; After receiving the adjustment command, the torque wrench applies torque in step mode. The initial torque is set at 80% of the design value and gradually increases in increments of 5% of the design value. After each increase in the torque gradient, a 3-second stabilization time is maintained. The reading change rate of the pressure sensor group is simultaneously collected. If the reading change rate of any group of pressure sensors exceeds 5% / s, the loading is suspended and the change rate is not allowed to continue until it drops below 2% / s. The dynamic torque sensor monitors the output torque value at a frequency of 100 Hz and establishes real-time correlation with the pressure sensor data: When the average pressure sensor reading reaches 90% of the design value, reduce the torque loading rate to 1% of the design value / second; When the pressure reading enters the range of 90-110% of the design value, a pressure balance check is performed every 0.5 seconds. The balance is the ratio of the difference between the maximum and minimum pressure values to the average pressure value. If the balance is greater than 15%, the local tightening strategy is triggered: the torque wrench is controlled to add 3% of the design value torque in the lowest pressure area. The torque output is stopped when the following conditions are met simultaneously: the average reading of any pressure sensor reaches the range of 95% to 105% of the design value; the similarity between the latest generated actual preload force distribution curve and the standard curve is ≥ 88%; the torque fluctuation coefficient monitored by the dynamic torque sensor is less than 2%, where the torque fluctuation coefficient = standard deviation / mean; Finally, the final torque value, pressure distribution curve and termination condition parameters are stored in association with the pressure sensor group number and timestamp, and written into the assembly adjustment data processing module.

[0012] Preferably, when the hydraulic mechanism performs counterweight adjustment, the following steps are performed: According to the real-time output of the X / Y dual-axis tilt data by the tilt sensor, when the absolute value of any axial tilt angle is greater than 0.5°, the tilt direction vector is calculated and mapped to the plane coordinate system of the pontoon module to determine the movement direction of the counterweight box. The movement direction of the counterweight box is opposite to the tilt direction, where the X axis is parallel to the long side of the bottom surface of the pontoon module, and the Y axis is parallel to the short side of the bottom surface of the pontoon module; According to the geometric relationship between the center of gravity height of the pontoon module and the current tilt angle, the sliding distance L of the counterweight box is automatically calculated, L=L0×tanθ, L0 is the reference distance from the center of mass of the counterweight box to the center of gravity of the pontoon module, and θ is the actual detected tilt angle; Start the counterweight drive on the slide rail and move the counterweight box in the predetermined direction. During the movement of the counterweight box, the hydraulic jack on the tilted side is synchronously controlled to slowly release pressure while maintaining the pressure on the jack on the opposite side.

[0013] Preferably, the process of generating a review instruction specifically includes: For the laser ranging data, pressure sensor data and tilt sensor data at the same timestamp, the Pearson correlation coefficient between the three groups of data is calculated respectively, and the minimum value of the three groups of coefficients is taken as the correlation coefficient. When the correlation coefficient is greater than 0.7, the first three horizontal deviation adjustment amounts, torque adjustment amounts and counterweight adjustment amounts under the current pontoon number are further extracted, and the three types of adjustment amounts are weighted averaged respectively. The weighted coefficients are 0.5, 0.3 and 0.2 in reverse chronological order to obtain the historical benchmark values of each type of adjustment amount; the current adjustment amount is compared with the historical benchmark value of the corresponding category. When the water When the deviation adjustment difference is greater than 15% of the reference value, or the torque adjustment difference is greater than 15% of the reference value, or the counterweight displacement difference is greater than 15% of the reference value, a manual review instruction containing an abnormality type identifier is automatically generated. The instruction displays the current sensor data waveform, historical adjustment curve and three-dimensional space posture model in an associated manner. After the review instruction is generated, the operation authority of the underwater hydraulic adjustment mechanism of the current pontoon module is locked, and the data packet containing the operation record of the previous 30 minutes is forcibly backed up. The data packet is marked with the timestamp synchronization status according to the data channels of the laser rangefinder, pressure sensor and inclination sensor.

[0014] Preferably, after the review instruction is generated, a multi-level verification is performed: Level 1 verification: comparing the current abnormal adjustment amount with the adjustment amount feature vectors in the previous five historical abnormal events, when the cosine similarity is greater than 0.85, the corresponding historical processing plan is automatically loaded and the operation authority lock is released; Secondary verification: When the similarity is ≤0.85, the adjacent pontoon module data verification is started: the three-layer sensor data of the left and right adjacent pontoon modules with the same time stamp are extracted, and the spatial correlation between the abnormal module data and the adjacent module data is calculated. If the correlation coefficient is greater than 0.6, it is marked as a systematic anomaly; Level 3 verification: For the pontoon module marked with systemic abnormality, the hydraulic mechanisms of the four surrounding modules are controlled to synchronously perform 0.5mm reverse displacement compensation to form an isolation buffer zone and wait for manual confirmation.

[0015] Preferably, wind speed and direction sensors and tide level gauges are set around the temporary support platform to collect environmental parameters in real time and input them into the assembly adjustment data processing module; when the wind speed is greater than 5m / s, the underwater hydraulic adjustment mechanism automatically increases the lifting speed of the underwater hydraulic adjustment mechanism by 10% when lifting; when the tidal water level change rate is detected to be greater than 50mm / min, the buoyancy dynamic compensation algorithm is started: according to the current water level change △h, the movement of the counterweight box is corrected, and the calculation formula is L=L0×tanθ×(1+0.005△h); the environmental parameter data and the adjustment operation records are stored in association with timestamps to form a mapping relationship database of environmental interference-adjustment response.

[0016] The present invention has at least the following beneficial effects: First, the present invention achieves multi-dimensional dynamic control of horizontal deviation, torque, and counterweight through the collaborative control of a three-layer sensor network of geometric positioning, mechanical perception, and posture monitoring, combined with three-dimensional index data storage and historical adjustment optimization algorithms. This significantly improves the precision control capability of the entire assembly process and suppresses error accumulation. Secondly, the present invention further adopts a multi-stage hydraulic jacking strategy and a dynamic torque loading mechanism. Through preload compensation, balanced jacking mode and pressure balance detection, it achieves refined coordinated control of horizontal deviation and bolt preload force, avoids overload or local stress concentration, and ensures the reliability of the assembly structure. Thirdly, the present invention is based on the adaptive adjustment of counterweight and multi-level verification and validation mechanism of tilt vector mapping. By identifying the abnormality type, verifying the data of adjacent modules and generating an isolation buffer zone, it improves the accuracy of assembly abnormality identification under complex working conditions and strengthens systemic risk management. Fourthly, the present invention also integrates a dynamic compensation algorithm for environmental parameters, which corrects the lifting speed and counterweight adjustment amount in real time according to wind speed and tidal water level, builds an environmental interference-adjustment response mapping database, and enhances the adaptive management and control capabilities in complex marine environments. Fifth, the present invention realizes rapid tracing and hierarchical response of abnormal events through multi-level verification mechanism and intelligent loading of historical processing solutions, reduces the frequency of manual intervention, and improves assembly efficiency and the level of management and control automation.

[0017] Other advantages, objectives and features of the present invention will be reflected in part from the following description and will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The figure is a flow chart of a method for real-time assembly quality control of a pontoon terminal according to a technical solution of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be described in further detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0020] It should be understood that terms such as “having”, “including” and “comprising” used herein do not preclude the existence or addition of one or more other elements or combinations thereof.

[0021] like Figure 1 As shown, the present invention provides a real-time assembly quality control method for a pontoon terminal, comprising: A three-layer sensor network consisting of geometric positioning, mechanical sensing, and attitude monitoring is set up on the assembly surface of the pontoon module. The geometric positioning layer includes laser rangefinders installed at the four corners of the pontoon module. The mechanical sensing layer includes pressure sensors arranged at equal intervals on the edges of the docking surfaces of adjacent pontoon modules. The attitude monitoring layer includes an inclination sensor installed at the bottom of the pontoon module. An assembly quality standard module and an assembly adjustment module are deployed within the real-time assembly quality control platform. When the laser rangefinder detects a horizontal deviation exceeding 5 mm, when the similarity between the actual warning force distribution curve generated by the pressure sensor and the standard curve is less than 85%, or when the inclination sensor detects a tilt angle exceeding 0.5°, the assembly adjustment module is triggered to control the underwater hydraulic adjustment mechanism to perform horizontal deviation adjustment, the torque wrench to perform torque adjustment, and the counterweight drive to perform counterweight adjustment. Among them, an assembly adjustment data processing module is established inside the assembly quality control platform. The assembly process data processing module stores data according to the three-dimensional index of the pontoon number, timestamp and sensor type. The index dimensions include the laser rangefinder number, the pressure sensor group number and the inclination sensor orientation code; at the same time, the assembly process data processing module extracts the previous three adjustments of the pontoon module at the same position and calculates the average value of the adjustment. When the correlation of the three layers of sensor data at the same timestamp is greater than 0.7 and any current adjustment value of the horizontal deviation adjustment, torque adjustment and counterweight adjustment of the pontoon module at the same position exceeds 15% of the average adjustment value of the previous three times, a review instruction is generated.

[0022] In the above technical solution, the geometric positioning layer uses an Omron ZX2-LD50 laser rangefinder, the pressure sensor can be a Honeywell FSS series micro pressure sensor array, each consisting of eight sensors, and the tilt sensor can be a SICKTMS88 dynamic inclinometer. Regarding installation location, the laser rangefinder should be installed in a waterproof compartment at the top corner of the pontoon module, 200 mm from the edge. The pressure sensor array is placed in a groove 50 mm inward from the edge of the docking surface, and the tilt sensor is installed in the center of the pontoon bottom. During operation, the laser rangefinder performs synchronous distance measurement at all four corners every 200 milliseconds, the pressure sensor array collects pressure distribution data every 5 seconds, and the tilt sensor updates attitude data every 1 second. The underwater hydraulic adjustment mechanism can be controlled by an ABB ACS880 frequency converter, the torque wrench can be an Atlas Copco QST30 series smart wrench, and the counterweight adjustment drive can be a Rexroth HMS01.1-W waterproof hydraulic cylinder. The horizontal deviation threshold was set to 5 mm, the pressure curve similarity threshold was set to 85%, and the tilt angle threshold was set to 0.5°.

[0023] In the above technical solution, a multi-dimensional quality monitoring system is constructed through a three-layer sensor network. The geometric positioning layer deploys laser rangefinders at the four corners of the pontoon to form a rectangular positioning reference network, which captures the horizontal displacement deviation between modules in real time. This diagonal point arrangement method can eliminate the interference of single-point errors. When the difference between any two points exceeds 5mm, the horizontal adjustment is triggered, which significantly improves the measurement efficiency compared with traditional total stations. The mechanical perception layer uses a pressure sensor array, and arranges sensor groups at the edge of the docking surface to form a closed monitoring loop. By comparing the similarity between the actual pressure distribution curve and the standard curve, hidden defects such as uneven bolt tightening and contact surface misalignment can be accurately identified. The similarity threshold is set to 85% to avoid false alarms while ensuring sensitivity.

[0024] In the above technical solution, the attitude monitoring layer installs high-precision tilt sensors at key load-bearing nodes at the bottom. The 0.5° threshold setting, verified by fluid dynamics simulations, effectively prevents the risk of capsizing due to center of gravity shift. The three-layer network data is temporally and spatially correlated through a three-dimensional indexing system. The combined encoding of laser rangefinder numbers (e.g., L1-L4), pressure sensor group numbers (P01-P12), and tilt sensor position codes (E / W / S / N) improves the retrieval speed of massive data. A dynamic weighting algorithm is used to calculate the moving average of the first three adjustments. A 15% deviation threshold can both identify abnormal operations and adapt to normal operating fluctuations. A review mechanism is activated when the correlation between the three-layer data exceeds 0.7, and a cross-validation strategy is employed to significantly reduce the false positive rate.

[0025] According to the above technical solution, the working process of the real-time assembly quality control method of the pontoon terminal is as follows: When assembly begins, a three-layer sensor network is activated simultaneously: a laser rangefinder continuously scans the contours of adjacent modules, generating a millimeter-level positioning data stream; a pressure sensor array collects stress waves from the contact surface every 0.5 seconds to construct a dynamic pressure cloud map; and an inclination sensor uploads posture parameters at a 10Hz frequency. After the real-time data is encrypted and transmitted to the control platform, the assembly quality standard module immediately initiates a multi-threaded comparison, with the horizontal deviation analysis thread taking priority. If a diagonal laser ranging error exceeding 5mm is detected, the hydraulic adjustment mechanism is immediately triggered to perform micron-level correction.

[0026] Simultaneously, the engine performs dynamic time warping (DTW) to match the real-time pressure distribution with the standard curve. When the similarity falls below the critical 85% threshold, the intelligent torque wrench automatically calculates the compensating torque value based on the difference area, achieving self-balancing adjustment of the contact surface stress. The moment the attitude monitoring layer detects an abnormal tilt angle, it activates the counterweight adaptive system, which performs counterweight adjustment through the counterweight drive. All operation logs generated during the adjustment process are time-stamped and stored in the assembly adjustment data processing module for subsequent traceability.

[0027] When there's a strong correlation (correlation > 0.7) between the three sensor layers and a sudden increase in the adjustment value for a single parameter, the review instruction generation module automatically retrieves the historical adjustment records for that location and uses a sliding window algorithm to calculate the trend baseline. If the current value deviates from the baseline by more than 15%, the system suspends automatic adjustments and initiates a review.

[0028] The above technical solution, through the construction of a three-in-one monitoring system combining geometry, mechanics, and posture, elevates assembly precision control from a single dimension to a spatially coupled dimension. The laser rangefinder layout enables dual flatness verification, the pressure sensor array overcomes the bottleneck in detecting hidden defects on the contact surface, and bottom inclination monitoring fundamentally mitigates the risk of structural instability. The three-layer data spatiotemporal correlation technology creates a synergistic effect between the monitoring results of different sensors, significantly improving fault identification rates compared to traditional single-layer monitoring methods.

[0029] The combined application of a dynamic mean threshold and a correlation review mechanism ensures adjustment sensitivity while avoiding over-adjustment. The three-dimensional index database not only enables efficient data management but also provides structured data support for subsequent digital twin modeling. Verified in actual engineering, this solution reduced pontoon docking adjustment time by 55%, increased assembly qualification from 82% to 98.6%, and reduced hydraulic mechanism actuation frequency by 30%, significantly extending equipment life. This solution provides an innovative solution for the intelligent construction of large-scale offshore floating structures and possesses significant technological leadership.

[0030] In one of the technical solutions, the pontoon module is set on a temporary support platform on the dock, and the underwater hydraulic adjustment mechanism includes four hydraulic jack units arranged on the temporary support platform. The four hydraulic jack units are just opposite the four corners of the bottom of the pontoon module, and the piston rod on the top of any hydraulic jack unit is connected to the reinforcing steel plate set at the bottom of the pontoon module through a ball joint; the torque wrench is driven by an electric servo, and its shell is fixed to the surface of the pontoon module within 200mm from the center of the connector by a magnetic clamp, and the front end of the torque wrench is provided with a hexagonal socket adapter matching the specifications of the pontoon module connector, the coaxiality deviation of the hexagonal socket adapter and the axis of the connector bolt does not exceed 0.1mm, and the rear end of the hexagonal socket adapter is integrated with a dynamic torque sensor; the counterweight adjustment is achieved by a pair of counterweight boxes symmetrically arranged at the bottom of the pontoon module, and the pair of counterweight boxes are respectively arranged on a pair of parallel slide rails and controlled by corresponding counterweight drives.

[0031] The above-mentioned technical solution constructs a high-precision, high-reliability physical adjustment system. The underwater hydraulic adjustment mechanism utilizes hydraulic jack units arranged symmetrically at the four corners. The top of each jack is connected to the reinforced steel plate at the bottom of the pontoon via a ball joint. This allows the lifting process to transmit vertical loads while also adapting to a ±3° deflection angle, avoiding local stress concentration caused by rigid connections. The torque wrench's magnetic clamp generates a 0.8T suction force through an electromagnet array, enabling positioning and installation within 30 seconds. The coaxiality of the hexagonal socket adapter and the bolt axis, combined with real-time feedback from a dynamic torque sensor, keeps the deviation between the actual applied torque and the target value within ±1.5%. The counterweight adjustment system utilizes a symmetrical dual-rail layout, with a wear-resistant ceramic coating sprayed on the rail surface to reduce the moving friction coefficient. The counterweight drive utilizes a waterproof hydraulic cylinder with a positioning resolution of 0.1mm, ensuring rapid recovery of the inclination angle after center of gravity adjustment.

[0032] In the above technical solution, when the horizontal deviation command is issued, the four jack units start the coordinated leveling mode: the jack piston rod performs micro-adjustment, and the ball joint compensates for the slight angular offset caused by the deformation of the buoy in real time. Preferably, during the leveling process, the strain gauges on the reinforced steel plate continuously monitor the contact pressure to prevent overload damage to the structure. During the torque adjustment stage, the magnetic clamp is adsorbed to the predetermined position after being energized. When the pressure distribution curve is abnormal, a compensating torque waveform is automatically generated, and the drive motor performs multi-stage variable speed rotation to ensure that the bolt group is uniformly stressed. After the counterweight adjustment is started, the counterweight drive calculates the moving distance of the counterweight box based on the vector data fed back by the inclination sensor, and positions it through the absolute encoder, so that the counterweight box moves synchronously in the opposite direction along the slide rail to dynamically balance the center of gravity of the module.

[0033] The above-mentioned technical solution significantly improves the reliability and response speed of the actuator through the deep integration of precision mechanical design and intelligent control algorithms. The hydraulic jack's ball joint connection and four-corner coordinated control technology ensure a smooth and impact-free pontoon leveling process, reducing structural stress peaks. The magnetic torque wrench shortens the time required for single bolt adjustments and eliminates the risk of human error. The counterweight adjustment utilizes a symmetrical drive design, achieving a center of gravity adjustment accuracy of 0.05°. Combined with wear-resistant slide rails, the equipment maintenance cycle can be extended to over 5,000 hours. The entire actuator system is fully adapted to the harsh working conditions of high humidity and high salt spray at the terminal.

[0034] In one of the technical solutions, when the underwater hydraulic adjustment mechanism performs horizontal deviation adjustment, it is implemented according to the following multi-stage control strategy: When the horizontal deviation value reaches 5mm for the first time, the preload compensation mode is activated: the two diagonally arranged jack units are controlled to perform initial displacement compensation at a synchronous lifting speed of 1mm / s. At the same time, the readings of the two sets of sensors closest to the adjustment point in the pressure sensor group are monitored in real time. When the pressure value of any set drops by more than 10% of its initial value, the lifting is immediately stopped and the system switches to the balanced lifting mode. In the balanced lifting mode, the four jack units are lifted alternately according to a spiral progressive algorithm: each time the jack on one side is lifted by 0.5 mm, the jack on the opposite side is switched to lift. The alternating frequency is synchronized with the frequency of updating the laser rangefinder data. The module geometric center offset is recalculated after each lifting cycle. When the offset calculated value converges to the range of ±1 mm for three consecutive iterative calculations, the adjustment is determined to be complete.

[0035] In this technical solution, a refined multi-level control strategy optimizes the hydraulic adjustment mechanism from rapid response to high-precision leveling. When the preload compensation mode is first triggered, a synchronized lifting strategy for the diagonal jacks is employed. By controlling the synchronized movement of the two diagonal units, this strategy eliminates initial deviations while preventing overload at a single point. Designing diagonal compensation, rather than unilateral lifting, maintains initial force balance within the pontoon module and reduces the risk of deformation of the supporting structure. Real-time monitoring of the pressure sensor group sets a 10% threshold to trigger mode switching. This prevents contact surface preload failure due to excessive lifting and identifies potential structural slippage, reducing the false positive rate by 25% compared to traditional single-threshold control methods. The balanced lifting mode incorporates a spiral progressive algorithm, alternating the lifting of the four corner jacks to create a progressive adjustment path. The alternating frequency is synchronized with the laser rangefinder data update cycle, ensuring the latest deformation data before each action and preventing error accumulation. The geometric center offset is recalculated after each cycle, and a three-iteration convergence determination mechanism is employed to maintain final leveling accuracy within ±1mm, significantly reducing energy consumption. The staged control strategy ensures both adjustment efficiency (preload mode quickly eliminates large deviations) and fine-tuning accuracy (balance mode fine correction), forming a complete closed-loop control chain.

[0036] According to the above technical solution, when the horizontal deviation first reaches the 5mm threshold, the system automatically enters preload compensation mode: Selected diagonal jack groups are simultaneously raised at a speed of 1mm / s, while real-time pressure sensor data in the corresponding areas is collected. If the pressure value of any group drops by more than 10% of the initial value during the lifting process, the lifting is immediately interrupted and the system switches to balanced lifting mode to prevent excessive release of contact surface preload. In balanced lifting mode, the control system drives each jack unit in a predetermined spiral sequence (e.g., left front → right rear → right front → left rear), pausing after each 0.5mm lift to allow the laser rangefinder to update the four corner coordinates. The calculation module reconstructs the 3D plane model based on the new data and calculates the geometric center offset vector. When the offset within the target range of ±1mm for three consecutive adjustment cycles, the system determines that leveling is complete and locks the hydraulic mechanism. Throughout the entire process, the pressure sensor data is cross-validated with the laser rangefinder data to ensure that both the mechanical state and geometric deformation meet the requirements.

[0037] The multi-level control strategy described above is particularly well-suited for the dynamic conditions encountered during dock assembly, significantly improving the reliability and accuracy of leveling. The preload compensation mode, through dual pressure-displacement monitoring, quickly eliminates large deviations while minimizing the risk of overshoot. A spiral progressive algorithm, combined with an iterative convergence mechanism, improves final leveling accuracy to the millimeter level, reducing stress fluctuations during adjustment and significantly extending the life of the hydraulic mechanism. The design of three iterative convergence criteria effectively filters out measurement noise.

[0038] In one technical solution, when the torque wrench performs torque adjustment, the following steps are performed: In the three sets of pressure sensors at the interface of adjacent pontoon modules, each set of sensors acquires contact pressure values at a sampling frequency of 10 Hz. The actual preload force distribution curve is generated using the cubic spline interpolation method and discretized into 50 equally spaced data points. The standard preload curve is stored in the assembly quality standard module. During the comparison, the dynamic time warping algorithm is used to calculate the similarity of the two curves, and a torque adjustment instruction is generated when the calculated result is less than 85%; After receiving the adjustment command, the torque wrench applies torque in step mode. The initial torque is set at 80% of the design value and gradually increases in increments of 5% of the design value. After each increase in the torque gradient, a 3-second stabilization time is maintained. The reading change rate of the pressure sensor group is simultaneously collected. If the reading change rate of any group of pressure sensors exceeds 5% / s, the loading is suspended and the change rate is not allowed to continue until it drops below 2% / s. The dynamic torque sensor monitors the output torque value at a frequency of 100 Hz and establishes real-time correlation with the pressure sensor data: When the average pressure sensor reading reaches 90% of the design value, reduce the torque loading rate to 1% of the design value / second; When the pressure reading enters the range of 90-110% of the design value, a pressure balance check is performed every 0.5 seconds. The balance is the ratio of the difference between the maximum and minimum pressure values to the average pressure value. If the balance is greater than 15%, the local tightening strategy is triggered: the torque wrench is controlled to add 3% of the design value torque in the lowest pressure area. The torque output is stopped when the following conditions are met simultaneously: the average reading of any pressure sensor reaches the range of 95% to 105% of the design value; the similarity between the latest generated actual preload force distribution curve and the standard curve is ≥ 88%; the torque fluctuation coefficient monitored by the dynamic torque sensor is less than 2%, where the torque fluctuation coefficient = standard deviation / mean; Finally, the final torque value, pressure distribution curve and termination condition parameters are stored in association with the pressure sensor group number and timestamp, and written into the assembly adjustment data processing module.

[0039] This technical solution achieves precise control of bolt preload and dynamic balancing of pressure distribution by building an intelligent, graded torque loading system. Cubic spline interpolation is used to reconstruct the discrete pressure data (sampled at 10Hz), generating an actual preload distribution curve with 50 equally spaced data points. This method preserves the original data characteristics while eliminating random noise. A dynamic time warping (DTW) algorithm is used for standard curve comparison, adaptively adapting to the pressure variation rhythm at different assembly stages. A similarity threshold of 85% ensures effective detection of early anomalies while avoiding false triggering. The staged torque loading strategy sets the initial torque at 80% of the design value. By increasing the pressure incrementally by 5% and monitoring the rate of change, it gradually eliminates gaps in the joint and compensates for material creep. A 5% / s pressure rate of change threshold serves as a safety buffer to prevent structural damage caused by sudden stress changes. A real-time correlation mechanism between the dynamic torque sensor and pressure data automatically reduces the loading rate when the average pressure reaches 90% of the design value, enabling seamless transition from coarse to fine control and improving torque control accuracy to within ±1.5%. The combination of pressure balance detection (maximum / minimum difference ratio) and local tightening strategy can strengthen weak areas in a targeted manner and improve the uniformity of force on the bolt group.

[0040] According to the above technical solution, when the system detects that the actual preload curve similarity falls below 85%, the torque adjustment process is triggered. First, a high-fidelity pressure distribution curve is generated using cubic spline interpolation. This is then matched against the standard curve using DTW to locate areas of discrepancy. Starting at 80% of the design value, the torque wrench is loaded in 5% increments, maintaining a stabilization time of 3 seconds per step. If the pressure change rate exceeds 5% / s (e.g., sudden bolt slippage), loading is immediately paused and the fluctuation is allowed to decay. When the mean pressure reaches 90% of the design value, the loading rate is reduced to 1% / s, and the precision adjustment phase begins. The system calculates pressure balance every 0.5 seconds. If it exceeds 15% (e.g., if a bolt on one side is undertightened), an additional 3% torque is automatically applied to the low-pressure area. The dynamic torque sensor also adjusts the sleeve's position in real time. Termination conditions are met when the mean pressure meets the standard (95-105%), the curve similarity is ≥88%, and the torque fluctuation coefficient is <2%. The system records the final parameters and stores them in a linked format. The entire process data is archived by pressure group number and timestamp, forming a complete process traceability chain.

[0041] The above technical solution significantly improves the quality of bolt connections through multi-dimensional closed-loop control. The phased gradient loading strategy makes the preload establishment process smooth and controllable, avoiding the risk of overload or underload, and extending the fatigue life of the bolts compared to the traditional one-time loading method. The combination of pressure balance detection and local tightening mechanism reduces stress unevenness on the connection surface and effectively prevents seal failure caused by eccentric loading. The cross-validation of dynamic torque sensor and pressure data ensures that the torque output value truly reflects the stress state of the structure and improves the assembly qualification rate.

[0042] In one of the technical solutions, when the hydraulic mechanism performs counterweight adjustment, the following steps are performed: According to the real-time output of the X / Y dual-axis tilt data by the tilt sensor, when the absolute value of any axial tilt angle is greater than 0.5°, the tilt direction vector is calculated and mapped to the plane coordinate system of the pontoon module to determine the movement direction of the counterweight box. The movement direction of the counterweight box is opposite to the tilt direction, where the X axis is parallel to the long side of the bottom surface of the pontoon module, and the Y axis is parallel to the short side of the bottom surface of the pontoon module; According to the geometric relationship between the center of gravity height of the pontoon module and the current tilt angle, the sliding distance L of the counterweight box is automatically calculated, L=L0×tanθ, L0 is the reference distance from the center of mass of the counterweight box to the center of gravity of the pontoon module, and θ is the actual detected tilt angle; Start the counterweight drive on the slide rail and move the counterweight box in the predetermined direction. During the movement of the counterweight box, the hydraulic jack on the tilted side is synchronously controlled to slowly release pressure while maintaining the pressure on the jack on the opposite side.

[0043] This technical solution relies on real-time X / Y tilt data from an inclinometer. The system uses plane coordinate mapping technology to convert the tilt direction into a vector path for the counterweight box's movement, ensuring that the adjustment direction is opposite to the center of gravity offset, thus avoiding directional errors typically associated with traditional manual judgment. The counterweight box's movement distance is calculated using a geometric relationship model (L = L0 × tanθ), where L0 is the reference distance from the counterweight box's center of mass to the center of gravity of the pontoon module. This parameter is optimized through finite element simulation, and the calculated result matches the actual center of gravity compensation requirements by over 98%. This reference distance needs to be re-optimized through finite element simulation for different counterweight box and pontoon modules. The synchronous control strategy embodies the coordinated operation of the counterweight box's movement and hydraulic jack pressure relief: the jack on the tilting side relieves pressure at a rate of 0.02 MPa / s, while maintaining pressure on the opposite side. This prevents sudden module instability and reduces hydraulic system energy loss. The counterweight drive utilizes dual closed-loop speed-pressure control. The counterweight box's movement speed is dynamically adjusted based on the real-time rate of tilt change, preventing overshoot and significantly improving adjustment safety.

[0044] According to the above technical solution, when the inclination sensor detects a tilt exceeding 0.5° on any axis, the system immediately initiates the counterweight adjustment process. First, the X / Y axis tilt data is analyzed to generate a tilt vector. The vector direction is mapped to the coordinate system of the pontoon bottom surface, determining whether the counterweight box should move along the long side (X-axis) or short side (Y-axis) slide rail. After calculating the movement distance using L = L0 × tanθ, the counterweight drive is activated, driving the counterweight box in the opposite direction of the tilt. During this movement, the hydraulic jack on the tilting side gradually releases pressure according to a preset curve. The pressure release rate is linearly correlated with the counterweight box's displacement (0.5% of the rated pressure for every 10mm of movement). Simultaneously, a pressure sensor monitors the load changes on the opposite jack in real time, ensuring that pressure fluctuations do not exceed ±3%. When the tilt sensor feedback angle drops within 0.3°, the system enters the fine-tuning phase, reducing the counterweight box's movement speed by 20% to 50% until the tilt angle stabilizes within a 0.1° error band. The entire process data is synchronously recorded along with the jack pressure value and the counterweight position coordinates, forming a complete adjustment log.

[0045] The above technical solution significantly improves center-of-gravity adjustment efficiency through precise geometric modeling and multi-system coordinated control. Vector mapping technology achieves a 99.8% accuracy rate in counterweight direction determination, eliminating the risk of secondary tilt caused by directional misjudgment in traditional methods. The dynamic distance calculation formula, combined with the reference distance parameter, improves the single-adjustment compliance rate and shortens adjustment time. The synchronized control strategy of hydraulic pressure relief and counterweight movement effectively suppresses structural vibration during the adjustment process, ensuring smooth module posture transitions and reducing the number of hydraulic system actuations.

[0046] In one of the technical solutions, the process of generating a review instruction specifically includes: For the laser ranging data, pressure sensor data and tilt sensor data at the same timestamp, the Pearson correlation coefficient between the three groups of data is calculated respectively, and the minimum value of the three groups of coefficients is taken as the correlation coefficient. When the correlation coefficient is greater than 0.7, the first three horizontal deviation adjustment amounts, torque adjustment amounts and counterweight adjustment amounts under the current pontoon number are further extracted, and the three types of adjustment amounts are weighted averaged respectively. The weighted coefficients are 0.5, 0.3 and 0.2 in reverse chronological order to obtain the historical benchmark values of each type of adjustment amount; the current adjustment amount is compared with the historical benchmark value of the corresponding category. When the water When the deviation adjustment difference is greater than 15% of the reference value, or the torque adjustment difference is greater than 15% of the reference value, or the counterweight displacement difference is greater than 15% of the reference value, a manual review instruction containing an abnormality type identifier is automatically generated. The instruction displays the current sensor data waveform, historical adjustment curve and three-dimensional space posture model in an associated manner. After the review instruction is generated, the operation authority of the underwater hydraulic adjustment mechanism of the current pontoon module is locked, and the data packet containing the operation record of the previous 30 minutes is forcibly backed up. The data packet is marked with the timestamp synchronization status according to the data channels of the laser rangefinder, pressure sensor and inclination sensor.

[0047] This technical solution establishes a multi-dimensional data correlation analysis and dynamic historical benchmark assessment system, enabling intelligent generation of review instructions and precise risk management. The Pearson correlation coefficient is used to calculate the correlation between data from the laser rangefinder, pressure sensor, and inclination sensor. The minimum of these three coefficients is used as the judgment basis. This effectively identifies unusual divergences between sensor data and avoids the risk of misjudgment caused by a single sensor failure. A correlation threshold of 0.7 is set to ensure data interoperability under normal operating conditions while providing margin for unexpected interference, thereby reducing false positives. Historical benchmark values are calculated using a time-weighted average algorithm, with the most recent three adjustments assigned decay weights of 0.5, 0.3, and 0.2, respectively. This emphasizes the value of the latest operating conditions while preserving the continuity of historical trends. The 15% deviation threshold, validated through Monte Carlo simulation, can capture abnormal operational events with a 95% confidence level. After a review instruction is generated, the system automatically locks operational permissions and backs up the encrypted data package. Timestamp synchronization ensures temporal consistency across multiple data sources, providing a reliable basis for subsequent analysis.

[0048] According to the above technical solution, when the correlation coefficient of the three-layer sensor data exceeds 0.7, the system initiates a review and determination process. First, the system extracts the three previous adjustment records of the current pontoon module from the 3D index database and performs a time-weighted average calculation on the three types of data: horizontal deviation, torque, and counterweight displacement. For example, if the previous three horizontal adjustments were 8mm, 7mm, and 6mm, the weighted baseline value is 8 × 0.5 + 7 × 0.3 + 6 × 0.2 = 7.3mm. If the current adjustment reaches 8.4mm (exceeding the baseline by 15%), a review mechanism is triggered. The system automatically generates a review instruction with an anomaly type tag and simultaneously retrieves the current sensor data waveform, historical adjustment trend curve, and 3D posture model, displaying them as an overlay on the control interface. After the operation permission is locked, the hydraulic mechanism enters hold mode and the data backup module is activated. The operation records for the previous 30 minutes are categorized by sensor channel and the time stamp synchronization error of each channel is verified (controlled within ±10ms). After manual review confirms the cause of the anomaly, the system resumes automatic adjustment or switches to manual intervention mode according to the authorized instruction.

[0049] According to the above technical solution, the present invention significantly improves the accuracy of identifying abnormal operating conditions through multi-source data fusion and historical trend analysis. The Pearson coefficient minimum determination strategy improves the sensitivity of cross-sensor anomaly detection, effectively avoiding systemic risks caused by localized failures. The time-weighted reference value calculation method takes into account both data timeliness and continuity, improving the accuracy of abnormal deviation detection. The linkage design of permission locking and data backup mechanism reduces the risk of structural damage caused by misoperation.

[0050] In one of the technical solutions, after the review instruction is generated, a multi-level verification is performed: Level 1 verification: compare the feature vectors of the current abnormal adjustment amount with the adjustment amount in the previous five historical abnormal events. When the cosine similarity is greater than 0.85, the corresponding historical processing plan is automatically loaded and the operation authority is unlocked; Secondary verification: When the similarity is ≤0.85, the adjacent pontoon module data verification is started: the three-layer sensor data of the left and right adjacent pontoon modules with the same time stamp are extracted, and the spatial correlation between the abnormal module data and the adjacent module data is calculated. If the correlation coefficient is greater than 0.6, it is marked as a systematic anomaly; Level 3 verification: For the pontoon module marked with systemic abnormality, the hydraulic mechanisms of the four surrounding modules are controlled to synchronously perform 0.5mm reverse displacement compensation to form an isolation buffer zone and wait for manual confirmation.

[0051] The above solution achieves intelligent classification and hierarchical handling of abnormal events by building a multi-level verification mechanism. The first-level verification utilizes feature vector comparison technology based on historical abnormal events, converting current abnormal adjustment variables (such as horizontal deviation and torque) into multidimensional feature vectors. Cosine similarity calculation (threshold 0.85) is then used to match similar cases in the historical database. Upon a successful match, the system automatically invokes a validated solution, shortening recovery time for common abnormalities and avoiding repeated manual analysis while preserving the reuse value of empirical knowledge. The second-level verification incorporates a spatial correlation analysis mechanism. If a match is not found in the first-level verification, the system extracts three-layer sensor data from adjacent pontoon modules and calculates the correlation coefficient (threshold 0.6) between the abnormal module and the adjacent modules. This effectively distinguishes local anomalies from systemic risks, such as multi-module linkage offset caused by tidal forces, and avoids global imbalances caused by isolated treatment. The third-level verification targets systemic anomalies by controlling surrounding modules to perform 0.5mm reverse compensation to form an isolation buffer zone. This small displacement offsets the spread of the anomaly, creating a safe window for manual intervention and reducing production losses compared to traditional emergency shutdown methods.

[0052] According to the above technical solution, when a review instruction is generated, the system first performs a first-level check: it extracts feature vectors of similar anomalies (such as adjustment amplitude, rate of change, and spatial distribution pattern) from a historical database and calculates their cosine similarity with the current event. If the similarity of a historical event exceeds 0.85 (e.g., matching a loose bolt signature), the system automatically loads the current torque compensation solution, unlocks permissions, and resumes the automatic adjustment process. If the first-level check fails, a second-level check is initiated: it retrieves sensor data from the left and right adjacent pontoon modules and analyzes the spatiotemporal correlation between the abnormal module and its adjacent modules in terms of horizontal deviation, pressure distribution, and tilt angle. If the correlation coefficient exceeds 0.6 (e.g., multiple modules are tilting synchronously), it is determined to be a systemic anomaly, triggering a third-level check. At this point, the system controls the hydraulic mechanisms of the four pontoon modules surrounding the abnormal module to perform a 0.5mm reverse lift, forming a physical isolation zone. It also locks access to the relevant area, awaiting manual intervention.

[0053] The above technical solution is particularly suitable for multi-module collaborative operation scenarios. Its reverse compensation strategy achieves risk isolation through tiny displacements, reducing material loss compared to traditional emergency stop solutions. The continuous learning function of the historical feature library enables the system to have adaptive optimization capabilities, thereby improving the safety and fault tolerance of large-scale assembly projects.

[0054] In one of the technical solutions, wind speed and direction sensors and tide level gauges are set up around the temporary support platform to collect environmental parameters in real time and input them into the assembly adjustment data processing module; when the wind speed is greater than 5m / s, the underwater hydraulic adjustment mechanism automatically increases the lifting speed of the underwater hydraulic adjustment mechanism by 10% during lifting; when the tidal water level change rate is detected to be greater than 50mm / min, the buoyancy dynamic compensation algorithm is started: according to the current water level change △h, the movement of the counterweight box is corrected, and the calculation formula is L=L0×tanθ×(1+0.005△h); the environmental parameter data and the adjustment operation records are stored in association with timestamps to form a mapping relationship database of environmental interference-adjustment response.

[0055] The above technical solution realizes the adaptive adjustment of the assembly process to the complex marine environment through the integration of environmental perception and dynamic compensation algorithm. Wind speed and direction sensors and tide level gauges are deployed around the temporary support platform to form an environmental monitoring ring network. When the wind speed is greater than 5m / s, the hydraulic jacking speed is automatically increased by 10%. Short-term acceleration is used to offset the impact of wind load disturbance on leveling accuracy and reduce horizontal deviation fluctuations. When the tidal water level change rate is greater than 50mm / min, the buoyancy dynamic compensation algorithm is activated. The buoyancy box movement formula is corrected in real time by introducing the water level change △h. The correction coefficient is calibrated by the water tank test. For different water environments, the correction coefficient needs to be recalibrated by the water tank test to accurately compensate for the overturning moment caused by the sudden change of buoyancy, so that the center of gravity adjustment accuracy remains within ±0.1° during the active tidal period. The present invention further stores the association between environmental parameters and adjustment operations, and constructs a mapping database of environmental interference-adjustment response. Wind speed and tide data are aligned with hydraulic jacking and counterweight adjustment records by timestamp, and the implicit relationship between environmental factors and adjustment amounts is mined through machine learning to provide data support for subsequent assembly process optimization.

[0056] According to the above technical solution, when the wind speed sensor detects wind speeds exceeding 5 m / s, the system immediately triggers hydraulic lifting speed adjustments: for example, increasing the lifting rate from 1 mm / s to 1.1 mm / s. Simultaneously, the sampling frequency of the jack pressure closed-loop control is increased to 50 Hz to ensure pressure stability during high-speed lifting. A tidal level gauge collects water level changes at a 1 Hz frequency. When the rate of change exceeds 50 mm / min, the buoyancy compensation algorithm automatically activates, calculating the current Δh value in real time and correcting the ballast box movement. For example, if the water level rises by 200 mm, the ballast box movement is corrected to 1 + 0.005 × 200 = 2 times, and the required ballast box movement is adjusted to 2 times to quickly compensate for the increased buoyancy. All environmental parameters and corresponding adjustment operations (such as lifting speed, ballast weight displacement, and torque) are synchronously stored with millisecond-level timestamps. The system performs data correlation analysis every 3-10 minutes, automatically annotating adjustment features under typical operating conditions such as sudden wind speed changes and tidal surges, and generating an environment-response knowledge graph for intelligent decision-making.

[0057] The number of devices and processing scale described here are used to simplify the description of the present invention. Applications, modifications and variations of the real-time assembly quality control method of the floating pontoon terminal of the present invention are obvious to those skilled in the art.

[0058] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A real-time assembly quality control method for a pontoon terminal, characterized in that: include: A three-layer sensor network consisting of geometric positioning, mechanical sensing, and attitude monitoring is set up on the assembly surface of the pontoon module. The geometric positioning layer includes laser rangefinders installed at the four corners of the pontoon module. The mechanical sensing layer includes pressure sensors arranged at equal intervals on the edges of the docking surfaces of adjacent pontoon modules. The attitude monitoring layer includes an inclination sensor installed at the bottom of the pontoon module. An assembly quality standard module and an assembly adjustment module are deployed within the real-time assembly quality control platform. When the laser rangefinder detects a horizontal deviation exceeding 5 mm, when the similarity between the actual warning force distribution curve generated by the pressure sensor and the standard curve is less than 85%, or when the inclination sensor detects a tilt angle exceeding 0.5°, the assembly adjustment module is triggered to control the underwater hydraulic adjustment mechanism to perform horizontal deviation adjustment, the torque wrench to perform torque adjustment, and the counterweight drive to perform counterweight adjustment. Among them, an assembly adjustment data processing module is established inside the assembly quality control platform. The assembly process data processing module stores data according to the three-dimensional index of the pontoon number, timestamp and sensor type. The index dimensions include the laser rangefinder number, the pressure sensor group number and the inclination sensor orientation code; at the same time, the assembly process data processing module extracts the previous three adjustments of the pontoon module at the same position and calculates the average value of the adjustment. When the correlation of the three layers of sensor data at the same timestamp is greater than 0.7 and any current adjustment value of the horizontal deviation adjustment, torque adjustment and counterweight adjustment of the pontoon module at the same position exceeds 15% of the average adjustment value of the previous three times, a review instruction is generated.

2. The method for real-time assembly quality control of a pontoon dock according to claim 1, characterized in that: The pontoon module is arranged on a temporary support platform on the dock, and the underwater hydraulic adjustment mechanism includes four hydraulic jack units arranged on the temporary support platform, and the four hydraulic jack units are just opposite the four corners of the bottom of the pontoon module. The piston rod on the top of any hydraulic jack unit is connected to the reinforcing steel plate arranged at the bottom of the pontoon module through a ball joint; the torque wrench is driven by an electric servo, and its shell is fixed to the surface of the pontoon module within 200mm from the center of the connector by a magnetic clamp, and the front end of the torque wrench is provided with a hexagonal socket adapter matching the specifications of the pontoon module connector, the coaxiality deviation of the hexagonal socket adapter and the axis of the connector bolt does not exceed 0.1mm, and the rear end of the hexagonal socket adapter is integrated with a dynamic torque sensor; the counterweight adjustment is achieved by a pair of counterweight boxes symmetrically arranged at the bottom of the pontoon module, and the pair of counterweight boxes are respectively arranged on a pair of parallel slide rails and controlled by corresponding counterweight drives.

3. The method for real-time assembly quality control of a pontoon dock according to claim 2, characterized in that: When the underwater hydraulic adjustment mechanism performs horizontal deviation adjustment, it is implemented according to the following multi-stage control strategy: When the horizontal deviation value reaches 5mm for the first time, the preload compensation mode is activated: the two diagonally arranged jack units are controlled to perform initial displacement compensation at a synchronous lifting speed of 1mm / s. At the same time, the readings of the two sets of sensors closest to the adjustment point in the pressure sensor group are monitored in real time. When the pressure value of any set drops by more than 10% of its initial value, the lifting is immediately stopped and the system switches to the balanced lifting mode. In the balanced lifting mode, the four jack units are lifted alternately according to a spiral progressive algorithm: each time the jack on one side is lifted by 0.5 mm, the jack on the opposite side is switched to lift. The alternating frequency is synchronized with the frequency of updating the laser rangefinder data. The module geometric center offset is recalculated after each lifting cycle. When the offset calculated value converges to the range of ±1 mm for three consecutive iterative calculations, the adjustment is determined to be complete.

4. The method for real-time assembly quality control of a pontoon dock according to claim 3, characterized in that: When adjusting the torque using a torque wrench, follow these steps: In the three sets of pressure sensors at the interface of adjacent pontoon modules, each set of sensors acquires contact pressure values at a sampling frequency of 10 Hz. The actual preload force distribution curve is generated using the cubic spline interpolation method and discretized into 50 equally spaced data points. The standard preload curve is stored in the assembly quality standard module. During the comparison, the dynamic time warping algorithm is used to calculate the similarity of the two curves, and a torque adjustment instruction is generated when the calculated result is less than 85%; After receiving the adjustment command, the torque wrench applies torque in step mode. The initial torque is set at 80% of the design value and gradually increases in increments of 5% of the design value. After each increase in the torque gradient, a 3-second stabilization time is maintained. The reading change rate of the pressure sensor group is simultaneously collected. If the reading change rate of any group of pressure sensors exceeds 5% / s, the loading is suspended and the change rate is not allowed to continue until it drops below 2% / s. The dynamic torque sensor monitors the output torque value at a frequency of 100 Hz and establishes real-time correlation with the pressure sensor data: When the average pressure sensor reading reaches 90% of the design value, reduce the torque loading rate to 1% of the design value / second; When the pressure reading enters the range of 90-110% of the design value, a pressure balance check is performed every 0.5 seconds. The balance is the ratio of the difference between the maximum and minimum pressure values to the average pressure value. If the balance is greater than 15%, the local tightening strategy is triggered: the torque wrench is controlled to add 3% of the design value torque in the lowest pressure area. The torque output is stopped when the following conditions are met simultaneously: the average reading of any pressure sensor reaches the range of 95% to 105% of the design value; the similarity between the latest generated actual preload force distribution curve and the standard curve is ≥ 88%; the torque fluctuation coefficient monitored by the dynamic torque sensor is less than 2%, where the torque fluctuation coefficient = standard deviation / mean; Finally, the final torque value, pressure distribution curve and termination condition parameters are stored in association with the pressure sensor group number and timestamp, and written into the assembly adjustment data processing module.

5. The method for real-time assembly quality control of a pontoon dock according to claim 4, characterized in that: When the hydraulic mechanism performs counterweight adjustment, perform the following steps: According to the real-time output of the X / Y dual-axis tilt data by the tilt sensor, when the absolute value of any axial tilt angle is greater than 0.5°, the tilt direction vector is calculated and mapped to the plane coordinate system of the pontoon module to determine the movement direction of the counterweight box. The movement direction of the counterweight box is opposite to the tilt direction, where the X axis is parallel to the long side of the bottom surface of the pontoon module, and the Y axis is parallel to the short side of the bottom surface of the pontoon module; According to the geometric relationship between the center of gravity height of the pontoon module and the current tilt angle, the sliding distance L of the counterweight box is automatically calculated, L=L0×tanθ, L0 is the reference distance from the center of mass of the counterweight box to the center of gravity of the pontoon module, and θ is the actual detected tilt angle; Start the counterweight drive on the slide rail and move the counterweight box in the predetermined direction. During the movement of the counterweight box, the hydraulic jack on the tilted side is synchronously controlled to slowly release pressure while maintaining the pressure on the jack on the opposite side.

6. The method for real-time assembly quality control of a pontoon dock according to claim 5, characterized in that: The process of generating a review instruction specifically includes: For the laser ranging data, pressure sensor data and tilt sensor data at the same timestamp, the Pearson correlation coefficient between the three groups of data is calculated respectively, and the minimum value of the three groups of coefficients is taken as the correlation coefficient. When the correlation coefficient is greater than 0.7, the first three horizontal deviation adjustment amounts, torque adjustment amounts and counterweight adjustment amounts under the current pontoon number are further extracted, and the three types of adjustment amounts are weighted averaged respectively. The weighted coefficients are 0.5, 0.3 and 0.2 in reverse chronological order to obtain the historical benchmark values of each type of adjustment amount; the current adjustment amount is compared with the historical benchmark value of the corresponding category. When the water When the deviation adjustment difference is greater than 15% of the reference value, or the torque adjustment difference is greater than 15% of the reference value, or the counterweight displacement difference is greater than 15% of the reference value, a manual review instruction containing an abnormality type identifier is automatically generated. The instruction displays the current sensor data waveform, historical adjustment curve and three-dimensional space posture model in an associated manner. After the review instruction is generated, the operation authority of the underwater hydraulic adjustment mechanism of the current pontoon module is locked, and the data packet containing the operation record of the previous 30 minutes is forcibly backed up. The data packet is marked with the timestamp synchronization status according to the data channels of the laser rangefinder, pressure sensor and inclination sensor.

7. The method for real-time assembly quality control of a pontoon dock according to claim 6, characterized in that: Perform multi-level verification after generating the review instruction: Level 1 verification: Compare the current abnormal adjustment amount with the adjustment amount feature vectors of the previous five historical abnormal events. When the cosine similarity is greater than 0.85, the corresponding historical processing plan is automatically loaded and the operation permission is unlocked; Secondary verification: When the similarity is ≤0.85, the adjacent pontoon module data verification is started: the three-layer sensor data of the left and right adjacent pontoon modules with the same time stamp are extracted, and the spatial correlation between the abnormal module data and the adjacent module data is calculated. If the correlation coefficient is greater than 0.6, it is marked as a systematic anomaly; Level 3 verification: For the pontoon module marked with systemic abnormality, the hydraulic mechanisms of the four surrounding modules are controlled to synchronously perform 0.5mm reverse displacement compensation to form an isolation buffer zone and wait for manual confirmation.

8. The method for real-time assembly quality control of a pontoon dock according to claim 7, characterized in that: Wind speed and direction sensors and tide level gauges are set up around the temporary support platform to collect environmental parameters in real time and input them into the assembly adjustment data processing module; when the wind speed is greater than 5m / s, the underwater hydraulic adjustment mechanism automatically increases the jacking speed of the underwater hydraulic adjustment mechanism by 10% during jacking; when the tidal water level change rate is detected to be greater than 50mm / min, the buoyancy dynamic compensation algorithm is started: according to the current water level change △h, the movement of the counterweight box is corrected, and the calculation formula is L=L0×tanθ×(1+0.005△h); the environmental parameter data and the adjustment operation records are associated and stored according to the timestamp to form a mapping relationship database between environmental interference and adjustment response.

Images