Intelligent operating system for girder erecting machines based on the installation of superstructure components of high-pile wharves

Abstract

This invention discloses an intelligent operating system for girder erection machines based on the installation of superstructure components of high-pile wharves. The system includes: a wharf digital twin modeling module, an identity recognition and feature extraction module, a multi-degree-of-freedom motion state perception module, a marine environmental disturbance monitoring module, a task planning and path generation module, a spatiotemporal synchronization and fusion processing module, an adaptive motion control execution module, a visual servo fine-tuning module for the installation process, and a safety boundary dynamic verification module. This invention achieves high-precision collaborative control of prefabricated components of high-pile wharves throughout the entire process of lifting, traveling, lateral movement, and precise positioning by constructing a multi-source heterogeneous sensor fusion system and a task-driven three-dimensional spatial dynamic modeling mechanism.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology for engineering machinery, specifically an intelligent operating system for a girder erecting machine based on the installation of superstructure components of a high-pile wharf. Background Technology

[0002] With the accelerated development of modern port construction towards large-scale, deep-water, and intelligent directions, the superstructure of high-pile wharves generally adopts a construction mode of factory prefabrication and on-site assembly. This mode, relying on standardized design and industrialized production, significantly improves component quality and construction efficiency, and has become the mainstream construction method for current port engineering. Its core components, such as berthing components, crossbeams, longitudinal beams, track beams, and various panels, are characterized by their complexity, large individual weight, and stringent requirements for three-dimensional installation accuracy, placing extremely high demands on on-site hoisting and positioning equipment. Against this backdrop, port construction girder erection machines, as key construction equipment, integrate multiple functions such as lifting, traveling, lateral movement, and precise positioning to adapt to the complex and ever-changing wharf operating environment.

[0003] Among them, the intelligent operating system for the girder erection machine based on the installation of precast components of the high-pile wharf superstructure aims to achieve precise control of heavy precast components throughout the entire process through digital sensing, intelligent decision-making, and automatic control technologies. This system needs to integrate functions such as high-precision positioning, real-time attitude adjustment, multi-degree-of-freedom cooperative motion control, and human-machine interactive visualization to ensure that millimeter-level installation operations can still be completed under adverse conditions such as dynamic sea conditions, wind load disturbances, and structural obstructions. However, existing girder erection equipment largely relies on manual experience and lacks real-time perception and closed-loop feedback capabilities for component spatial orientation, equipment operating status, and environmental interference factors.

[0004] In existing technologies, traditional bridge erecting machines generally employ discrete control systems, with lifting, traveling, and lateral movement functions controlled separately by independent subsystems. The lack of coordination logic between these actions leads to cumbersome operating procedures and delayed responses. Furthermore, on-site positioning relies heavily on external measurement methods such as total stations or laser rangefinders, resulting in low data update frequency and susceptibility to weather and line-of-sight obstructions, making it difficult to support continuous and efficient automated installation operations. In addition, the operating interfaces are mostly mechanical instruments or simple digital displays, lacking three-dimensional visualization guidance and risk warning mechanisms. Under complex working conditions, misjudgments can easily lead to collisions, misalignments, or even safety accidents. These shortcomings expose systemic bottlenecks in existing bridge erecting equipment when facing the high-precision, high-efficiency, and high-safety demands of modern port construction, including insufficient intelligence, poor operational robustness, and low human-machine collaboration efficiency. Therefore, an intelligent operating system integrating perception, decision-making, control, and visualization is urgently needed. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides an intelligent operating system for beam erecting machines based on the installation of superstructure components of high-pile wharves. By constructing a multi-source heterogeneous sensor fusion system and a task-driven three-dimensional spatial dynamic modeling mechanism, it achieves high-precision collaborative control of prefabricated components of high-pile wharves throughout the entire process of lifting, traveling, lateral movement, and precise positioning.

[0006] The technical solution to achieve the above objectives is: The intelligent operating system for the girder erecting machine based on the installation of superstructure components of high-pile wharves includes: The wharf digital twin modeling module is used to import and parse the building information model data of the high-pile wharf, extract the design installation coordinates, geometric dimensions, center of gravity position, lifting point arrangement and mutual assembly constraint relationship of all prefabricated components to be installed, and form a wharf component installation task map containing spatial topology and process logic sequence. The identity recognition and feature extraction module is used to acquire the point cloud data of the precast component's outline and the surface identification code image before the precast component enters the beam erecting machine's working area. Through point cloud registration algorithm and optical character recognition algorithm, it matches the component's unique identity in the dock component installation task map and outputs the component's actual geometric parameters, mass distribution characteristics and hoisting posture requirements. The multi-degree-of-freedom motion state sensing module is used to collect in real time the posture data, structural deformation and dynamic load distribution information of the entire girder erecting machine in the X-axis traveling direction, Y-axis lateral movement direction, Z-axis lifting direction and rotation direction around each axis; The marine environmental disturbance monitoring module is used to obtain the wind force level, wind direction angle, wave period, wave height amplitude and tidal level change rate of the current operation area in real time; The task planning and path generation module is used to generate a full set of operation path instructions, including initial lifting attitude setting, aerial transport trajectory planning, obstacle avoidance path adjustment strategy and final placement approximation curve, based on the target installation coordinates provided by the dock digital twin modeling module, the component type information output by the identity recognition and feature extraction module, and the real-time environmental parameters fed back by the marine environmental disturbance monitoring module, by calling the pre-stored component installation process knowledge base. The spatiotemporal synchronization and fusion processing module is used to align the heterogeneous data streams collected by the identity recognition and feature extraction module, the multi-degree-of-freedom motion state perception module, and the marine environmental disturbance monitoring module with timestamps based on a unified time base, and to use the Kalman filter algorithm to suppress noise and estimate the state of the pose data, outputting a high-confidence joint state vector of the girder erecting machine-component-environment. The adaptive motion control execution module is used to receive the set of operation path instructions issued by the task planning and path generation module, and combine it with the joint state vector output by the spatiotemporal synchronization and fusion processing module. By solving the composite control law containing feedforward compensation term and feedback correction term, it outputs precise control current signals to the traveling drive motor, lateral servo mechanism, hoisting winch system and lifting device attitude adjustment hydraulic cylinder of the beam erecting machine, respectively, to drive the various execution mechanisms to act in coordination. The installation process visual servo fine-tuning module is used to activate the binocular stereo vision system deployed under the lifting device when the prefabricated component is lowered to a height of 500 mm from the target installation surface. The system captures the relative pose deviation between the pre-embedded sleeve at the bottom of the component and the reserved reinforcing bar on the dock support surface in real time, generates a six-degree-of-freedom deviation correction amount with micron-level resolution, and injects the correction amount into the end control loop of the adaptive motion control execution module to perform sub-millimeter-level dynamic correction in the final placement stage. The safety boundary dynamic verification module is used to calculate the structural safety margin and anti-overturning stability coefficient of the girder erecting machine in real time during the entire operation, based on the structural strain data fed back by the multi-degree-of-freedom motion state perception module, the instantaneous wind and wave load provided by the marine environmental disturbance monitoring module, and the motion trajectory planned by the task planning and path generation module. When any indicator is lower than the preset threshold, the graded early warning mechanism is immediately triggered and deceleration, hovering or emergency retreat operations are automatically executed.

[0007] Preferably, the building information model data imported into the dock digital twin modeling module is in the standard format of industrial basic category, and the model accuracy level is not lower than LOD400. Each prefabricated component contains a unique component code, design and installation coordinate system, allowable installation tolerance range, and assembly interference inspection rules for adjacent components.

[0008] Preferably, in the identity recognition and feature extraction module, the component's outline point cloud data and surface identification code image are obtained by using a laser scanning array and a high-definition industrial camera array deployed at the front of the dock. in, The laser scanning array consists of no less than eight line lasers, with a scanning frequency of no less than 20 Hz per second and a point cloud density of no less than 5,000 points per square meter. The high-definition industrial camera array consists of four global shutter cameras with a resolution of no less than five million pixels, a frame rate of no less than thirty frames per second, and a fixed lens focal length of 16 mm.

[0009] Preferably, the multi-degree-of-freedom motion state sensing module includes: Tilt sensors installed at both ends and in the middle of the main beam are used to monitor the pitch and roll attitude of the main beam in the XY plane. An absolute photoelectric displacement encoder is installed at the bottom of the traveling outriggers to record the X-axis displacement of the girder erecting machine along the dock track in real time. A high-precision displacement encoder is configured on the guide rail of the traverse carriage to measure the lateral displacement of the Y-axis; Strain gauge arrays are arranged at the lower flange of the main beam and the connection nodes between the legs and the main beam to monitor the local stress concentration of the structure under dynamic loads. An inertial measurement unit is installed on the top of the spreader to acquire the linear acceleration and angular velocity of the spreader in space in real time; in, The tilt sensor has a range of ±90 degrees and a resolution of not less than 0.01 degrees. The strain gauge array is arranged in a full-bridge manner at the lower flange of the main beam and the connection node of the support leg, with a sampling frequency of not less than one kilohertz per second. The inertial measurement unit includes a three-axis accelerometer and a three-axis gyroscope. The accelerometer has a range of ±10 times the gravitational acceleration, and the gyroscope has a zero-bias stability better than 0.1 degrees per hour.

[0010] Preferably, the marine environmental disturbance monitoring module includes: wind speed and direction instruments, wave radar and tide gauges deployed on the dock working platform and key parts of the girder erecting machine; The anemometer is used to obtain the wind speed and direction level in the current work area; The wave radar is used to obtain the wind direction angle, wave period, and wave height amplitude of the current working area; The tide gauge is used to obtain the rate of change of tidal water level in the current working area; in, The anemometer has a measurement range of 0 to 70 meters per second and a wind direction resolution of one degree. The wave radar adopts an X-band continuous wave system, with an effective detection range of not less than 500 meters and a wave height measurement accuracy of better than 5%. The tide gauge uses a pressure sensor with a range of 10 meters of water column and an accuracy of 0.2% of the full scale.

[0011] Preferably, the component installation process knowledge base built into the task planning and path generation module is divided into a berthing component sub-library, a crossbeam sub-library, a longitudinal beam sub-library, a track beam sub-library, and a panel sub-library according to the type of prefabricated component. Each sub-library stores the standard hoisting attitude angle range, recommended transportation speed curve, maximum allowable crosswind offset, and typical placement approximation strategy template for the corresponding component.

[0012] Preferably, the spatiotemporal synchronization and fusion processing module uses the IEEE 1588 precision time protocol to achieve nanosecond-level time synchronization of each sensor node. The state vector dimension of the Kalman filter algorithm is twelve-dimensional, including the position, velocity, attitude angle and angular velocity of the beam erecting machine in three-dimensional space. The observation vector is composed of the displacement encoder reading, the output of the inertial measurement unit and the relative pose fed back by the visual servo module.

[0013] Preferably, the adaptive motion control execution module employs a composite control law, in which the feedforward compensation term is calculated based on the target trajectory provided by the dock digital twin modeling module using inverse dynamics, the feedback correction term adopts a sliding mode variable structure control strategy, and the boundary layer thickness of the switching function is dynamically adjusted according to the wind and wave levels fed back by the marine environmental disturbance monitoring module.

[0014] Preferably, in the visual servo fine-tuning module of the installation process, the baseline length of the binocular stereo vision system is 1.5 meters, the working distance range is 0.5-3.0 meters, the depth measurement repeatability error is less than 0.1 millimeters, and the six-degree-of-freedom pose deviation data is output at a frequency of fifty frames per second during the component placement stage.

[0015] Preferably, in the safety boundary dynamic verification module, the preset structural safety margin threshold is 1.5 and the overturning stability coefficient threshold is 1.2; The graded early warning mechanism includes three levels of response: Level 1 provides audible and visual alerts and restricts acceleration commands; Level 2 automatically reduces the operating speed to 50% of the rated value; and Level 3 cuts off the main drive power and activates the mechanical braking device.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention solves the core problems in the existing high-pile wharf prefabricated component installation process, such as reliance on manual experience judgment, fragmentation of multi-source information, weak ability to suppress environmental disturbances, and difficulty in achieving stable millimeter-level positioning accuracy, by constructing an intelligent operating system covering the entire chain of "digital twin-sensory fusion-intelligent decision-making-precise execution-safe closed loop". This invention uses the wharf BIM model as a unified data source, achieving seamless integration from design intent to on-site execution; through the spatiotemporal synchronization and fusion of multi-source heterogeneous sensor data, it significantly improves the robustness of the girder erecting machine's state perception under severe sea conditions such as strong winds and swells; the process knowledge base embedded in the task planning module ensures the professionalization and standardization of installation strategies for different types of components; the dual closed-loop mechanism of adaptive motion control and visual servo fine-tuning enables large components weighing tens of tons to achieve sub-millimeter-level dynamic correction capability in the final placement stage; and the dynamic verification module of safety boundaries fundamentally guarantees the operational safety of large special equipment in complex marine environments. Overall, this invention improves the installation efficiency of prefabricated components for high-pile wharves by more than 40%, increases the first-time installation qualification rate to 99.8%, significantly reduces rework costs and safety risks, and provides key technical support for the construction of modern smart ports. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a module diagram of the intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf, as per the present invention. Figure 2 This is a schematic diagram of the multi-level interaction relationship and data flow between the girder erecting machine, precast components, and marine environment in this invention; Figure 3 This is a logical flowchart of the process of identifying prefabricated components, extracting features, and matching the digital twin task map of the dock in this invention. Figure 4 This is a closed-loop collaborative control logic framework diagram of task planning and adaptive motion control execution in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 , 2 As shown, the intelligent operating system for the girder erecting machine based on the installation of the superstructure components of the high-pile wharf includes: wharf digital twin modeling module 1, identity recognition and feature extraction module 2, multi-degree-of-freedom motion state perception module 3, marine environmental disturbance monitoring module 4, task planning and path generation module 5, spatiotemporal synchronization and fusion processing module 6, adaptive motion control execution module 7, installation process visual servo fine-tuning module 8, and safety boundary dynamic verification module 9.

[0020] The wharf digital twin modeling module 1 is used to import and parse the building information model data of the high-pile wharf, extract the design installation coordinates, geometric dimensions, center of gravity position, lifting point arrangement and mutual assembly constraints of all prefabricated components to be installed, and form a wharf component installation task map that includes spatial topology and process logic sequence.

[0021] In this embodiment, the imported building information model data is in the standard format of industrial basics, and the model accuracy level is not lower than LOD400. Each prefabricated component contains a unique component code, design and installation coordinate system, allowable installation tolerance range, and assembly interference check rules for adjacent components. All components are assigned a unique component code as a unique identifier for subsequent identity matching and task scheduling. At the same time, the installation tolerance zone in the model is defined in the form of a six-degree-of-freedom tolerance domain in three-dimensional space, covering the tolerance of the three translational directions (X, Y, and Z) and the rotation angle around the three axes. In addition, the system extracts the spatial topological connection relationship between components from the model to form a dock component installation task map with a process logic sequence. This task map is stored in a directed acyclic graph structure, where nodes represent individual prefabricated components and edges represent assembly dependencies, ensuring that the construction logic of installing the support structure first and then the upper panel is strictly executed.

[0022] The identity recognition and feature extraction module 2 is used to acquire the point cloud data of the precast component's outline and the surface identification code image before the precast component enters the beam erecting machine's working area. After the acquired point cloud data is processed by denoising, filtering and downsampling, it is matched with the unique identity identifier in the dock component installation task map by the point cloud registration algorithm and optical character recognition algorithm, and the actual geometric parameters, mass distribution characteristics and hoisting posture requirements of the component are output.

[0023] like Figure 3 As shown, the point cloud data of the component's outline is obtained by a laser scanning array deployed at the front of the wharf. After denoising, filtering and downsampling, the obtained point cloud data is input into the point cloud registration algorithm. The point cloud registration algorithm adopts an improved iterative nearest point method. Using the theoretical point cloud of the corresponding component type in the wharf digital twin modeling module as a template, the algorithm calculates the optimal rigid body transformation matrix between the current measured point cloud and the template point cloud, thereby determining the actual spatial attitude and geometric deviation of the component. A high-definition industrial camera array captures images of QR codes or barcodes on the surface of components from different perspectives. After being decoded by an optical character recognition algorithm, the unique identification code of the component is obtained. The system matches the code with the dock component installation task map to confirm its target installation location and component type. After a successful match, the system outputs the actual geometric parameters of the component (including length, width, height, and cross-sectional profile), mass distribution characteristics (estimated by combining point cloud volume integral with preset concrete density), and hoisting posture requirements (e.g., the crossbeam needs to be hoisted horizontally, and the berthing component needs to be tilted at a certain angle to avoid the pile cap). in, The laser scanning array consists of no fewer than eight line lasers, with a scanning frequency of no less than 20 Hz per second and a point cloud density of no less than 5,000 points per square meter. The high-definition industrial camera array consists of four global shutter cameras with a resolution of no less than five million pixels, a frame rate of no less than thirty frames per second, and a fixed lens focal length of 16 mm.

[0024] The multi-degree-of-freedom motion state sensing module 3 is used to collect in real time the position and posture data, structural deformation and dynamic load distribution information of the entire girder erecting machine in the X-axis traveling direction, Y-axis lateral movement direction, Z-axis lifting direction and rotation direction around each axis.

[0025] In this embodiment, the multi-degree-of-freedom motion state sensing module 3 includes: Tilt sensors installed at both ends and in the middle of the main beam are used to monitor the pitch and roll attitude of the main beam in the XY plane. An absolute photoelectric displacement encoder is installed at the bottom of the traveling outriggers to record the X-axis displacement of the girder erecting machine along the dock track in real time. A high-precision displacement encoder is configured on the guide rail of the traverse carriage to measure the lateral displacement of the Y-axis; Strain gauge arrays are arranged at the lower flange of the main beam and the connection nodes between the legs and the main beam to monitor the local stress concentration of the structure under dynamic loads. An inertial measurement unit is installed on the top of the spreader to acquire the linear acceleration and angular velocity of the spreader in space in real time; All the sensor data mentioned above are uploaded to the central processing unit in real time via industrial Ethernet, forming a complete positional description of the entire girder erecting machine in the three translational degrees of freedom (X, Y, and Z) and the three rotational degrees of freedom (around three axes). in, The tilt sensor has a range of ±90 degrees and a resolution of no less than 0.01 degrees. The strain gauge array is arranged in a full-bridge manner at the lower flange of the main beam and the connection nodes of the legs, with a sampling frequency of not less than one kilohertz per second. The inertial measurement unit includes a three-axis accelerometer and a three-axis gyroscope. The accelerometer has a range of ±10 times the gravitational acceleration, and the gyroscope has zero-bias stability better than 0.1 degrees per hour.

[0026] The marine environmental disturbance monitoring module 4 is used to obtain the wind force level, wind direction angle, wave period, wave height amplitude and tidal water level change rate of the current operation area in real time.

[0027] In this embodiment, the marine environmental disturbance monitoring module 4 includes: wind speed and direction instruments, wave radar and tide gauges deployed on the dock working platform and key parts of the girder erecting machine; An anemometer is used to obtain the wind speed and direction level in the current work area; Wave radar is used to obtain the wind direction angle, wave period, and wave height amplitude of the current working area; Tide gauges are used to obtain the rate of change of tidal water level in the current working area; in, The anemometer has a measurement range of 0 to 70 meters per second, a wind direction resolution of one degree, and outputs instantaneous wind speed and direction data once per second. The wave radar uses an X-band continuous wave system, with an effective detection range of no less than 500 meters and a wave height measurement accuracy of better than 5%. The tide gauge uses a pressure sensor with a range of ten meters of water column and an accuracy of two-thousandths of the full scale.

[0028] The task planning and path generation module 5 is used to generate a full set of operation path instructions, including the initial lifting attitude setting, aerial transport trajectory planning, obstacle avoidance path adjustment strategy, and final placement approximation curve, based on the target installation coordinates provided by the dock digital twin modeling module 1, the component type information output by the identity recognition and feature extraction module 2, and the real-time environmental parameters fed back by the marine environmental disturbance monitoring module 4.

[0029] In this embodiment, the component installation process knowledge base built into the task planning and path generation module 5 is divided into sub-libraries for berthing components, crossbeams, longitudinal beams, track beams, and panels according to the type of prefabricated components. Each sub-library stores the standard lifting attitude angle range for the corresponding component (for example, for crossbeams, the lifting device must be kept horizontal and the attitude angle deviation must not exceed 0.5 degrees), recommended transportation speed curve (usually an S-shaped acceleration and deceleration curve, with a maximum operating speed not exceeding 15 meters per minute), maximum allowable crosswind offset (calculated based on the component's windward area and center of gravity height), and typical placement approximation strategy templates (such as using a uniform descent of 1 meter per minute for the last 50 centimeters). The system combines the target installation coordinates, current marine environmental parameters (especially wind speed and wave height), and the current position and attitude of the girder erector, and uses an AI search algorithm to plan a collision-free, low-energy, and disturbance-resistant aerial transport trajectory in three-dimensional space. This trajectory is represented by a series of discrete path points, each containing X, Y, and Z coordinates and the corresponding spreader attitude angle. At the same time, the system predicts the vertical displacement of the dock platform within the next ten seconds based on the wave cycle and dynamically adjusts the placement timing to ensure that the components are finally docked when the dock platform is in a wave trough or a stable phase.

[0030] The spatiotemporal synchronization and fusion processing module 6 is used to align the heterogeneous data streams collected by the identity recognition and feature extraction module 2, the multi-degree-of-freedom motion state perception module 3, and the marine environmental disturbance monitoring module 4 with timestamps based on a unified time base, and to use the Kalman filter algorithm to suppress noise and estimate the state of the pose data, outputting a high-confidence joint state vector of the girder erecting machine-component-environment.

[0031] In this embodiment, the spatiotemporal synchronization and fusion processing module 6 uses the IEEE 1588 precision time protocol to achieve nanosecond-level time synchronization of each sensor node. The state vector dimension of the Kalman filter algorithm is twelve-dimensional, including the position, velocity, attitude angle and angular velocity of the girder erecting machine in three-dimensional space. The observation vector is composed of displacement encoder readings (providing direct measurement of x, y and z), inertial measurement unit output (providing integral information of acceleration and angular velocity), and relative pose feedback from the visual servo module (providing high-precision correction during the fine-tuning stage). The Kalman filter continuously suppresses sensor noise and compensates for dynamic delay through a prediction-update loop, outputting a high-confidence joint state vector. This vector serves as the core input to the subsequent control module, with an update frequency of 100 Hz.

[0032] The adaptive motion control execution module 7 is used to receive the set of operation path instructions issued by the task planning and path generation module 5, and combine it with the joint state vector output by the spatiotemporal synchronization and fusion processing module 6. By solving the composite control law containing feedforward compensation terms and feedback correction terms, it outputs precise control current signals to the traveling drive motor, lateral servo mechanism, hoisting winch system and lifting device attitude adjustment hydraulic cylinder of the beam erecting machine, respectively, to drive the coordinated action of each execution mechanism.

[0033] like Figure 4 As shown, the adaptive motion control execution module 7 employs a composite control law. Its feedforward compensation term is based on the target trajectory provided by the dock digital twin modeling module 1, which performs inverse dynamics calculation to calculate the theoretical driving torque required by each actuator under ideal conditions without disturbance. The feedback correction term adopts a sliding mode variable structure control strategy. The boundary layer thickness of the switching function is dynamically adjusted according to the wind and wave level fed back by the marine environmental disturbance monitoring module 4: when the wind speed exceeds 10 meters per second or the wave height exceeds 0.5 meters, the boundary layer thickness automatically increases to reduce control signal chattering and improve system robustness. The calculated control quantities are converted into the PWM duty cycle of the travel drive motor, the position command of the lateral servo mechanism, the speed setpoint of the hoisting system, and the valve core opening command of the hydraulic cylinder for adjusting the attitude of the spreader. All control signals are output to the bottom actuator via the fieldbus at a frequency of 200 Hz per second to drive the coordinated motion of each degree of freedom of the girder erecting machine.

[0034] The installation process visual servo fine-tuning module 8 is used to activate the binocular stereo vision system deployed under the lifting device when the prefabricated component is lowered to a height of 500 mm from the target installation surface. The system captures the relative pose deviation between the pre-embedded sleeve at the bottom of the component and the reserved reinforcing bar on the dock support surface in real time, generates a six-degree-of-freedom deviation correction amount with micron-level resolution, and injects the correction amount into the end control loop of the adaptive motion control execution module for sub-millimeter-level dynamic correction during the final placement stage. For example, when a deviation of 0.8 mm is detected in the X direction, the system immediately sends a reverse fine-tuning command to the lateral servo mechanism, so that the component moves laterally by 0.8 mm while continuing to descend, ensuring that the sleeve and reinforcing bar are precisely aligned. This fine-tuning process continues until the component is fully placed and the contact force sensor detects the support reaction force.

[0035] In this embodiment, the binocular stereo vision system has a baseline length of 1.5 meters, a working distance range of 0.5-3.0 meters, a depth measurement repeatability error of less than 0.1 millimeters, and outputs six-degree-of-freedom pose deviation data at a frequency of fifty frames per second during the component placement stage.

[0036] The safety boundary dynamic verification module 9 is used to calculate the structural safety margin and anti-overturning stability coefficient of the girder erecting machine in real time during the entire operation, based on the structural strain data fed back by the multi-degree-of-freedom motion state perception module 3, the instantaneous wind and wave load provided by the marine environmental disturbance monitoring module 4, and the motion trajectory planned by the task planning and path generation module 5. When any indicator is lower than the preset threshold, the graded early warning mechanism is immediately triggered and deceleration, hovering or emergency retreat operations are automatically executed.

[0037] In this embodiment, the preset structural safety margin threshold is 1.5, and the overturning stability coefficient threshold is 1.2. The graded early warning mechanism includes three levels of response: Level 1 is an audible and visual warning and a restriction of acceleration commands; Level 2 is to automatically reduce the operating speed to 50% of the rated value; and Level 3 is to cut off the main drive power and activate the mechanical braking device. When any indicator falls below the threshold, the tiered early warning mechanism is immediately triggered: if the indicator is slightly below the threshold (the decrease is less than 10%), the system executes a Level 1 response, issuing an audible and visual warning and prohibiting any acceleration commands; if the indicator continues to drop below 80% of the threshold, a Level 2 response is executed, automatically limiting all operating speeds to 50% of the rated value; if the indicator deteriorates rapidly or a sudden strong gust occurs (wind speed increases by more than 15 meters per second), a Level 3 response is executed, immediately cutting off the main drive power and activating the mechanical braking device to put the girder erector into an emergency hovering state until environmental conditions return to a safe range.

[0038] The aforementioned modules work collaboratively to form a complete closed loop, encompassing digital model import, component identification, environmental perception, path planning, state fusion, adaptive control, and safety verification. The entire operation requires no manual intervention; all decisions are made automatically based on real-time data and preset rules. Before each installation operation, the system automatically loads the corresponding component's process template, dynamically integrates multi-source sensing information during the operation, introduces visual servoing at the final stage to achieve sub-millimeter-level fine-tuning, and implements three levels of safety protection throughout the process, ensuring that prefabricated components weighing tens of tons can be installed stably, efficiently, and safely with millimeter-level precision under complex sea conditions.

[0039] The intelligent operation method of the girder erecting machine based on the installation of precast components of the superstructure of the high-pile wharf includes the following steps: Step S1: Import and parse the building information model data of the high-pile wharf, extract the design installation coordinates, geometric dimensions, center of gravity position, lifting point arrangement and mutual assembly constraints of all prefabricated components to be installed, and form a wharf component installation task map containing spatial topology and process logic sequence; Step S2: Before the prefabricated component enters the beam erecting machine's working area, acquire the component's outline point cloud data and surface identification code image through a laser scanning array and a high-definition industrial camera array, match its unique identity in the wharf component installation task map using a point cloud registration algorithm and an optical character recognition algorithm, and output the component's actual geometric parameters, mass distribution characteristics and lifting posture requirements; Step S3: Through tilt sensors, displacement encoders, strain gauge arrays and inertial measurement units installed on the main beam, outriggers, lifting trolley and lifting gear, collect in real time the beam erecting machine's posture data, structural deformation and dynamic load distribution information in the X-axis traveling direction, Y-axis lateral direction, Z-axis lifting direction and rotational directions around each axis; Step S4: By deploying anemometers, wave radars, and tide gauges on the dock work platform and key parts of the girder erecting machine, the wind force level, wind direction angle, wave period, wave height amplitude, and tidal level change rate of the current working area are obtained in real time; Step S5, based on the identification of the prefabricated components, target installation coordinates, and real-time environmental parameters, a pre-stored component installation process knowledge base is called to generate a full-process operation path instruction set including initial lifting attitude setting, aerial transport trajectory planning, obstacle avoidance path adjustment strategy, and final placement approximation curve; Step S6, the girder erecting machine motion status data and marine environmental data are processed. The environmental disturbance data and precast component feature data are timestamped according to a unified time base, and the Kalman filter algorithm is used to suppress noise and estimate the state of the pose data, outputting a high-confidence joint state vector of the girder erecting machine-component-environment; in step S7, the operation path instruction set is received, and combined with the joint state vector, the composite control law containing feedforward compensation term and feedback correction term is solved, and precise control current signals are output to the girder erecting machine's travel drive motor, lateral servo mechanism, hoisting winch system and lifting device attitude adjustment hydraulic cylinder respectively, driving each actuator to act in coordination; Step S8: When the precast component descends to a height of 500 mm from the target installation surface, the binocular stereo vision system is activated to capture the relative pose deviation between the pre-embedded sleeve at the bottom of the component and the reserved reinforcing bar on the dock support surface in real time. A six-degree-of-freedom deviation correction amount with micron-level resolution is generated and injected into the end control loop to achieve sub-millimeter-level dynamic correction in the final placement stage. Step S9: Throughout the operation, based on structural strain data, instantaneous wind and wave loads, and planned motion trajectory, the structural safety margin and overturning stability coefficient of the girder erecting machine under the current working condition are calculated in real time. When any indicator is lower than the preset threshold, the graded early warning mechanism is immediately triggered and deceleration, hovering, or emergency retraction operations are automatically executed.

[0040] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent operating system for beam erecting machines based on the installation of superstructure components of high-pile wharves, characterized in that, include: The wharf digital twin modeling module is used to import and parse the building information model data of the high-pile wharf, extract the design installation coordinates, geometric dimensions, center of gravity position, lifting point arrangement and mutual assembly constraint relationship of all prefabricated components to be installed, and form a wharf component installation task map containing spatial topology and process logic sequence. The identity recognition and feature extraction module is used to acquire the point cloud data of the precast component's outline and the surface identification code image before the precast component enters the beam erecting machine's working area. Through point cloud registration algorithm and optical character recognition algorithm, it matches the component's unique identity in the dock component installation task map and outputs the component's actual geometric parameters, mass distribution characteristics and hoisting posture requirements. The multi-degree-of-freedom motion state sensing module is used to collect in real time the posture data, structural deformation and dynamic load distribution information of the entire girder erecting machine in the X-axis traveling direction, Y-axis lateral movement direction, Z-axis lifting direction and rotation direction around each axis; The marine environmental disturbance monitoring module is used to obtain the wind force level, wind direction angle, wave period, wave height amplitude and tidal level change rate of the current operation area in real time; The task planning and path generation module is used to generate a full set of operation path instructions, including initial lifting attitude setting, aerial transport trajectory planning, obstacle avoidance path adjustment strategy and final placement approximation curve, based on the target installation coordinates provided by the dock digital twin modeling module, the component type information output by the identity recognition and feature extraction module, and the real-time environmental parameters fed back by the marine environmental disturbance monitoring module, by calling the pre-stored component installation process knowledge base. The spatiotemporal synchronization and fusion processing module is used to align the heterogeneous data streams collected by the identity recognition and feature extraction module, the multi-degree-of-freedom motion state perception module, and the marine environmental disturbance monitoring module with timestamps based on a unified time base, and to use the Kalman filter algorithm to suppress noise and estimate the state of the pose data, outputting a high-confidence joint state vector of the girder erecting machine-component-environment. The adaptive motion control execution module is used to receive the set of operation path instructions issued by the task planning and path generation module, and combine it with the joint state vector output by the spatiotemporal synchronization and fusion processing module. By solving the composite control law containing feedforward compensation term and feedback correction term, it outputs precise control current signals to the traveling drive motor, lateral servo mechanism, hoisting winch system and lifting device attitude adjustment hydraulic cylinder of the beam erecting machine, respectively, to drive the various execution mechanisms to act in coordination. The installation process visual servo fine-tuning module is used to activate the binocular stereo vision system deployed under the lifting device when the prefabricated component is lowered to a height of 500 mm from the target installation surface. The system captures the relative pose deviation between the pre-embedded sleeve at the bottom of the component and the reserved reinforcing bar on the dock support surface in real time, generates a six-degree-of-freedom deviation correction amount with micron-level resolution, and injects the correction amount into the end control loop of the adaptive motion control execution module to perform sub-millimeter-level dynamic correction in the final placement stage. The safety boundary dynamic verification module is used to calculate the structural safety margin and anti-overturning stability coefficient of the girder erecting machine in real time during the entire operation, based on the structural strain data fed back by the multi-degree-of-freedom motion state perception module, the instantaneous wind and wave load provided by the marine environmental disturbance monitoring module, and the motion trajectory planned by the task planning and path generation module. When any indicator is lower than the preset threshold, the graded early warning mechanism is immediately triggered and deceleration, hovering or emergency retreat operations are automatically executed.

2. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The building information model data imported into the dock digital twin modeling module is in the standard format of industrial basic category, and the model accuracy level is not lower than LOD400. Each prefabricated component contains a unique component code, design and installation coordinate system, allowable installation tolerance range, and assembly interference inspection rules for adjacent components.

3. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, In the identity recognition and feature extraction module, the component's outline point cloud data and surface identification code image are obtained by using a laser scanning array and a high-definition industrial camera array deployed at the front of the dock. in, The laser scanning array consists of no less than eight line lasers, with a scanning frequency of no less than 20 Hz per second and a point cloud density of no less than 5,000 points per square meter. The high-definition industrial camera array consists of four global shutter cameras with a resolution of no less than five million pixels, a frame rate of no less than thirty frames per second, and a fixed lens focal length of 16 mm.

4. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The multi-degree-of-freedom motion state sensing module includes: Tilt sensors installed at both ends and in the middle of the main beam are used to monitor the pitch and roll attitude of the main beam in the XY plane. An absolute photoelectric displacement encoder is installed at the bottom of the traveling outriggers to record the X-axis displacement of the girder erecting machine along the dock track in real time. A high-precision displacement encoder is configured on the guide rail of the traverse carriage to measure the lateral displacement of the Y-axis; Strain gauge arrays are arranged at the lower flange of the main beam and the connection nodes between the legs and the main beam to monitor the local stress concentration of the structure under dynamic loads. An inertial measurement unit is installed on the top of the spreader to acquire the linear acceleration and angular velocity of the spreader in space in real time; in, The tilt sensor has a range of ±90 degrees and a resolution of not less than 0.01 degrees. The strain gauge array is arranged in a full-bridge manner at the lower flange of the main beam and the connection node of the support leg, with a sampling frequency of not less than one kilohertz per second. The inertial measurement unit includes a three-axis accelerometer and a three-axis gyroscope. The accelerometer has a range of ±10 times the gravitational acceleration, and the gyroscope has a zero-bias stability better than 0.1 degrees per hour.

5. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The marine environmental disturbance monitoring module includes: wind speed and direction instruments, wave radar and tide gauges deployed on the dock working platform and key parts of the girder erecting machine. The anemometer is used to obtain the wind speed and direction level in the current work area; The wave radar is used to obtain the wind direction angle, wave period, and wave height amplitude of the current working area; The tide gauge is used to obtain the rate of change of tidal water level in the current working area; in, The anemometer has a measurement range of 0 to 70 meters per second and a wind direction resolution of one degree. The wave radar adopts an X-band continuous wave system, with an effective detection range of not less than 500 meters and a wave height measurement accuracy of better than 5%. The tide gauge uses a pressure sensor with a range of ten meters of water column and an accuracy of two-thousandths of the full scale.

6. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The component installation process knowledge base built into the task planning and path generation module is divided into sub-libraries for berthing components, crossbeams, longitudinal beams, track beams, and panels according to the type of prefabricated components. Each sub-library stores the standard hoisting attitude angle range, recommended transportation speed curve, maximum allowable crosswind offset, and typical placement approximation strategy template for the corresponding component.

7. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The spatiotemporal synchronization and fusion processing module uses the IEEE 1588 precision time protocol to achieve nanosecond-level time synchronization of each sensor node. The state vector dimension of the Kalman filter algorithm is twelve-dimensional, including the position, velocity, attitude angle and angular velocity of the beam erecting machine in three-dimensional space. The observation vector is composed of the displacement encoder reading, the output of the inertial measurement unit and the relative pose fed back by the visual servo module.

8. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, The adaptive motion control execution module employs a composite control law, in which the feedforward compensation term is calculated based on the target trajectory provided by the dock digital twin modeling module using inverse dynamics, the feedback correction term adopts a sliding mode variable structure control strategy, and the boundary layer thickness of the switching function is dynamically adjusted according to the wind and wave levels fed back by the marine environmental disturbance monitoring module.

9. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf as described in claim 1, characterized in that, In the visual servo fine-tuning module of the installation process, the baseline length of the binocular stereo vision system is 1.5 meters, the working distance range is 0.5-3.0 meters, the depth measurement repeatability error is less than 0.1 mm, and the six-degree-of-freedom pose deviation data is output at a frequency of fifty frames per second during the component placement stage.

10. The intelligent operating system for the girder erecting machine based on the installation of superstructure components of a high-pile wharf according to claim 1, characterized in that, In the aforementioned safety boundary dynamic verification module, the preset structural safety margin threshold is 1.5, and the overturning stability coefficient threshold is 1.

2. The graded early warning mechanism includes three levels of response: Level 1 provides audible and visual alerts and restricts acceleration commands; Level 2 automatically reduces the operating speed to 50% of the rated value; and Level 3 cuts off the main drive power and activates the mechanical braking device.

Images