Adaptive anti-overturning control method and device of a bridge erecting machine

Abstract

The application relates to the technical field of bridge construction equipment control, and discloses a self-adaptive anti-overturning control method and equipment for a bridge erecting machine, which comprises the following steps: collecting structural vibration response data and environmental wind load data in a hole passing operation of the bridge erecting machine in real time; adopting a modal parameter identification algorithm in online operation to identify current modal parameters of a main beam in real time based on the data; determining whether a trigger condition of active regulation is met according to the current modal parameters; when the trigger condition is met, calculating an anti-overturning compensation amount according to the environmental wind load data and the current modal parameters; generating a cooperative control instruction according to the compensation amount, and respectively sending the cooperative control instruction to a crown block moving mechanism and a supporting leg hydraulic damping adjusting mechanism to execute cooperative actions; and continuously acquiring feedback signals for dynamic correction in the execution process until the trigger condition is no longer met. The application can realize closed-loop self-adaptive anti-overturning control in the whole hole passing operation process, and is suitable for safe construction of the bridge erecting machine in a complex wind field environment.

Description

Technical Field

[0001] This invention relates to the field of bridge construction equipment control technology, specifically to an adaptive anti-overturning control method and equipment for a bridge erecting machine. Background Technology

[0002] As a key large-scale piece of equipment in bridge construction, the overturning risk control during bridge crossing operations directly affects construction safety and project progress. Due to the long structure, heavy weight and load of the bridge erecting machine, and complex operating environment, it is susceptible to multiple factors such as wind load and changes in outrigger reaction force during bridge crossing operations, leading to frequent overturning accidents. CN117196320B proposes a method for assessing the overturning risk of bridge erecting machines during bridge crossings. By deploying laser rangefinders to obtain the real-time interval distances between each outrigger and the end of the guide beam during the bridge erecting machine's crossing, the theoretical and actual overturning resistance coefficients are obtained. This leads to the overturning influence coefficient, which is used as an influencing factor to obtain the critical overturning resistance coefficient under critical overturning conditions. This allows for the assessment of the risk status of the bridge erecting machine during bridge crossings and real-time adjustments to ensure crossing safety. This method features a simple and easy-to-install device that enables real-time automatic monitoring of stability, real-time accurate risk assessment, and early warning. The accuracy of the assessment is improved through influence coefficient correction.

[0003] Existing bridge erecting machine evaluation methods have the following problems: the overturning resistance coefficient calculation is based on a static mechanical model, which is difficult to cope with the time-varying characteristics of wind loads and the continuous changes in structural response; the early warning mechanism can only trigger an alarm or adjustment when the actual overturning resistance coefficient is lower than the theoretical value, and lacks an active identification and judgment mechanism for structural risk status.

[0004] Based on the above situation, there is an urgent need for an adaptive anti-overturning control method and equipment for bridge erecting machines to solve the problems of not being able to achieve real-time control and lacking an active identification and judgment mechanism for structural risk status. Summary of the Invention

[0005] The purpose of this invention is to address the problems of existing bridge erecting machine evaluation methods, which calculate the overturning resistance coefficient based on a static mechanical model, making it difficult to cope with the time-varying characteristics of wind loads and the continuous changes in structural response; and the early warning mechanism can only trigger an alarm or adjustment when the actual overturning resistance coefficient is lower than the theoretical value, lacking an active identification and judgment mechanism for structural risk status. This invention solves the problems of not being able to achieve real-time control and lacking an active identification and judgment mechanism for structural risk status.

[0006] The technical solution of the present invention is as follows: On the one hand, an adaptive anti-overturning control method for bridge erecting machines is provided, including: Dynamic data acquisition: Real-time acquisition of structural vibration response data and environmental wind load data of the bridge erecting machine during the span crossing operation; Real-time modal parameter identification: Based on the structural vibration response data, an online modal parameter identification algorithm is used to identify the current modal parameters of the main beam of the bridge erecting machine in real time; Triggering condition determination: Based on the current modal parameters, determine whether the triggering conditions for active control are met; Overturning compensation calculation: When the triggering condition is met, the overturning compensation is calculated based on the environmental wind load data and the current modal parameters; Coordinated control and closed-loop correction: Coordinated control commands are generated based on the anti-overturning compensation amount and sent to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. The anti-overturning compensation amount includes compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. During the operation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, the current structural vibration response data is continuously acquired as a feedback signal, and the anti-overturning compensation amount is dynamically corrected based on the feedback signal until the triggering condition is no longer met.

[0007] This approach uses modal parameters as the basis for triggering active control, changing the passive response mode of traditional schemes that only triggers alarms when the overturning coefficient drops below a threshold. It advances the control intervention point to the early risk warning stage. By decomposing the overturning compensation into compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, it achieves the synergistic effect of center of gravity position control and vibration energy dissipation. During execution, it introduces structural vibration response as a feedback signal for dynamic correction until the triggering condition is eliminated, thus forming a closed-loop control mechanism with adaptive capabilities. This improves the timeliness and reliability of overturning control during bridge erection machine crossing operations and solves the problems of not being able to achieve real-time control and lacking an active identification and judgment mechanism for structural risk status.

[0008] Furthermore, this solution does not exclusively limit the specific method for determining the triggering condition. One feasible solution is that determining whether the triggering condition for active control is met includes: The current modal parameters are compared with a preset dynamic feature library, which stores multiple modal parameter patterns and corresponding risk state types. If the current modal parameter matches any risk state type in the dynamic feature library, then the triggering condition for the active control is determined to be met. The risk state type includes at least one or more of the following: vortex-induced resonance precursor state and flutter critical state.

[0009] When this scheme is adopted, it can achieve accurate matching based on the risk mode patterns accumulated from historical or simulation data, effectively reducing the probability of misjudgment and omission. It identifies risk state types with different physical mechanisms, enabling the control system to adopt differentiated control strategies according to the specific risk induction mechanism, further improving the pertinence and effectiveness of active control and ensuring structural safety.

[0010] Furthermore, this solution does not exclusively limit the specific types of the structural vibration response data and environmental wind load data. One feasible solution is that the structural vibration response data at least includes the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam of the bridge erecting machine. The multi-degree-of-freedom dynamic acceleration and dynamic angular velocity are input into the real-time modal parameter identification step to construct the system state matrix and output matrix required by the online modal parameter identification algorithm. The system state matrix represents the distribution of inertial forces on each degree of freedom of the main beam, and the output matrix represents the distribution of acceleration and angular velocity responses on each degree of freedom of the main beam, so as to separate the dominant vibration frequency, damping ratio and mode shape of the main beam under the current wind load from the dynamic acceleration and dynamic angular velocity. The environmental wind load data includes at least real-time wind speed and real-time wind direction; The real-time wind speed and real-time wind direction, together with the current modal parameters, are input into the overturning compensation calculation step. This step is used to calculate the dynamic bending moment increment generated by the current wind load acting on the main beam based on the structural dynamics equations. Combined with the damping ratio in the current modal parameters, which characterizes the vibration energy dissipation capacity, the overturning compensation required to offset the dynamic bending moment increment is calculated.

[0011] This approach comprehensively characterizes the translational and rotational responses of the main beam through the joint acquisition of multi-degree-of-freedom dynamic acceleration and angular velocity, providing more complete structural dynamics information for modal parameter identification. The construction of the system state matrix and output matrix gives the extraction of modal parameters clear physical meaning and mathematical basis. Furthermore, combining real-time wind speed and direction with the damping ratio from the modal parameters for dynamic bending moment increment calculation fully considers the time-varying characteristics of wind loads and the structure's own energy dissipation capacity, making the determination of compensation amounts more accurate.

[0012] Furthermore, this solution does not exclusively limit the specific structure of the modal parameters. One feasible solution is that the modal parameters include at least the dominant vibration frequency, damping ratio, and mode shape; the dominant vibration frequency, damping ratio, and mode shape together serve as the basis for determining whether the active control triggering condition is met in the triggering condition determination. The online modal parameter identification algorithm is an online random subspace identification algorithm. It uses recursive QR decomposition or sliding window technology to realize real-time processing of the dynamic acceleration and dynamic angular velocity data, so as to update the modal parameters online.

[0013] This approach can separately reflect the structure's stiffness characteristics, energy dissipation capacity, and vibration spatial distribution, forming a complete description system of the structure's dynamic characteristics. The online random subspace recognition algorithm, combined with recursive QR decomposition or sliding window technology, significantly reduces the data storage requirements and computational complexity of online computing while ensuring recognition accuracy. This allows for the engineering implementation of real-time modal parameter recognition on edge computing controllers. The recursive algorithm only incrementally updates existing recognition results when new data arrives, eliminating the need to repeatedly process all historical data and improving computational efficiency.

[0014] Furthermore, in order to improve adaptability in complex wind field environments, one feasible solution is as follows: the calculation of the anti-overturning compensation amount is specifically a feedforward calculation of the anti-overturning compensation amount, that is, the instantaneous value and trend of wind load are calculated based on the real-time wind speed and real-time wind direction at the current moment, combined with the dominant vibration frequency and damping ratio in the current modal parameters, the dynamic bending moment increment of the main beam at future moments is predicted, and the anti-overturning compensation amount required to offset the dynamic bending moment increment is calculated in advance based on the prediction results.

[0015] This approach calculates the instantaneous value and trend of wind load based on real-time wind speed and direction, and predicts the dynamic bending moment increment of the main beam in the future by combining the dominant vibration frequency and damping ratio, thereby pre-calculating the compensation amount. This allows for early initiation of control actions before the wind load change trend fully impacts the structure, effectively overcoming the inherent lag of pure feedback control. Simultaneously, by combining the structure's own frequency and damping characteristics for dynamic bending moment increment prediction, the prediction results more closely match the actual dynamic response behavior of the structure. Addressing the inherent randomness and time-varying nature of wind loads, the feedforward calculation method effectively handles disturbances caused by fluctuating wind speed components and instantaneous changes in wind direction, enhancing the adaptability of the control system in complex wind field environments.

[0016] Furthermore, to enhance the overturning resistance of the bridge erecting machine, one feasible solution is as follows: the generation of coordinated control commands, which are then sent to the overhead crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively, includes: The anti-overturning compensation amount is decomposed into a crane displacement compensation component and a hydraulic damping adjustment component. Control commands for the crane moving mechanism are generated based on the crane displacement compensation components. Control commands for the outrigger hydraulic damping adjustment mechanism are generated based on the hydraulic damping adjustment component.

[0017] This scheme utilizes gantry displacement compensation to alter the center of gravity, thereby counteracting the overturning moment. Furthermore, it employs hydraulic damping adjustment of the outriggers to modify energy dissipation, thus suppressing vibration amplitude, resulting in complementary and coordinated control. The two components correspond to displacement and damping, respectively. Simultaneous control of both gantry displacement and damping adjustment further shortens the control response time and enhances the overturning resistance of the bridge erecting machine.

[0018] Furthermore, in order to further improve the control effect, one feasible solution is: the crane displacement compensation component and the hydraulic damping adjustment component are allocated according to a preset coordination weighting coefficient; the coordination weighting coefficient is used to adjust the degree of participation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism when performing coordinated control. The collaborative weighting coefficient is dynamically adjusted according to the currently identified modal parameters: when the modal parameters are in a preset low-frequency range, the weight of the crane displacement compensation component is increased to suppress low-frequency swaying through center of gravity adjustment; when the modal parameters are in a preset high-frequency range, the weight of the hydraulic damping adjustment component is increased to dissipate high-frequency vibration energy through active damping.

[0019] This scheme increases the crane displacement compensation weight in the low-frequency range to suppress low-frequency swaying, and increases the hydraulic damping adjustment weight in the high-frequency range to dissipate high-frequency vibration energy. It allows for optimized configuration based on the structure's current vibration characteristics, effectively avoiding the pitfalls of minimal damping adjustment at low frequencies and slow response of center-of-gravity adjustment at high frequencies under fixed weights. The strategy of emphasizing crane displacement compensation in the low-frequency range utilizes the large stroke adjustment capability of center-of-gravity adjustment for low-frequency swaying, while the strategy of emphasizing hydraulic damping adjustment in the high-frequency range leverages the rapid response and energy dissipation advantages of the hydraulic system for high-frequency vibrations, further improving the control effect.

[0020] Furthermore, this solution is not limited to a single method for dynamically correcting the overturning compensation amount. One feasible solution is to dynamically correct the overturning compensation amount based on the feedback signal, including: The current feedback structural vibration response data is compared with the expected control target to calculate the deviation; the correction is then calculated using a feedback control law based on the deviation. The correction amount is added to the current anti-overturning compensation amount to generate the updated compensation amount; The coordinated control instructions are regenerated based on the updated compensation amount.

[0021] This approach endows the control system with closed-loop characteristics. The compensation amount calculated by the feedforward is continuously corrected during execution based on the actual response of the structure, effectively compensating for the loss of control accuracy caused by model errors and prediction deviations in feedforward control. The feedforward component provides rapid pre-action, while the feedback component provides continuous accuracy correction. The correction process is carried out throughout the dynamic iteration of the entire control phase until the triggering condition is no longer met.

[0022] On the other hand, an adaptive anti-overturning device for a bridge erecting machine is provided for performing the above-mentioned control method, including: The inertial measurement unit is installed at the mid-span of the main beam and the top of the outriggers of the bridge erecting machine to collect the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam in real time. A dynamic anemometer is installed on the upper part of the main beam of the bridge erecting machine to collect ambient wind speed and direction in real time. The edge computing controller is communicatively connected to the inertial measurement unit and the dynamic anemometer, respectively, and is used to receive the dynamic acceleration, dynamic angular velocity, wind speed and wind direction data, and to identify the current modal parameters of the main beam online based on the data, determine whether the triggering conditions for active control are met, and calculate the anti-overturning compensation amount when the triggering conditions are met. The overhead crane drive mechanism is communicatively connected to the edge computing controller and is used to drive the overhead crane to move according to the first control command sent by the edge computing controller in order to change the center of gravity position of the bridge erecting machine. The outrigger hydraulic damping adjustment mechanism is communicatively connected to the edge computing controller and is used to adjust the damping characteristics of the outrigger hydraulic system according to the second control command sent by the edge computing controller.

[0023] When this scheme is adopted, closed-loop processing is achieved, effectively reducing communication latency. The crane drive mechanism and the outrigger hydraulic damping adjustment mechanism are two independent actuators that respond to the corresponding control commands respectively, realizing the simultaneous execution of center of gravity adjustment and damping enhancement.

[0024] Furthermore, to facilitate real-time tracking and immediate adjustment, one feasible solution is that the edge computing controller includes: The feedback data receiving interface is communicatively connected to the inertial measurement unit and is used to continuously receive the dynamic acceleration and dynamic angular velocity fed back in real time by the inertial measurement unit during the operation of the crane drive mechanism and the outrigger hydraulic damping adjustment mechanism. The comparison and calculation unit is connected to the feedback data receiving interface and is used to compare the current feedback dynamic acceleration and dynamic angular velocity with the preset expected control target to generate a deviation signal. A correction calculation unit, connected to the comparison operation unit, is used to calculate the correction amount based on the deviation signal using a feedback control law. The compensation update unit, connected to the correction calculation unit, is used to superimpose the correction amount onto the current anti-overturning compensation amount calculated by the edge computing controller to generate the updated compensation amount. The instruction regeneration unit, connected to the compensation update unit, is used to regenerate the cooperative control instruction based on the updated compensation amount.

[0025] In this scheme, the continuous communication connection between the feedback data receiving interface and the inertial measurement unit ensures the real-time acquisition of feedback signals throughout the entire actuator operation, providing uninterrupted data flow support for closed-loop correction. The comparison and calculation unit compares the feedback data with the expected control target to generate a deviation signal, quantifying the control target into a calculable deviation, thus providing a mathematical basis for the quantitative calculation of the correction amount.

[0026] Compared with existing technologies, the advantages of this invention are: First, by using modal parameters as the basis for triggering active control, the passive response mode of the traditional scheme, which only triggers an alarm when the overturning coefficient drops below the threshold, is changed, and the control intervention point is advanced to the risk precursor stage. By decomposing the overturning compensation amount into compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, the synergistic effect of center of gravity position control and vibration energy dissipation is realized. During the execution process, the structural vibration response is introduced as a feedback signal for dynamic correction until the triggering condition is eliminated, thus forming a closed-loop control mechanism with adaptive capabilities. This improves the timeliness and reliability of overturning control during the bridge erecting machine's span operation and solves the problems of not being able to achieve real-time control and lacking an active identification and judgment mechanism for structural risk status. Second, it can achieve accurate matching based on risk mode patterns accumulated from historical or simulation data, effectively reducing the probability of misjudgment and missed judgment; it identifies risk state types with different physical mechanisms, enabling the control system to adopt differentiated control strategies according to specific risk induction mechanisms, further improving the pertinence and effectiveness of active control and ensuring structural safety. Third, the combined acquisition of multi-degree-of-freedom dynamic acceleration and angular velocity can comprehensively characterize the translational and rotational responses of the main beam, providing more complete structural dynamics information for modal parameter identification. The construction of the system state matrix and output matrix gives the extraction of modal parameters clear physical meaning and mathematical basis. Furthermore, combining real-time wind speed and direction with the damping ratio in the modal parameters for dynamic bending moment increment calculation fully considers the time-varying characteristics of wind load and the energy dissipation capacity of the structure itself, making the determination of compensation amount more accurate. Attached Figure Description

[0027] Figure 1 A flowchart of an adaptive anti-overturning control method for a bridge erecting machine provided in Embodiment 1 of the present invention; Figure 2 A flowchart illustrating the determination of triggering conditions provided in Embodiment 1 of the present invention; Figure 3 A flowchart for establishing a dynamic feature library is provided for Embodiment 1 of the present invention; Figure 4 A flowchart for compensation amount decomposition and collaborative control instruction generation provided in Embodiment 1 of the present invention; Figure 5 A flowchart of feedback signal comparison and dynamic correction of compensation amount provided in Embodiment 1 of the present invention; Figure 6 A schematic diagram of an adaptive anti-overturning device for a bridge erecting machine provided in Embodiment 2 of the present invention; Figure 7 for Figure 6 Enlarged view of point A in the image; Figure 8 for Figure 6 Enlarged view of point B in the image.

[0028] Figure label: 100. Inertial Measurement Unit; 200. Dynamic Anemometer; 300. Crane Moving Mechanism; 400. Outrigger Hydraulic Damping Adjustment Mechanism. Detailed Implementation

[0029] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0030] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0031] Example 1: Please refer to Figure 1 An adaptive anti-overturning control method for a bridge erecting machine, comprising: S100, Dynamic Data Acquisition: Real-time acquisition of structural vibration response data and environmental wind load data of the bridge erecting machine during the span-crossing operation; S200, Real-time Modal Parameter Identification: Based on structural vibration response data, an online modal parameter identification algorithm is used to identify the current modal parameters of the main beam of the bridge erecting machine in real time. The modal parameters include at least the dominant vibration frequency, damping ratio, and mode shape. S300, Triggering condition determination: Based on the current modal parameters, determine whether the triggering conditions for active control are met; S400, Overturning Compensation Calculation: When the triggering conditions are met, the overturning compensation is calculated based on the environmental wind load data and the current modal parameters; S500, Cooperative Control and Closed-Loop Correction: Cooperative control commands are generated based on the anti-overturning compensation amount and sent to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. The anti-overturning compensation amount includes compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. During the operation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, the current structural vibration response data is continuously acquired as feedback signals, and the anti-overturning compensation amount is dynamically corrected based on the feedback signals until the triggering conditions are no longer met.

[0032] This approach uses modal parameters as the basis for triggering active control, changing the passive response mode of traditional schemes that only triggers alarms when the overturning coefficient drops below a threshold. It advances the control intervention point to the early risk warning stage. By decomposing the overturning compensation into compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, it achieves the synergistic effect of center of gravity position control and vibration energy dissipation. During execution, it introduces structural vibration response as a feedback signal for dynamic correction until the triggering condition is eliminated, thus forming a closed-loop control mechanism with adaptive capabilities. This improves the timeliness and reliability of overturning control during bridge erection machine crossing operations and solves the problems of not being able to achieve real-time control and lacking an active identification and judgment mechanism for structural risk status.

[0033] Reference Figure 2 This solution does not exclusively limit the specific method for determining the triggering conditions. One feasible solution is to determine whether the triggering conditions for active regulation are met, including: S310. Compare the current modal parameters with the preset dynamic feature library, which stores multiple modal parameter patterns and corresponding risk state types. S320. If the current modal parameters match any risk state type in the dynamic feature library, then the triggering condition for active control is met. Among them, the risk state types include at least one or more of the vortex-induced resonance precursor state and flutter critical state.

[0034] When this scheme is adopted, it can achieve accurate matching based on the risk mode patterns accumulated from historical or simulation data, effectively reducing the probability of misjudgment and omission. It identifies risk state types with different physical mechanisms, enabling the control system to adopt differentiated control strategies according to the specific risk induction mechanism, further improving the pertinence and effectiveness of active control and ensuring structural safety.

[0035] It should be noted that the dynamic feature library is a database that is pre-established and continuously updated through a combination of offline and online methods before the control method of this embodiment is put into actual use. Its core function is to provide a comparison basis between modal parameter patterns and risk state types for the trigger condition determination step. Considering the special characteristics of the bridge erecting machine's span crossing operation, such as the rapid time-varying structural state, the directional nature of risk evolution, and the fact that control actions themselves can change the dynamic response of the structure.

[0036] Reference Figure 3 This embodiment uses a state evolution graph method based on the via process to establish a dynamic feature library, specifically including the following steps: S311. Stage Division and State Variable Definition of the Span Crossing Process: The entire span crossing process of the bridge erecting machine is divided into several characteristic stages according to operating procedures and structural mechanical characteristics. The span crossing process typically includes: initial counterweight adjustment stage, main beam forward movement stage, front outrigger conversion stage, middle outrigger forward movement stage, rear outrigger release stage, and beam placement stage. Each stage is uniquely defined by a set of measurable state variables, which at least include: the longitudinal position of the gantry crane along the length of the main beam, the support reaction force of each outrigger, and the cantilever length of the main beam. These state variables constitute a multi-dimensional state space. At every instant during the entire span crossing process, the bridge erecting machine is located at a specific point in this state space, and the entire span crossing process is a continuous trajectory from the start point to the end point of the state space.

[0037] S312. Offline calibration of nominal evolutionary trajectories: For each segmented feature stage, the baseline values ​​of modal parameters corresponding to each feature stage are obtained through the following two methods: Approach A involves obtaining stage benchmark values ​​based on multibody dynamics simulation. A multibody dynamics model of the bridge erecting machine is established, treating core components such as the main beam, outriggers, and traveling trolley as flexible bodies. Physical properties such as the elastic modulus, density, and moment of inertia of each component are defined. Following the span operation procedures, boundary conditions and load conditions for each stage are sequentially applied in the simulation environment. Boundary conditions include the support state and constraint method of each outrigger, and load conditions include the concentrated force corresponding to the crane position, the self-weight distribution force of the main beam, and wind load. Transient dynamics simulations are then performed. Modal parameters corresponding to each stage are extracted from the simulation results, including at least the dominant vibration frequency, damping ratio, and mode shape, serving as the simulation benchmark values ​​for the modal parameters of that stage.

[0038] Approach B involves obtaining stage benchmark values ​​based on historical span-crossing data. In previous successful span-crossing operations of the bridge erecting machine, the dynamic acceleration and dynamic angular velocity data of the main beam were recorded throughout the entire process using an inertial measurement unit, along with state variables such as gantry position, outrigger reaction force, wind speed, and wind direction. The data from each successful span-crossing was segmented according to the stage divisions established in the first step. A data-driven random subspace identification algorithm was used to extract the modal parameters corresponding to each stage: first, a Hankel matrix was constructed from each set of data; then, the Hankel matrix was decomposed using QR decomposition and the projection matrix was calculated; next, singular value decomposition was performed on the projection matrix to obtain the system state matrix; finally, the modal parameters were obtained by solving for the eigenvalues ​​and eigenvectors of the system state matrix. Statistical analysis was performed on the data from the same stage of multiple successful span-crossing operations, calculating the mean and standard deviation of the modal parameters for each stage, which served as the statistical benchmark values ​​for the modal parameters of that stage.

[0039] The simulated baseline values ​​of modal parameters obtained from method A and the statistical baseline values ​​of modal parameters obtained from method B are weighted and fused to obtain the baseline values ​​and allowable deviation ranges of modal parameters for each stage. The allowable deviation range is determined by multiplying the standard deviation of the statistical baseline values ​​of modal parameters by a preset confidence coefficient. The baseline values ​​of modal parameters for each stage are connected sequentially according to the time order of the bridge crossing to form a nominal evolution trajectory with the bridge crossing process as the horizontal axis and the modal parameters as the vertical axis. This nominal evolution trajectory represents the expected change path of the modal parameters at each stage when the bridge erecting machine crosses the bridge under ideal conditions.

[0040] S313. Real-time trajectory matching and risk classification determination: In actual hole-passing operations, the control system executes dynamic data acquisition and real-time modal parameter identification steps in real time to obtain the actual modal parameters at the current moment. Simultaneously, based on the currently acquired state variables such as the crane's longitudinal position and the support reaction forces of each leg, the system locates the current stage of the hole-passing process in the state space and retrieves the corresponding modal parameter benchmark values ​​and allowable deviation ranges from the dynamic feature library.

[0041] The actual modal parameters at the current moment are compared with the baseline values ​​at the corresponding positions on the nominal evolution trajectory. If the actual modal parameters are within the allowable deviation range, the structural state is considered normal, and no active control is triggered. If the actual modal parameters deviate from the allowable deviation range but the deviation is less than a preset first deviation threshold, it is considered a slight trajectory deviation, triggering the first level of active control. If the actual modal parameters deviate from the allowable deviation range and the deviation is greater than or equal to the preset first deviation threshold, or if the deviation trend of the actual modal parameters points towards the unstable branch direction of the nominal evolution trajectory, it is considered a severe trajectory deviation, triggering the second level of active control. The deviation trend is determined by calculating the deviation direction and velocity of the actual modal parameters relative to the nominal evolution trajectory at multiple consecutive sampling moments.

[0042] S314. Based on Bayesian inference, the feature library is adaptively updated. After each via operation, the actual modal parameter trajectory and the nominal evolution trajectory of the entire via process are compared stage by stage.

[0043] S315. If the via is successful and there is a systematic deviation between the actual trajectory and the nominal trajectory, then the nominal evolution trajectory is corrected online using Bayesian inference.

[0044] By correcting the nominal evolution trajectory online, the nominal evolution trajectory gradually converges to the actual modal parameter evolution law of the bridge erecting machine, and the judgment boundary of the feature library is continuously optimized accordingly. Simultaneously, the deviation direction and magnitude of the actual trajectory from the nominal trajectory at each update are recorded as a deviation pattern record. The accumulated deviation pattern records from multiple crossings constitute a deviation pattern library. When the current deviation pattern appearing in a crossing matches a historical deviation pattern in the deviation pattern library, the future development trend of the current deviation is predicted based on the subsequent actual development results of that historical deviation pattern, achieving advanced risk prediction.

[0045] The dynamic feature library is stored in the non-volatile storage unit of the edge computing controller. It contains the baseline values ​​of modal parameters and the allowable deviation range for each stage, complete data of the nominal evolution trajectory, and the deviation pattern library.

[0046] This scheme does not exclusively limit the specific types of structural vibration response data and environmental wind load data. One feasible scheme is that the structural vibration response data should at least include the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam of the bridge erecting machine. Among them, the real-time modal parameter identification step of multi-degree-of-freedom dynamic acceleration and dynamic angular velocity input is used to construct the system state matrix and output matrix required for online modal parameter identification algorithm. The system state matrix represents the distribution of inertial force on each degree of freedom of the main beam, and the output matrix represents the distribution of acceleration and angular velocity response on each degree of freedom of the main beam, so as to separate the dominant vibration frequency, damping ratio and mode shape of the main beam under the current wind load from the dynamic acceleration and dynamic angular velocity. Environmental wind load data should include at least real-time wind speed and real-time wind direction; The calculation step for anti-overturning compensation includes inputting real-time wind speed and direction along with current modal parameters. This is used to calculate the dynamic bending moment increment generated by the current wind load acting on the main beam based on the structural dynamics equations. Combined with the damping ratio in the current modal parameters, which characterizes the vibration energy dissipation capacity, the calculation of the anti-overturning compensation required to offset the dynamic bending moment increment is also performed.

[0047] This approach comprehensively characterizes the translational and rotational responses of the main beam through the joint acquisition of multi-degree-of-freedom dynamic acceleration and angular velocity, providing more complete structural dynamics information for modal parameter identification. The construction of the system state matrix and output matrix gives the extraction of modal parameters clear physical meaning and mathematical basis. Furthermore, combining real-time wind speed and direction with the damping ratio from the modal parameters for dynamic bending moment increment calculation fully considers the time-varying characteristics of wind loads and the structure's own energy dissipation capacity, making the determination of compensation amounts more accurate.

[0048] Optionally, the system state matrix is ​​the core matrix describing the dynamic characteristics of the system in the state-space model. In structural dynamics, the main girder of the bridge erecting machine can be modeled as a multi-degree-of-freedom linear vibration system, and its dynamic behavior can be described by the following continuous-time state-space equations: ẋ(t) = A·x(t) + B·u(t) Where x(t) is the state vector, ẋ(t) is the first derivative of the state vector x(t) with respect to time t, containing the velocity and acceleration of each degree of freedom of the main beam; u(t) is the input vector, which in this embodiment is the wind load excitation; A is the system state matrix. The system state matrix A is an n×n square matrix (n is the number of state variables), and its elements are jointly determined by the mass matrix, damping matrix, and stiffness matrix of the main beam. In the real-time modal parameter identification step, the collected multi-degree-of-freedom dynamic acceleration and dynamic angular velocity data are processed by an online random subspace identification algorithm. First, the row space projection of the Hankel matrix composed of system outputs is calculated by QR decomposition. Then, singular value decomposition is performed on the projection to obtain the observable matrix of the system and the Kalman filter estimate of the state sequence. Finally, the system state matrix A is determined by the observation matrix and / or the state sequence. The eigenvalues ​​and eigenvectors of the system state matrix A directly determine the modal parameters of the main beam: the real and imaginary parts of the eigenvalues ​​of A correspond to the damping ratio and natural frequency (dominant vibration frequency) of each mode, respectively, and the eigenvectors of A correspond to each mode shape. By identifying the system state matrix A online and extracting its eigenvalues ​​and eigenvectors, the dominant vibration frequency, damping ratio, and mode shape of the main beam under the current wind load can be separated from the original dynamic response data.

[0049] Optionally, the output matrix is ​​a matrix in the state-space model that describes how the system states are mapped to observable outputs. In the structural dynamics state-space model, the output equation can be expressed as: y(t) = C·x(t); Where y(t) is the output vector, representing the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity acquired by the inertial measurement unit (IMU), and C is the output matrix. The output matrix C is a p×n matrix (p is the number of output quantities, n is the number of state quantities), and its elements determine the distribution weight of the acceleration and angular velocity responses of each degree of freedom in the overall output. The IMU is installed at multiple measuring points at the mid-span of the main beam and the top of the legs. Each measuring point outputs translational acceleration in three directions and rotational angular velocity in three directions, forming the output vector y(t). The output matrix C essentially characterizes the spatial correspondence between the sensor placement positions at each measuring point and the mode shapes of the main beam: defining how the system state (modal coordinates) is mapped to the acceleration and angular velocity responses at each measuring point. In the random subspace identification algorithm, the output matrix C, along with the system state matrix A, is obtained from the singular value decomposition of the Hankel matrix. By using the output matrix C, the acceleration and angular velocity responses collected from different measurement points can be separated according to the participation weight of each mode, thereby accurately extracting the contribution of each mode to the overall response and providing a reliable modal parameter basis for subsequent trigger condition determination.

[0050] Understandably, the dynamic bending moment increment refers to the time-varying increment of the dynamic bending moment generated by the main girder of the bridge erecting machine under the action of pulsating wind load, relative to the static equilibrium state. Specifically, the wind load acting on the main girder can be decomposed into two parts: mean wind (static part) and pulsating wind (dynamic part). The mean wind generates a static bending moment, placing the main girder at a certain static equilibrium position; the pulsating wind causes the main girder to vibrate dynamically around this equilibrium position, and the additional bending moment generated as a function of time is the dynamic bending moment increment. In the overturning compensation calculation steps of this embodiment, the calculation process of the dynamic bending moment increment is as follows: First, based on the real-time wind speed and real-time wind direction, combined with the aerodynamic coefficient and windward area of ​​the main girder section, the time history of the pulsating wind load acting on each section of the main girder at the current moment is calculated according to the wind load model in structural dynamics; then, based on the force diagram of the main girder and the structural dynamic equations, the pulsating wind load is integrated along the length of the main girder to calculate the dynamic bending moment increment generated by it at the root or key section of the main girder. The magnitude of the dynamic bending moment increment is proportional to the square of the real-time wind speed and related to the deflection angle of the wind direction relative to the main beam axis. It is also significantly influenced by the damping ratio in the current modal parameters of the main beam: a larger damping ratio results in a stronger ability of the structure to dissipate vibrational energy, and a smaller dynamic bending moment increment under the same wind load. This embodiment calculates the dynamic bending moment increment and uses it as the basis for feedforward compensation, thereby quantifying the specific impact of the dynamic changes in wind load on the overturning stability of the main beam.

[0051] This scheme does not exclusively limit the specific structure of the modal parameters. One feasible scheme is that the modal parameters include at least the dominant vibration frequency, damping ratio, and mode shape; the dominant vibration frequency, damping ratio, and mode shape are used together as the basis for determining whether the active control triggering condition is met in the triggering condition determination. The online modal parameter identification algorithm is an online random subspace identification algorithm. It uses recursive QR decomposition or sliding window technology to realize real-time processing of dynamic acceleration and dynamic angular velocity data in order to update modal parameters online.

[0052] This approach can separately reflect the structure's stiffness characteristics, energy dissipation capacity, and vibration spatial distribution, forming a complete description system of the structure's dynamic characteristics. The online random subspace recognition algorithm, combined with recursive QR decomposition or sliding window technology, significantly reduces the data storage requirements and computational complexity of online computing while ensuring recognition accuracy. This allows for the engineering implementation of real-time modal parameter recognition on edge computing controllers. The recursive algorithm only incrementally updates existing recognition results when new data arrives, eliminating the need to repeatedly process all historical data and improving computational efficiency.

[0053] It should be noted that the recursive QR decomposition and sliding window techniques are two optional implementation methods for the real-time processing of the above-mentioned online random subspace recognition algorithm. Their purpose is to enable the algorithm to continuously update the modal parameter recognition results as new data is continuously collected, avoiding the problems of large computational load, high storage requirements and long output delay caused by the traditional batch processing method, which requires reprocessing all historical data each time.

[0054] The basic principle of the recursive QR decomposition is as follows: In the data-driven random subspace recognition algorithm, the Hankel matrix is ​​first constructed from the system's output response data, and the projection matrix from the future row space to the past row space of this Hankel matrix is ​​calculated through QR decomposition. When a new frame of data arrives, the Hankel matrix has several new columns of data. The traditional method requires re-performing the QR decomposition on the entire updated Hankel matrix. The recursive QR decomposition, however, utilizes the properties of matrix partitioning and orthogonal transformation to process only the newly added data: Let R be the QR decomposition obtained at time k. k A matrix that, when a new data block arrives, will combine the new data with existing R... k The matrices are concatenated to form a new matrix to be decomposed, and then Givens rotation is applied to orthogonally triangulate the concatenated matrix. Givens rotation is an orthogonal transformation method that sets specified elements in a matrix to zero through planar rotation. Each rotation affects only two rows of elements in the matrix, resulting in low computational cost and good numerical stability. After a series of Givens rotations, the concatenated matrix is ​​re-decomposed into an upper triangular matrix R. k+1 and orthogonal matrix Q k+1The product of the two components is used to update the QR decomposition without reprocessing all historical data. Subsequently, based on the updated projection matrix, a subspace tracking algorithm (such as projection approximation subspace tracking) is used to continuously track and calculate the left singular value vector of the projection matrix. Finally, the modal parameters of the system are solved using the least squares method. Through this recursive update mechanism, whenever a new frame of data arrives, the algorithm only needs to process the computational amount corresponding to the newly added data, reducing the computational complexity from O(N) for batch processing. 3 The order of magnitude is reduced to O(N) 2 The system achieves online real-time tracking of time-varying modal parameters on a scale of 10000 to 10000 (where N is the total amount of data).

[0055] The basic principle of the sliding window technique is as follows: During algorithm execution, a fixed-length time window is maintained, containing structural vibration response data from the most recent M moments. Each time a new frame of data is acquired, it is added to the end of the window, while the earliest frame from the beginning is removed, ensuring the window always contains the latest M frames. Then, using the M frames of data within the current window as input, a random subspace recognition algorithm is run to identify the modal parameters at the current moment. As data continues to flow in, the window slides forward, and recognition is re-executed after each slide, thus achieving continuous tracking and recognition of time-varying modal parameters. The choice of window length needs to strike a balance between recognition accuracy and computational efficiency: a window that is too short will result in insufficient data for recognition, reducing the statistical reliability of the recognition results; a window that is too long will fail to reflect the time-varying characteristics of the structural modal parameters in a timely manner. For the application scenario of bridge erecting machine crossing the span, the window length can be pre-calibrated based on the dominant vibration period of the main beam and the time-varying characteristics of the wind load, for example, set to 5 to 10 times the dominant vibration period. The sliding window technique also has the function of eliminating spurious modes: by continuously monitoring the stability of modal parameters at each order during the window sliding process, those modes that only appear in a few windows and lack persistence are identified as spurious modes and eliminated.

[0056] Understandably, recursive QR decomposition and sliding window techniques are two different technical approaches to solving the same problem (online real-time recognition): Recursive QR decomposition focuses on the evolution of the calculation method, reducing the amount of calculation per operation by recursively modifying the QR decomposition steps themselves. The sliding window technique focuses on limiting the data range by controlling the data size for each calculation by limiting the algorithm's input data to the latest data window of a fixed length.

[0057] Both approaches can be used independently in conjunction with the random subspace identification algorithm, or one can be chosen based on the computing resources and usage habits of the actual edge computing controller. Using either approach, the online random subspace identification algorithm can achieve continuous real-time updates of modal parameters on the edge computing controller with engineering-feasible computing resources.

[0058] To improve adaptability in complex wind field environments, one feasible solution is to calculate the overturning compensation amount specifically by feedforward calculation. This involves calculating the instantaneous value and trend of the wind load based on the real-time wind speed and direction, combining the dominant vibration frequency and damping ratio in the current modal parameters, predicting the dynamic bending moment increment of the main beam in the future, and pre-calculating the overturning compensation amount required to offset the dynamic bending moment increment based on the prediction results.

[0059] This approach calculates the instantaneous value and trend of wind load based on real-time wind speed and direction, and predicts the dynamic bending moment increment of the main beam in the future by combining the dominant vibration frequency and damping ratio, thereby pre-calculating the compensation amount. This allows for early initiation of control actions before the wind load change trend fully impacts the structure, effectively overcoming the inherent lag of pure feedback control. Simultaneously, by combining the structure's own frequency and damping characteristics for dynamic bending moment increment prediction, the prediction results more closely match the actual dynamic response behavior of the structure. Addressing the inherent randomness and time-varying nature of wind loads, the feedforward calculation method effectively handles disturbances caused by fluctuating wind speed components and instantaneous changes in wind direction, enhancing the adaptability of the control system in complex wind field environments.

[0060] Reference Figure 4 To improve the overturning resistance of the bridge erecting machine, one feasible solution is to generate coordinated control commands and send them to the overhead crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, respectively, including: S510. Decompose the anti-overturning compensation into the crane displacement compensation component and the hydraulic damping adjustment component. S520: Generate control commands for the overhead crane moving mechanism based on the overhead crane displacement compensation components. S530: Generate control commands for the outrigger hydraulic damping adjustment mechanism based on the hydraulic damping adjustment component.

[0061] This scheme utilizes gantry displacement compensation to alter the center of gravity, thereby counteracting the overturning moment. Furthermore, it employs hydraulic damping adjustment of the outriggers to modify energy dissipation, thus suppressing vibration amplitude, resulting in complementary and coordinated control. The two components correspond to displacement and damping, respectively. Simultaneous control of both gantry displacement and damping adjustment further shortens the control response time and enhances the overturning resistance of the bridge erecting machine.

[0062] To further improve the control effect, one feasible solution is to allocate the crane displacement compensation component and the hydraulic damping adjustment component according to a preset coordination weighting coefficient; the coordination weighting coefficient is used to adjust the degree of participation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism when performing coordinated control. The collaborative weighting coefficient is dynamically adjusted based on the currently identified modal parameters: when the modal parameters are in the preset low-frequency range, the weight of the crane displacement compensation component is increased to suppress low-frequency swaying through center of gravity adjustment; when the modal parameters are in the preset high-frequency range, the weight of the hydraulic damping adjustment component is increased to dissipate high-frequency vibration energy through active damping.

[0063] This scheme increases the crane displacement compensation weight in the low-frequency range to suppress low-frequency swaying, and increases the hydraulic damping adjustment weight in the high-frequency range to dissipate high-frequency vibration energy. It allows for optimized configuration based on the structure's current vibration characteristics, effectively avoiding the pitfalls of minimal damping adjustment at low frequencies and slow response of center-of-gravity adjustment at high frequencies under fixed weights. The strategy of emphasizing crane displacement compensation in the low-frequency range utilizes the large stroke adjustment capability of center-of-gravity adjustment for low-frequency swaying, while the strategy of emphasizing hydraulic damping adjustment in the high-frequency range leverages the rapid response and energy dissipation advantages of the hydraulic system for high-frequency vibrations, further improving the control effect.

[0064] Reference Figure 5 This scheme does not exclusively limit the dynamic correction method for overturning compensation. One feasible scheme is to dynamically correct the overturning compensation based on feedback signals, including: S540. Compare the current feedback structural vibration response data with the expected control target and calculate the deviation; calculate the correction amount using the feedback control law based on the deviation. S550: Add the correction amount to the current anti-overturning compensation amount to generate the updated compensation amount; S560: Regenerate the coordinated control instructions based on the updated compensation amount.

[0065] This approach endows the control system with closed-loop characteristics. The compensation amount calculated by the feedforward is continuously corrected during execution based on the actual response of the structure, effectively compensating for the loss of control accuracy caused by model errors and prediction deviations in feedforward control. The feedforward component provides rapid pre-action, while the feedback component provides continuous accuracy correction. The correction process is carried out throughout the dynamic iteration of the entire control phase until the triggering condition is no longer met.

[0066] Example 2: Reference Figure 6 , Figure 7 and Figure 8 An adaptive anti-overturning device for a bridge erecting machine, used to execute the above-mentioned control method, includes: The inertial measurement unit 100 is installed at the mid-span of the main beam and the top of the outriggers of the bridge erecting machine to collect the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam in real time. The dynamic anemometer 200 is installed on the upper part of the main beam of the bridge erecting machine to collect ambient wind speed and direction in real time. The edge computing controller is communicatively connected to the inertial measurement unit 100 and the dynamic anemometer 200, respectively. It is used to receive dynamic acceleration, dynamic angular velocity, wind speed and wind direction data, and to identify the current modal parameters of the main beam online based on the data, determine whether the triggering conditions for active control are met, and calculate the anti-overturning compensation amount when the triggering conditions are met. The overhead crane drive mechanism 300 is communicatively connected to the edge computing controller and is used to drive the overhead crane to move according to the first control command sent by the edge computing controller in order to change the center of gravity position of the bridge erecting machine. The outrigger hydraulic damping adjustment mechanism 400 is communicatively connected to the edge computing controller and is used to adjust the damping characteristics of the outrigger hydraulic system according to the second control command sent by the edge computing controller.

[0067] When this scheme is adopted, closed-loop processing is achieved, effectively reducing communication delay. The crane drive mechanism 300 and the outrigger hydraulic damping adjustment mechanism 400 are two independent actuators that respond to the corresponding control commands respectively, realizing the simultaneous execution of center of gravity adjustment and damping enhancement.

[0068] To facilitate real-time tracking and immediate adjustment, one feasible solution is as follows: the edge computing controller includes: The feedback data receiving interface is connected to the inertial measurement unit for continuously receiving the dynamic acceleration and dynamic angular velocity fed back by the inertial measurement unit in real time during the operation of the crane drive mechanism and the outrigger hydraulic damping adjustment mechanism. The comparison and calculation unit is connected to the feedback data receiving interface and is used to compare the current feedback dynamic acceleration and dynamic angular velocity with the preset expected control target to generate a deviation signal. The correction calculation unit, connected to the comparison operation unit, is used to calculate the correction amount based on the deviation signal using a feedback control law. The compensation update unit, connected to the correction calculation unit, is used to superimpose the correction amount onto the current anti-overturning compensation amount calculated by the edge computing controller to generate the updated compensation amount. The instruction regeneration unit, connected to the compensation update unit, is used to regenerate the coordinated control instruction based on the updated compensation amount.

[0069] In this scheme, the continuous communication connection between the feedback data receiving interface and the inertial measurement unit ensures the real-time acquisition of feedback signals throughout the entire actuator operation, providing uninterrupted data flow support for closed-loop correction. The comparison and calculation unit compares the feedback data with the expected control target to generate a deviation signal, quantifying the control target into a calculable deviation, thus providing a mathematical basis for the quantitative calculation of the correction amount.

[0070] How this solution works: The inertial measurement unit 100 and dynamic anemometer 200 collect real-time multi-degree-of-freedom dynamic acceleration, dynamic angular velocity, and ambient wind speed and direction of the main beam as raw input data. Based on the above data, an online random subspace recognition algorithm is used to identify the current modal parameters of the main beam, such as the dominant vibration frequency, damping ratio, and mode shape. The current modal parameters are compared with the risk state types in the preset dynamic feature library. If they match, it is determined that the active control trigger condition is met. When the trigger condition is met, the dynamic bending of the main beam at future moments is predicted using a feedforward calculation method based on the real-time wind speed and direction and the dominant vibration frequency and damping ratio in the current modal parameters. The torque increment is calculated in advance, and the anti-overturning compensation is decomposed into the crane displacement compensation component and the hydraulic damping adjustment component, and control commands are generated respectively. These commands are simultaneously sent to the crane moving mechanism 300 and the outrigger hydraulic damping adjustment mechanism 400 to perform coordinated control. During the operation of the actuator, the dynamic acceleration and dynamic angular velocity fed back in real time by the inertial measurement unit are continuously acquired as feedback signals. The feedback data is compared with the expected control target to obtain the deviation. The correction amount is calculated using the feedback control law and then superimposed on the current compensation amount to generate the updated compensation amount. Based on this, the coordinated control command is regenerated. The cycle iterates until the triggering condition is no longer met.

[0071] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An adaptive anti-overturning control method for a bridge erecting machine, characterized in that, include: Dynamic data acquisition: Real-time acquisition of structural vibration response data and environmental wind load data of the bridge erecting machine during the span crossing operation; Real-time modal parameter identification: Based on the structural vibration response data, an online modal parameter identification algorithm is used to identify the current modal parameters of the main beam of the bridge erecting machine in real time; Triggering condition determination: Based on the current modal parameters, determine whether the triggering conditions for active control are met; Overturning compensation calculation: When the triggering condition is met, the overturning compensation is calculated based on the environmental wind load data and the current modal parameters; Coordinated control and closed-loop correction: Coordinated control commands are generated based on the anti-overturning compensation amount and sent to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. The anti-overturning compensation amount includes compensation components corresponding to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively. During the operation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism, the current structural vibration response data is continuously acquired as a feedback signal, and the anti-overturning compensation amount is dynamically corrected based on the feedback signal until the triggering condition is no longer met.

2. The adaptive anti-overturning control method for a bridge erecting machine according to claim 1, characterized in that, The determination of whether the triggering conditions for active control are met includes: The current modal parameters are compared with a preset dynamic feature library, which stores multiple modal parameter patterns and corresponding risk state types. If the current modal parameter matches any risk state type in the dynamic feature library, then the triggering condition for the active control is determined to be met. The risk state type includes at least one or more of the following: vortex-induced resonance precursor state and flutter critical state.

3. The adaptive anti-overturning control method for a bridge erecting machine according to claim 1, characterized in that: The structural vibration response data includes at least the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam of the bridge erecting machine; The step of real-time modal parameter identification using multi-degree-of-freedom dynamic acceleration and dynamic angular velocity inputs is used to construct the system state matrix and output matrix required by the online modal parameter identification algorithm. The system state matrix represents the distribution of inertial forces on each degree of freedom of the main beam, and the output matrix represents the distribution of acceleration and angular velocity responses on each degree of freedom of the main beam, so as to separate the dominant vibration frequency, damping ratio and mode shape of the main beam under the current wind load from the dynamic acceleration and dynamic angular velocity. The environmental wind load data includes at least real-time wind speed and real-time wind direction; The real-time wind speed and real-time wind direction, together with the current modal parameters, are input into the overturning compensation calculation step. This step is used to calculate the dynamic bending moment increment generated by the current wind load acting on the main beam based on the structural dynamics equations. Combined with the damping ratio in the current modal parameters, which characterizes the vibration energy dissipation capacity, the overturning compensation required to offset the dynamic bending moment increment is calculated.

4. The adaptive anti-overturning control method for a bridge erecting machine according to claim 3, characterized in that: The modal parameters include at least the dominant vibration frequency, damping ratio, and mode shape; the dominant vibration frequency, damping ratio, and mode shape together serve as the basis for determining whether the active control triggering condition is met in the triggering condition determination. The online modal parameter identification algorithm is an online random subspace identification algorithm. It uses recursive QR decomposition or sliding window technology to realize real-time processing of the dynamic acceleration and dynamic angular velocity data, so as to update the modal parameters online.

5. The adaptive anti-overturning control method for a bridge erecting machine according to claim 3, characterized in that: The calculation of the overturning compensation amount is specifically a feedforward calculation of the overturning compensation amount, that is, based on the real-time wind speed and real-time wind direction at the current moment, the instantaneous value and trend of the wind load are calculated, combined with the dominant vibration frequency and damping ratio in the current modal parameters, the dynamic bending moment increment of the main beam at future moments is predicted, and the overturning compensation amount required to offset the dynamic bending moment increment is calculated in advance based on the prediction results.

6. The adaptive anti-overturning control method for a bridge erecting machine according to claim 1, characterized in that, The generation of coordinated control commands, which are then sent to the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism respectively, includes: The anti-overturning compensation amount is decomposed into a crane displacement compensation component and a hydraulic damping adjustment component. Control commands for the crane moving mechanism are generated based on the crane displacement compensation components. Control commands for the outrigger hydraulic damping adjustment mechanism are generated based on the hydraulic damping adjustment component.

7. The adaptive anti-overturning control method for a bridge erecting machine according to claim 6, characterized in that: The crane displacement compensation component and the hydraulic damping adjustment component are allocated according to a preset coordination weighting coefficient; the coordination weighting coefficient is used to adjust the degree of participation of the crane moving mechanism and the outrigger hydraulic damping adjustment mechanism when performing coordinated control. The collaborative weighting coefficient is dynamically adjusted according to the currently identified modal parameters: when the modal parameters are in a preset low-frequency range, the weight of the crane displacement compensation component is increased to suppress low-frequency swaying through center of gravity adjustment; When the modal parameters are in a preset high-frequency range, the weight of the hydraulic damping adjustment component is increased to dissipate high-frequency vibration energy through active damping.

8. The adaptive anti-overturning control method for a bridge erecting machine according to claim 1, characterized in that, Dynamic correction of the anti-overturning compensation amount based on the feedback signal includes: The current feedback structural vibration response data is compared with the expected control target to calculate the deviation; the correction is then calculated using a feedback control law based on the deviation. The correction amount is added to the current anti-overturning compensation amount to generate the updated compensation amount; The coordinated control instructions are regenerated based on the updated compensation amount.

9. An adaptive anti-overturning device for a bridge erecting machine, used to execute the control method as described in any one of claims 1-8, characterized in that, include: The inertial measurement unit is installed at the mid-span of the main beam and the top of the outriggers of the bridge erecting machine to collect the multi-degree-of-freedom dynamic acceleration and dynamic angular velocity of the main beam in real time. A dynamic anemometer is installed on the upper part of the main beam of the bridge erecting machine to collect ambient wind speed and direction in real time. The edge computing controller is communicatively connected to the inertial measurement unit and the dynamic anemometer, respectively, and is used to receive the dynamic acceleration, dynamic angular velocity, wind speed and wind direction data, and to identify the current modal parameters of the main beam online based on the data, determine whether the triggering conditions for active control are met, and calculate the anti-overturning compensation amount when the triggering conditions are met. The overhead crane drive mechanism is communicatively connected to the edge computing controller and is used to drive the overhead crane to move according to the first control command sent by the edge computing controller in order to change the center of gravity position of the bridge erecting machine. The outrigger hydraulic damping adjustment mechanism is communicatively connected to the edge computing controller and is used to adjust the damping characteristics of the outrigger hydraulic system according to the second control command sent by the edge computing controller.

10. The adaptive anti-overturning device for a bridge erecting machine according to claim 9, characterized in that, The edge computing controller includes: The feedback data receiving interface is communicatively connected to the inertial measurement unit and is used to continuously receive the dynamic acceleration and dynamic angular velocity fed back in real time by the inertial measurement unit during the operation of the crane drive mechanism and the outrigger hydraulic damping adjustment mechanism. The comparison and calculation unit is connected to the feedback data receiving interface and is used to compare the current feedback dynamic acceleration and dynamic angular velocity with the preset expected control target to generate a deviation signal. A correction calculation unit, connected to the comparison operation unit, is used to calculate the correction amount based on the deviation signal using a feedback control law. The compensation update unit, connected to the correction calculation unit, is used to superimpose the correction amount onto the current anti-overturning compensation amount calculated by the edge computing controller to generate the updated compensation amount. The instruction regeneration unit, connected to the compensation update unit, is used to regenerate the cooperative control instruction based on the updated compensation amount.

Images