Multi-machine cooperative control method and system for bridge cable hoisting device

Abstract

The invention relates to the technical field of cooperative control, in particular to a multi-machine cooperative control method and system for a bridge cable hoisting device, and the method comprises the following steps: collecting tension change characteristics through a distributed sensor array, extracting indexes through a sliding window algorithm to generate a response time window, and generating a cooperative time window through time sequence overlapping and speed trend adjustment. Calculating a compensation value by combining inertial parameters and acceleration to generate a trigger window, detecting tension and displacement difference, inputting the tension and displacement difference into a dynamic threshold function for screening and fusion, outputting a synchronous trigger condition, and generating a synchronous control instruction by encoding a device number and a tension change rate after timestamp alignment. According to the method, multi-node tension synchronous acquisition and normalization processing, dynamic response generation through a sliding window, boundary matching action triggering adjustment through an overlapping algorithm, response lag correction through inertia compensation, tension and displacement difference real-time detection, dynamic threshold screening and triggering condition fusion and timestamp alignment packaging instruction consistency are carried out; and the cooperation precision and the system robustness are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cooperative control, and particularly to a multi-machine cooperative control method and system for a bridge cable hoisting device. Background Art

[0002] The technical field of cooperative control involves information interaction and coordinated operations among multiple execution units, aiming to complete a unified control goal through dynamic cooperation among subsystems. The core matters in this technical field include multi-machine cooperative scheduling, distribution and execution of control strategies, and adaptive application of coordination mechanisms in different operation scenarios, which are widely applied in complex systems that require multi-device cooperative operations such as industrial manufacturing, logistics handling, autonomous driving, and construction. Its systematicness is reflected in the comprehensive design and coordination of aspects such as control algorithms, communication structures, synchronization mechanisms, and task allocation strategies to ensure the orderly cooperation of multiple control objects in space and time.

[0003] Among them, the multi-machine cooperative control method for traditional bridge cable hoisting devices refers to the synchronous control and load coordination problems required by multiple lifting devices involved in the bridge construction process during cable hoisting tasks, and a master-slave control method or a centralized control system is used for management. The master-slave control method drives the slave devices to execute actions by setting the instructions of a master control unit. The control contents include instruction matching of hoisting speed and error correction of position feedback. The centralized control system collects the sensor data of multiple devices through a central processing unit, generates control instructions according to the set cable hoisting path and synchronous time nodes, and distributes them to the hoisting devices. In addition, the control signals are transmitted through wired communication, and the cable tension and displacement information are collected in real time by sensors for control adjustment.

[0004] Traditional master-slave control relies on a single instruction source. There is a time lag in error correction between the master and slave devices when the cable tension suddenly changes, which is likely to cause the superposition of multiple-level errors. The centralized control system adopts a fixed time node synchronization strategy, which is difficult to adapt to the changes in the acceleration of the lifting points and the differences in mechanical inertia, resulting in inaccurate triggering times of device actions. The wired communication architecture has significant signal transmission delays in long-distance multi-node deployments, affecting the real-time control efficiency. The fixed threshold synchronization condition lacks the ability to adjust when the load dynamically changes, increasing the risk of uneven tension distribution and restricting the control stability of the multi-machine cooperative system in complex paths and disturbance environments. Summary of the Invention

[0005] To solve the technical problems that traditional master-slave control relies on a single instruction source, there is a time lag in error correction between master and slave devices when the cable tension suddenly changes, which is prone to cause multi-level error superposition. The centralized control system adopts a fixed-time-node synchronization strategy, which is difficult to adapt to the changes in the hoisting point acceleration and the mechanical inertia differences, resulting in inaccurate triggering times of device actions. The wired communication architecture has significant signal transmission delays in long-distance multi-node deployments, affecting the real-time control efficiency. The fixed-threshold synchronization condition lacks the ability to adjust when the load dynamically changes, exacerbating the risk of uneven tension distribution and restricting the control stability of the multi-machine cooperative system in complex paths and disturbance environments, the embodiments of the present invention provide a multi-machine cooperative control method and system for bridge cable hoisting devices. The technical solutions are as follows: On the one hand, a multi-machine cooperative control method for a bridge cable hoisting device is provided, and the method includes: S1: Collect the starting time of cable tension change, the duration of tension fluctuation, and the steady-state response time of cable nodes through an optical fiber sensor array, perform normalization processing, input them into a sliding window algorithm to extract tension response characteristics, and generate an effective response time window; S2: Based on the effective response time window, calculate the intersection of node windows through a time series overlapping algorithm, adjust the time window boundary in combination with the displacement speed change trend of the device, and output a multi-node cooperative time window; S3: According to the multi-node cooperative time window, combine the structural inertia parameters and the hoisting point acceleration data, and use inertial compensation operation to calculate the structural response compensation value to generate a device action trigger window; S4: Based on the device action trigger window, detect the tension conduction time difference and the hoisting point displacement difference between adjacent devices, input them into a dynamic threshold function that self-matches according to the real-time working conditions, perform extreme value screening and interval fusion operations, and output a synchronization trigger condition; S5: According to the synchronization trigger condition, perform timestamp alignment operation. When the displacement difference is within the tolerance range, encapsulate the device number, trigger time, and tension change rate through instruction encoding, and output a multi-device synchronization control instruction.

[0006] As a further solution of the present invention, the starting time of the tension change is when the tension change exceeds 5% of the static mean value and the duration exceeds 0.2 seconds, and the duration of the tension fluctuation is when the tension fluctuation amplitude is lower than 10% of the initial fluctuation amplitude and lasts for at least 0.3 seconds; The adjustment of the window boundary is dynamically fine-tuned based on the positive and negative 1σ change trends of the weighted mean value of the original displacement speed; The inertial compensation operation corrects the original response signal to restore the real physical response state; The structural response compensation value corrects the hoisting response deviation of the device; The dynamic threshold function adjusts the synchronous trigger condition in real time to achieve consistent device responses. The effective response time window includes the tension change rate, the fluctuation duration, and the steady-state time constant. The multi-node cooperation time window specifically refers to the time window overlap degree, the displacement speed variance, and the cooperation convergence threshold. The device action trigger window includes the inertia compensation factor, the acceleration deviation threshold, and the phase synchronization margin. The synchronous trigger condition specifically refers to the peak conduction time difference, the displacement difference standard deviation, and the dynamic fusion threshold. The multi-device synchronous control instruction includes the device identifier, the trigger timestamp, and the tension change rate sequence. The cooperation convergence threshold refers to the critical value at which the multi-node time window reaches stability after adjustment. The dynamic fusion threshold is a fusion boundary value calculated by a real-time working condition adaptive algorithm based on multiple indicators such as the tension conduction time difference and the suspension point displacement difference.

[0007] As a further solution of the present invention, the specific steps of S1 include: S101: Collect the starting moment of the tension change, the fluctuation duration, and the steady-state response time of the cable nodes through fiber optic sensors, perform time domain alignment on the multi-node time series, remove abnormal sampling points caused by environmental noise, and generate a time synchronization data set. S102: Based on the time synchronization data set, establish a normalization processing model for the node tension parameters, use the Min-Max normalization method to eliminate the dimension difference, map the original sampling values to a unified interval, and generate a normalized tension sequence. S103: Call the normalized tension sequence, input it into the sliding window algorithm to calculate the extreme value, variance value, and slope change rate of the tension within each window, splice the window indicators to construct a multi-dimensional feature matrix, calculate the similarity value between adjacent window feature vectors, and generate an effective response time window when the similarity of consecutive several windows exceeds the similarity threshold. The sliding window algorithm sets the window width to 5 sampling periods and the step size to 1 period. The similarity threshold determines whether it belongs to the same response time period by judging the feature similarity of two adjacent windows.

[0008] As a further solution of the present invention, the specific steps of S2 include: S201: Based on the effective response time window, use the time series overlapping algorithm to compare the time window frame sequences of adjacent nodes, identify the time difference between the start and end time tags, screen the time slices with an absolute difference less than the synchronous error tolerance threshold, and generate the node overlapping time interval. The synchronous error tolerance threshold is set according to the node clock synchronization accuracy and the tolerance standard. S202: Invoke the node overlapping time interval, extract the displacement velocity sampling point sequence within the interval, calculate the rate of change of velocity based on the time difference and velocity difference of the sampling points, extract the absolute value of the regression slope, and generate the displacement velocity change coefficient. S203: Combine the displacement velocity change coefficient and the node overlapping time interval, calculate the start and end boundaries adjusted by the time extension and compression values, invoke the dynamic boundary compensation algorithm to repair the boundary gap, and generate a multi-node collaborative time window.

[0009] As a further solution of the present invention, to calculate the rate of change of velocity, the formula is used: ; where represents the rate of change of velocity parameter value of the th node, with the unit of m / s 2 , represents the velocity value of the th node at the th sampling point, with the unit of m / s, represents the velocity value of the th node at the th sampling point, with the unit of m / s, represents the time difference between the th and the th sampling points of the th node, with the unit of s, represents the average velocity value of the th node at the sampling points, with the unit of m / s, is the weighted adjustment coefficient set by the th node based on the number of sampling points and sensitivity, is the time correction parameter set by the th node according to the sampling delay, with the unit of s, represents the number of pairs of sampling points participating in the calculation within the current interval.

[0010] As a further solution of the present invention, the specific steps of S3 include: S301: Based on the multi-node collaborative time window, select the mass distribution coefficient, damping ratio, and stiffness matrix, combine the triaxial acceleration and the change in velocity, calculate the node inertial force difference rate and establish an inertial change mapping matrix, and perform point-by-point difference calculation by invoking the offset of its adjacent nodes to generate an inertial response difference sequence. S302: According to the inertial response difference sequence, invoke the mass distribution coefficient as the distribution reference, combine the vertical acceleration value and the node response change rate, calculate the node response offset degree using the response change distribution function, and make a judgment in combination with the offset judgment threshold to obtain the response offset identification interval. S303: For the response offset identification interval, combine the vertical acceleration change value of the suspension point and the node stiffness term, calculate the response offset change rate, and judge it node by node with the trigger change threshold to generate an equipment action trigger window; The upper and lower bounds of the trigger change threshold are set according to the response rate fluctuation range of the node in the normal state.

[0011] As a further solution of the present invention, the node response offset degree is calculated using the formula: ; where represents the response offset degree of the j-th node, represents the average mass distribution coefficient of the j-th node within the current time window, with the unit of kg·m, represents the average vertical acceleration value of the j-th node within the current time window, with the unit of m / s 2 , represents the average velocity change amount of the j-th node within the current time window, with the unit of m / s, represents the average stiffness coefficient of the j-th node within the current time window, with the unit of N·m, represents the average acceleration change value of the j-th node within the current time window, with the unit of m / s 2 , represents the average inertial force difference value of the j-th node within the current time window, with the unit of N, is the node inertia offset dynamic factor, is the response compensation coefficient, a dimensionless response index generated based on the node structure damping ratio and the suspension point acceleration change, is the acceleration change standard deviation with the unit of m / s 2 .

[0012] As a further solution of the present invention, the specific steps of S4 include: S401: Based on the equipment action trigger window, detect the tension change situation between adjacent devices, collect the continuous response data of the tension sensor at the starting moment of the equipment action, extract the starting moment of the leading peak in the tension waveform of each pair of devices as the conduction starting point, calculate the time difference of the leading peak propagation to adjacent devices among multiple groups of devices, and generate the tension response time difference; S402: Call the tension response time difference, combine the three-axis displacement monitoring data before and after the equipment suspension point action, intercept the suspension point displacement sequence within the time window corresponding to the tension conduction starting point, calculate the maximum displacement change amount between adjacent suspension points, and compare it with the tension response time difference interval to obtain the suspension point dynamic displacement difference; S403: Input the dynamic displacement difference and the tension response time difference of the suspension points into a dynamic threshold function that self-matches and adjusts based on the real-time working conditions, extract the extreme point intervals with concentrated fluctuation frequencies in the two sets of data, and perform a fusion operation to obtain the synchronous trigger condition.

[0013] As a further solution of the present invention, the specific steps of S5 include: S501: Based on the equipment displacement values recorded in the synchronous trigger condition, calculate the displacement difference between the equipment at the current moment, compare this difference with the synchronous tolerance range, screen the time nodes where the differences all fall within the tolerance range, and generate a synchronous time screening interval; S502: Call the time nodes recorded in the synchronous time screening interval, extract the equipment timestamp values, convert them to a unified reference time by linear interpolation, and reconstruct them in combination with the equipment numbers to generate an equipment timestamp alignment sequence; S503: According to the reference time in the equipment timestamp alignment sequence, read the tension change value, calculate the first-order time difference to obtain the change rate, and encapsulate it with the equipment number and time point to generate a multi-equipment synchronous control instruction.

[0014] On the other hand, a multi-machine collaborative control system for a bridge cable hoisting device is provided. The multi-machine collaborative control system for the bridge cable hoisting device is used to execute the above multi-machine collaborative control method for the bridge cable hoisting device. The system includes: A tension monitoring module, which is used to obtain tension data through a sensor array, record the starting time of change, the duration of fluctuation, and the stabilization time, call the sliding window algorithm to normalize the tension sequence, extract the abnormal intervals and cluster them, output an effective response time window, and transfer it to the window extraction module; A window extraction module, which is used to receive the effective response time window, call the time series overlapping algorithm to calculate the multi-node intersection, construct a trend function based on the displacement velocity derivative, perform a fine-tuning operation on the intersection region boundary, obtain a multi-node collaborative time window, and transfer it to the collaborative recognition module; A collaborative recognition module, which is used to receive the multi-node collaborative time window, obtain the suspension point acceleration and node inertia parameters, call the inertia compensation model to perform a collaborative operation to adjust the node response amplitude, screen the time periods that reach the response rate threshold, and output an equipment action trigger window, and transfer it to the action compensation module; An action compensation module, which is used to receive the equipment action trigger window, obtain the tension conduction time and the suspension point displacement difference, perform reduction and interval fitting, call the dynamic threshold function model to screen the extreme values and apply threshold constraints, output a synchronous trigger condition, and transfer it to the synchronous instruction module; The synchronization instruction module is used to receive the synchronization trigger condition, call a laser rangefinder to judge the node displacement difference limit value, collect the device number, trigger time, and tension change rate according to the control error limit domain, and obtain a multi-device synchronization control instruction through encapsulation by the encoding module.

[0015] The beneficial effects brought by the technical solution provided by the embodiments of the present invention at least include: By using a distributed sensor array to collect the starting moment, fluctuation duration, and steady-state response time of the tension change of multiple nodes, extracting characteristic indexes through normalization processing and a sliding window algorithm, constructing a dynamic response time window, adjusting the window boundary by combining the time series overlapping algorithm with the displacement speed change trend, realizing the dynamic matching of the multi-device action trigger window, participating in the inertial compensation operation with the structural inertia parameters and the suspension point acceleration data, reducing the influence of the mechanical structure dynamic response lag, real-time detecting the tension conduction time difference and displacement difference, performing extreme value screening and interval fusion through a dynamic threshold function, generating a synchronization trigger condition adapted to the working condition change, and combining the timestamp alignment mechanism with the data encapsulation technology to ensure the spatio-temporal synchronization of multi-device instructions, and improving the synchronization accuracy and dynamic coordination ability of multi-machine cooperation under complex load conditions. Description of the Drawings

[0016] Figure 1 It is a schematic diagram of the working process of the present invention; Figure 2 It is a system flowchart of the present invention. Detailed Embodiments

[0017] The technical solutions in the present invention will be described below with reference to the drawings.

[0018] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations, or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "example" is intended to present concepts in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0019] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meaning they express is the same. "(of)", "corresponding", and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meaning they express is the same.

[0020] In the embodiments of the present invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0021] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0022] Please refer to Figure 1 , the embodiments of the present invention provide a multi-machine cooperative control method for a bridge cable hoisting device. The processing flow of this method may include the following steps: S1: Collect the starting moment of the tension change, the duration of the tension fluctuation, and the steady-state response time of the cable nodes through an optical fiber sensor array, perform normalization processing, input the sliding window algorithm to extract the tension response characteristics, and generate an effective response time window; S2: Based on the effective response time window, calculate the intersection of the node windows through the time series overlapping algorithm, and adjust the time window boundary in combination with the displacement speed change trend of the device to output a multi-node cooperative time window; S3: According to the multi-node cooperative time window, combine the structural inertia parameters and the suspension point acceleration data, and use the inertial compensation operation to calculate the structural response compensation value to generate a device action trigger window; S4: Based on the device action trigger window, detect the tension conduction time difference and the suspension point displacement difference between adjacent devices, input them into the dynamic threshold function that self-matches according to the real-time working conditions, and perform extreme value screening and interval fusion operations to output a synchronization trigger condition; S5: According to the synchronization trigger condition, perform timestamp alignment operation. When the displacement difference is within the tolerance range, encapsulate the device number, trigger moment, and tension change rate through instruction encoding to output a multi-device synchronization control instruction; The effective response time window includes the tension change rate, the fluctuation duration, and the steady-state time constant. The multi-node cooperative time window specifically refers to the time window overlap degree, the displacement speed variance, and the cooperative convergence threshold. The device action trigger window includes the inertial compensation factor, the acceleration deviation threshold, and the phase synchronization margin. The synchronization trigger condition specifically refers to the peak value of the conduction time difference, the standard deviation of the displacement difference, and the dynamic fusion threshold. The multi-device synchronization control instruction includes the device identifier, the trigger timestamp, and the tension change rate sequence; The cooperative convergence threshold refers to the critical value when the multi-node time window reaches stability after adjustment; The dynamic fusion threshold is a fusion boundary value calculated by the real-time working condition adaptive algorithm based on multiple indicators such as the tension conduction time difference and the suspension point displacement difference.

[0023] Specifically, the steps of S1 are as follows: S101: Fiber optic sensors are used to collect the starting time of tension changes, fluctuation duration, and stabilization response time of cable nodes. Time domain alignment is performed on multi-node time series to eliminate abnormal sampling points caused by environmental noise and generate a time-synchronized dataset. The starting moment of tension change is when the tension change exceeds 5% of the static mean and lasts for more than 0.2 seconds. The duration of tension fluctuation is when the tension fluctuation amplitude is less than 10% of the initial fluctuation amplitude and lasts for at least 0.3 seconds. The fiber optic sensor is used to collect the starting time of the tension change, the duration of the fluctuation, and the stabilization response time of the cable node. First, a distributed fiber optic sensing unit is deployed on each node, and the tension change curve is recorded at intervals of 0.1 seconds. The static tension mean value during the initial monitoring period is set to The criteria for determining the onset of tension change and the duration of fluctuation are based on the dynamic response characteristics of the cable structure and the error tolerance commonly used in engineering. They are set as tension exceeding the static mean by 5% and continuously for more than 0.2 seconds, and fluctuation amplitude being 10% lower than the initial disturbance and continuously for more than 0.3 seconds, respectively, to ensure the identification of effective disturbance and stable state. By comparing the tension value at each sampling moment to see whether it is the first time that it is continuously higher than the static mean by 5%, that is, , and whether the duration exceeds 0.2 seconds, that is, 3 or more points in a row meet the above tension conditions, and determine the starting time of the tension change. For example, the tension values from the 3rd to the 5th second in the table below are 104.5N, 106.2N and 105.7N respectively. After meeting the conditions, the starting time of the tension change is determined to be 0.3 seconds. Then calculate the fluctuation duration in sequence, and find whether the tension change amplitude gradually decreases from the starting time backward. When the tension fluctuation amplitude drops to less than 10% of the initial fluctuation amplitude, for example, in this example, the initial fluctuation amplitude is , 10% is 0.75N. When the tension value fluctuation between three consecutive sampling points is lower than 0.75N, for example, the tension values at points 6 to 9 are 103.9N, 101.2N, 99.5N, and 98.9N respectively, and the maximum change amount between them is less than 0.75N, then from 0.6 seconds to 0.9 seconds, the condition of becoming stable is satisfied. Thus, the fluctuation duration can be determined as 0.6 seconds. The steady-state response time is the total duration from after the tension change to before becoming stable, that is, from 0.3 seconds to 0.9 seconds. After obtaining the initial point, termination point, and response interval of the node time series, horizontal time-domain alignment is performed. By comparing the starting offset of the tension change points of the nodes, a unified starting reference point is set. For example, the starting time is uniformly set to 0 seconds. The linear interpolation or time offset method is used to synchronize the node tension sequences. During the environmental noise elimination process, referring to the upper and lower limits of the tension fluctuation during the initial steady state, the standard deviation is set as σ. When a sampling point deviates from the mean value of the nearest three points by more than 3σ, for example, the standard deviation of the starting point is 0.3N, then the threshold is ±0.9N. At the time point, it is compared whether the tension value exceeds the upper and lower limits. For example, if the sampling point is 102.5N and the mean value of its corresponding adjacent points is 99.0N, then the deviation is 3.5N > 0.9N, and this point is excluded. The above operations are repeated until the establishment of the time-domain synchronization data set for all nodes is completed.

[0024] Table 1: Tension Monitoring Data Table

[0025] As shown in Table 1, the tension value starts to increase significantly from 0.3 seconds during the sampling period and gradually stabilizes at 0.9 seconds. It is calculated that the starting moment of the tension change is 0.3 seconds, the fluctuation duration is 0.6 seconds, and the steady-state response time is 0.6 seconds during the period from 0.3 seconds to 0.9 seconds. Overall, the time synchronization data set corresponding to the node tension change is obtained.

[0026] S102: Based on the time synchronization data set, establish a normalization processing model for node tension parameters. The Min - Max normalization method is used to eliminate the dimension difference, map the original sampling values to a unified interval, and generate a normalized tension sequence; Based on the time synchronization data set, establish a normalization processing model for node tension parameters. First, the tension data of each node are reorganized according to a unified time-domain sequence, and the tension sampling values at each time point after alignment of the nodes are extracted to form a node tension time series matrix. Subsequently, the minimum and maximum values of the tension sequence of each node are determined as the benchmark boundaries for normalization. For example, the tension value range of a certain node within 0 to 0.9 seconds is 98.5N to 106.2N, then the normalization lower limit of this node is set to 98.5, and the upper limit is set to 106.2. Then, the normalization calculation is performed on the tension value at each time point. The tension value is subtracted by the minimum value and then divided by the difference between the maximum value and the minimum value to obtain the normalization result. For example, if the tension value is 104.5N, then its normalization value is , the normalization operation of the sampling points is completed in sequence to obtain a standardized tension sequence. In the case where there are different tension ranges at multiple nodes, the differences between absolute values are eliminated through the normalization operation, so that the tension sequences are all compressed into the corresponding intervals between 0 and 1, forming a unified data input format suitable for subsequent sliding window analysis. During this process, if an abnormal high value or low value occurs at a certain node due to environmental disturbances, resulting in the stretching of the maximum and minimum boundaries, thereby compressing the main data distribution range, this can be addressed by analyzing the distribution characteristics of the tension sequence of each node, excluding extreme values that deviate from the mean ±3σ and resetting the upper and lower boundary limits. For example, if the tension at a certain point in the initial data is 120.0N, which is significantly higher than the main distribution range of 99.0N to 106.2N of adjacent sampling points and deviates more than 3 times the standard deviation from the mean, it is excluded, and the minimum and maximum values are recalculated, thereby generating a stable and effective standardized tension sequence.

[0027] S103: Call the standardized tension sequence and input it into the sliding window algorithm to calculate the extreme difference value, variance value, and slope change rate within each window. Concatenate the window metrics to construct a multi-dimensional feature matrix, and calculate the similarity value between the feature vectors of adjacent windows. When the similarities of several consecutive windows all exceed the similarity threshold, an effective response time window is generated; The sliding window algorithm sets the window width to 5 sampling periods and the step size to 1 period; The similarity threshold determines whether it belongs to the same response time period by judging the feature similarity of two adjacent windows; Call the standardized tension sequence and input it into the sliding window algorithm. Set the window width to 5 sampling periods and the step size to 1 period. Sequentially extract the tension sequence data segments within each window and calculate the difference between the maximum value and the minimum value as the range. Calculate the mean of the sum of the squared differences between the sampling values and their mean as the variance. At the same time, divide the tension difference between the end point and the start point within the current window by the time interval as the linear slope, and perform a division operation on the time step by the slope difference between the two adjacent windows to obtain the slope change rate. For example, if the 5 tension values within a certain window are 0.15, 0.22, 0.35, 0.40, and 0.50, then the range is 0.50 - 0.15 = 0.35, the mean is 0.324, the variance is the average of the squared differences between each value and the mean, and the calculated value is approximately 0.0189. The slope is (0.50 - 0.15) / 0.4 = 0.875. Subsequently, after moving one sampling period, the window data is updated, and the characteristics of the new window are recalculated. In this way, a feature matrix containing multi-dimensional indicators such as range, variance, and slope change rate is formed. Calculate the similarity pairwise between the feature vectors, which can be numerically expressed through the cosine angle formula. If the similarity of three or more consecutive adjacent windows exceeds the preset threshold, for example, the set threshold is 0.96, indicating a very high similarity, indicating that the consecutive windows are formed during the same response process. Under this condition, mark this segment of the window as the effective response time window. The threshold setting process can be carried out through the similarity comparison analysis between the response intervals and non-response intervals in the previous training data to screen the critical value that can effectively distinguish the two states. The specific method is to calculate the average similarity and variance of the window pairs in multiple disturbance events to find the optimal threshold point that can minimize the misrecognition rate. For example, in 50 groups of samples, when the threshold is set to 0.96, the response recognition rate reaches 97%, and the misrecognition rate is lower than 2%. Then, this threshold can be fixed as the judgment standard, and finally, the time window marks that meet the continuous similarity exceeding the threshold are obtained and output as the effective response time period.

[0028] Specifically, the steps of S2 are as follows: S201: Based on the effective response time window, use the time series overlapping algorithm to compare the time window frame sequences of adjacent nodes, identify the time difference between the start and end time tags, screen the time slices with an absolute difference in intervals less than the synchronous fault tolerance threshold, and generate the node overlapping time interval; The synchronous fault tolerance threshold is set according to the node clock synchronization accuracy and tolerance standard; Based on the effective response time window, first set the synchronization start time period of multiple nodes in the system. For example, set the communication sampling interval between the main control node and the slave node of the lifting equipment to 10 ms, establish a collection time window range of [0 ms, 5000 ms], and obtain the time frame sequence data of each node within this window range. Each time frame records control quantities such as displacement and speed collected at that moment. Subsequently, extract the node time frame sequences within the current window segment every 1 ms from the start time in a sliding window manner, and use cross-comparison operations to perform paired analysis on the time frame sequences of two adjacent nodes A and B. During the specific execution process, for the nth frame time tag in node A as , find the frame in node B with the smallest absolute value of the time difference from it , calculate the time difference , and extract the set of frame pairs that satisfy from the paired data, where is the synchronization fault tolerance threshold. This threshold is set to 10 ms with reference to the clock synchronization accuracy, and at the same time, refer to the system tolerance standard such as the allowable error of ±5% for the tension adjustment of the bridge steel cable to convert it into the acceptable error range corresponding to the time difference, forming the first group of screened time segments. By comparing multiple frame pairs one by one in the above manner and marking the timestamp combinations that meet the conditions, further extract the start and end time tags of consecutive time frames that meet the pairing conditions as the overlapping intervals. Judge the length of the overlapping intervals. If the time span is less than 200 ms, then discard it to avoid the influence of short-time disturbances on the results, and retain the overlapping segments with a duration exceeding 200 ms, forming multiple groups of effective overlapping intervals. Finally, output the overlapping intervals as the time alignment reference area between nodes for subsequent extraction of speed and displacement change data. Taking the example of the simultaneous lifting of the main cable and the auxiliary cable of the bridge, when the sampling time point of the main cable node is 1000 ms and the sampling time point of the auxiliary cable node is 1005 ms, the time difference between the two is 5 ms, which is lower than the 10 ms threshold, then this group of time frames is considered to be part of the overlapping time interval. By comparing multiple similar frame pairs, [1000 ms, 1400 ms] can be extracted as a complete overlapping interval, which is used for sampling data extraction and analysis processing in subsequent steps.

[0029] Table 2: Example sampling frame data table during the overlapping interval extraction process

[0030] As shown in Table 2, through the judgment condition that the time difference between the timestamps of nodes A and B is less than the synchronization fault tolerance threshold of 10 ms, multiple frame pairs that meet the requirements are successfully extracted, and the corresponding time period is the node overlapping time interval.

[0031] S202: Invoke the node overlapping time interval, extract the displacement velocity sampling point sequence within the interval, calculate the rate of change of velocity based on the time difference and velocity difference of the sampling points, extract the absolute value of the regression slope, and generate the displacement velocity change coefficient. Invoke the node overlapping time interval. First, clarify the start and end boundary time points of each overlapping time interval. Taking the interval of the main cable and auxiliary cable of the node from 1000ms to 1400ms as an example, if the sampling interval is set to 10ms, a total of 41 groups of velocity sampling data can be extracted, numbered sequentially as to , the node number is set to , and the corresponding velocity value array is . A velocity change unit is formed between every two adjacent velocity values, and the time difference is , and the velocity difference is . It is necessary to calculate the total change amount term in the velocity change trend. Use , where . The product of each velocity difference and the time difference gives the velocity increment contribution. For example, assuming some sampling values are as follows: , then the first two velocity differences are 0.02 and 0.03, and the time difference is 0.01s for both, and the corresponding products are and . Calculate the cumulative result of multiplying the velocity differences by the time differences for 40 groups in this way, denoted as term A; calculate the velocity fluctuation amplitude term. First, obtain the average velocity , which is the arithmetic average of the velocity values. If the velocity sequence is , then . Then calculate the sum of the squares of the absolute values of the deviations of the velocity from the average velocity . If some terms are 0.12, 0.10, 0.07, etc., and the total deviation sum is 0.45, take the square root after squaring to get approximately 0.67, and multiply by the adjustment coefficient to form the correction term B. Let , then term B is ; calculate the total time denominator term, which is , that is . Let , and the total sum is 0.45s; calculate the rate of change of velocity using the formula: ; Among them, represents the rate of change of velocity parameter value of the rd node, with the unit of m / s2, represents the velocity value of the th node at the th sampling point, with the unit of m / s, represents the velocity value of the th node at the th sampling point, with the unit of m / s, Represents the th node at the and the th sampling point, with the time difference in seconds, Represents the th node's average velocity at the sampling point, in m / s, Is the th node's weighted adjustment coefficient set based on the number of sampling points and sensitivity, Is the th node's time correction parameter set according to the sampling delay, in seconds, Represents the number of pairs of sampling points participating in the calculation within the current interval.

[0032]

[0033] Precisely extract the change in the node velocity response. Subsequently, with the time point as the independent variable and the corresponding velocity as the dependent variable, calculate its linear regression slope after normalization. Let the time value sequence be , and the velocity value be , its average time , average velocity , then the covariance term is the sum of each accumulation. Let the partial sample data be as follows: , , and the corresponding product term is etc. The total covariance is approximately 3200, and the time variance is , if it is 20000, then the slope , convert the unit to , and its absolute value is the displacement velocity change coefficient for this section . According to the interval division standard, less than 0.001 is low variation, 0.001 - 0.003 is medium variation, and higher than 0.003 is high variation. Therefore, it is determined that this overlapping section is a medium variation section for subsequent boundary adjustment.

[0034] S203: Combine the displacement velocity change coefficient and the node overlapping time interval, calculate the time extension and compression values to adjust the start and end boundaries, call the dynamic boundary compensation algorithm to repair the boundary gap, and generate a multi-node collaborative time window; Based on the obtained velocity change rate parameter and the displacement velocity change coefficient , continue to perform dynamic time window adjustment processing on the overlapping time interval [1000ms, 1400ms]. First, according to Judge the change level of the belonging interval. The current value is in the medium change segment (0.001–0.003). According to the set strategy, the time window is extended bidirectionally, and the extension ratio is 5%. Therefore, the original interval length is , and the corresponding extension amount is . Extend both ends of the time window 20 ms forward and backward respectively to form a preliminarily adjusted time window . Next, it is necessary to judge whether the current extended window overlaps with the adjacent time period. Suppose the previous overlapping interval is [920 ms, 990 ms], and the minimum gap between the two intervals is . A negative value indicates that the two windows overlap, which does not meet the minimum gap threshold setting (30 ms). Therefore, it is necessary to correct the left starting point of the current window. The specific processing is to adjust the starting point of the left boundary to be 30 ms away from the right boundary of the previous window, that is, the left starting point is updated to . However, considering the real-time situation is data extension, it is only necessary to roll back forward until there is no overlap, that is, the left end of the current window is corrected to 990 ms, and the temporarily adjusted window is [990 ms, 1420 ms].

[0035] Similarly, check whether the right side of the current window overlaps with the subsequent window. Suppose the subsequent interval is [1430 ms, 1600 ms], and the interval is . It still does not meet the minimum gap condition, and the right end of the current window needs to be shrunk to . Finally, the corrected effective time window is .

[0036] The above correction mechanism will be executed cyclically in the overlapping section to ensure that there is no data coverage or boundary conflict between time windows, and to ensure the continuity and independence of multi-segment data. After the adjustment is completed, the dynamic window is numbered and marked, and the information such as the corresponding node number, time interval, displacement velocity change coefficient and its change level of each segment is stored in a structured manner, providing basic data support for subsequent data alignment analysis based on the window and multi-node behavior evolution recognition.

[0037] Specifically, the steps of S3 are as follows: S301: Based on the multi-node collaborative time window, select the mass distribution coefficient, damping ratio and stiffness matrix, combine the triaxial acceleration and velocity change amount, calculate the node inertial force difference rate and establish an inertial change mapping matrix, and call the offset of its adjacent nodes for point-by-point difference calculation to generate an inertial response difference sequence; Based on the multi-node collaborative time window, for each node, it is necessary to first extract the basic information of its mass distribution coefficient, damping ratio and stiffness matrix. Suppose the mass distribution coefficient of the first node is 1.2 kg, the damping ratio is 0.04, the main diagonal element in the stiffness matrix is set to 180 N / m, and the secondary diagonal element is set to 80 N / m. The selected time point sequence is , the three-axis accelerations are respectively denoted as , with the unit of m / s². Combining the obtained velocity change rate parameter , for each time sampling point, the current acceleration and the velocity difference at the adjacent time point are respectively extracted, and the execution process of calculating the node inertial force difference rate is as follows: First, call the mass value of 1.2 kg of the current node and the sampled acceleration value at the current moment. For example, at 1020 ms, , the inertial force is calculated as . Subsequently, at the same time point, call the velocity change rate value of 0.615 m / s², convert it to the velocity increment at the corresponding moment and the velocity at the previous moment to calculate the inertial response change value. For example, the velocities of this node at 1010 ms and 1020 ms are 0.83 m / s and 0.85 m / s respectively, and the difference is 0.02 m / s. At a time interval of 0.01 s, the corresponding inertial change force is . Then, compare the inertial response differences obtained by the two calculation methods, and the inertial force difference rate at the current time point is , that is, 66%. Record this difference value into the difference rate sequence, and complete the repeated calculation for each sampling time point in sequence. After generating the complete difference rate sequence, compare the inertial response differences at the same time point of its adjacent nodes (such as node 2, with a mass of 1.1 kg, the corresponding acceleration is 0.64 m / s², and the velocity change rate is 0.58 m / s²) for point-to-point difference. If the inertial force of node 2 at the same moment is , the corresponding inertial change is , then its difference rate is , and the difference value of the difference rates between the two nodes is . Construct the difference results at the time points into an inertial response difference sequence, and the results are shown in the following table: Table 3: Inertial Response Difference Sequence Table

[0038] As shown in Table 3, through the calculation of the inertial rate difference, the inertial response differences of the differentiated nodes at the same time point can be intuitively reflected.

[0039] S302: According to the inertial response difference sequence, call the mass distribution coefficient as the distribution benchmark, combine the vertical acceleration value and the node response change rate, use the response change distribution function to calculate the node response offset degree, and combine the offset judgment threshold to make a judgment to obtain the response offset recognition interval; Based on the generated inertial response difference sequence, taking node 1 in the time window [1000 ms, 1400 ms] as an example, first extract the parameters within the current window of this node: the average mass distribution coefficient , the average vertical acceleration , the average velocity change , average stiffness coefficient , average acceleration change value , average inertial force difference value , inertial displacement dynamic factor (obtained by looking up the table according to the node damping ratio of 0.04 and the acceleration fluctuation range of 0.1 - 0.2 m / s²), response compensation coefficient (calculated from the damping ratio of 0.04 and the standard deviation of acceleration change as , but corrected to 0.5 according to engineering experience), standard deviation of acceleration change , calculate the response offset degree, using the formula: ; Wherein, represents the response offset degree of the j-th node, represents the average mass distribution coefficient of the j-th node within the current time window, with the unit of kg·m, represents the average vertical acceleration value of the j-th node within the current time window, with the unit of m / s 2 , represents the average velocity change of the j-th node within the current time window, with the unit of m / s, represents the average stiffness coefficient of the j-th node within the current time window, with the unit of N·m, represents the average acceleration change value of the j-th node within the current time window, with the unit of m / s 2 , represents the average inertial force difference value of the j-th node within the current time window, with the unit of N, is the node inertial displacement dynamic factor, is the response compensation coefficient, a dimensionless response index generated based on the node structural damping ratio and the acceleration change at the suspension point, is the standard deviation of acceleration change with the unit of m / s 2 .

[0040] Substitute into the formula for calculation: ; Compare the calculation result with the offset judgment threshold interval [2.0, 5.0]. 7.60 significantly exceeds the upper threshold, determining that there is a response offset at this node within the time window. It is necessary to mark it as an abnormal section. By calculating and summarizing the offset degree sequence node by node, a response offset identification interval list is generated for subsequent device action trigger judgment.

[0041] S303: For the response offset identification interval, combine the vertical acceleration change value of the suspension point and the node stiffness term, calculate the response offset change rate and judge it node by node with the trigger change threshold to generate a device action trigger window; The triggering change threshold is set for the upper and lower bounds according to the fluctuation range of the response rate of the node in the normal state; Based on the identified response offset interval, taking node 1 as an example, it is identified as having a response offset within the time window from 1050 ms to 1080 ms. It is necessary to further determine whether it meets the device action triggering condition. First, collect the vertical acceleration change value of the suspension point during this time period. Taking the real-time measurement data as an example, the vertical acceleration of the suspension point is 0.70 m / s² at 1050 ms and 0.90 m / s² at 1080 ms. After three-point sampling, the middle moment at 1070 ms is 0.84 m / s². Calculate the acceleration change value of the suspension point as the difference between the maximum value and the minimum value, that is , and the stiffness term of this node is known from the previous section to be 180 N / m. So the corresponding response offset change rate is the product of the stiffness term and the acceleration change value, and the calculated result is , compare with the preset triggering change threshold setting value of 30 N / s (its setting is based on the up and down fluctuation range of the original normal response rate from 20 to 28 N / s, and the maximum value is set to 30 N / s after expanding the 10% safety margin upward). Since 36.0 is greater than the threshold, record that the current node triggers a response action. The time point is set to the starting point 1050 ms of the offset interval and continues until the offset rate drops below the threshold. Then perform the same processing on node 2. Its acceleration change value rises from 0.69 m / s² at 1060 ms to 0.87 m / s² at 1080 ms, and the change amount is 0.18 m / s². The node stiffness is 160 N / m, and the corresponding response change rate is calculated as , which is less than 30 N / s and does not meet the triggering condition. Do not record this node as an object in the action window. Perform the above process on all nodes in turn to screen the triggered nodes and their corresponding triggering time periods, and finally construct a list of device action triggering windows; In addition, to verify the stability of the judgment boundary, further analyze the setting basis of the triggering threshold. Set the continuous monitoring response rate value sequence of a certain node in the non-disturbed state as N / s, compare its maximum value of 27.2 with the average value of about 24.1. On this basis, set 30 N / s as the safety boundary value to ensure that when the response rate exceeds this threshold, the device response record is triggered. Finally, the nodes and time periods that meet the above calculation criteria and occur within the offset section are recorded as device action triggering windows and are used for subsequent device control logic judgment.

[0042] Specifically, the steps of S4 are as follows: S401: Based on the device action triggering window, detect the change in tension generated between adjacent devices, collect the continuous response data of the tension sensor at the starting moment of the device action, extract the starting moment of the leading peak in the tension waveform of each pair of devices as the conduction starting point, calculate the time difference of the leading peak propagation between multiple pairs of devices to adjacent devices, and generate the tension response time difference; Based on the device action trigger window, first, taking the device response triggered by node 1 at 1050 ms as an example, collect the original voltage output sequence of the tension sensor between the devices it is connected to in the time period from 1050 ms to 1080 ms. Record the tension voltage value corresponding to each sampling moment, and convert the voltage to tension values. Use the factory calibration parameters of the sensor. For example, the sensor sensitivity is 2.5 mV / V, the bridge voltage is 5 V, and the conversion ratio coefficient is 2.5×5 = 12.5 N / V. Taking the collected voltages of 0.28 V, 0.33 V, 0.39 V, 0.35 V, 0.29 V, etc. as examples, the corresponding tension values are 3.5 N, 4.13 N, 4.88 N, 4.38 N, 3.63 N respectively. Determine the peak interval through the continuous tension curve, and extract the starting moment of the leading peak as the first peak rising point that exceeds the mean plus twice the standard deviation. For example, in the current sampling section, the tension mean is 4.0 N and the standard deviation is 0.3 N, then the judgment threshold is 4.0 + 2×0.3 = 4.6 N. The corresponding rising section is that the tension value of 4.88 N converted from 0.39 V first exceeds this threshold, and record its moment as 1056 ms as the starting point of the leading peak; in the same time window, obtain the tension response sequences of adjacent node devices, perform the same voltage-tension conversion, and extract the starting point of its leading peak through the mean + 2 times standard deviation method. For example, the tension values of adjacent nodes are 3.2 N, 3.9 N, 4.1 N, 4.7 N, 4.3 N, its mean is 3.84 N, and the standard deviation is 0.53 N, then the threshold is 4.9 N, and the maximum value of 4.7 N does not exceed it, so it is judged that no clear leading peak is detected, and there is no record of the starting point of the leading waveform for this node. Then calculate the response propagation time difference between the nodes with the starting point of the leading peak. Taking the leading peak time of 1056 ms of node 1 as the reference, if the adjacent node does not record the leading peak, skip the calculation of this group of propagation differences, and continue to execute the above voltage conversion and peak starting point extraction process for the subsequent adjacent devices. When it is found that multiple nodes have the starting point of the peak, record their starting times respectively and sort them in the order of nodes to generate a propagation time difference sequence. If the starting time of node 3 is 1059 ms, then its propagation time difference from node 1 is 1059−1056 = 3 ms, and so on to complete the calculation of the propagation time difference between multiple groups of devices, and obtain the tension response time differences as shown in the following table: Table 4: Table of Tension Response Time Differences between Device Nodes

[0043] As shown in Table 4, no effective propagation time difference is generated between node 1 and node 2 because no clear tension wave peak rising point of the adjacent device is detected. There are propagation differences of 3 ms and 6 ms between node 3 and node 4 and node 1 respectively. This response time difference is used in the subsequent dynamic displacement change analysis step, and it is necessary to further superimpose the lifting point action data for judgment.

[0044] There is no explicit formula in this paragraph, but the core calculation logic of the tension response time difference is based on the time difference operation between the starting points of the wave crests, that is ; Among them, is the time of the first rising point of the leading wave crest in the tension curve of the reference node (such as node 1), is the time of the first rising point of the corresponding leading wave crest of its adjacent nodes (such as node 3 and node 4). If the adjacent nodes do not meet the condition of "peak value higher than the mean plus 2 times the standard deviation", it is determined that there is no clear leading peak, and the response time difference is recorded as zero. If the condition is met, the time subtraction operation is directly performed, and the result value is in milliseconds (ms). This calculation method can be widely applied to the tension response data sequence with the same sampling period to ensure a unified standard for time difference determination.

[0045] The key parameter "standard deviation threshold factor" in the above processing process takes a value of 2, which comes from the engineering experience and is often used in the two - standard - deviation method for identifying the peak value fluctuation of the response. The setting basis is the stability requirement of the tension fluctuation range, and the conversion coefficient of the tension sensor should be obtained by converting based on the factory calibration parameters of the sensor. It is not allowed to directly introduce assumed values. For example, if the sensor sensitivity is 2.5 mV / V and the bridge voltage is 5 V, then the tension value corresponding to each 1 V voltage is N / V, and when the voltage is 0.39 V, the tension value is N, and the accurate value participates in the subsequent determination operation.

[0046] This result shows that when calculating the tension response time difference, only when the equipment node has a clear starting point of the tension wave crest can an effective propagation time difference be obtained. If the adjacent equipment does not form a wave crest signal with sufficient intensity, it should be excluded in the propagation judgment to avoid forming an interference response propagation path.

[0047] S402: Call the tension response time difference, combine the three - axis displacement monitoring data before and after the movement of the equipment hanging point, intercept the hanging point displacement sequence within the time window corresponding to the starting point of tension conduction, calculate the maximum displacement change between adjacent hanging points, and compare it with the tension response time difference interval to obtain the dynamic displacement difference of the hanging point; After calling the aforementioned tension response time difference, taking the propagation time difference of 3 ms between node 1 and node 3 as an example, first, starting from the starting time of 1056 ms of the leading peak of node 1 tension, intercept the time interval covered by this propagation time difference, that is, the hanging point three - axis displacement data sequence from 1056 ms to 1059 ms. When the sampling frequency is 1 kHz, a total of 4 sets of displacement data points need to be obtained, which are the X, Y, and Z direction displacement values corresponding to 1056 ms, 1057 ms, 1058 ms, and 1059 ms respectively. For example, the three - axis displacements of the hanging point of node 1 within this time interval are , while the corresponding displacement values of node 3 are , within this time period, perform point-to-point difference calculations on the displacement sequences of two nodes in the X, Y, and Z-axis directions at the same moment, and extract the maximum difference in each direction as the dynamic displacement difference result. The maximum difference in the X-axis direction is , the maximum difference in the Y-axis direction is , the maximum difference in the Z-axis direction is , where the X-axis difference is the maximum dynamic difference and is used as the basis for subsequent judgments; then, according to the propagation time difference of 3 ms, it is necessary to determine whether the dynamic displacement difference is significant. The dynamic displacement significance judgment threshold interval is , as a reference, the current 0.14 m exceeds the upper limit, and it is recorded that there is a significant synchronous displacement change between the nodes; similarly, process the propagation time difference of 6 ms between node 1 and node 4, extract the suspension point three-axis displacement data from 1056 ms to 1062 ms, and perform corresponding difference calculations on the displacement points of node 1 and node 4 within this time window. For example, the three-axis displacement sequence of node 4 is , compare it point by point with node 1, and the maximum X-axis displacement difference is , the Y-axis difference is , the Z-axis difference is , the maximum direction difference is still 0.14 m in the X-axis direction, which also exceeds the upper limit of the judgment interval, and it is recorded that there is also a significant displacement difference between node 1 and node 4. In this way, perform displacement difference extraction and judgment on the node pairs with tension propagation time differences, construct the suspension point dynamic displacement difference sequence of the corresponding node pairs, and perform the next-stage fusion judgment based on this.

[0048] S403: According to the suspension point dynamic displacement difference and the tension response time difference, input them into the dynamic threshold function that self-matches and adjusts according to the real-time working conditions, extract the extreme point intervals where the fluctuation frequencies are concentrated in the two sets of data and perform a fusion operation to obtain the synchronous trigger condition; According to the previously extracted suspension point dynamic displacement difference and the tension response time difference, substitute the two as input items into the dynamic threshold fusion judgment process. Taking the response difference between node 1-3 as an example, the time difference is 3 ms, and the dynamic displacement difference is 0.14 m. First, extract the number of extreme points with the highest change frequency in the X-axis displacement curve between node 1 and node 3 during this time period. For example, within the time window from 1056 ms to 1059 ms, there are two maximum points and one minimum point in the X-axis displacement fluctuation of node 1, and there are three maximum points and one minimum point in the X-axis displacement fluctuation of node 3. Combine the fluctuation extreme value numbers of the two to a total of 7, and define the fluctuation frequency as the number of extreme points divided by the time window length, that is , compare this value with the frequency threshold corresponding to a time difference of 3 ms. The frequency threshold is set to 1.5 per ms. Since the current frequency is higher than this threshold, the first condition is met. Next, perform normalization on the dynamic displacement difference of 0.14 m. Use a normalization coefficient of 0.2 m as the normalization reference value. The normalized value is , and then compare it with the normalized judgment reference value of 0.6. Since 0.7 is higher than 0.6, the second condition is met. Finally, combine the above two judgments as the synchronous trigger determination condition. When both conditions are met, it is determined that there is a synchronous trigger relationship between Node 1 and Node 3 within this time window, and record it as the synchronous trigger node pair and the time period from 1056 ms to 1059 ms. Similarly, process the data between Node 1 and Node 4. The number of extreme points is three for Node 1 and four for Node 4, with a total of seven extreme points within a 6 ms time window. The frequency is , which is lower than the set frequency threshold of 1.5 per ms. It is judged that the synchronous trigger frequency condition is not met. Even though the normalized displacement difference is as high as 0.7, the complete condition cannot be satisfied. Finally, this node pair is not recorded as a synchronous trigger combination. According to this process, perform frequency and displacement difference judgments on each node pair with a propagation time difference one by one, sort out the node pairs that meet the conditions and the corresponding time segments, form the final synchronous trigger condition list, and use it for linkage judgment and action priority setting in further control logic.

[0049] Specifically, the steps of S5 are as follows: S501: Based on the device displacement values recorded in the synchronous trigger conditions, calculate the displacement difference between devices at the current moment. Compare this difference with the synchronous tolerance range, and screen out the time nodes where the differences all fall within the tolerance range to generate a synchronous moment screening interval; Based on the device displacement values recorded in the synchronous trigger conditions, first read the displacement data of the participating nodes in the X, Y, and Z axis directions within the synchronous time period of each node pair. By comparing the displacement values pairwise point by point within this time period, calculate the spatial displacement difference between devices one by one. The calculation method is: calculate the absolute difference of the displacements in the three axis directions at each moment respectively, and obtain the comprehensive displacement difference at this moment through the three-dimensional Euclidean distance formula. For example, for Node 1 and Node 3 at 1056 ms, if their displacements are (0.82, 0.35, -0.10) m and (0.88, 0.36, -0.11) m respectively, then their spatial difference is: [[ID=z12]] [[ID=z13]] ; [[ID=z15]] And so on, process the displacement values at each moment point within the entire synchronous time period, and compare each difference with the set synchronous tolerance range. The tolerance interval is set to [[ID=z17]] , the maximum displacement error of the mechanical structure of the device is set as ±4 cm for this interval, which serves as the upper and lower limits of the allowable tolerance for dynamic synchronization to ensure a reasonable screening process. During screening, the spatial differences at each sampling moment are compared one by one to determine whether they fall within the tolerance interval. The moments when the three-dimensional Euclidean displacement differences fall within this interval are selected, summarized and recorded, and finally a synchronization moment screening interval is formed. For example, in the time period from 1056 ms to 1059 ms between Node 1 and Node 3, the displacement differences at the four sampling moments of 1057 ms, 1058 ms, and 1059 ms are 0.062 m, 0.067 m, and 0.061 m respectively, all of which meet the tolerance conditions. The moments of these three points are recorded and entered into the synchronization moment screening interval.

[0050] S502: Call the time nodes recorded in the synchronization moment screening interval, extract the device timestamp values, convert them to a unified reference moment by linear interpolation, and reconstruct them in combination with the device numbers to generate a device timestamp alignment sequence; Call the time nodes recorded in the synchronization moment screening interval, extract the timestamps of each node at the synchronization moment, and convert the device timestamps to a unified reference moment according to the time reference of the synchronization control system by linear interpolation. For example, at the three time points of 1057 ms, 1058 ms, and 1059 ms, the displacements of Node 1 and Node 3 both meet the synchronization tolerance standard. Assume that the original sampling interval of Node 1 is 1 ms, while Node 3 has a time sampling error of ±0.2 ms. It is necessary to interpolate and convert the timestamps of Node 3 to 1057 ms, 1058 ms, and 1059 ms to maintain the temporal consistency. The linear interpolation process is based on the sampled timestamps of Node 3 and their corresponding displacements to calculate the estimated timestamps at the interpolation points. For example, Node 3 samples at 1056.8 ms and 1057.8 ms, and the corresponding displacements are (0.91, 0.39, -0.10) m and (0.94, 0.42, -0.08) m. Then the calculation method of the interpolation displacement at 1057 ms is as follows: ; Substitute and calculate in the three-axis directions respectively. After obtaining the interpolation point displacements, append the interpolation timestamp 1057 ms to form a timestamp sequence under the unified reference, and at the same time append the device number, that is, generate a form of: (node number, unified timestamp, corresponding displacement value), and finally form a complete device timestamp alignment sequence.

[0051] S503: According to the reference moments in the device timestamp alignment sequence, read the tension change values, calculate the first-order time difference to obtain the change rate, and encapsulate them with the device numbers and time points to generate multi-device synchronization control instructions; Align the moment values of the nodes in the sequence according to the device timestamp under a unified time reference, call the tension change data of the nodes, take each node as an index, read its tension value sequence under the unified reference moment, and calculate the tension difference between two adjacent time points divided by the time interval to obtain the first-order time difference value as the tension change rate. For example, the tension of node 1 is 134 N at 1057 ms and 139 N at 1058 ms, then the tension change rate is , encapsulate the tension change rate, node number, and timestamp into a synchronization control instruction, such as (node 1, 1058 ms, tension change rate 5 N / ms). Process the moment nodes in the alignment sequence in this way to form a complete set of synchronization instructions for subsequent action control triggering. The control instruction set should be stored in a unified format to plan the execution order of the action modules.

[0052] Table 5: Table of Spatial Displacement Differences and Synchronization Screening Results between Devices

[0053] Table 5 lists the three-axis displacements and spatial differences of the time points of node 1 and node 3 during the synchronization period, and marks whether the tolerance condition within 0.08 m is met.

[0054] As Figure 2 shown, the multi-machine collaborative control system of the bridge cable hoisting device, the system includes: Tension monitoring module, used to obtain tension data through the sensor array, record the start of change, fluctuation duration, and stabilization time, call the sliding window algorithm to normalize the tension sequence, extract the abnormal interval and cluster, output the effective response time window, and transfer it to the window extraction module; Window extraction module, used to receive the effective response time window, call the time series overlapping algorithm to calculate the multi-node intersection, construct a trend function based on the displacement velocity derivative, perform fine-tuning operations on the intersection region boundary, obtain the multi-node collaborative time window, and transfer it to the collaborative recognition module; Collaborative recognition module, used to receive the multi-node collaborative time window, obtain the hoisting point acceleration and node inertia parameters, call the inertia compensation model to perform collaborative operations to adjust the node response amplitude, screen the time period reaching the response rate threshold, output the device action trigger window, and transfer it to the action compensation module; Action compensation module, used to receive the device action trigger window, obtain the tension conduction time and hoisting point displacement difference, perform reduction and interval fitting, call the dynamic threshold function model to screen the extreme values and apply threshold constraints, output the synchronization trigger condition, and transfer it to the synchronization instruction module; The synchronization instruction module is used to receive the synchronization trigger condition, call the laser rangefinder to judge the node displacement difference limit value, collect the device number, trigger time, and tension change rate according to the control error limit domain, and obtain the multi-device synchronization control instruction through encapsulation by the encoding module.

[0055] The above is only a specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A multi-machine collaborative control method for a bridge cable hoisting device, characterized in that, It includes the following steps: S1: Collect the starting moment of the cable node tension change, the duration of the tension fluctuation, and the steady-state response time through the fiber optic sensor array, perform normalization processing, input it into the sliding window algorithm to extract the tension response characteristics, and generate an effective response time window; S2: Based on the effective response time window, calculate the intersection of the node windows through the time series overlapping algorithm, adjust the time window boundary in combination with the displacement speed change trend of the device, and output a multi-node collaborative time window; S3: According to the multi-node collaborative time window, combine the structural inertia parameters and the suspension point acceleration data, and use the inertial compensation operation to calculate the structural response compensation value to generate a device action trigger window; S4: Based on the device action trigger window, detect the tension conduction time difference and the suspension point displacement difference between adjacent devices, input them into the dynamic threshold function that self-matches based on the real-time working conditions, perform extreme value screening and interval fusion operations, and output the synchronous trigger condition; S5: According to the synchronous trigger condition, perform timestamp alignment operation. When the displacement difference is within the tolerance range, encapsulate the device number, trigger moment, and tension change rate through instruction encoding, and output a multi-device synchronous control instruction.

2. The multi-machine cooperative control method of the bridge cable hoisting device according to claim 1, characterized in that The inertial compensation operation corrects the original response signal to restore the real physical response state; The structural response compensation value corrects the lifting response deviation of the device; The dynamic threshold function dynamically adjusts the synchronous trigger condition in real time to achieve device response consistency; The effective response time window includes the tension change rate, the fluctuation duration, and the steady-state time constant. The multi-node collaborative time window specifically refers to the time window overlap degree, the displacement speed variance, and the collaborative convergence threshold. The device action trigger window includes the inertial compensation factor, the acceleration deviation threshold, and the phase synchronization margin. The synchronous trigger condition specifically refers to the peak value of the conduction time difference, the standard deviation of the displacement difference, and the dynamic fusion threshold. The multi-device synchronous control instruction includes the device identifier, the trigger timestamp, and the tension change rate sequence; The collaborative convergence threshold refers to the critical value when the multi-node time window reaches stability after adjustment; The dynamic fusion threshold is a fusion boundary value calculated by the real-time working condition adaptive algorithm based on multiple indicators such as the tension conduction time difference and the suspension point displacement difference.

3. The multi-machine collaborative control method for the bridge cable hoisting device according to claim 1, characterized in that, The specific steps of S1 include: S101: Collect the starting moment of the cable node tension change, the fluctuation duration, and the steady-state response time through the fiber optic sensor, perform time domain alignment on the multi-node time series, eliminate abnormal sampling points caused by environmental noise, and generate a time synchronization data set; S102: Based on the time synchronization data set, establish a normalization processing model for the node tension parameters, use the Min-Max normalization method to eliminate the dimension difference, map the original sampling value to a unified interval, and generate a normalized tension sequence; S103: Call the normalized tension sequence, input it into the sliding window algorithm to calculate the tension extreme value, variance value, and slope change rate within each window, splice the window indicators to construct a multi-dimensional feature matrix, calculate the similarity value between adjacent window feature vectors, and generate an effective response time window when the similarity of several consecutive windows exceeds the similarity threshold; The similarity threshold determines whether it belongs to the same response time period by judging the feature similarity of two adjacent windows.

4. The multi-machine cooperative control method of the bridge cable hoisting device according to claim 3, characterized in that, The specific steps of S2 include: S201: Based on the effective response time window, use the time series overlapping algorithm to compare the time window frame sequences of adjacent nodes, identify the time difference between the start and end time tags, filter out the time slices with the absolute difference in intervals less than the synchronization fault tolerance threshold, and generate the node overlapping time interval; The synchronization fault tolerance threshold is set according to the node clock synchronization accuracy and tolerance standard; S202: Invoke the node overlapping time interval, extract the sequence of displacement velocity sampling points within the interval, calculate the rate of change of velocity based on the time difference and velocity difference of the sampling points, extract the absolute value of the regression slope, and generate the displacement velocity change coefficient; S203: Combine the displacement velocity change coefficient and the node overlapping time interval, calculate the time extension and compression values to adjust the start and end boundaries, and invoke the dynamic boundary compensation algorithm to repair the boundary gap, generating the multi-node collaborative time window.

5. The multi-machine cooperative control method of the bridge cable hoisting device according to claim 4, characterized in that, The rate of change of velocity is calculated using the formula: ； Among them, represents the value of the speed change rate parameter of the th node, with the unit of m / s 2 , represents the speed value of the th node at the th sampling point, with the unit of m / s, represents the speed value of the th node at the th sampling point, with the unit of m / s, represents the time difference between the th node and the th and the th sampling points, with the unit of s, represents the average speed value of the th node at the sampling point, with the unit of m / s, is the weighted adjustment coefficient of the th node based on the number of sampling points and sensitivity setting, is the time correction parameter of the th node according to the sampling delay setting, with the unit of s, represents the number of pairs of sampling points participating in the calculation within the current interval.

6. The multi-machine collaborative control method of the bridge cable hoisting device according to claim 4, characterized in that, The specific steps of S3 include: S301: Based on the multi-node collaborative time window, select the mass distribution coefficient, damping ratio, and stiffness matrix, combine the three-axis acceleration and the change in velocity, calculate the node inertial force difference rate, establish the inertial change mapping matrix, and perform point-by-point difference calculation using the adjacent node offset to generate the inertial response difference sequence; S302: According to the inertial response difference sequence, use the mass distribution coefficient as the distribution reference, combine the vertical acceleration value and the node response change rate, calculate the node response offset degree using the response change distribution function, and make a judgment in combination with the offset judgment threshold to obtain the response offset identification interval; S303: For the response offset identification interval, combine the vertical acceleration change value of the suspension point and the node stiffness term, calculate the response offset change rate, and make a judgment with the trigger change threshold for each node to generate the device action trigger window; The trigger change threshold is set for the upper and lower bounds according to the response rate fluctuation range of the node in the normal state.

7. The multi-machine cooperative control method of the bridge cable hoisting device according to claim 6, characterized in that, The node response offset degree is calculated using the formula: ； Among them, represents the response offset of the j-th node, represents the average mass distribution coefficient of the j-th node within the current time window, with the unit of kg·m, represents the average vertical acceleration value of the j-th node within the current time window, with the unit of m / s 2 , represents the average velocity change of the j-th node within the current time window, with the unit of m / s, represents the average stiffness coefficient of the j-th node within the current time window, with the unit of N·m, represents the average acceleration change value of the j-th node within the current time window, with the unit of m / s 2 , represents the average inertial force difference value of the j-th node within the current time window, with the unit of N, is the node inertia displacement dynamic factor, is the response compensation coefficient, a dimensionless response index generated based on the node structural damping ratio and the change in the acceleration of the suspension point, is the standard deviation of the acceleration change, with the unit of m / s 2 .

8. The multi-machine collaborative control method of the bridge cable hoisting device according to claim 6, characterized in that, The specific steps of S4 include: S401: Based on the device action trigger window, detect the change in tension between adjacent devices, collect the continuous response data of the tension sensor at the starting moment of the device action, extract the starting moment of the leading peak in the tension waveform of each pair of devices as the conduction starting point, and calculate the time difference for the leading peak to propagate to adjacent devices among multiple groups of devices to generate the tension response time difference; S402: Invoke the tension response time difference, combine the three-axis displacement monitoring data before and after the device suspension point action, intercept the suspension point displacement sequence within the time window corresponding to the tension conduction starting point, calculate the maximum displacement change amount between adjacent suspension points, and compare it with the tension response time difference interval to obtain the suspension point dynamic displacement difference; S403: According to the suspension point dynamic displacement difference and the tension response time difference, input them into the dynamic threshold function that self-matches and adjusts according to the real-time working conditions, extract the extreme point interval with concentrated fluctuation frequencies in the two sets of data, and perform a fusion operation to obtain the synchronization trigger condition.

9. The multi-machine cooperative control method of the bridge cable hoisting device according to claim 8, characterized in that, The specific steps of S5 include: S501: Calculate the displacement difference between devices at the current moment based on the device displacement values recorded in the synchronous trigger condition, compare the displacement difference with the synchronous tolerance range, screen the time nodes where the differences all fall within the tolerance range, and generate a synchronous time screening interval. S502: Call the time nodes recorded in the synchronous time screening interval, extract the device timestamp values, convert them to a unified reference time according to the linear interpolation method, and reconstruct them in combination with the device numbers to generate a device timestamp alignment sequence. S503: According to the reference time in the device timestamp alignment sequence, read the tension change value, calculate the first-order time difference to obtain the change rate, and package it with the device number and time point to generate a multi-device synchronous control instruction.

10. The multi-machine cooperative control system of the bridge cable hoisting device is characterized in that, The system is used to implement the multi-machine cooperative control method of the bridge cable hoisting device described in any one of claims 1-9. The system includes: A tension monitoring module, which is used to obtain tension data through a sensor array, record the starting time of change, the duration of fluctuation, and the steady state time, call the sliding window algorithm to normalize the tension sequence, extract the abnormal interval and cluster it, output the effective response time window, and transfer it to the window extraction module. A window extraction module, which is used to receive the effective response time window, call the time series overlapping algorithm to calculate the multi-node intersection, construct a trend function based on the displacement velocity derivative, perform a fine-tuning operation on the boundary of the intersection area, obtain the multi-node cooperative time window, and transfer it to the cooperative recognition module. A cooperative recognition module, which is used to receive the multi-node cooperative time window, obtain the lifting point acceleration and node inertia parameters, call the inertia compensation model to perform cooperative operation to adjust the node response amplitude, screen the time period that reaches the response rate threshold, and output the device action trigger window, and transfer it to the action compensation module. An action compensation module, which is used to receive the device action trigger window, obtain the tension conduction time and the lifting point displacement difference, perform reduction and interval fitting, call the dynamic threshold function model to screen the extreme values and apply threshold constraints, output the synchronous trigger condition, and transfer it to the synchronous instruction module. A synchronous instruction module, which is used to receive the synchronous trigger condition, call the laser rangefinder to judge the node displacement difference limit value, collect the device number, trigger time, and tension change rate according to the control error limit domain, and package them through the encoding module to obtain the multi-device synchronous control instruction.

Images