Positioning control method and system for intelligent sliding assembly of thin-walled steel structure tank wall plate

Abstract

This invention discloses a positioning control method and system for intelligent sliding assembly of thin-walled steel tank panels, relating to the field of panel assembly control technology. The method includes: S1, collecting assembly construction monitoring data and performing preprocessing on the data; S2, constructing a continuous monitoring sequence within a segment and generating a panel assembly positioning status dataset; S3, generating a panel assembly positioning coupling evaluation value, determining positioning deviation sections, and extracting positioning change trigger points; S4, marking normal and abnormal lifting points, constructing and training a lifting point adjustment parameter estimation model based on normal lifting points, inputting the assembly construction monitoring data of abnormal lifting points into the model, outputting corresponding adjustment parameters, and sending them to the corresponding control unit for execution. This solves the problem in existing tank panel sliding assembly where the fluctuation of the expansion ring fitting gap and the uneven stiffness of the portal steel clamp connection lead to difficulty in identifying synchronous deviation evolution sections and causing positioning correction lag.

Description

Technical Field

[0001] This invention relates to the field of panel assembly control technology, specifically to a positioning control method and system for intelligent sliding assembly of thin-walled steel structure tank panels. Background Technology

[0002] In recent years, as large-scale storage tank projects have developed towards larger capacity, rapid construction, and process traceability, the demand for positioning accuracy, synchronization stability, and process data management during the hydraulic lifting and sliding assembly of tank wall panels has been continuously increasing. Construction sites are gradually introducing multi-source sensing monitoring, digital data acquisition, and programmed control methods to achieve unified management of drive execution, attitude measurement, and assembly geometric parameters, and to form reusable status data and control command links during lifting and assembly. Various related technical solutions have emerged focusing on positioning control, measurement feedback, and automatic adjustment.

[0003] For example, the invention with publication number CN107250933B relates to a positioning control device that can register control signals or devices to be sampled through simple operation. The positioning control device, which controls the drive of a motor while displaying the operation screen of a motion control program, includes: a software development assistance unit that includes a motion control program and generates and outputs a list of detectors for control signals or devices used in the motion control program; and a sampling display unit that outputs the detector list to a controller and displays the data sampled in the controller. The software development assistance unit generates a list of detectors for one or more control signals or devices according to the operation on the operation screen of the motion control program.

[0004] For example, the invention disclosed in CN112305989B is an automatic adjustment control system for segmental precast beam formwork, including digital twin technology, a CNC electric screw drive device and a CNC measurement system, a matching beam segment, a bottom formwork trolley and a longitudinal transfer track. The bottom formwork trolley is connected to the longitudinal transfer track via the CNC electric screw drive device. A bottom template is connected above the bottom formwork trolley. The matching beam segment is set on the bottom template. A fixed end template is set on the rear side of the matching beam segment. The CNC measurement system is set on the fixed end template and is connected to the digital twin technology. A beam top measurement point is set on the upper surface of the matching beam segment.

[0005] However, the aforementioned existing technologies mainly focus on the registration and sampling of general positioning control signals, programmed drive control, or geometric measurement and automatic adjustment of component templates. They do not adequately cover the positioning control requirements of "multi-point synchronization, attitude coupling, and assembly constraints working together" in the hydraulic lifting and sliding assembly scenarios of tank wall panels. Especially in the case of segmented splicing of expansion rings and welding them to the tank wall with portal steel clamps, the gap between the expansion ring and the wall panel, roundness deviation, or local quality fluctuations of the portal steel clamp weld can easily cause uneven connection stiffness. During the lifting process, the load path undergoes instantaneous redistribution, resulting in sudden jumps in the displacement of local lifting points and inducing a rapid increase in synchronization error. Such early anomalies often only manifest as fluctuations in pressure readings or slight deformation of the tank wall. Without a mechanism for quantifying the positioning status of lifting segments, judging deviation sections, and identifying trigger points, as well as a parameter estimation and multi-actuator collaborative correction control link for abnormal lifting points, segment-level positioning deviation evolution can easily occur, affecting the circumferential seam connection and assembly accuracy, thus making it difficult to meet the positioning control requirements of intelligent sliding assembly of tank wall panels.

[0006] Therefore, in order to address the above problems, there is an urgent need for a positioning control method and system for intelligent sliding assembly of thin-walled steel structure tank panels. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a positioning control method and system for intelligent sliding assembly of thin-walled steel tank wall panels. This method solves the problem that the superposition of fluctuations in the fit gap of the expansion ring and uneven stiffness of the gantry steel clamp connection in the existing sliding assembly of tank wall panels leads to difficulty in identifying the synchronous deviation evolution section and causing positioning correction to lag.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a positioning control method for intelligent sliding assembly of thin-walled steel structure tank panels, comprising: S1, collecting assembly construction monitoring data, and performing timestamp alignment, abnormal data removal, missing interval completion, and standardization processing on the assembly construction monitoring data to generate pre-processed assembly construction monitoring data; S2, constructing a continuous monitoring sequence within a segment based on the pre-processed assembly construction monitoring data according to the lifting segment number, and uniformly quantifying the pressure deviation change at each lifting point, the relative displacement difference change at the lifting point, and the overall average lifting height of the panel to generate the panel assembly positioning status. Dataset; S3, based on the panel assembly positioning status dataset, evaluate the panel assembly positioning coupling status, generate panel assembly positioning coupling evaluation values, and determine the positioning status deviation sections based on the panel assembly positioning coupling evaluation values, and extract positioning change trigger points; S4, based on the positioning change trigger points, evaluate the synchronization deviation status of each lifting point, mark normal lifting points and abnormal lifting points, construct and train the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points, input the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, output the corresponding adjustment parameters, and send them to the corresponding control unit for execution.

[0009] Further, the following steps are taken to collect assembly construction monitoring data and perform timestamp alignment, abnormal data removal, missing interval completion, and standardization processing on the data to generate pre-processed assembly construction monitoring data: Real-time collection of tank wall panel assembly construction monitoring data, including collection timestamp, lifting segment number, working pressure of a single hydraulic jack, total output pressure of the hydraulic system, cumulative lifting height of a single lifting point, relative displacement difference of the lifting point, vertical offset of the top of the wall panel, gap between the expansion ring and the wall panel, spacing of the portal steel clips, and width of the circumferential joint gap; for the collection... The collected assembly construction monitoring data were sorted according to the collection timestamp, and the multi-source asynchronous sampling records were mapped to a unified time axis using the nearest neighbor time alignment algorithm. The median filtering algorithm was used to smooth the short-period fluctuation segments in the assembly construction monitoring data sequence. The box plot outlier detection algorithm was used to identify and remove abnormal sampling points in the assembly construction monitoring dataset, and the piecewise linear interpolation algorithm was combined to continuously complete the missing intervals. The Z-Score standardization algorithm was used to perform numerical standardization processing on the assembly construction monitoring data to eliminate the dimensional differences between different physical quantities.

[0010] Furthermore, based on the pre-processed assembly construction monitoring data, a continuous monitoring sequence within each lifting segment is constructed according to the lifting segment number. The specific steps for generating the wall panel assembly positioning status dataset are as follows: Read the pre-processed assembly construction monitoring data, group the data according to the lifting segment number, and within each lifting segment, arrange the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel in chronological order based on the collection timestamp to obtain a continuous monitoring sequence within the segment; based on the continuous monitoring sequence within the segment, count the number of lifting points in the current lifting segment, and use the total hydraulic system output pressure as the basis for the dataset. The pressure output is divided by the number of lifting points to obtain the pressure average benchmark. A relative deviation calculation algorithm is used to calculate the deviation of the working pressure of each individual hydraulic jack from the pressure average benchmark. A first-order difference algorithm is performed on the deviation values ​​at adjacent time stamps to obtain the pressure deviation change corresponding to each lifting point. A first-order difference algorithm is also performed on the relative displacement difference of each lifting point at adjacent time stamps to obtain the relative displacement difference change corresponding to each lifting point. The overall average lifting height of the wall panel is calculated based on the cumulative lifting height of each lifting point. The continuous monitoring sequence within the segment, the number of lifting points, the pressure deviation change corresponding to each lifting point, the relative displacement difference change corresponding to each lifting point, and the overall average lifting height of the wall panel are stored in a structured manner according to the lifting segment number to generate a wall panel assembly positioning status dataset.

[0011] Furthermore, based on the panel assembly positioning status dataset, the specific steps for evaluating the panel assembly positioning coupling status and generating the panel assembly positioning coupling evaluation value are as follows: Read the panel assembly positioning status dataset; take the absolute value of the pressure deviation change corresponding to each lifting point within the current lifting segment, multiply it by the square of the relative displacement difference change of the corresponding lifting point, and then perform a summation operation to obtain the pressure displacement coupling cumulative amount; take the absolute value of the relative displacement difference change corresponding to each lifting point within the current lifting segment, perform a summation operation, and add a minima term to obtain the displacement normalization base; divide the pressure displacement coupling cumulative amount by the displacement normalization base, add one to the ratio, and then take the natural logarithm to obtain the basic coupling evaluation term; take the absolute value of the ratio of the vertical offset of the top of the panel within the current lifting segment to the overall average lifting height of the panel, and add one to obtain the attitude amplification term; multiply the basic coupling evaluation term and the attitude amplification term to obtain the panel assembly positioning coupling evaluation value for the current lifting segment.

[0012] Furthermore, based on the panel assembly positioning coupling evaluation value, the specific steps for determining the positioning deviation section and extracting the positioning change trigger point are as follows: Extract the panel assembly positioning coupling evaluation value continuously calculated according to the acquisition timestamp within the current lifting segment, and calculate the change in the panel assembly positioning coupling evaluation value under adjacent timestamps to obtain the evaluation value change sequence; and construct the change judgment interval based on the median and interquartile range of the evaluation value change sequence. The time segment where the evaluation value change exceeds the upper limit of the judgment interval under N consecutive acquisition timestamps is determined as the positioning deviation section, and the timestamp of the first time that the positioning deviation section exceeds the upper limit of the judgment interval is determined as the positioning change trigger point.

[0013] Further, the specific steps for evaluating the synchronization deviation status of each lifting point based on the positioning change trigger point are as follows: Read all positioning deviation segments and corresponding positioning change trigger points according to the lifting segment number, and extract the relative displacement difference change of all lifting points within the current lifting segment at the timestamp corresponding to the positioning change trigger point. Take the median of the relative displacement difference change as the baseline value of the relative displacement difference change within the current lifting segment. Subtract the relative displacement difference change of the i-th lifting point from the baseline value of the relative displacement difference change within the current lifting segment and take the absolute value to obtain the displacement deviation. Take the absolute value of the pressure deviation change corresponding to the i-th lifting point, add one, and take the natural logarithm. Add one to the resulting logarithm to obtain the pressure deviation modulation term. Multiply the displacement deviation by the pressure deviation modulation term to obtain the synchronization deviation evaluation value of the i-th lifting point.

[0014] Furthermore, the specific steps for marking normal and abnormal lift points are as follows: For each lift point, calculate the synchronization deviation evaluation value corresponding to each sampling point within the corresponding positioning state deviation segment, and extract the maximum value of the synchronization deviation evaluation value; compare the maximum value of the synchronization deviation evaluation value with the deviation threshold in real time. When the maximum value of the synchronization deviation evaluation value is less than or equal to the deviation threshold, mark the corresponding lift point as a normal lift point; when the maximum value of the synchronization deviation evaluation value is greater than the deviation threshold, mark the corresponding lift point as an abnormal lift point.

[0015] Furthermore, the specific steps for constructing and training the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points are as follows: For each lifting segment, extract the assembly construction monitoring data of all normal lifting points to construct a normal construction state sample set; use the support vector regression algorithm to construct the lifting point adjustment parameter estimation model, and use the normal construction state sample set as the sample input to establish a regression mapping relationship for the gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap, and perform regression iteration training on the model; when the back calculation error of the regression mapping is less than the error threshold in K consecutive regression iterations, it is determined that the lifting point adjustment parameter estimation model training is complete.

[0016] Furthermore, the specific steps for inputting the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, outputting the corresponding adjustment parameters, and issuing them to the corresponding control units are as follows: The assembly construction monitoring data of the abnormal lifting points of the corresponding lifting segment are sequentially input into the trained lifting point adjustment parameter estimation model, outputting the corresponding wall panel bonding gap adjustment amount, portal steel clamp arrangement spacing adjustment amount, and circumferential seam butt joint gap width adjustment amount. Combined with the current hydraulic cylinder piston stroke state of the abnormal lifting point, the hydraulic cylinder piston stroke adjustment amount is determined. Each adjustment amount is written into the control execution instruction set, and hydraulic cylinder piston stroke adjustment instructions are issued to the hydraulic pump station control unit of the corresponding abnormal lifting point, while simultaneously issuing structural adjustment instructions to the expansion ring assembly and portal steel clamp actuator at the corresponding positions. After the control execution is completed, the assembly construction monitoring data within the corresponding lifting segment continues to be collected, forming a closed-loop control system.

[0017] The second aspect of this invention provides a positioning control system for intelligent sliding assembly of thin-walled steel structure tank panels, comprising: a construction data acquisition and preprocessing module, a panel assembly positioning state construction module, a panel assembly positioning evolution interpretation module, and a panel assembly positioning correction control module. The construction data acquisition and preprocessing module collects assembly construction monitoring data and performs timestamp alignment, abnormal data removal, missing interval completion, and standardization processing on the data to generate preprocessed assembly construction monitoring data. The panel assembly positioning state construction module constructs a continuous monitoring sequence within each lifting segment based on the preprocessed assembly construction monitoring data and according to the lifting segment number. It also monitors the pressure deviation change at each lifting point, the relative displacement difference change at each lifting point, and the overall average lifting height of the panel. The system performs unified quantitative characterization to generate a panel assembly positioning status dataset. The panel assembly positioning evolution judgment module evaluates the panel assembly positioning coupling status based on this dataset, generates a panel assembly positioning coupling evaluation value, determines the positioning status deviation section based on the evaluation value, and extracts the positioning change trigger point. The panel assembly positioning correction control module evaluates the synchronization deviation status of each lifting point based on the positioning change trigger point, marks normal and abnormal lifting points, constructs and trains a lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points, inputs the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, outputs the corresponding adjustment parameters, and sends them to the corresponding control unit for execution.

[0018] The present invention has the following beneficial effects:

[0019] (1) Positioning control method and system for intelligent sliding assembly of thin-walled steel tank wall panels: By constructing a wall panel assembly positioning status dataset with lifting segments as the organizational unit, the pressure deviation change, the relative displacement difference of the lifting point and the overall average lifting height of the wall panel are uniformly incorporated into the same structured expression, forming a segment-oriented comparable input basis, reducing the impact of the difference in dimensions and scale between different lifting points and different segments on the consistency of positioning interpretation.

[0020] (2) Positioning control method and system for intelligent sliding assembly of thin-walled steel tank wall panels: By using the positioning coupling evaluation value of the wall panel assembly, the "pressure-displacement-attitude" coupling deviation within the segment is quantified in a centralized manner, and further the positioning state deviation segment and positioning change trigger point are determined based on the evaluation value change sequence, so that the positioning deviation is transformed from a discrete abnormal phenomenon into a continuously evolving segment that can be located.

[0021] (3) Positioning control method and system for intelligent sliding assembly of thin-walled steel tank wall panels. By introducing a reference value for the change of relative displacement difference of lifting points at the positioning change trigger point, and constructing a synchronous deviation evaluation value to refine the synchronous deviation status of a single lifting point, a layered progressive judgment from "segment-level deviation judgment" to "lifting point-level abnormal positioning" is realized, which facilitates the rapid locking of abnormal lifting points.

[0022] (4) Positioning control method and system for intelligent sliding assembly of thin-walled steel tank wall panels: By using a lifting point adjustment parameter estimation model trained based on normal lifting point samples, the state characteristics of abnormal lifting points are directly mapped into executable structural and stroke adjustment amounts. This ensures that the expansion ring fitting, portal steel clamp arrangement, circumferential seam connection and hydraulic cylinder stroke correction actions have consistent parameter output diameter and distribution path, thereby enabling the positioning deviation to form a closed-loop correction and continuous convergence towards abnormal lifting points within the same lifting segment. Attached Figure Description

[0023] Figure 1 Flowchart of the positioning control method for intelligent sliding assembly of thin-walled steel structure tank panels;

[0024] Figure 2 Structural diagram of the positioning control system for intelligent sliding assembly of thin-walled steel tank wall panels;

[0025] Figure 3 This is a chart for determining the maximum value of the synchronization deviation assessment within the positioning status deviation section.

[0026] Figure 4 A partial sectional view of the assembly relationship of the sliding assembly of the tank wall panels;

[0027] Figure 5 This is a schematic diagram of the partial circumferential arrangement of the sliding assembly of the tank wall panels. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figures 1-5 This invention provides a technical solution: a positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels, comprising: S1, collecting assembly construction monitoring data, and performing timestamp alignment, abnormal data removal, missing interval completion, and standardization processing on the assembly construction monitoring data to generate pre-processed assembly construction monitoring data; S2, constructing a continuous monitoring sequence within a segment based on the pre-processed assembly construction monitoring data according to the lifting segment number, and uniformly quantifying the pressure deviation change at each lifting point, the relative displacement difference change at the lifting point, and the overall average lifting height of the wall panel to generate a wall panel assembly positioning status dataset; S3, based on the wall panel assembly positioning status dataset, evaluate the wall panel assembly positioning coupling status, generate wall panel assembly positioning coupling evaluation value, and determine the positioning status deviation section based on the wall panel assembly positioning coupling evaluation value, and extract the positioning change trigger point; S4, based on the positioning change trigger point, evaluate the synchronization deviation status of each lifting point, mark normal lifting points and abnormal lifting points, construct and train the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points, input the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, output the corresponding adjustment parameters, and send them to the corresponding control unit for execution.

[0030] Specifically, the steps for collecting assembly construction monitoring data and performing timestamp alignment, abnormal data removal, missing interval completion, and standardization on the assembly construction monitoring data to generate pre-processed assembly construction monitoring data are as follows: Real-time collection of tank wall panel assembly construction monitoring data, including collection timestamp, lifting segment number, working pressure of a single hydraulic jack, total output pressure of the hydraulic system, cumulative lifting height of a single lifting point, relative displacement difference of the lifting point, vertical offset of the top of the wall panel, gap between the expansion ring and the wall panel, spacing of the portal steel clips, and width of the circumferential joint gap. The data acquisition timestamp is written by the clock source of the construction data acquisition terminal and synchronously fixed with the sampling frame; the lifting segment number is written by the lifting operation control console into the segment field of the sampling frame when entering the corresponding lifting segment; the working pressure of a single hydraulic jack is acquired and output by the pressure transmitter installed at the inlet of the corresponding hydraulic jack, and written into the sampling frame after being bound to the acquisition timestamp in the acquisition terminal; the total output pressure of the hydraulic system is acquired and output by the pressure sensor of the main pressure output channel of the hydraulic pump station, and written into the sampling frame after being bound to the acquisition timestamp in the acquisition terminal; the cumulative lifting height of a single lifting point is acquired by the displacement sensor of the corresponding lifting point, which acquires the piston stroke displacement of the hydraulic cylinder, and the cumulative lifting height is obtained by accumulating the piston stroke displacement according to the acquisition timestamp in the acquisition terminal and then written into the sampling frame; the relative displacement difference of the lifting points is synchronously sampled by the displacement sensors of each lifting point. The results are generated as a difference output in the acquisition terminal and written into the sampling frame. The relative displacement difference of the lifting point is based on the median of the cumulative lifting height of all lifting points in the current lifting segment at the same acquisition timestamp. The difference between the cumulative lifting height of each lifting point and the displacement reference value is calculated and the absolute value is taken. Thus, the relative displacement difference of the lifting point represents the deviation of the lifting point from the horizontal center of the overall lifting height in the segment. The vertical offset of the top of the wall panel is obtained by periodically measuring the measuring point on the top of the wall panel with a total station and the offset reading is written into the sampling frame. The gap between the expansion ring and the wall panel is read by the gap measuring gauge at the fitting position of the expansion ring and recorded into the sampling frame. The spacing of the portal steel clips is measured by a steel tape measure between the reference edges of adjacent portal steel clips and recorded into the sampling frame. The width of the circumferential joint gap is measured by the weld gap gauge at the circumferential joint position and recorded into the sampling frame. For the collected assembly construction monitoring data, the monitoring records are sorted according to the collection timestamp. During the sorting process, records with the same collection timestamp are merged into the same sampling frame. During the merging process, the lifting segment number is retained to distinguish the sampling sequences of different lifting segments.The nearest neighbor time alignment algorithm is used to map multi-source asynchronous sampling records to a unified time axis. During the mapping process, a target time axis sequence is constructed using the acquisition timestamps. For each target time point on the target time axis, the record with the smallest absolute time difference is retrieved from the acquisition timestamps of each data source. The working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, the vertical offset of the top of the wall panel, the gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap of the corresponding record are written into the target time point to form an aligned sampling frame. The median filtering algorithm is used to process the assembly construction monitoring data sequence. Short-period fluctuations within the data are smoothed. In this smoothing process, a fixed-length sliding window is constructed using the acquisition timestamp as a sequence index. The window length ranges from 5 to 21 consecutive acquisition timestamps. During the window sliding process, the median working pressure of a single hydraulic jack within the window is calculated and replaced with the corresponding value at the window center point. The median total output pressure of the hydraulic system within the window is calculated and replaced with the corresponding value at the window center point. The median cumulative lifting height of a single lifting point within the window is calculated and replaced with the corresponding value at the window center point. The median relative displacement difference of the lifting points within the window is calculated and replaced with the corresponding value at the window center point. The median vertical offset of the top of the wall panel within the window is calculated and... The values ​​corresponding to the center point of the window are replaced. The median value of the gap between the expansion ring and the wall panel within the window is calculated and replaced with the corresponding value of the center point of the window. The median value of the spacing between the portal steel clips within the window is calculated and replaced with the corresponding value of the center point of the window. The median value of the width of the circumferential joint gap within the window is calculated and replaced with the corresponding value of the center point of the window. A box plot outlier detection algorithm is used to identify and remove abnormal sampling points in the assembly construction monitoring dataset. During the identification process, the lower quartile, upper quartile, and interquartile range are calculated for each data name. The lower quartile minus one and a half interquartile range is used as the lower threshold, and the upper quartile plus one and a half interquartile range is used as the upper threshold. Outliers are removed. Sampling points exceeding the lower threshold and those exceeding the upper threshold are marked as abnormal sampling points, and the corresponding sampling frames are removed based on the abnormal sampling point markings. In addition, piecewise linear interpolation algorithm is used to continuously complete the local missing intervals. During the completion process, the left and right boundary sampling frames of the missing interval are determined based on the collection timestamp. For each target time point in the missing interval, linear interpolation is performed according to the corresponding values ​​of the left and right boundary sampling frames to obtain the completed values ​​of single hydraulic jack working pressure, total hydraulic system output pressure, cumulative lifting height of a single lifting point, relative displacement difference of lifting points, vertical offset of the top of the wall panel, gap between the expansion ring and the wall panel, spacing of the portal steel clips, and width of the circumferential seam butt joint gap.The Z-Score standardization algorithm was used to perform numerical standardization processing on the assembly construction monitoring data. In the standardization process, the mean and standard deviation of the sampling frames within each lifting segment were calculated using the lifting segment number as the grouping key. The following transformations were performed: for the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of lifting points, the vertical offset of the top of the wall panel, the gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap. This eliminated dimensional differences between different physical quantities.

[0031] In this implementation plan, by processing the assembly construction monitoring data under a unified sampling frame structure, the data collection timestamp, lifting segment number, working pressure of a single hydraulic jack, total output pressure of the hydraulic system, cumulative lifting height of a single lifting point, relative displacement difference of the lifting point, vertical offset of the top of the wall panel, gap between the expansion ring and the wall panel, spacing of the portal steel clips, and width of the circumferential joint gap are made into a time-consistent and comparable data basis within the same lifting segment. This provides a unified data standard for the subsequent quantitative characterization of pressure deviation changes, relative displacement difference changes of lifting points, and overall average lifting height of the wall panel. Furthermore, it establishes the interpretation of positioning status deviation sections and positioning change trigger points based on the continuous evolution characteristics within the segment, reducing interpretation ambiguities caused by differences in sampling rhythm and field scale, and improving the reusability and consistency of abnormal lifting point identification and adjustment parameter estimation in the positioning control process.

[0032] Specifically, based on the preprocessed assembly construction monitoring data, a continuous monitoring sequence within each lifting segment is constructed according to the lifting segment number. The changes in pressure deviation at each lifting point, the changes in relative displacement difference at each lifting point, and the overall average lifting height of the wall panel are uniformly quantified to generate a wall panel assembly positioning status dataset. The specific steps are as follows: Read the preprocessed assembly construction monitoring data, group the monitoring data according to the lifting segment number, and during the grouping process, use the lifting segment number as the grouping key to write the sampling frames under the same lifting segment number into the same segment record set. Within each lifting segment, based on the collection timestamp, the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, and the relative displacement difference at each lifting point are analyzed. The values ​​and vertical offsets at the top of the wall panel are arranged chronologically. During the arrangement process, for sampling frames with repeated acquisition timestamps, a single record is retained according to the acquisition timestamp, and the fields are kept in one-to-one correspondence to obtain a continuous monitoring sequence within the segment. Based on the continuous monitoring sequence within the segment, the number of lifting points in the current lifting segment is counted. In the counting process, the number of pressure records corresponding to the lifting point identifier field is used as the number of lifting points. The number of lifting points is used to limit the synchronous calculation range of the current lifting segment. The total output pressure of the hydraulic system is divided by the number of lifting points as the pressure average benchmark. When constructing the pressure average benchmark, the total output pressure of the hydraulic system is read point by point according to the acquisition timestamp, and a pressure average benchmark sequence of the same length as the continuous monitoring sequence within the segment is generated. Since the same lifting... Under synchronous lifting conditions, all lifting points within a segment share the same hydraulic pump station pressure and bear the same gravity path of the wall panel segment. The total output pressure of the hydraulic system can characterize the instantaneous load level at the segment level, and the number of lifting points can characterize the load distribution scale among the lifting points. The pressure average benchmark formed by the two is used to characterize the average bearing pressure level that a single lifting point should correspond to under synchronous lifting conditions. When a local connection stiffness difference occurs at a certain lifting point due to changes in the fit gap between the expansion ring and the wall panel, changes in the spacing of the portal steel clamps, or changes in the width of the circumferential joint gap, the working pressure of a single hydraulic jack at that lifting point will continuously deviate from the pressure average benchmark and amplify synchronously with the change in the relative displacement difference of the lifting point. Thus, the pressure average benchmark can be used as a segment... A unified reference boundary for internal synchronization deviation identification is established, providing a consistent comparison baseline for subsequent pressure deviation change extraction. A relative deviation calculation algorithm is used to calculate the deviation of the working pressure of each individual hydraulic jack relative to the pressure mean benchmark. During the calculation process, the working pressure of each individual hydraulic jack at the same acquisition time stamp is calculated by performing a difference operation with the pressure mean benchmark at the corresponding acquisition time stamp and then dividing by the pressure mean benchmark to obtain a set of deviation values ​​consistent with the number of acquisition points. A first-order difference algorithm is then performed on the deviation values ​​at adjacent acquisition time stamps to obtain the pressure deviation change corresponding to each acquisition point. When performing the first-order difference, the deviation values ​​of the same acquisition point at adjacent acquisition time stamps are differentially analyzed before and after, while keeping the acquisition time stamp index unchanged.For the relative displacement difference of each lifting point, a first-order difference algorithm is performed at adjacent acquisition timestamps. During the first-order difference, the relative displacement difference of the same lifting point at adjacent acquisition timestamps is differentially analyzed while maintaining the acquisition timestamp index unchanged, yielding the change in the relative displacement difference for each lifting point. Then, the overall average lifting height of the wall panel is calculated based on the cumulative lifting height of all lifting points. During the calculation, the cumulative lifting height of all lifting points at the same acquisition timestamp is summed and divided by the number of lifting points to obtain the result consistent with the continuous monitoring sequence within the segment. The average lifting height sequence of the entire wall panel is generated. The continuous monitoring sequence within each segment, the number of lifting points, the pressure deviation change corresponding to each lifting point, the relative displacement difference change corresponding to each lifting point, and the overall average lifting height of the wall panel are structured and stored according to the lifting segment number. During the structured storage process, the lifting segment number is used as the index key to write segment-level record entries, the collection timestamp is used as the time sequence index to write the sequence field, and the lifting point number is used as the lifting point index to write the pressure deviation change field and the relative displacement difference change field, generating a wall panel assembly positioning status dataset.

[0033] In this implementation plan, by organizing the assembly construction monitoring data by lifting segment number and forming a continuous monitoring sequence within the segment, the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel are kept directly correlated within the same data collection time stamp framework. Based on this, a unified quantitative representation standard is established for the pressure deviation change, the relative displacement difference change of the lifting point, and the overall average lifting height of the wall panel. This enables the wall panel assembly positioning status dataset to stably support the comparative interpretation of the synchronous deviation evolution characteristics within the lifting segment, reducing the incomparability of states caused by differences in recording density between different lifting segments, and enhancing the consistency and anomaly directionality of the subsequent wall panel assembly positioning coupling evaluation values ​​in identifying positioning status deviation segments and positioning change trigger points.

[0034] Specifically, based on the wall panel assembly positioning status dataset, the specific steps for evaluating the wall panel assembly positioning coupling status and generating the wall panel assembly positioning coupling evaluation value are as follows: Read the wall panel assembly positioning status dataset. During the reading process, locate the current lifting segment record entry according to the lifting segment number, and extract the pressure deviation change and relative displacement difference change of each lifting point within the current lifting segment according to the collection timestamp. Extract the vertical offset of the top of the wall panel and the overall average lifting height of the wall panel. Take the absolute value of the pressure deviation change corresponding to each lifting point within the current lifting segment. This absolute value is used to represent the pressure deviation change... The positive and negative directions of the quantity are unified as the deviation amplitude measurement, and after multiplying it by the square of the change in the relative displacement difference of the corresponding lifting point, a summation operation is performed. The squaring operation is used to give higher amplitude to the larger fluctuations of the change in the relative displacement difference of the lifting point and to relatively compress the small fluctuations. The multiplication is used to characterize the coupling amplification relationship between the pressure deviation change and the change in the relative displacement difference of the lifting point under the same acquisition time stamp. The summation operation is used to aggregate the segment-level coupling accumulation strength within the number of lifting points to obtain the pressure displacement coupling accumulation. The pressure displacement coupling accumulation is used to characterize the cumulative degree of synchronous disturbance of multiple lifting points within the current lifting segment. After taking the absolute value of the relative displacement difference change of each lifting point within the current lifting segment, a summation operation is performed. The absolute value is used to unify the directional differences in the relative displacement difference change of the lifting points into an amplitude measure. The summation operation forms an overall scale reference for displacement fluctuation within the segment. Adding a minterm yields the displacement normalization base, which provides a unified normalization reference for the scale differences in displacement fluctuation across different lifting segments. The minterm is a very small but non-zero positive real number used to avoid numerical instability caused by division by zero during the calculation process; its value ranges from 10⁻⁸ to 10⁻⁵. The cumulative pressure-displacement coupling is divided by the displacement normalization base. The division operation is used to map the cumulative coupling strength to a comparable ratio under the relative displacement scale. The resulting ratio is incremented by one. The increment is used to shift the ratio to the positive domain to meet the domain constraints of the subsequent natural logarithmic transformation and to suppress the numerical amplification caused by the minimum ratio. Then, the natural logarithm is taken. The natural logarithm is used to compress the coupling ratio over a large range to reduce the dominance of extreme jumps in interpretation and enhance the distinguishability of moderate amplitude changes. The basic coupling evaluation term is obtained. The basic coupling evaluation term is used to characterize the relative coupling level between the pressure deviation change in the current lifting segment and the relative displacement difference change of the lifting point.The absolute value of the ratio of the vertical offset of the top of the wall panel within the current lifting segment to the average lifting height of the entire wall panel is used to map the vertical offset of the top of the wall panel into a dimensionless attitude offset ratio relative to the average lifting height of the entire wall panel. Taking the absolute value unifies the directional differences of the attitude offset into an offset amplitude measure. Adding one to the absolute value shifts the attitude offset ratio to a positive zero amplification factor, ensuring that the baseline is not distorted when the attitude offset is close to zero, resulting in an attitude amplification term. The attitude amplification term is used to apply differentiated amplification to the segment states with different attitude offset amplitudes under the same coupling level. The basic coupling evaluation term is multiplied by the attitude amplification term. This multiplication is used to link and superimpose the pressure displacement coupling level and the wall panel attitude offset ratio to form a comprehensive representation under the same dimensional framework, obtaining the wall panel assembly positioning coupling evaluation value for the current lifting segment. The wall panel assembly positioning coupling evaluation value is used as the judgment input for subsequent determination of positioning state deviation segments and extraction of positioning change trigger points.

[0035] The specific calculation formula for the panel assembly positioning coupling evaluation value is as follows:

[0036] ;

[0037] In the formula, This represents the evaluation value for the coupling of panel assembly and positioning. This indicates the number of lift points within the current lift segment. This represents the pressure deviation change corresponding to the i-th lift point. This represents the change in the relative displacement difference corresponding to the i-th lift point. Indicates minterms, This indicates the vertical offset of the top of the wall panel within the current lifting segment. This indicates the average overall lifting height of the wall panels within the current lifting segment.

[0038] In this implementation scheme, the pressure deviation change, the relative displacement difference of the lifting point, the vertical offset of the top of the wall panel, and the overall average lifting height of the wall panel are converged into a wall panel assembly positioning coupling evaluation value within the same lifting segment. This transforms the interpretation of the positioning status from a single measurement fluctuation into a comprehensive deviation index that can be compared laterally. Thus, even under conditions of changes in the number of lifting points, differences in displacement scale, and differences in attitude amplitude, it can still maintain a stable characterization of the segment-level coupling instability trend. This provides a consistent interpretation benchmark for subsequent identification of positioning status deviation segments and positioning change trigger points, reduces the probability of misjudgment caused by local pressure fluctuations or local displacement jumps dominating the judgment alone, and improves the directionality of positioning correction control in screening abnormal lifting points and estimating adjustment parameters.

[0039] Specifically, the steps for determining the positioning deviation section based on the panel assembly positioning coupling evaluation value and extracting the positioning change trigger point are as follows: Extract the panel assembly positioning coupling evaluation values ​​continuously calculated according to the acquisition timestamp within the current lifting segment. During the extraction process, locate the current lifting segment record entry according to the lifting segment number, and read the panel assembly positioning coupling evaluation values ​​from smallest to largest acquisition timestamp to form an evaluation value sequence arranged strictly in ascending order according to the acquisition timestamp; calculate the change in the panel assembly positioning coupling evaluation value under adjacent timestamps. During the calculation process, perform a difference operation on the panel assembly positioning coupling evaluation values ​​corresponding to two adjacent acquisition timestamps, subtract the previous time-time evaluation value from the later time-time evaluation value, and write it into the corresponding later time-time acquisition timestamp index to obtain the evaluation value change sequence; construct the change judgment interval based on the median and interquartile range of the evaluation value change sequence. During the construction process, first calculate the median of the evaluation value change sequence and record it as the change median, then calculate the evaluation value change sequence... The upper and lower quartiles are calculated, and the interquartile range is obtained accordingly. The median of the change plus one interquartile range is used as the upper limit of the judgment interval, and the median of the change minus one interquartile range is used as the lower limit of the judgment interval. The collection timestamp positions in the sequence of changes in the evaluation value that exceed the upper limit of the judgment interval are marked as out-of-limit points. The time segment where the change in the evaluation value exceeds the upper limit of the judgment interval under N consecutive collection timestamps is determined as the positioning status deviation segment, where N ranges from 3 to 8 consecutive collection timestamps. In the determination process, the collection timestamps are traversed backward from the out-of-limit point mark as the starting point, and the number of consecutive out-of-limit points is counted. When the number of consecutive out-of-limit points reaches N, the interval from the corresponding starting collection timestamp to the corresponding ending collection timestamp is written into the positioning status deviation segment, and the timestamp that first exceeds the upper limit of the judgment interval in the positioning status deviation segment is determined as the positioning change trigger point. In the determination process, the collection timestamp corresponding to the first out-of-limit point in the positioning status deviation segment is written into the positioning change trigger point field.

[0040] In this implementation scheme, by establishing a consistent segment interpretation standard for the change characteristics of the panel assembly positioning coupling evaluation value based on the acquisition timestamp as the main line within the lifting segment, the positioning deviation segment and the positioning change trigger point can be stably located from the continuous temporal evolution. This enables the extraction of subsequent synchronous deviation evaluation values ​​to have clear time anchor points and segment boundaries, reduces trigger drift caused by single-point fluctuations, improves the temporal consistency and positioning targeting of abnormal lifting point screening, and provides a reproducible trigger basis for adjusting parameter estimation.

[0041] Specifically, the steps for evaluating the synchronization deviation status of each lifting point based on the positioning change trigger point are as follows: Read all positioning deviation segments and their corresponding positioning change trigger points according to the lifting segment number. During the reading process, locate the current lifting segment record entry using the lifting segment number, and write the start and end timestamps of the positioning deviation segment into the segment judgment interval field. Write the timestamp of the positioning change trigger point into the trigger acquisition timestamp field. Extract the relative displacement difference of all lifting points within the current lifting segment at the timestamp corresponding to the positioning change trigger point. During the extraction process, evaluate the relative displacement difference of each lifting point at the same acquisition timestamp. The change in value is written into the displacement change set according to the lift point index. The displacement change set is used to characterize the lateral distribution of synchronous displacement disturbance within the segment at the trigger time. The median of the change in relative displacement difference of lift points is taken as the benchmark value of the change in relative displacement difference of lift points within the current lift segment. The median is used as a robust center reference for the displacement change set within the segment to weaken the pull of individual lift point jumps on the benchmark, so that the benchmark value can represent the typical displacement change level of most lift points at the trigger time. The absolute value of the difference between the change in relative displacement difference of the lift point corresponding to the i-th lift point and the benchmark value of the change in relative displacement difference of lift points within the current lift segment is taken to obtain the displacement deviation. The difference is used to characterize the i-th lift point. The deviation of each lift point from the typical displacement change level of the segment is taken as an absolute value to unify the deviation direction as a measure of deviation magnitude. The displacement deviation is used to characterize the degree of synchronous displacement mismatch of the i-th lift point at the trigger moment. The absolute value of the pressure deviation change corresponding to the i-th lift point is taken to unify the positive and negative directions of the pressure deviation change as a measure of deviation magnitude. One is added to shift the magnitude of the pressure deviation change to the positive domain to meet the domain constraint of the subsequent natural logarithmic transformation. The natural logarithm is then taken to compress the dynamic range of the pressure deviation change magnitude to suppress the sole dominance of extreme pressure jumps on the evaluation results and enhance the performance of moderate-amplitude pressure fluctuations. The discrimination index is obtained, and the resulting logarithmic value is incremented by one. This increment is used to form a modulation scale based on one and to ensure that the modulation term remains continuous when the pressure deviation change is close to zero, resulting in a pressure deviation modulation term. The pressure deviation modulation term is used to map the amplitude information of the pressure deviation change in a nonlinear manner into a modulation coefficient for the displacement deviation. The displacement deviation is multiplied by the pressure deviation modulation term. This multiplication is used to couple and superimpose the synchronous displacement mismatch degree of the i-th lift point with the amplitude characteristics of the corresponding pressure deviation change to form a comprehensive deviation quantification result for the same lift point, resulting in a synchronous deviation evaluation value for the i-th lift point. The synchronous deviation evaluation value is used as the input for subsequent discrimination between normal and abnormal lift points.

[0042] The specific formula for calculating the synchronization deviation assessment value is as follows:

[0043] ;

[0044] In the formula, This represents the synchronization deviation assessment value at the i-th lift point. This represents the change in the relative displacement difference corresponding to the i-th lift point. This represents the baseline value of the relative displacement difference change of the lifting points within the corresponding lifting segment. This represents the pressure deviation change corresponding to the i-th boost point.

[0045] In this embodiment, Table 1 shows the statistical results of the synchronization deviation evaluation values ​​of the same lifting point within the positioning deviation section at five sampling points, along with related input data. Specifically: Sampling point 1: The change in the relative displacement difference of the lifting point is 0.32, the baseline value for the change in the relative displacement difference of the lifting point is 0.28, the pressure deviation change is 0.08, and the corresponding synchronization deviation evaluation value is 0.0431. Sampling point 2: The change in the relative displacement difference of the lifting point is 0.35, the baseline value for the change in the relative displacement difference of the lifting point is 0.28, the pressure deviation change is 0.12, and the corresponding synchronization deviation evaluation value is 0.0779. Sampling point 3: The change in the relative displacement difference of the lifting point is 0.41, the baseline value for the change in the relative displacement difference of the lifting point is 0.28, the pressure deviation change is 0.20, and the corresponding synchronization deviation evaluation value is 0.1537. Sampling point 4: The change in relative displacement difference of the lifting point is 0.48, the baseline value is 0.28, the pressure deviation change is 0.35, and the corresponding synchronous deviation evaluation value is 0.2600. Sampling point 5: The change in relative displacement difference of the lifting point is 0.56, the baseline value is 0.28, the pressure deviation change is 0.50, and the corresponding synchronous deviation evaluation value is 0.3935. The data in Table 1 are used to quantify and compare the coupling enhancement relationship between the displacement deviation and pressure deviation changes of the lifting point within the positioning deviation section. This provides direct input for extracting the maximum synchronous deviation evaluation value within the positioning deviation section, comparing it with the deviation threshold to mark abnormal lifting points, and further outputting structural adjustment parameters such as the hydraulic cylinder piston stroke adjustment amount and the width of the expansion ring, portal steel clamp, and circumferential seam connection gap.

[0046] Table 1. Data on Synchronization Deviation Assessment Values

[0047]

[0048] like Figure 3The figure shows the variation curves of the synchronization deviation evaluation values ​​for five sampling points within the positioning deviation section for the same lift point. The line graph uses the sampling point number as the x-axis and the synchronization deviation evaluation value as the y-axis. Blue line points represent the calculated synchronization deviation evaluation value for that lift point at the corresponding sampling point, while red line points mark the maximum synchronization deviation evaluation value. The corresponding synchronization deviation evaluation value is also labeled above the line points to provide a visual comparison of the evolution of synchronization deviation within the deviation section. Specifically, the synchronization deviation evaluation values ​​for sampling points 1 to 5 are 0.0431, 0.0779, 0.1537, 0.2600, and 0.3935, respectively, showing a continuous upward trend. The synchronization deviation evaluation value for sampling point 5, 0.3935, is the maximum value within the section, corresponding to the red-marked point. This value serves as the input for subsequently extracting the maximum synchronization deviation evaluation value for that lift point within the positioning deviation section and comparing it with the deviation threshold to mark abnormal lift points. Figure 3 By using a visualization method that combines line graphs, highlights maximum values, and numerical annotations, the cumulative amplification process of synchronous deviation of the lifting point within the positioning deviation section is presented in a structured manner. This provides a directly applicable evaluation basis for subsequent determination of abnormal lifting points based on positioning change trigger points, and further output of structural adjustment amounts such as hydraulic cylinder piston stroke adjustment, expansion ring and wall panel fitting gap, portal steel clip arrangement spacing, and circumferential seam butt joint gap width.

[0049] In this implementation scheme, by using the location change trigger point as the unified alignment time, the nonlinear modulation relationship between the intra-segment benchmark level and the pressure deviation change of the relative displacement difference of the lifting point is incorporated into the synchronization deviation evaluation value. This allows the synchronization deviation status to form a directly sortable lateral comparison scale within the same lifting segment, thereby quickly locking the lifting point with a significant deviation at the trigger time, reducing the interference of benchmark drift on the anomaly indication, improving the stability and consistency of the abnormal lifting point marking, and providing a more discriminative sample division basis for the subsequent establishment of a lifting point adjustment parameter estimation model based on the normal lifting point assembly construction monitoring data.

[0050] Specifically, the steps for marking normal and abnormal lifting points are as follows: For each lifting point, calculate the synchronization deviation assessment value corresponding to each sampling point within the corresponding positioning deviation section. During the calculation process, locate the current lifting segment record entry according to the lifting segment number, and limit the sampling point range based on the start and end timestamps of the positioning deviation section. Within the limited range, read the pressure deviation change, relative displacement difference change, and relative displacement difference change benchmark value corresponding to the lifting point point one by one according to the collection timestamp from smallest to largest. Then, generate the synchronization deviation assessment value of the lifting point within the positioning deviation section according to the specific calculation formula. The system generates a sequence of deviation assessment values ​​and extracts the maximum value of the synchronization deviation assessment value. During the extraction process, the sequence of synchronization deviation assessment values ​​is compared point by point, and the maximum value is written into the maximum deviation field of the boost point. The maximum deviation field of the boost point is used to characterize the strongest synchronization deviation level of the boost point in the positioning status deviation section. The maximum value of the synchronization deviation assessment value and the deviation threshold are compared in real time. When the maximum value of the synchronization deviation assessment value is less than or equal to the deviation threshold, the corresponding boost point is marked as a normal boost point, and the normal mark is written into the boost point status field. When the maximum value of the synchronization deviation assessment value is greater than the deviation threshold, the corresponding boost point is marked as an abnormal boost point, and the abnormal mark is written into the boost point status field.

[0051] In this implementation plan, by forming a sequence of synchronous deviation evaluation values ​​indexed by the collection timestamp within the deviation section of the positioning status, and using the maximum value of the synchronous deviation evaluation value as the basis for judging the lifting point level, the division between normal lifting points and abnormal lifting points can cover the most unfavorable fluctuation situation within the deviation section. This reduces the risk of occasional fallback of a single sampling point masking the identification of anomalies, improves the verifiability and consistency of lifting point status markings within the same lifting segment, and provides a more stable sample boundary for subsequent estimation of lifting point adjustment parameters based on the assembly construction monitoring data of normal lifting points.

[0052] Specifically, the steps for constructing and training the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points are as follows: For each lifting segment, extract the assembly construction monitoring data of all normal lifting points. During the extraction process, locate the current lifting segment record entry according to the lifting segment number, and traverse the data records corresponding to the normal lifting points according to the collection timestamp from smallest to largest. The working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel are spliced ​​together as a state feature vector according to the same collection timestamp. The gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap are spliced ​​together as a construction parameter vector according to the same collection timestamp. The state feature vector and the construction parameter vector are written into the sample records according to the collection timestamp to construct a normal construction state sample set. The support vector regression algorithm is used to construct the lifting point adjustment parameter estimation model. During the construction process, all state feature vectors in the normal construction state sample set are used to form a feature matrix, and the gap between the expansion ring and the wall panel is used as the basis for the model. The model is trained using three sets of regression targets: a first target vector based on the spacing of the portal steel clamps, a second target vector based on the spacing of the circumferential joint gap, and a third target vector based on the width of the circumferential joint gap. Three sets of support vector regression mapping relationships are established, with the mapping relationships expressed using a radial basis function kernel to represent the nonlinear association and an insensitive loss function to constrain the regression residuals. The model is trained iteratively, with candidate value sets set for the penalty parameter, kernel width parameter, and insensitive interval parameter during training. Cross-validation is performed on the normal construction state sample set, and the candidate value sets are iterated through group by group while training the regression mapping relationships. The back-calculation error is calculated based on the cross-validation results, and the parameter combination with the smallest back-calculation error is selected as the regression mapping parameter for the current lifting segment. A sequential minimum optimization algorithm is used to iteratively update the regression mapping relationship, and back-calculation is performed on the normal construction state sample set after each iteration to obtain the back-calculation error of the regression mapping. When the back-calculation error of the regression mapping is less than the error threshold in K consecutive regression iterations, the training of the lifting point adjustment parameter estimation model is considered complete, where K ranges from 5 to 20 consecutive regression iterations.

[0053] In this implementation plan, by using the assembly and construction monitoring data of normal lifting points to form a state feature vector and construction parameter vector with a one-to-one correspondence between the collection timestamps within the lifting segment, and using support vector regression mapping relationship to jointly estimate the gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap, the estimated construction parameters can maintain corresponding constraints in the same data context as the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel. This improves the comparability and stability of the adjustment parameter output of abnormal lifting points, reduces the risk of inter-segment drift caused by manual experience setting, and enhances the feasibility of multi-parameter coordinated adjustment when generating subsequent control execution instruction sets.

[0054] Specifically, the steps for inputting the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, outputting the corresponding adjustment parameters, and sending them to the corresponding control unit for execution are as follows: The assembly construction monitoring data of abnormal lifting points in the corresponding lifting segment are sequentially input into the trained lifting point adjustment parameter estimation model. During the input process, the lifting segment record entry to which the abnormal lifting point belongs is located according to the lifting segment number. The working pressure of the single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of the single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel corresponding to the abnormal lifting point are read from smallest to largest according to the collection timestamp. The abnormal state feature vectors are concatenated and written into the model input sequence one by one, triggering regression estimation output. The corresponding wall panel bonding gap adjustment, portal steel clip arrangement spacing adjustment, and circumferential joint gap width adjustment are output. During the output process, the wall panel bonding gap estimation value obtained by the model is subtracted from the current expansion ring and wall panel bonding gap at the abnormal lifting point to obtain the wall panel bonding gap adjustment amount. The wall panel bonding gap adjustment amount is then mapped to the radial tightening amount adjustment command of the expansion ring assembly. By adjusting the radial tightening displacement of the expansion ring assembly, the bonding and pressing degree of the expansion ring on the wall panel is changed to achieve controllable bonding gap between the expansion ring and the wall panel. Correction; The difference between the estimated portal steel clamp spacing obtained from the model and the current portal steel clamp spacing at the abnormal lifting point is used to obtain the portal steel clamp spacing adjustment amount. This adjustment amount is then mapped to the steel clamp position adjustment command of the portal steel clamp actuator. By adjusting the movement of the portal steel clamp in the circumferential arrangement position, the spacing between adjacent steel clamps is updated to achieve controllable correction of the portal steel clamp spacing. Simultaneously, the portal steel clamp position adjustment command is bound to the portal steel clamp clamping force setting value to maintain consistent clamping state. The difference between the estimated circumferential seam butt gap width obtained from the model and the current circumferential seam butt gap width at the abnormal lifting point is used to obtain the circumferential seam... The width adjustment amount of the circumferential seam gap is mapped to the circumferential tension adjustment command of the expansion ring assembly. By adjusting the circumferential tension displacement of the expansion ring assembly, the relative proximity of the wall panel docking edges is changed to achieve controllable correction of the circumferential seam gap width. In addition, the hydraulic cylinder piston stroke adjustment amount is determined by combining the current hydraulic cylinder piston stroke state at the abnormal lifting point. In the determination process, the current single lifting stroke displacement at the abnormal lifting point is used as the stroke adjustment benchmark, and the stroke adjustment direction is determined according to the magnitude of the synchronous deviation evaluation value of the abnormal lifting point. The stroke adjustment benchmark and the stroke adjustment direction are combined to generate the hydraulic cylinder piston stroke adjustment amount.Each adjustment quantity is written into the control execution instruction set. During the writing process, an instruction number is generated for each adjustment quantity and bound to the lifting segment number, collection timestamp, and abnormal lifting point index. The radial tightening adjustment instruction of the expansion ring corresponding to the wall panel fitting gap adjustment quantity is written into the radial adjustment field of the expansion ring. The steel clip position adjustment instruction corresponding to the portal steel clip arrangement spacing adjustment quantity is written into the steel clip position adjustment field and the steel clip clamping force setting value field. The circumferential tension adjustment instruction of the expansion ring corresponding to the circumferential joint gap width adjustment quantity is written into the circumferential adjustment field of the expansion ring. The hydraulic cylinder piston stroke adjustment quantity is written into the stroke adjustment field. Hydraulic cylinder piston stroke adjustment instructions are issued to the hydraulic pump station control unit corresponding to the abnormal lifting point. During the issuance process, the stroke adjustment field is mapped to the target stroke setting value of the hydraulic pump station control unit and written into the issuance timestamp. At the same time, the radial tightening adjustment instruction and the circumferential tension adjustment instruction of the expansion ring are issued to the expansion ring assembly at the corresponding position. The instruction, during its issuance, maps the radial adjustment field of the expansion ring to the radial tightening displacement setting value of the expansion ring, and maps the circumferential adjustment field of the expansion ring to the circumferential tensioning displacement setting value of the expansion ring, and writes this value to the issuance timestamp. Simultaneously, it issues steel clamp position adjustment instructions and steel clamp clamping force setting values ​​to the corresponding portal steel clamp actuators. During issuance, the steel clamp position adjustment field is mapped to the circumferential movement setting value of the steel clamp, and the steel clamp clamping force setting value field is mapped to the target clamping force value of the steel clamp, and written to the issuance timestamp. After the control is executed, it continues to collect assembly construction monitoring data within the corresponding lifting segment. During the data collection process, it continuously updates and writes back to the assembly construction monitoring data record entries according to the collection timestamp, based on the working pressure of a single hydraulic jack, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, the vertical offset of the top of the wall panel, the gap between the expansion ring and the wall panel, the spacing of the portal steel clamps, and the width of the circumferential joint butt gap, forming a closed-loop control mechanism.

[0055] In this implementation plan, a one-to-one correspondence is established between the assembly construction monitoring data of abnormal lifting points and the adjustment amounts of wall panel fitting gap, portal steel clamp arrangement spacing, circumferential joint butt gap width, and hydraulic cylinder piston stroke. The adjustment parameters are directly associated with the executable set values ​​of the hydraulic pump station control unit, expansion ring assembly, and portal steel clamp actuator. This allows the handling of the positioning change trigger point to form a reproducible closed-loop adjustment link within the same lifting segment. This improves the timeliness and consistency of abnormal lifting point handling and parameter implementation consistency, reduces the probability of synchronous deviation spread caused by local connection stiffness fluctuations during construction, and enhances the stability of wall panel sliding assembly positioning control under continuous operation conditions between segments.

[0056] like Figure 4The diagram shows a partial cross-sectional view of the expansion ring assembly during the sliding assembly of the tank wall panel. The expansion ring is arranged along the inner side of the tank wall and fits against the inner surface of the wall panel, with a fitting gap between the outer edge of the expansion ring and the wall panel. The clamp is fitted with the expansion ring to constrain the relative position of the expansion ring and the wall panel. The diagram also shows the lower connection between the wall panel and the bottom plate, as well as the upper transition between the wall panel and the spherical dome, reflecting the local geometric assembly structure between the expansion ring, clamp, wall panel, bottom plate, and spherical dome. This provides a structural basis for subsequently extracting the fitting gap between the expansion ring and the wall panel and implementing corresponding structural adjustments.

[0057] like Figure 5 As shown, the local distribution relationship of the expansion ring assembly and clamps along the circumferential direction of the tank wall is illustrated. The expansion ring extends circumferentially along the inner side of the wall, and multiple clamps are distributed at intervals along the circumferential direction and cooperate with the expansion ring. The circumferential spacing between adjacent clamps corresponds to the arrangement spacing of the portal steel clamps, which is used to reflect the constraint arrangement state of the expansion ring and the wall within the circumferential range. This provides a direct structural basis for extracting the arrangement spacing of the portal steel clamps and implementing corresponding structural adjustments at the positioning change trigger point.

[0058] like Figure 2 As shown, the second aspect of the present invention provides a positioning control system for intelligent sliding assembly of thin-walled steel structure tank panels, including: a construction data acquisition and preprocessing module, a panel assembly positioning state construction module, a panel assembly positioning evolution judgment module, and a panel assembly positioning correction control module. The construction data acquisition and preprocessing module is used to acquire assembly construction monitoring data and perform timestamp alignment, abnormal data removal, missing interval completion, and standardization processing on the assembly construction monitoring data to generate preprocessed assembly construction monitoring data. The panel assembly positioning state construction module is used to construct a continuous monitoring sequence within a segment based on the preprocessed assembly construction monitoring data according to the lifting segment number, and to monitor the pressure deviation change at each lifting point, the relative displacement difference change at the lifting point, and the overall average lifting position of the panel. The lifting height is uniformly quantified to generate a panel assembly positioning status dataset. The panel assembly positioning evolution judgment module evaluates the panel assembly positioning coupling status based on this dataset, generates a panel assembly positioning coupling evaluation value, and determines the positioning deviation section based on this value, extracting positioning change trigger points. The panel assembly positioning correction control module evaluates the synchronization deviation status of each lifting point based on the positioning change trigger points, marks normal and abnormal lifting points, constructs and trains a lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points, inputs the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, outputs the corresponding adjustment parameters, and sends them to the corresponding control unit for execution.

[0059] In this implementation plan, by integrating the assembly construction monitoring data into the wall panel assembly positioning status dataset, the wall panel assembly positioning coupling evaluation value is further used to uniformly guide the identification of positioning status deviation sections and the extraction of positioning change trigger points. At the positioning change trigger points, the synchronous deviation evaluation value is used to distinguish between normal lifting points and abnormal lifting points. Then, the assembly construction monitoring data of normal lifting points is used to form a usable benchmark for the lifting point adjustment parameter estimation model. This enables the positioning control process to have closed-loop consistency and parameter traceability from monitoring to interpretation to adjustment, reducing the risk of synchronous error within the lifting segment spreading from local fluctuations to the entire segment, and improving the stability of the wall panel sliding assembly positioning control under continuous operation conditions.

[0060] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0061] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall plates, characterized in that, Includes the following steps: S1 collects assembly construction monitoring data and performs timestamp alignment, abnormal data removal, missing interval completion and standardization processing on the assembly construction monitoring data to generate pre-processed assembly construction monitoring data; S2. Based on the pre-processed assembly construction monitoring data, a continuous monitoring sequence within the segment is constructed according to the lifting segment number. The pressure deviation change, relative displacement difference change and overall average lifting height of each lifting point are uniformly quantified and characterized to generate a panel assembly positioning status dataset. S3. Based on the wall panel assembly positioning status dataset, evaluate the wall panel assembly positioning coupling status, generate the wall panel assembly positioning coupling evaluation value, and determine the positioning status deviation section based on the wall panel assembly positioning coupling evaluation value, and extract the positioning change trigger point. S4. Based on the positioning change trigger point, evaluate the synchronization deviation status of each lifting point, mark the normal lifting point and the abnormal lifting point, construct and train the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of the normal lifting point, input the assembly construction monitoring data of the abnormal lifting point into the lifting point adjustment parameter estimation model, output the corresponding adjustment parameters, and send them to the corresponding control unit for execution.

2. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for collecting assembly construction monitoring data and performing timestamp alignment, outlier removal, missing interval completion, and standardization on the data to generate preprocessed assembly construction monitoring data are as follows: Real-time collection of tank wall panel assembly construction monitoring data, including collection timestamp, lifting segment number, working pressure of a single hydraulic jack, total output pressure of the hydraulic system, cumulative lifting height of a single lifting point, relative displacement difference of the lifting point, vertical offset of the top of the wall panel, gap between the expansion ring and the wall panel, spacing of the portal steel clips, and width of the circumferential seam butt joint gap. For the collected assembly construction monitoring data, the monitoring records are sorted according to the collection timestamp, and the nearest neighbor time alignment algorithm is used to map the multi-source asynchronous sampling records to a unified time axis; the median filtering algorithm is used to smooth the short-period fluctuation segments in the assembly construction monitoring data sequence; the box plot outlier detection algorithm is used to identify and remove abnormal sampling points in the assembly construction monitoring dataset, and the piecewise linear interpolation algorithm is combined to continuously complete the local missing intervals; the Z-Score standardization algorithm is used to perform numerical standardization processing on the assembly construction monitoring data to eliminate the dimensional differences between different physical quantities.

3. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 2, characterized in that: The specific steps for generating the panel assembly positioning status dataset based on the preprocessed assembly construction monitoring data are as follows: A continuous monitoring sequence within each lifting segment is constructed according to the segment number. The changes in pressure deviation at each lifting point, the changes in relative displacement difference at each lifting point, and the overall average lifting height of the panel are uniformly quantified and characterized. Read the pre-processed assembly construction monitoring data, group the construction monitoring data according to the lifting segment number, and within each lifting segment, arrange the working pressure of a single hydraulic jack, the total output pressure of the hydraulic system, the cumulative lifting height of a single lifting point, the relative displacement difference of the lifting point, and the vertical offset of the top of the wall panel in chronological order according to the collection timestamp to obtain a continuous monitoring sequence within the segment. Based on the continuous monitoring sequence within the segment, the number of lifting points within the current lifting segment is counted, and the total output pressure of the hydraulic system divided by the number of lifting points is used as the pressure average benchmark. The deviation of the working pressure of each individual hydraulic jack relative to the pressure average benchmark is calculated using a relative deviation calculation algorithm, and a first-order difference algorithm is performed on the deviation values ​​under adjacent time stamps to obtain the pressure deviation change corresponding to each lifting point. A first-order difference algorithm is performed on the relative displacement difference of each lift point at adjacent timestamps to obtain the change in the relative displacement difference of each lift point. The overall average lifting height of the wall panel is calculated based on the cumulative lifting height of each lifting point. The continuous monitoring sequence within the segment, the number of lifting points, the pressure deviation change corresponding to each lifting point, the relative displacement difference change corresponding to each lifting point, and the overall average lifting height of the wall panel are stored in a structured manner according to the lifting segment number to generate a wall panel assembly positioning status dataset.

4. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for evaluating the wall panel assembly positioning coupling state based on the wall panel assembly positioning state dataset and generating the wall panel assembly positioning coupling evaluation value are as follows: Read the panel assembly positioning status dataset, take the absolute value of the pressure deviation change corresponding to each lifting point in the current lifting segment, multiply it by the square of the relative displacement difference change of the corresponding lifting point, and then perform a summation operation to obtain the pressure displacement coupling cumulative amount; take the absolute value of the relative displacement difference change corresponding to each lifting point in the current lifting segment, perform a summation operation, and add a minima term to obtain the displacement normalization base; divide the pressure displacement coupling cumulative amount by the displacement normalization base, add one to the ratio, and then take the natural logarithm to obtain the basic coupling evaluation term; take the absolute value of the ratio of the vertical offset of the top of the panel in the current lifting segment to the average lifting height of the entire panel, and add one to obtain the attitude amplification term; multiply the basic coupling evaluation term and the attitude amplification term to obtain the panel assembly positioning coupling evaluation value of the current lifting segment.

5. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for determining the positioning deviation section based on the wall panel assembly positioning coupling evaluation value and extracting the positioning change trigger point are as follows: Extract the panel assembly positioning coupling evaluation values ​​continuously calculated according to the acquisition timestamp within the current lifting segment, and calculate the change in the panel assembly positioning coupling evaluation values ​​under adjacent timestamps to obtain the evaluation value change sequence; and construct the change judgment interval based on the median and interquartile range of the evaluation value change sequence. The time segment where the evaluation value change exceeds the upper limit of the judgment interval under N consecutive acquisition timestamps is determined as the positioning state deviation segment, and the timestamp of the first time that the positioning state deviation segment exceeds the upper limit of the judgment interval is determined as the positioning change trigger point.

6. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for evaluating the synchronization deviation status of each lifting point based on the positioning change trigger point are as follows: Read all positioning status deviation sections and corresponding positioning change trigger points according to the lifting segment number, and extract the relative displacement difference change of all lifting points in the current lifting segment under the timestamp corresponding to the positioning change trigger point. Take the median of the relative displacement difference change of the lifting points as the benchmark value of the relative displacement difference change of the lifting points in the current lifting segment. The absolute value of the difference between the relative displacement difference of the i-th lifting point and the baseline value of the relative displacement difference of the lifting points in the current lifting segment is taken to obtain the displacement deviation. The absolute value of the pressure deviation change corresponding to the i-th lifting point is taken, one is added, and the natural logarithm is taken. The logarithm is then added to one to obtain the pressure deviation modulation term. The displacement deviation is multiplied by the pressure deviation modulation term to obtain the synchronization deviation evaluation value of the i-th lifting point.

7. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for marking normal and abnormal lift points are as follows: For each lift point, calculate the synchronization deviation evaluation value corresponding to each sampling point within the corresponding positioning status deviation section, and extract the maximum value of the synchronization deviation evaluation value; compare the maximum value of the synchronization deviation evaluation value with the deviation threshold in real time. When the maximum value of the synchronization deviation evaluation value is less than or equal to the deviation threshold, mark the corresponding lift point as a normal lift point; when the maximum value of the synchronization deviation evaluation value is greater than the deviation threshold, mark the corresponding lift point as an abnormal lift point.

8. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 1, characterized in that: The specific steps for constructing and training the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points are as follows: For each lifting segment, extract the assembly construction monitoring data of all normal lifting points to construct a sample set of normal construction status; A model for estimating the adjustment parameters of the lifting point is constructed using the support vector regression algorithm. A sample set of normal construction conditions is used as the sample input. Regression mapping relationships are established for the gap between the expansion ring and the wall panel, the spacing of the portal steel clips, and the width of the circumferential joint gap. The model is trained by regression iteration. When the back calculation error of the regression mapping is less than the error threshold in K consecutive regression iterations, the model for estimating the adjustment parameters of the lifting point is considered to have been trained successfully.

9. The positioning control method for intelligent sliding assembly of thin-walled steel structure tank wall panels according to claim 8, characterized in that: The specific steps for inputting the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, outputting the corresponding adjustment parameters, and sending them to the corresponding control unit for execution are as follows: The assembly construction monitoring data of the abnormal lifting point of the corresponding lifting segment is sequentially input into the training completed lifting point adjustment parameter estimation model, and the corresponding wall panel bonding gap adjustment amount, portal steel clip arrangement spacing adjustment amount and circumferential seam butt gap width adjustment amount are output. Combined with the current hydraulic cylinder piston stroke status of the abnormal lifting point, the hydraulic cylinder piston stroke adjustment amount is determined. Each adjustment value is written into the control execution instruction set, and hydraulic cylinder piston stroke adjustment instructions are issued to the hydraulic pump station control unit of the corresponding abnormal lifting point, and structural adjustment instructions are issued to the expansion ring assembly and gantry steel clamp actuator at the corresponding position at the same time. After the control is executed, the assembly construction monitoring data in the corresponding lifting segment is collected to form a closed loop of control.

10. A positioning control system for intelligent sliding assembly of thin-walled steel structure tank wall panels, characterized in that, include: The system includes a construction data acquisition and preprocessing module, a wall panel assembly and positioning status construction module, a wall panel assembly and positioning evolution interpretation module, and a wall panel assembly and positioning correction and control module, among which: The construction data acquisition and preprocessing module is used to collect assembly construction monitoring data, and to perform timestamp alignment, abnormal data removal, missing interval completion and standardization processing on the assembly construction monitoring data to generate preprocessed assembly construction monitoring data. The wall panel assembly positioning status construction module is used to construct a continuous monitoring sequence within a segment based on the pre-processed assembly construction monitoring data according to the segment number, and to uniformly quantify the pressure deviation change, the relative displacement difference change of each lifting point, and the overall average lifting height of the wall panel, thereby generating a wall panel assembly positioning status dataset. The panel assembly positioning evolution interpretation module is used to evaluate the panel assembly positioning coupling state based on the panel assembly positioning state dataset, generate the panel assembly positioning coupling evaluation value, determine the positioning state deviation section based on the panel assembly positioning coupling evaluation value, and extract the positioning change trigger point. The panel assembly positioning correction control module is used to evaluate the synchronous deviation status of each lifting point based on the positioning change trigger point, mark normal lifting points and abnormal lifting points, construct and train the lifting point adjustment parameter estimation model based on the assembly construction monitoring data of normal lifting points, input the assembly construction monitoring data of abnormal lifting points into the lifting point adjustment parameter estimation model, output the corresponding adjustment parameters, and send them to the corresponding control unit for execution.

Images