Intelligent control method and system for construction of long-span steel reinforced concrete structure

Abstract

The present application relates to the technical field of process control, in particular to a large-span steel reinforced concrete structure construction intelligent control method and system, comprising the following steps: obtaining control parameters and generating a support point support corresponding relationship, synchronously collecting elevation strain support force and execution height, eliminating timeout measurement points or support points, forming a support influence relationship by micro-motion exploration, determining compensation measurement point direction and deviation, screening co-directional support and converting and correcting height adjustment amount, generating support adjustment instructions and checking remaining elevation deviation. In the present application, the measurement point support distance relationship is coupled with the micro-motion exploration response, the actual force transmission influence on site is converted into adjustment basis, the compensation direction is more consistent with the structure deformation state, the timeout is excluded and the force limit travel limit amplitude constraint is used, the hysteresis data and overload overshoot interference are reduced, the elevation deviation is gradually converged, and continuous closed-loop basis is provided for staged construction, and the construction process control stability is improved.

Description

Technical Field

[0001] This invention relates to the field of process control technology, and in particular to intelligent control methods and systems for the construction of large-span steel-concrete structures. Background Technology

[0002] The field of process control technology involves the use of industrial computers, controllers, sensors, and actuators to collect, judge, and output process parameters, structural states, and control commands during construction. Among these, the traditional intelligent control method for large-span steel-concrete composite structure construction refers to the use of BIM modeling, construction simulation, monitoring instruments, and on-site control equipment in the construction of large-span steel structures or steel-concrete composite structures. This involves recording the steel hoisting, temporary support, concrete pouring, support unloading, and structural deformation processes in stages. Monitoring data is transmitted from displacement sensors, strain sensors, load sensors, and data acquisition devices to an industrial control computer or PLC. The control program then generates lifting, unloading, adjustment, or alarm commands according to preset procedures, thresholds, and control logic.

[0003] Existing technologies operate based on construction phase records and preset thresholds. Monitoring data is collected by acquisition devices and then output by the control program as lifting, unloading, or alarm commands. On-site judgment relies on process experience and fixed limits. The actual force transmission between the measuring point and temporary support is difficult to reflect in a timely manner. Data arrival delays still participate in the judgment. When the support force is close to the limit or the execution height is close to the boundary, adjustment lag is still likely to occur, resulting in insufficient accuracy of elevation compensation direction, increased risk of local stress, decreased control continuity, and difficulty in stable convergence of residual deviation. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide an intelligent control method for the construction of large-span steel-concrete composite structures.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent control method for the construction of large-span steel-concrete composite structures, comprising the following steps:

[0006] Obtain the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment and arrival time limit, and generate the correspondence between the measuring point support according to the distance between the structural control measuring point and the temporary support point;

[0007] Simultaneously collect the current elevation, member strain, support force, and height of the support actuator, exclude the structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit, and generate a list of valid collection objects;

[0008] Based on the list of valid data collection objects, for the support actuators that are not excluded and whose support force is lower than the upper limit of the support force and whose height has not reached the travel boundary, the micro-motion test height increment is performed, and the height, support force, elevation and strain change are calculated to generate the support influence relationship;

[0009] Compare the current elevation with the target elevation, and determine the compensation measuring point, compensation direction and compensation elevation deviation from the structural control measuring points whose absolute deviation exceeds the measured elevation deviation according to the absolute deviation;

[0010] Based on the support influence relationship and the corresponding relationship of the measuring point support, the support actuators that are consistent with the compensation direction are selected, the compensation elevation deviation is converted into a height adjustment amount, the height adjustment amount is corrected according to the single adjustment height limit and the support force limit, a support adjustment command is generated, and the remaining elevation deviation is checked.

[0011] As a further aspect of the present invention, the step of generating the correspondence between the measuring point and the temporary support point according to the distance between the structural control measuring point and the temporary support point includes:

[0012] Obtain the coordinates of the structural control measuring points, the coordinates of the temporary support points, and the construction stage identifier; read the target elevation, the elevation deviation, and the arrival time limit from the construction status configuration table of the digital twin platform; read the upper limit of the support force, the stroke boundary, the upper limit of the single adjustment height, and the micro-motion test height increment from the parameter register table of the support actuator controller; and generate a parameter snapshot.

[0013] Calculate the spatial distance between each of the structural control measuring points and the temporary support points, arrange the temporary support points according to the spatial distance, and sequentially associate the temporary support points whose spatial distance is not greater than a preset association distance threshold with the structural control measuring points to generate the measuring point support correspondence.

[0014] As a further aspect of the present invention, the step of generating the list of valid collection objects specifically includes:

[0015] The current elevation and strain of the rod are collected from the elevation sensor and strain sensor deployed at the control measurement point of the structure. The support force and the height of the support actuator are collected from the pressure sensor at the temporary support point and the displacement encoder of the support actuator. Each collected value is written with the device identifier, sampling time and receiving time to generate a synchronous acquisition record.

[0016] Align the synchronous acquisition records with the reference time of the same sampling period, calculate the arrival time of the reference time corresponding to the reception time of each synchronous acquisition record, and when the arrival time exceeds the arrival time limit, write the corresponding structural control measurement point or the temporary support point into the exclusion list to generate a list of valid acquisition objects.

[0017] As a further aspect of the present invention, the step of generating the supporting influence relationship specifically includes:

[0018] Based on the list of valid collected objects, the support actuators that are not included in the exclusion list, whose current support force is lower than the upper limit of the support force, and whose current height has not reached the travel boundary are screened to determine the test support actuators. A micro-motion test command is sent to each of the test support actuators so that the test support actuators can complete the height adjustment according to the micro-motion test height increment.

[0019] The height, support force, current elevation, and strain of the members before and after adjustment are collected. The changes in height, support force, elevation, and strain are calculated respectively. The test support actuator is associated with the structural control measuring point according to the corresponding relationship of the measuring point support. A mapping record between the direction of influence, the magnitude of change, and the corresponding identifier is established to generate the support influence relationship.

[0020] As a further aspect of the present invention, the step of establishing the mapping record specifically includes:

[0021] Obtain the original micro-motion test records, and label the original micro-motion test records according to the construction stage, micro-motion test height increment, height change, support force change, elevation change and strain change. Use the original micro-motion test records to train an anomaly recognition model based on the isolated forest model, support vector machine model or long short-term memory network model, and generate anomaly recognition model parameters.

[0022] The current micro-motion trial record is input into the anomaly recognition model based on the isolated forest model, support vector machine model, or long short-term memory network model to obtain the anomaly probability. When the anomaly probability exceeds a preset anomaly probability threshold, the corresponding mapping record is deleted. When the anomaly probability does not exceed the preset anomaly probability threshold, the corresponding mapping record is written into the support influence relationship.

[0023] As a further aspect of the present invention, the steps for determining the compensation measuring point, the compensation direction, and the compensation elevation deviation are specifically as follows:

[0024] Obtain the current elevation and target elevation corresponding to the unexcluded structural control measuring points; calculate the elevation deviation and absolute value of the deviation for each structural control measuring point; compare the absolute value of the deviation with the elevation deviation; when the absolute value of the deviation exceeds the elevation deviation, write the corresponding structural control measuring point into the table of measuring points to be compensated.

[0025] The structural control measuring points in the table of measuring points to be compensated are arranged in descending order of the absolute value of the deviation. The structural control measuring point ranked first is determined as the compensation measuring point. The compensation direction is determined according to the positive or negative state of the measured elevation deviation, and the absolute value of the deviation is determined as the compensation elevation deviation.

[0026] As a further aspect of the present invention, the step of arranging the structural control measuring points according to the absolute value of the deviation from largest to smallest is as follows:

[0027] Obtain the original compensation records, label the original compensation records according to the construction stage, absolute value of deviation, number of associated temporary support points, strain change of members and compensation completion status, use the original compensation records to train a compensation priority model based on gradient boosting decision tree model, and generate priority anomaly identification model parameters;

[0028] The current construction stage, the absolute value of the deviation, the number of associated temporary support points, and the strain change of the member are input into the compensation priority model based on the gradient lifting decision tree model to obtain the compensation priority. The compensation priority is used as the sorting basis within the same absolute value range of the deviation to generate a sequence of compensation measurement points.

[0029] As a further aspect of the present invention, the steps of generating the support adjustment command and verifying the remaining elevation deviation are specifically as follows:

[0030] Read the support influence relationship and the measurement point support correspondence relationship, filter the support actuators that are consistent with the compensation direction and associated with the compensation measurement point, and arrange them in descending order of the absolute value of the elevation change that is consistent with the compensation direction to generate a candidate support actuator sequence;

[0031] The unit height elevation response is determined based on the micro-motion test height increment and the corresponding elevation change. The compensated elevation deviation is converted into an initial height adjustment amount. The initial height adjustment amount is compared with the single adjustment height upper limit. The initial height adjustment amount is then corrected by combining the current support force, the support force change amount, and the support force upper limit, thereby generating the support adjustment command.

[0032] After executing the support adjustment command, the current elevation is collected, and the current elevation is compared with the target elevation to obtain the remaining elevation deviation.

[0033] As a further aspect of the present invention, the step of correcting the initial height adjustment amount specifically includes:

[0034] Obtain the original support adjustment record, which records the real-time height adjustment amount, unit height elevation response amount, support force change amount, member strain change amount, and remaining elevation deviation after adjustment. Use the original support adjustment record to train a machine learning residual deviation prediction model, and make the machine learning residual deviation prediction model output the expected remaining elevation deviation based on the candidate height adjustment amount.

[0035] Within the range not exceeding the upper limit of the single adjustment height, a candidate height adjustment amount sequence is generated according to a preset height interval. The expected support force is determined based on the current support force, the change in support force, and the candidate height adjustment amount. The candidate height adjustment amount whose expected support force is lower than the upper limit of the support force is input into the machine learning residual deviation prediction model to obtain the expected residual elevation deviation corresponding to the candidate height adjustment amount. The candidate height adjustment amount with the smallest absolute value of the expected residual elevation deviation is determined as the corrected height adjustment amount.

[0036] The intelligent control system for the construction of large-span steel-concrete composite structures includes:

[0037] The control parameter acquisition module acquires the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment and arrival time limit, and generates the correspondence between the measuring point support according to the distance between the structural control measuring point and the temporary support point.

[0038] The data synchronization acquisition module synchronously acquires the current elevation, member strain, support force, and height of the support actuator, excluding the structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit;

[0039] The micro-motion test generation module performs the micro-motion test height increment on the support actuator that has not been eliminated and whose support force is lower than the upper limit of the support force and whose height has not reached the travel boundary, calculates the height, support force, elevation and strain change, and generates the support influence relationship;

[0040] The deviation compensation determination module compares the current elevation with the target elevation, and determines the compensation measurement point, compensation direction and compensation elevation deviation from the structural control measurement points whose absolute deviation value exceeds the measured elevation deviation according to the absolute deviation value.

[0041] The adjustment command generation module selects the support actuators that are consistent with the compensation direction based on the support influence relationship and the corresponding relationship of the measuring point support, converts the compensation elevation deviation into a height adjustment amount, corrects the height adjustment amount according to the single adjustment height limit and the support force limit, generates a support adjustment command, and verifies the remaining elevation deviation.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0043] In this invention, a correspondence between measuring points and supports is established around the target elevation deviation. The elevation strain support force and execution height are collected simultaneously, so that the control judgment has a state basis at the same time. Time-out data is eliminated to reduce the interference of lagging measurements on the compensation direction. The support influence relationship is generated by micro-motion probing, and the actual force transmission response is transformed into a calculable adjustment basis. The compensation measuring points and directions are determined according to the deviation magnitude, so that the elevation correction focuses on the main imbalance position. Then, the height adjustment amount is corrected by combining the upper limit of the single adjustment height and the upper limit of the support force at the stroke boundary, so as to avoid local support overload and overshoot. The support adjustment command is linked with the remaining deviation check, so that the elevation deviation continues to converge and the stability of construction control is improved. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is the main flowchart of the closed-loop control of the construction mode of the present invention;

[0046] Figure 2 This is a schematic diagram illustrating the intelligent control layout effect at the construction site of the present invention;

[0047] Figure 3 This is a schematic diagram illustrating the data interaction effect of the digital twin platform of the present invention;

[0048] Figure 4 This is a schematic diagram showing the effect of elevation deviation compensation before and after the present invention;

[0049] Figure 5 This is a schematic diagram illustrating the screening effect of the candidate height adjustment amount in this invention. Detailed Implementation

[0050] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0051] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0052] Please see Figures 1 to 5 This invention provides an intelligent control method for the construction of large-span steel-concrete composite structures. In practical applications, such as during the construction of large-span steel-concrete composite structures, a digital twin platform continuously receives BIM models, finite element models of the construction phase, on-site IoT monitoring data, and feedback data from adaptive support equipment. Structural control measuring points are deployed at mid-span, support transition zones, component assembly nodes, and temporary support influence zones. Temporary support points are deployed along the force transmission path during the construction phase. The support execution mechanism receives height adjustment commands through a controller and returns support force, height, and execution confirmation status. This method performs closed-loop processing around construction phase switching, temporary support removal, component installation sequence, and morphological changes caused by temperature. It incorporates target elevation, on-site elevation, member strain, support force, and support height into the same control link, allowing morphological deviations during construction to be collected, verified, and screened first. Then, the support influence relationship is established through micro-motion testing. Subsequently, compensation measuring points, compensation directions, and adjustment commands are determined. After execution, the elevation is re-collected to complete the remaining deviation verification. The method includes the following steps:

[0053] S1: Obtain the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment, and arrival time limit. Generate the correspondence between the measurement points and temporary support points based on the distance between the structural control measurement points and the temporary support points. The target elevation is derived from the forming control elevation field output by the construction stage model in the digital twin platform. The upper limit of support force is derived from the force boundary field in the parameter register table of the support actuator controller. The travel boundary is derived from the mechanically reachable boundary field corresponding to the displacement encoder. The upper limit of single adjustment height is derived from the adjustment constraint field of the support controller. The micro-motion test height increment is derived from the support influence identification task configuration field. The arrival time limit is derived from the IoT acquisition link validity configuration field. The structural control measurement points are the field monitoring objects that bear the current elevation and member strain. The temporary support points are the field support objects that bear the support force, support height, and actuator status. The measurement point support correspondence is a spatial association record between the structural control measurement points and the temporary support points. The record includes the measurement point identifier, support point identifier, construction stage identifier, distance sorting status, association validity status, and subsequent call status, which are read by the micro-motion test, deviation compensation, and support adjustment command generation stages.

[0054] S101: Obtain the coordinates of structural control points, temporary support points, and construction stage identifiers. Read the target elevation, elevation deviation, and arrival time limit from the construction status configuration table of the digital twin platform. Read the upper limit of support force, travel boundary, upper limit of single adjustment height, and micro-motion test height increment from the parameter register table of the support actuator controller, and generate a parameter snapshot. The coordinates of the structural control points are formed by BIM component positioning data, on-site verification positioning data, and construction coordinate transformation records. The coordinates of the temporary support points are formed by the temporary support layout model, support base positioning records, and support top connection records. The construction stage identifier is used to describe the component installation status, temporary support status, system transformation status, and temperature effect status. Before entering parameter reading, it is matched with the current construction task record. After successful matching, it serves as the index field for reading the construction status configuration table and controller parameter register table. The parameter snapshot is a set of fields frozen within the same control cycle, including target elevation, elevation deviation, arrival time limit, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment, field source identifier, reading status identifier, and update version identifier. When the construction status configuration table is read abnormally, the previous valid configuration is read and written into the inherited status; when the controller returns a missing field, the corresponding supporting actuator enters the parameter pending confirmation state and does not participate in this round of micro-motion probing and formal adjustment.

[0055] S102: Calculate the spatial distance between each structural control measuring point and temporary support point. Arrange the temporary support points according to the spatial distance, and sequentially associate temporary support points with a spatial distance not greater than the preset association distance threshold with the structural control measuring points to generate the corresponding relationship between the measuring points and supports. Before processing the spatial distance, verify whether the coordinates of the structural control measuring points and the coordinates of the temporary support points are in the same construction coordinate system. If the coordinate system identifiers are inconsistent, call the construction coordinate transformation record in the digital twin platform to unify them. If the transformation record is missing, the corresponding measuring point or support point is written to the coordinate pending verification status and does not enter the effective association. The preset association distance threshold is the association boundary field in the construction status configuration table, which comes from the structural span, support layout scheme, force transmission path in the finite element model, and construction stage control strategy. After the distance is arranged, the structural control measuring points read the temporary support points according to the spatial relationship from near to far. At the same time, candidate association records are generated by combining the activation status of the temporary support points, the online status of the support actuator, the top connection confirmation status, and the controller feedback status. When a candidate temporary support point is within the associated boundary and is in a controllable state, the corresponding relationship of the measurement point support is written; when it is within the associated boundary but the controller is offline, the travel is unreadable, or the connection is not confirmed, it is written to the associated cache to be restored and re-verified in subsequent collection cycles.

[0056] S2: Synchronously acquire current elevation, member strain, support force, and support actuator height, excluding structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit, and generate a list of valid acquisition objects. The current elevation is output by elevation sensors deployed at the structural control measuring points and bound to the corresponding component position via a digital twin platform; member strain is output by strain sensors to describe changes in the component's stress state; support force is output by pressure sensors at temporary support points to describe the support's stress state; and the support actuator height is output by a displacement encoder to describe the support's expansion and contraction position. The synchronous acquisition process uses the sampling period reference time as the alignment basis, writing data records uploaded by different sensors and controllers into the acquisition buffer, and writing device identifier, sampling time, receiving time, communication link status, and field integrity status for each record. The list of valid acquisition objects is the set of objects allowed to enter the micro-motion probing, deviation identification, and adjustment command generation in this round of control, including measuring point identifier, support point identifier, acquisition field integrity status, arrival status, exclusion reason, previous valid status inheritance identifier, and subsequent call identifier. During sensor cold start, an initial state is generated based on the device's self-test status and the first valid data acquisition record; after entering stable data acquisition, the valid state is updated based on continuous data acquisition records.

[0057] S201: Elevation and strain of members are collected from elevation and strain sensors deployed at structural control points. Support force and height of the support actuator are collected from pressure sensors at temporary support points and displacement encoders at the support actuators. Each collected value is written with its device identifier, sampling time, and receiving time, generating a synchronous acquisition record. Before acquisition, the elevation, strain, pressure, and displacement encoders are bound to their device identifiers. A fixed correspondence is established between the device identifier and the BIM component code, structural control point identifier, or temporary support point identifier in the digital twin platform. Before the collected values ​​enter the synchronous acquisition record, field integrity verification, device online status verification, sampling time format verification, and source consistency verification are performed. When a field is missing, the synchronous acquisition record retains the device identifier, the name of the missing field, and the missing status. When the sampling time cannot be parsed, the corresponding record enters a time-pending verification state. When duplicate records exist for the same device within the same acquisition cycle, records with complete fields, confirmed link status, and passed device self-tests are prioritized, and duplicate records are written to the duplicate acquisition log. Synchronous acquisition records include acquisition object identifier, acquisition field name, acquisition field category, sampling time, reception time, communication link status, device self-test status, raw acquisition status, and preprocessing status.

[0058] S202: Synchronously acquire records according to the reference time of the same sampling period, calculate the arrival time of the reference time corresponding to the reception time of each synchronously acquired record, and when the arrival time exceeds the arrival time limit, write the corresponding structural control measurement point or temporary support point into the exclusion list to generate a list of valid acquisition objects. The reference time is generated by the acquisition scheduler when the current control cycle starts, the reception time is written by the digital twin platform data access interface when the record enters the buffer, and the arrival time is used as the acquisition link status field for validity judgment. The arrival time limit is read from the parameter snapshot and is used to limit whether the acquisition record can still belong to the current control cycle. When the arrival status does not exceed the configuration boundary and the field integrity verification passes, the corresponding structural control measurement point or temporary support point is written into the list of valid acquisition objects; when the arrival status exceeds the configuration boundary, the field is missing and cannot be recovered, the device identifier cannot be matched, the communication link returns an abnormality, or the acquired value is in a sensor distortion state, the corresponding object is written into the exclusion list. The exclusion list records the object identifier, exclusion field, exclusion reason, occurrence stage, previous valid status, and feedback path. When structural control measuring points are excluded, they are not included in the current elevation deviation ranking; when temporary support points are excluded, they are not included in the micro-motion testing and formal adjustment screening.

[0059] S3: Based on the list of valid data acquisition objects, perform micro-motion testing on support actuators that are not excluded and whose support force is below the upper limit and whose height has not reached the travel boundary. Calculate the changes in height, support force, elevation, and strain to generate support influence relationships. The support influence relationship is a data object describing the corresponding state between the micro-motion action of the support actuator and the response of the structural control measurement point. It includes the test support actuator identifier, associated structural control measurement point identifier, construction stage identifier, micro-motion test height increment field, height change status, support force change status, elevation change status, strain change status, influence direction, change amplitude level, anomaly identification status, and subsequent call identifier. The influence direction indicates the correspondence between the elevation change direction of the associated structural control measurement point and the support adjustment direction after the support actuator micro-motion. The change amplitude level indicates the response category corresponding to the structural response in the construction status configuration table. Micro-motion testing is only performed on support actuators in the list of valid data acquisition objects that meet the control boundaries. The judgment that the support force is below the upper limit is based on the current support force field and the force boundary field in the parameter snapshot. The judgment that the current height has not reached the travel boundary is based on the displacement encoder height field and the mechanically reachable boundary field. When the support force field is missing, the height field is missing, the controller fails to acknowledge receipt of the command, or there is an abnormality in the connection at the top of the support, the corresponding support actuator enters a state where probing is prohibited.

[0060] S301: Based on the list of valid data acquisition objects, filter support actuators that are not written into the exclusion list, whose current support force is lower than the upper limit of the support force, and whose current height has not reached the travel boundary. Determine the test support actuators and send micro-motion test commands to each test support actuator, enabling them to complete height adjustment according to the micro-motion test height increment. Test support actuators are support objects allowed to perform micro-motion actions in this round of control. The determination process reads the list of valid data acquisition objects, parameter snapshots, online status of support actuators, and the correspondence between measuring points and support. The screening process first removes temporary support points from the exclusion list, then reads the support force field and height field. Support actuators in the force boundary, mechanical boundary, unconfirmed controller, or unconfirmed top connection states are marked with a "prohibited test" flag. The screened support actuators are written into the test task cache and, together with the associated structural control measuring point flags, form a test task record. The micro-motion test command includes the support actuator identifier, construction stage identifier, test direction, micro-motion test height increment field, command generation status, command issuance status, and execution confirmation status. After receiving the instruction, the support actuator performs a local boundary check on the controller. If the check passes, the height adjustment is completed, and the execution confirmation status is returned to the digital twin platform. If the instruction is not confirmed, the pressure sensor malfunctions during execution, or the displacement encoder does not provide effective feedback, the trial task enters an incomplete state.

[0061] S302: Collect height, support force, current elevation, and member strain before and after adjustment. Calculate the changes in height, support force, elevation, and strain respectively. Associate the test support actuator with the structural control measuring points according to the support correspondence relationship, establish a mapping record between the direction of influence, the magnitude of change, and the corresponding identifier, and generate the support influence relationship. The pre-adjustment collection record comes from the list of valid collection objects before the micro-motion test command is issued. The post-adjustment collection record comes from the synchronous collection process triggered after the support actuator's feedback confirmation. The height change status is used to confirm whether the support actuator has completed the micro-motion action; the support force change status is used to confirm the impact of the micro-motion action on the temporary support force; the elevation change status is used to confirm the morphological response of the associated structural control measuring points; and the strain change status is used to confirm whether the member force response has entered the abnormal category. Each change status only forms a before-and-after comparison within the same type of field, without direct comparison across physical quantities. The mapping record includes the test support actuator identifier, structural control measuring point identifier, construction stage identifier, change status field, influence direction field, change magnitude level field, and anomaly identification field. The variation level is derived from the response level rules in the construction status configuration table, which are formed based on the stage response of the finite element model, historical micro-motion test records, and on-site calibration results. After the mapping record is generated, it enters the anomaly identification process.

[0062] S303: Obtain the original micro-motion test records. Label the original micro-motion test records according to the construction stage, micro-motion test height increment, height change, support force change, elevation change, and strain change. Use the original micro-motion test records to train an anomaly recognition model based on an isolated forest model, support vector machine model, or long short-term memory network model, generating anomaly recognition model parameters. The original micro-motion test records are derived from historical control logs of the construction process, support equipment debugging logs, and on-site verification records. The record content includes construction stage fields, test command fields, execution feedback fields, data acquisition change status fields, manual verification status fields, and equipment anomaly identification fields. The labeling process categorizes records into normal or anomaly categories based on construction stage consistency, command execution confirmation status, sensor field integrity, change status continuity, and on-site verification conclusions. The anomaly recognition model based on an isolated forest model, support vector machine model, or long short-term memory network model is used to identify the credible state of the micro-motion test mapping records. Internally, it includes an input standardization layer, feature encoding layer, temporal correlation layer, response consistency judgment layer, and anomaly probability output layer. The input standardization layer receives construction phase, trial instructions, and various change status fields, and converts them into unified coded statuses; the feature coding layer extracts the same-period correlation features between support height changes, support force changes, elevation changes, and strain changes; the temporal correlation layer reads the historical status of the same support actuator and the same structural control measurement point within adjacent control cycles; the response consistency judgment layer compares the historical calibration records, finite element stage response records, and equipment execution confirmation records; and the anomaly probability output layer generates anomaly category judgment fields.

[0063] S304: Input the current micro-motion test record into an anomaly identification model based on an isolated forest model, support vector machine model, or long short-term memory network model to obtain the anomaly probability. When the anomaly probability exceeds a preset anomaly probability threshold, delete the corresponding mapping record; when the anomaly probability does not exceed the preset anomaly probability threshold, write the corresponding mapping record into the support influence relationship. Before the current micro-motion test record enters the model, field integrity verification, construction stage matching verification, equipment status verification, and measurement point support correspondence validity verification are performed. The anomaly probability is an anomaly category determination field output by the model, used to indicate the category proximity between the current mapping record and historical anomaly test records. The preset anomaly probability threshold is a determination boundary field in the model configuration library, derived from historical annotation records, on-site verification rules, and construction stage control strategies, and is synchronously written to the configuration library with model version updates. After the anomaly probability field and the threshold field complete the category boundary judgment, a normal mapping state or an anomaly mapping state is formed. An abnormal mapping state triggers a mapping record deletion action. The deletion is limited to candidate writes for supporting influence relationships and does not delete the original micro-motion probe log. The deleted record retains the supporting execution mechanism identifier, structural control measurement point identifier, abnormal cause, model version identifier, and review status, and is written to the abnormal probe record database. A normal mapping state triggers a write action, the mapping record enters the supporting influence relationship, and generates a subsequent call identifier.

[0064] S4: Compare the current elevation with the target elevation. From the structural control points whose absolute deviation exceeds the elevation deviation, determine the compensation points, compensation direction, and compensation elevation deviation based on the absolute deviation value. The current elevation comes from the elevation field of structural control points not excluded from the list of valid data acquisition objects. The target elevation comes from the forming control elevation field bound to the current construction stage in the parameter snapshot. Elevation deviation is the morphological deviation of the current elevation relative to the target elevation; the absolute deviation value is a field representing the degree of deviation without distinguishing direction. Compensation points are the structural control point objects entering the support adjustment process. The compensation direction is the elevation correction direction generated by the height change of the support actuator. The compensation elevation deviation is a morphological deviation field used to generate the height adjustment amount. Deviation identification only reads structural control points not excluded from the list of valid data acquisition objects. If the current elevation field is missing, the elevation sensor calibration fails, the measurement point coordinates are not fully BIM bound, or the target elevation field does not match the construction stage, the measurement point enters a deviation pending confirmation state and is not written to the compensation point table. When the absolute value of the deviation exceeds the allowable boundary corresponding to the elevation deviation, the measuring point is written into the table of measuring points to be compensated; when it does not exceed the allowable boundary, the measuring point remains in the monitoring state and enters the subsequent acquisition cycle.

[0065] S401: Obtain the current and target elevations of the unexcluded structural control points. Calculate the elevation deviation and absolute value of each structural control point. Compare the absolute value of the deviation with the elevation deviation. If the absolute value of the deviation exceeds the elevation deviation, write the corresponding structural control point into the compensation point table. The unexcluded structural control points are provided by the list of valid acquisition objects. The current elevation field is provided by the synchronous acquisition record, and the target elevation field is provided by the parameter snapshot. After the elevation deviation status is generated, the direction field and deviation degree field are synchronously written. The direction field is used to determine the compensation direction later, and the deviation degree field is used for sorting and determining the compensation elevation deviation. The compensation point table is the input object of the deviation compensation process, containing the structural control point identifier, current elevation status, target elevation status, deviation direction, deviation degree status, number of associated temporary support points, strain change status of members, and acquisition reliability status. After the structural control points enter the compensation point table, immediately read the valid associated support points in the support correspondence relationship of the points for subsequent compensation priority generation and candidate support execution mechanism screening. When a measuring point enters a state awaiting compensation but all associated temporary support points are excluded, the measuring point is written to a state of limited compensation, and the reason for the limited compensation is fed back to the visualization state layer of the digital twin platform for on-site verification and support recovery scheduling.

[0066] S402: Structural control points in the compensation point list are arranged in descending order of absolute deviation value. The structural control point ranked first is identified as the compensation point. The compensation direction is determined based on the positive or negative status of the elevation deviation, and the absolute value of the deviation is identified as the compensation elevation deviation. Structural control points in the compensation point list are sorted according to their degree of deviation. When the degree of deviation is within the same deviation range, the sorting process reads the compensation priority output by the compensation priority model as the sorting basis within the range. The compensation direction is determined by the deviation direction field. When the current elevation is higher than the target elevation, the compensation direction is identified as lowering the elevation; when the current elevation is lower than the target elevation, the compensation direction is identified as raising the elevation. The compensation elevation deviation is written to the compensation task cache and bound to the compensation point identifier, construction stage identifier, and candidate support lookup table identifier. The sorting results form a compensation point sequence. Points ranked first in the sequence and with valid associated support points are entered into the compensation point field for the current control cycle. If a measurement point ranked high lacks a valid associated support point, or if all associated support actuators are in a state of support force boundary, travel boundary, or controller abnormality, the measurement point is written to a state of suspended compensation, and the ranking process continues to read the next valid compensation measurement point.

[0067] S403: Obtain the original compensation records and label them according to construction stage, absolute deviation value, number of associated temporary support points, strain change of members, and compensation completion status. Use the original compensation records to train a compensation priority model based on a gradient boosting decision tree model to generate priority anomaly identification model parameters. The original compensation records are derived from historical construction control logs, compensation task cache, support adjustment execution logs, and structural re-measurement records, carrying construction stage fields, deviation degree status, associated support status, member strain change status, adjustment execution status, and compensation completion status. The labeling process organizes the compensation completion status, structural stress stability status, support execution confirmation status, and re-measurement status into priority labeling fields, which are used to limit the order in which different structural control measurement points within the same deviation interval enter the compensation sequence. The compensation priority model based on a gradient boosting decision tree model is used to determine the sorting status of the measurement points to be compensated. Internally, it includes an input standardization layer, a stage feature encoding layer, a support association feature layer, a strain response feature layer, and a sorting status output layer. The input standardization layer receives the construction stage, deviation level, number of associated temporary support points, and strain change status of the members; the stage feature coding layer converts the system transformation status, component installation status, and temporary support status into internal model categories; the support association feature layer reads the status of valid associated support points; the strain response feature layer reads the strain change status of the members and identifies the stress change category; and the sorting status output layer outputs the compensation priority.

[0068] S404: Input the current construction stage, absolute value of deviation, number of associated temporary support points, and strain change of members into the compensation priority model based on the gradient lifting decision tree model to obtain the compensation priority. This compensation priority is used as the sorting criterion within the same absolute value range of deviation to generate a compensation measurement point sequence. Before the current measurement point to be compensated enters the model, the deviation degree status, the number of associated temporary support points, and the strain change of members are read from the measurement point to be compensated table, the measurement point support correspondence, and the synchronous acquisition record, respectively. The compensation priority is a sorting status field output by the model, used to indicate the order in which structural control measurement points enter the compensation process within the same deviation range. The absolute value range of deviation comes from the deviation grouping rules in the construction status configuration table. These rules are formed based on the construction stage morphological control strategy, finite element model response results, and historical compensation records, and are stored as sorting boundary fields. After the compensation measurement point sequence is generated, it is written to the compensation task cache. Each compensation measurement point in the sequence retains the compensation direction, compensation elevation deviation, effective associated support point status, sorting source identifier, and model version identifier. If the model parameter library is unreadable, input fields are missing, or the model output does not form a valid sorting state, the sorting process reverts to the deviation level sorting and writes the model's recovery state to the compensation task cache. The generated compensation measurement point sequence is then read by the support adjustment instruction generation stage.

[0069] S5: Based on the support influence relationship and the correspondence between the measuring points and the support, select the support actuators that are consistent with the compensation direction. Convert the compensation elevation deviation into a height adjustment amount, correct the height adjustment amount according to the upper limit of the single adjustment height and the upper limit of the support force, generate the support adjustment command, and verify the remaining elevation deviation. The support adjustment command is a closed-loop control data object sent to the support actuator controller, carrying the compensation measuring point identifier, candidate support actuator identifier, compensation direction, corrected height adjustment amount, support force boundary verification status, travel boundary verification status, command issuance status, execution confirmation status, and remaining elevation deviation verification status. The height adjustment amount is the height change control field that the support actuator should execute, derived from the conversion rule between the compensation elevation deviation and the unit height elevation response amount in the support influence relationship. The unit height elevation response amount is the elevation change response field corresponding to the micro-motion test height increment, used to represent the influence status of the support actuator's height action on the elevation change of the associated structural control measuring points in the current construction stage. The support adjustment command generation process is simultaneously constrained by the upper limit of the single adjustment height, the upper limit of the support force, and the travel boundary. After the instruction is generated, the support actuator controller returns a receipt confirmation, an execution confirmation, and a post-execution altitude status. The confirmation results are entered into the adjustment log and the remaining deviation verification process.

[0070] S501: Read the support influence relationship and the measurement point support correspondence relationship, filter the support actuators that are consistent with the compensation direction and associated with the compensation measurement points, and arrange them in descending order of the absolute value of the elevation change that is consistent with the compensation direction to generate a candidate support actuator sequence. The support influence relationship provides the influence direction and change magnitude level of the support actuator on the structural control measurement points, and the measurement point support correspondence relationship provides the effective association status between the compensation measurement points and the temporary support points. The filtering process first reads the compensation measurement point identifier, then searches for support actuators that are effectively associated with the compensation measurement points, and then reads the influence direction field in the support influence relationship. When the influence direction is consistent with the compensation direction, the support actuator is added to the candidate set; when the influence direction is opposite, the mapping record is in an abnormal state, the support actuator is excluded, the support actuator lacks execution confirmation capability, or the top connection status is not confirmed, the corresponding support actuator does not enter the candidate set. The candidate support actuator sequence is arranged according to the change magnitude level corresponding to the elevation change state. The change magnitude level comes from the support influence relationship and is not directly compared across dimensions with the support force change state or strain change state. During the sorting process, when the strain change status corresponding to the supporting execution agency enters the risk category, the sorting status is downgraded and written with the stress concern mark; when the candidate set is empty, the compensation task enters the review status and feeds back the status of compensation measurement points, compensation direction, reasons for missing candidates and the list of valid collection objects to the digital twin platform.

[0071] S502: Determine the unit height elevation response based on the micro-motion test height increment and the corresponding elevation change. Convert the compensated elevation deviation into the initial height adjustment amount. Compare the initial height adjustment amount with the single adjustment height upper limit, and adjust the initial height adjustment amount in conjunction with the current support force, support force change, and support force upper limit to generate a support adjustment command. The unit height elevation response amount is generated from the micro-motion test records of the same support actuator, the same compensation measuring point, and the same construction stage in the support influence relationship, reflecting the stage-by-stage correspondence between the support height action and the compensation measuring point elevation response. Before the compensated elevation deviation is converted into the height adjustment amount, it is confirmed that the support influence relationship is not in an abnormal state, the candidate support actuator is not at the travel boundary, the current support force field is valid, and the support actuator controller is in a command-receiving state. The initial height adjustment amount is the support control field formed based on the compensated elevation deviation and the unit height elevation response amount, and it is only allowed to be issued after constraint correction. The correction process first reads the single-adjustment height upper limit field to constrain the initial height adjustment amount in a single round. Then, it reads the current support force, support force change status, and support force upper limit fields to form the expected support force status. When the expected support force status is within the safe boundary, candidate height adjustment amounts are retained; when the expected support force status enters the out-of-limit category, candidate height adjustment amounts are discarded. The corrected height adjustment amount is then written into the support adjustment command.

[0072] S503: After executing the support adjustment command, the current elevation is collected, and the current elevation is compared with the target elevation to obtain the remaining elevation deviation. After the support adjustment command is executed, the acquisition scheduler triggers the elevation sensor corresponding to the compensation measuring point to re-collect the elevation and writes the collected value into the post-adjustment synchronous acquisition record. The current elevation after adjustment is compared with the target elevation in the parameter snapshot using the same type of elevation field to generate the remaining elevation deviation status. The remaining elevation deviation status includes the deviation direction, deviation degree status, acquisition reliability status, and a continuing compensation status indicator. When the remaining elevation deviation enters the allowable elevation deviation boundary, the compensation task is written to the completion status, and the height status, support force status, and compensation measuring point elevation status after the support actuator execution are written back to the digital twin platform; when the remaining elevation deviation is still in the pending compensation status, the compensation task enters the subsequent control cycle, and retains the support influence relationship, candidate support actuator sequence, and execution log of the current cycle as input. When the post-adjustment acquisition fails, the elevation sensor returns an abnormality, the target elevation field expires, or the support actuator execution confirmation is missing, the remaining elevation deviation is written to the pending review status, and the continuous issuance of adjustment commands is blocked. This verification process ensures a closed-loop connection between support adjustment actions, morphological responses, and subsequent control tasks.

[0073] S504: Obtain the original support adjustment records. These records document real-time height adjustment, unit height elevation response, support force change, member strain change, and the remaining elevation deviation after adjustment. Use these records to train a machine learning model for predicting the remaining elevation deviation, which then outputs the predicted remaining elevation deviation based on candidate height adjustment values. The original support adjustment records originate from historical support adjustment logs, post-adjustment retest records, support actuator feedback records, and member strain monitoring records. They include construction stage fields, candidate height adjustment fields, unit height elevation response fields, support force change status, member strain change status, execution confirmation status, and remaining elevation deviation status. Records undergo source verification, field integrity verification, construction stage matching verification, and abnormal execution status removal before entering the training cache. Records with missing execution confirmation, abnormal sensor acquisition links, or incomplete retest record closures are not included in the training cache but are written to the adjustment anomaly log. The machine learning residual deviation prediction model is used to predict the residual elevation deviation state corresponding to candidate elevation adjustment amounts before the support adjustment command is issued. Internally, it includes an input standardization layer, a support response feature layer, a force constraint feature layer, a stage state association layer, and a residual deviation output layer. During training, the model compares the predicted state with the actual residual elevation deviation state after adjustment, adjusts the inter-layer parameters based on deviation feedback, and writes the trained parameters into the prediction model library.

[0074] S505: Within the range not exceeding the upper limit of a single adjustment height, a candidate height adjustment sequence is generated according to a preset height interval. The expected support force is determined based on the current support force, the change in support force, and the candidate height adjustment values. Candidate height adjustment values ​​whose expected support force is lower than the upper limit are input into the machine learning residual deviation prediction model to obtain the expected residual elevation deviation corresponding to the candidate height adjustment values. The candidate height adjustment value with the smallest absolute value of the expected residual elevation deviation is determined as the corrected height adjustment value. The preset height interval is a candidate generation field in the support actuator controller parameter register table, derived from the actuator control accuracy, on-site calibration status, and construction stage control strategy. The candidate height adjustment sequence is generated from the initial height adjustment value after being constrained by the upper limit of a single adjustment height. Each candidate height adjustment value in the sequence carries a candidate identifier, support actuator identifier, compensation measuring point identifier, and construction stage identifier. Before entering the prediction model, the candidate height adjustment value undergoes a support force boundary check. The current support force field, the support force change status, and the candidate height adjustment value jointly generate the expected support force status. When the expected support force is within the safe boundary, candidate height adjustment values ​​are entered into the prediction model; when the expected support force is in the out-of-limit category, candidate height adjustment values ​​are removed from the sequence, and the reason for removal is written. After the prediction model outputs the expected remaining elevation deviation state, the correction process reads the deviation degree state corresponding to each candidate height adjustment value, selects the candidate height adjustment value with the smallest deviation degree state as the corrected height adjustment value, and writes it into the support adjustment command.

[0075] The intelligent control system for the construction of large-span steel-concrete composite structures includes:

[0076] The control parameter acquisition module acquires the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment and arrival time limit, and generates the correspondence between the measuring point support according to the distance between the structural control measuring point and the temporary support point.

[0077] The data synchronization acquisition module synchronously acquires the current elevation, member strain, support force, and height of the support actuator, excluding structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit;

[0078] The micro-motion test generation module performs micro-motion test height increments on support actuators that have not been eliminated and whose support force is lower than the upper limit of the support force and whose height has not reached the travel boundary, calculates the changes in height, support force, elevation and strain, and generates support influence relationships;

[0079] The deviation compensation determination module compares the current elevation with the target elevation and determines the compensation measurement point, compensation direction, and compensation elevation deviation from the structural control measurement points whose absolute deviation exceeds the elevation deviation, based on the absolute value of the deviation.

[0080] The adjustment command generation module selects support actuators that are consistent with the compensation direction based on the support influence relationship and the correspondence between the measuring points and the support. It converts the compensation elevation deviation into a height adjustment amount, corrects the height adjustment amount according to the upper limit of the single adjustment height and the upper limit of the support force, generates support adjustment commands, and verifies the remaining elevation deviation.

[0081] The above embodiments illustrate preferred embodiments of the present invention. Any equivalent adjustments to the technical solution based on software engineering methods are within the scope of protection, including but not limited to: implementing algorithm logic using different programming languages, refactoring functional modules into services, adjusting data interaction protocols, and optimizing resource scheduling strategies. Any implementation scheme derived from reasonable modifications to the data processing flow, service call chain, or system architecture layer without departing from the core technology of the present invention should be considered within the protection scope defined by the technical solution of the present invention.

Claims

1. An intelligent control method for the construction of large-span steel-concrete composite structures, characterized in that, Includes the following steps: Obtain the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment and arrival time limit, and generate the correspondence between the measuring point support according to the distance between the structural control measuring point and the temporary support point; Simultaneously collect the current elevation, member strain, support force, and height of the support actuator, exclude the structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit, and generate a list of valid collection objects; Based on the list of valid data collection objects, for the support actuators that are not excluded and whose support force is lower than the upper limit of the support force and whose height has not reached the travel boundary, the micro-motion test height increment is performed, and the height, support force, elevation and strain change are calculated to generate the support influence relationship; Compare the current elevation with the target elevation, and determine the compensation measuring point, compensation direction and compensation elevation deviation from the structural control measuring points whose absolute deviation exceeds the measured elevation deviation according to the absolute deviation; Based on the support influence relationship and the corresponding relationship of the measuring point support, the support actuators that are consistent with the compensation direction are selected, the compensation elevation deviation is converted into a height adjustment amount, the height adjustment amount is corrected according to the single adjustment height limit and the support force limit, a support adjustment command is generated, and the remaining elevation deviation is checked.

2. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 1, characterized in that, The steps for generating the correspondence between the measuring points and the temporary support points based on the distance between the structural control measuring points and the temporary support points include: Obtain the coordinates of the structural control measuring points, the coordinates of the temporary support points, and the construction stage identifier; read the target elevation, the elevation deviation, and the arrival time limit from the construction status configuration table of the digital twin platform; read the upper limit of the support force, the stroke boundary, the upper limit of the single adjustment height, and the micro-motion test height increment from the parameter register table of the support actuator controller; and generate a parameter snapshot. Calculate the spatial distance between each of the structural control measuring points and the temporary support points, arrange the temporary support points according to the spatial distance, and sequentially associate the temporary support points whose spatial distance is not greater than a preset association distance threshold with the structural control measuring points to generate the measuring point support correspondence.

3. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 1, characterized in that, The specific steps for generating the list of valid data collection objects are as follows: The current elevation and strain of the rod are collected from the elevation sensor and strain sensor deployed at the control measurement point of the structure. The support force and the height of the support actuator are collected from the pressure sensor at the temporary support point and the displacement encoder of the support actuator. Each collected value is written with the device identifier, sampling time and receiving time to generate a synchronous acquisition record. Align the synchronous acquisition records with the reference time of the same sampling period, calculate the arrival time of the reference time corresponding to the reception time of each synchronous acquisition record, and when the arrival time exceeds the arrival time limit, write the corresponding structural control measurement point or the temporary support point into the exclusion list to generate a list of valid acquisition objects.

4. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 1, characterized in that, The specific steps for generating the supporting influence relationship are as follows: Based on the list of valid collected objects, the support actuators that are not included in the exclusion list, whose current support force is lower than the upper limit of the support force, and whose current height has not reached the travel boundary are screened to determine the test support actuators. A micro-motion test command is sent to each of the test support actuators so that the test support actuators can complete the height adjustment according to the micro-motion test height increment. The height, support force, current elevation, and strain of the members before and after adjustment are collected. The changes in height, support force, elevation, and strain are calculated respectively. The test support actuator is associated with the structural control measuring point according to the corresponding relationship of the measuring point support. A mapping record between the direction of influence, the magnitude of change, and the corresponding identifier is established to generate the support influence relationship.

5. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 4, characterized in that, The specific steps for establishing the mapping record are as follows: Obtain the original micro-motion test records, and label the original micro-motion test records according to the construction stage, micro-motion test height increment, height change, support force change, elevation change and strain change. Use the original micro-motion test records to train an anomaly recognition model based on the isolated forest model, support vector machine model or long short-term memory network model, and generate anomaly recognition model parameters. The current micro-motion trial record is input into the anomaly recognition model based on the isolated forest model, support vector machine model, or long short-term memory network model to obtain the anomaly probability. When the anomaly probability exceeds a preset anomaly probability threshold, the corresponding mapping record is deleted. When the anomaly probability does not exceed the preset anomaly probability threshold, the corresponding mapping record is written into the support influence relationship.

6. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 1, characterized in that, The specific steps for determining the compensation measuring point, the compensation direction, and the compensation elevation deviation are as follows: Obtain the current elevation and target elevation corresponding to the unexcluded structural control measuring points; calculate the elevation deviation and absolute value of the deviation for each structural control measuring point; compare the absolute value of the deviation with the elevation deviation; when the absolute value of the deviation exceeds the elevation deviation, write the corresponding structural control measuring point into the table of measuring points to be compensated. The structural control measuring points in the table of measuring points to be compensated are arranged in descending order of the absolute value of the deviation. The structural control measuring point ranked first is determined as the compensation measuring point. The compensation direction is determined according to the positive or negative state of the measured elevation deviation, and the absolute value of the deviation is determined as the compensation elevation deviation.

7. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 6, characterized in that, The specific steps for arranging the structural control measuring points according to the absolute value of the deviation from largest to smallest are as follows: Obtain the original compensation records, label the original compensation records according to the construction stage, absolute value of deviation, number of associated temporary support points, strain change of members and compensation completion status, use the original compensation records to train a compensation priority model based on gradient boosting decision tree model, and generate priority anomaly identification model parameters; The current construction stage, the absolute value of the deviation, the number of associated temporary support points, and the strain change of the member are input into the compensation priority model based on the gradient lifting decision tree model to obtain the compensation priority. The compensation priority is used as the sorting basis within the same absolute value range of the deviation to generate a sequence of compensation measurement points.

8. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 1, characterized in that, The specific steps for generating the support adjustment command and verifying the remaining elevation deviation are as follows: Read the support influence relationship and the measurement point support correspondence relationship, filter the support actuators that are consistent with the compensation direction and associated with the compensation measurement point, and arrange them in descending order of the absolute value of the elevation change that is consistent with the compensation direction to generate a candidate support actuator sequence; The unit height elevation response is determined based on the micro-motion test height increment and the corresponding elevation change. The compensated elevation deviation is converted into an initial height adjustment amount. The initial height adjustment amount is compared with the single adjustment height upper limit. The initial height adjustment amount is then corrected by combining the current support force, the support force change amount, and the support force upper limit, thereby generating the support adjustment command. After executing the support adjustment command, the current elevation is collected, and the current elevation is compared with the target elevation to obtain the remaining elevation deviation.

9. The intelligent control method for construction of large-span steel-concrete composite structures according to claim 8, characterized in that, The specific steps for correcting the initial height adjustment amount are as follows: Obtain the original support adjustment record, which records the real-time height adjustment amount, unit height elevation response amount, support force change amount, member strain change amount, and remaining elevation deviation after adjustment. Use the original support adjustment record to train a machine learning residual deviation prediction model, and make the machine learning residual deviation prediction model output the expected remaining elevation deviation based on the candidate height adjustment amount. Within the range not exceeding the upper limit of the single adjustment height, a candidate height adjustment amount sequence is generated according to a preset height interval. The expected support force is determined based on the current support force, the change in support force, and the candidate height adjustment amount. The candidate height adjustment amount whose expected support force is lower than the upper limit of the support force is input into the machine learning residual deviation prediction model to obtain the expected residual elevation deviation corresponding to the candidate height adjustment amount. The candidate height adjustment amount with the smallest absolute value of the expected residual elevation deviation is determined as the corrected height adjustment amount.

10. An intelligent control system for the construction of large-span steel-concrete composite structures, characterized in that: The system is used to implement the intelligent control method for construction of large-span steel-concrete composite structures as described in any one of claims 1-9, and the system includes: The control parameter acquisition module acquires the target elevation, allowable elevation deviation threshold, upper limit of support force, travel boundary, upper limit of single adjustment height, micro-motion test height increment and arrival time limit, and generates the correspondence between the measuring point support according to the distance between the structural control measuring point and the temporary support point. The data synchronization acquisition module synchronously acquires the current elevation, member strain, support force, and height of the support actuator, excluding the structural control measuring points or temporary support points whose arrival time exceeds the arrival time limit; The micro-motion test generation module performs the micro-motion test height increment on the support actuator that has not been eliminated and whose support force is lower than the upper limit of the support force and whose height has not reached the travel boundary, calculates the height, support force, elevation and strain change, and generates the support influence relationship; The deviation compensation determination module compares the current elevation with the target elevation, and determines the compensation measurement point, compensation direction and compensation elevation deviation from the structural control measurement points whose absolute deviation value exceeds the measured elevation deviation according to the absolute deviation value. The adjustment command generation module selects the support actuators that are consistent with the compensation direction based on the support influence relationship and the corresponding relationship of the measuring point support, converts the compensation elevation deviation into a height adjustment amount, corrects the height adjustment amount according to the single adjustment height limit and the support force limit, generates a support adjustment command, and verifies the remaining elevation deviation.

Images