An engineering construction digital twin site perception linkage monitoring system and method

Abstract

The application relates to the field of on-site monitoring alarm technology and discloses a digital twin on-site sensing linkage monitoring system and method in engineering construction, which comprises the following steps: acquiring multi-source state data, current process execution data and construction object association data of a construction site; identifying a process switching stage and determining a state reconstruction time period; performing segmented analysis on data in the state reconstruction time period, extracting time sequence change characteristics; generating stage-based abnormal characteristics according to a preset stage response law; combining coupling response relationships among construction objects to screen out transient abnormalities and generate real abnormal results; mapping the real abnormal results to a digital twin scene, determining linkage monitoring objects and linkage action sequences and outputting linkage monitoring results; and the application can distinguish between normal transition changes and real abnormal changes in the process switching stage.

Description

Technical Field

[0001] This application relates to the field of on-site monitoring and alarm technology, and more specifically, to a digital twin on-site perception and linkage monitoring system and method for engineering construction. Background Technology

[0002] Digital twin monitoring technology is commonly used in current construction sites to map construction equipment, components under construction, workers, and work areas into a scene. It combines location data, equipment operation data, video data, and environmental monitoring data to visualize the status of the construction site and handle anomalies. This type of solution can display the status of construction objects and monitor risks under normal construction conditions.

[0003] However, in actual construction, changes in the state of the construction object do not always correspond to real anomalies. Especially during the stages of process switching, equipment transfer, hoisting and repositioning, temporary support adjustment, material transfer and switching, and work area switching, the construction object often undergoes changes in position, posture, load, and action switching that accompany the work transition process. These changes are phased and transitional in time and are normal responses in the construction process. Existing solutions usually update the state of the construction object in the digital twin scenario directly based on real-time acquisition results and perform anomaly identification according to unified monitoring rules. They lack separate identification and transition state screening processing for process switching stages, and are prone to directly identifying normal transitional changes during process switching as abnormal changes.

[0004] For example, when hoisting equipment transitions from the current hoisting process to the shifting process, the boom angle, hoisting height, equipment position, and associated work area will change continuously within a short period of time. Such changes are normal transitional responses during process switching. If the existing solution only performs anomaly identification based on the real-time changes within this time period, it is easy to directly retain normal transitional responses as abnormal results, and further trigger the expansion of subsequent linkage monitoring objects, the deviation of linkage range, or inaccurate alarm output. Similarly, when there are spatial adjacency relationships, work contracting relationships, or collaborative work relationships between multiple construction objects, short-term linkage changes caused by the process switching of a certain construction object are also easy to be mistakenly identified as abnormal changes.

[0005] Therefore, the existing technology has at least the following problems: In the process of digital twin monitoring of engineering construction, there is a lack of a dedicated identification mechanism for the process switching stage, a lack of a screening mechanism for the normal response pattern of the process switching process, and a lack of a coupling verification mechanism for the linkage changes of related construction objects. Therefore, it is difficult to effectively distinguish between normal transition changes and real abnormal changes in the process switching stage, which in turn affects the accuracy and stability of the abnormal identification results, the determination results of linkage monitoring objects, and the output results of linkage actions. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of existing technologies and achieve the above objectives, this application provides a hierarchical discrimination chain for normal transitional changes and actual abnormal changes during process switching phases. This chain involves performing process switching phase identification, state reconstruction time period determination, temporal change feature extraction, phase response pattern comparison, coupled response consistency verification, and linkage monitoring output processing on multi-source status data, current process execution data, and construction object-related data from the construction site. The chain reduces the interference of short-term disturbances during process switching phases on the anomaly identification results and linkage monitoring output results. The technical solution provided in this application is as follows: A digital twin-based on-site sensing and linkage monitoring method for engineering construction includes: Acquire multi-source status data of the construction site, current process execution data, and construction object related data; Identify the process switching phase based on the current process execution data, and determine the state reconstruction period corresponding to the process switching phase; The multi-source state data within the state reconstruction period corresponding to the process switching stage is segmented and analyzed to extract temporal change features. Based on the preset stage response rules corresponding to the process switching stage, the time-series change features are classified to determine the switching response features and continuous abnormal features corresponding to the process switching stage, and staged abnormal features are generated. The coupling response relationship between construction objects is determined based on the associated data of the construction objects, and the consistency of the staged anomaly features is verified based on the coupling response relationship between the construction objects to screen out transient anomalies caused by the process switching stage and generate real anomaly results. The real anomaly results are mapped to construction objects in the digital twin scene, and the linked monitoring objects and corresponding linkage action sequences are determined based on the mapping results, and the linked monitoring results are output. By first identifying the process switching stage and defining the state reconstruction period corresponding to that process switching stage, and then performing stage response identification and coupling consistency verification on the multi-source state changes within the state reconstruction period, the anomaly identification is based on the linkage relationship between the process switching process and the construction object, rather than directly judging based on the real-time state results without distinguishing the process background. This allows the normal transitional changes in the process switching stage to be separated from the abnormal changes that need to be retained, and provides real anomaly input for the subsequent determination of linkage monitoring objects.

[0007] Furthermore, the method for determining the state reconstruction period corresponding to the process switching stage includes: Based on the current process execution data, a process switching candidate time period is determined. When the process switching candidate time period meets the following conditions: process identifier change or current execution action change is established, equipment action code sequence has continuous action switching, work area identifier sequence has area entry, area exit or cross-area movement, and the duration of the candidate time period is not less than the shortest confirmation time, the process switching candidate time period is determined as the process switching stage. Based on the start and end times of the process switching phase, and in conjunction with the preceding and following time ranges, the state reconstruction period is determined. By performing validity verification on the candidate process switching periods, and combining the start and end times with the preceding and following time ranges, the state reconstruction period is determined, ensuring that subsequent state analysis corresponds to the complete change range of the current process transition. This avoids omitting the pre-switching preparation process from the analysis scope, or mixing data that has entered a stable state after the switch with the switch process data.

[0008] Furthermore, methods for extracting temporal variation features include: Read multi-source state data within the state reconstruction period to form a state time series; Generate a sequence of state changes based on the state time series corresponding to adjacent settling times. Based on the state change sequence, determine the pre-switching preparation segment, the switching execution segment, and the post-switching stabilization segment, and generate a segmented state sequence; Based on the segmented state sequences corresponding to each segment, the temporal change features corresponding to the position state type, attitude state type, load state type, action state type, and area state type are extracted. By converting the multi-source state data within the state reconstruction period into state time series, state change quantity series, and segmented state series, and further extracting the temporal change features corresponding to different state types, the original monitoring data is transformed from discrete acquisition results into structured change results that can be compared by segment and by state type. This provides a unified feature basis for subsequent sub-item comparisons based on the preset stage response rules.

[0009] Furthermore, methods for generating staged anomaly features include: Based on historical records of similar process switching, process conversion requirements in the construction plan, equipment action execution requirements, and on-site operation constraints, a preset stage response pattern is generated for the process switching stage. Based on the preset stage response pattern, the time-series change features are compared item by item to generate feature comparison results; Based on the feature comparison results, switching response features and continuous anomaly features are determined, and staged anomaly features are generated based on the continuous anomaly features. By comparing the time-series change features with the preset stage response rules, and organizing changes that do not meet the corresponding segmentation requirements, sequence requirements, direction requirements, continuous interval requirements, fallback requirements, or result requirements into staged anomaly features, the abnormal changes in the process of process switching are no longer limited to deviation judgments at a single moment, but are categorized into anomaly result objects corresponding to the process switching stage, segment position, and state type, thereby facilitating subsequent consistency verification based on the coupled response relationship between construction objects.

[0010] Furthermore, the method for determining the coupling response relationship between construction objects based on the associated data of the construction objects includes: Based on the spatial adjacency determination results, operation dependency determination results, and synergy determination results in the construction object association data, determine the coupling relationship type between construction objects; Then, based on the current combination of construction object categories, the current process switching type, and the current coupling relationship type, determine the response state type, response direction requirements, response time interval, and response release interval to generate the coupling response relationship.

[0011] Furthermore, the method for generating verification sub-units based on the coupling response relationship and the staged anomaly characteristics includes: Filter from the phased anomaly features corresponding to the current construction object and the phased anomaly features corresponding to related construction objects; When the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction object conform to the response time interval corresponding to the coupling response relationship, the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction object are paired, and the paired staged anomaly features are organized into verification sub-units.

[0012] Furthermore, methods for performing consistency checks include: For the staged anomaly characteristics in the verification subunit, perform state type consistency verification, direction consistency verification, time consistency verification, and deconsistency verification. When all consistency checks are successful, the corresponding staged anomaly characteristics are identified as transient anomalies caused by the process switching stage. If any consistency check fails, the corresponding staged anomaly feature is identified as a retained anomaly. By performing state type consistency check, direction consistency check, time consistency check, and deconsistency check on the staged anomaly features in the check subunit, short-term changes caused by spatial adjacency, job succession, or collaborative job relationships can be identified as transient anomalies corresponding to the process switching stage. Anomalies that do not meet the coupling linkage rules are retained, thereby completing further convergence of the true anomaly results.

[0013] Furthermore, the methods for identifying the objects to be monitored in a coordinated manner include: The monitoring objects of the anomaly center are determined based on the mapping results; Based on the work area identifier, scene coordinates, area boundary, and preset impact range corresponding to the monitoring object in the anomaly center, determine the area linkage monitoring object; Based on the task receiving object corresponding to the monitoring object in the anomaly center, determine the task receiving and linkage monitoring object; Based on the collaborative operation objects corresponding to the monitoring objects in the anomaly center, determine the collaborative monitoring objects; The abnormality center monitoring objects, regional linkage monitoring objects, receiving linkage monitoring objects, and collaborative linkage monitoring objects are deduplicated and sorted to generate the linkage monitoring objects.

[0014] Furthermore, methods for determining the sequence of linked actions include: Based on the actual anomaly results and the linked monitoring objects, the linkage actions are determined; each linkage action is sorted, duplicate actions are deduplicated and retained, and conflicting actions are filtered to generate the linkage action sequence; by using the actual anomaly results as the basis for generating linkage actions, and combining the linkage monitoring objects to perform sorting, deduplication and conflict filtering on the linkage actions, the output linkage action sequence corresponds to the actual monitoring needs of the anomaly center object and its related objects, thereby reducing the redundancy of linkage output caused by repeated triggering or action conflicts.

[0015] A digital twin-based on-site sensing and linkage monitoring system for engineering construction includes: The data acquisition module is used to acquire multi-source status data, current process execution data, and construction object related data from the construction site. The process switching identification module is used to identify the process switching stage based on the current process execution data, and determine the state reconstruction period corresponding to the process switching stage; The temporal feature extraction module is used to perform segmented analysis on the multi-source state data within the state reconstruction period corresponding to the process switching stage, and extract temporal change features. The stage anomaly identification module is used to classify the time-series change features according to the preset stage response rules corresponding to the process switching stage, determine the switching response features and continuous anomaly features corresponding to the process switching stage, and generate staged anomaly features. The coupling verification module is used to determine the coupling response relationship between construction objects based on the associated data of the construction objects, and to perform consistency verification on the staged abnormal features based on the coupling response relationship between the construction objects, thereby filtering out transient abnormalities caused by the process switching stage and generating real abnormal results. The linkage monitoring module is used to map the real abnormal results to the construction objects in the digital twin scene, and determine the linkage monitoring objects and corresponding linkage action sequences based on the mapping results, and output the linkage monitoring results.

[0016] Compared with related technologies, this application has the following advantages: This application, through joint processing of multi-source status data, current process execution data, and construction object-related data at the construction site, first identifies the process switching stage and determines the corresponding status reconstruction period. Then, it extracts the temporal change features from the multi-source status data within the status reconstruction period. Furthermore, by combining the preset stage response rules corresponding to the process switching stage with the coupling response relationship between construction objects, it identifies abnormal changes in the process switching stage and maps the identified real abnormal results to the digital twin scenario to output the linkage monitoring results. This forms an abnormal identification and linkage monitoring processing chain for process switching scenarios, improving the consistency between the digital twin monitoring results and the actual risk status on site.

[0017] First, by identifying the process switching phase and determining the corresponding state reconstruction period, anomaly identification is based on a dedicated analysis time range corresponding to the current process conversion process, rather than directly judging based on real-time state results without distinguishing the process background. This allows the preparation process before process switching, the switching execution process, and the stabilization process after switching to be included in a unified analysis scope.

[0018] Second, by segmenting and analyzing the multi-source state data during the state reconstruction period and extracting the temporal change features, and then classifying the temporal change features according to the preset stage response rules, changes that conform to the normal response rules of process switching can be identified as switching response features, and changes that do not conform to the preset stage response rules or continue to exist after process switching can be identified as continuous abnormal features, thereby realizing the staged distinction between normal transitional changes and abnormal changes in process switching.

[0019] Third, by determining the coupling response relationship between construction objects based on the associated data of construction objects, and performing consistency verification on the staged abnormal features based on the coupling response relationship, it is possible to further filter out the linkage transient changes caused by spatial adjacency, operation undertaking relationship or collaborative operation relationship, and avoid mistakenly retaining short-term changes of associated construction objects that conform to the linkage law during the process switching stage as real abnormal results.

[0020] Fourth, by mapping real anomaly results to construction objects in the digital twin scenario, and determining the linked monitoring objects and corresponding linkage action sequences based on the mapping results, the anomaly display, scope prompts, acceptance prompts, collaborative alarms, and restriction prompts can be executed based on real anomaly results, thereby improving the consistency between the determination results of linked monitoring objects and the output results of linkage actions and the actual risk status on site.

[0021] Fifth, this application connects process switching identification, stage response screening, coupling consistency verification, and digital twin linkage output into a continuous processing chain, which can reduce the impact of short-term disturbances during process switching on abnormal monitoring results and improve the accuracy and stability of digital twin linkage monitoring at the construction site. Attached Figure Description

[0022] Figure 1 This application provides a schematic diagram of a digital twin-based on-site perception and linkage monitoring method for engineering construction. Figure 2 The flowchart for process switching identification, temporal feature extraction, and staged anomaly feature generation provided in this application is as follows: Figure 3 This application provides a schematic diagram of a digital twin on-site perception and linkage monitoring system module for engineering construction. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Example 1

[0024] Please see Figure 1 As shown in the figure, this embodiment provides a digital twin-based on-site perception and linkage monitoring method for engineering construction, including the following steps: Step S1: Obtain multi-source status data of the construction site, current process execution data, and construction object association data.

[0025] In some implementation methods, the steps for acquiring multi-source status data of the construction site, current process execution data, and construction object related data include: Step 1-1: Determine the construction object and generate a construction object identifier. Specifically, first, determine the construction objects in the current construction site that need to be continuously sensed and monitored. The construction object is used to represent the objects that participate in the current construction activities and will affect the subsequent process switching judgment and anomaly identification. The construction object may include lifting equipment, transportation equipment, construction components, temporary support components, and workers.

[0026] Subsequently, a unique construction object identifier is generated for each construction object, and a correspondence is established between the construction object identifier and the construction object name, construction object category, and the work area where the construction object is located; the construction object identifier serves as a unified correspondence basis for subsequently obtaining multi-source status data, current process execution data, and construction object associated data.

[0027] Steps 1-2: Obtain location status information around the construction object identifier and form location status data. Specifically, location acquisition devices are deployed for each construction object; for fixed large construction equipment and temporary support components, a total station reflecting prism, a global navigation positioning device, or an ultra-wideband positioning tag can be used for location acquisition; for mobile transportation equipment, a vehicle-mounted global navigation positioning device, an ultra-wideband positioning tag, or a visual recognition positioning device can be used for location acquisition; for workers, a wearable ultra-wideband positioning tag can be used for location acquisition.

[0028] The location status information specifically refers to the plane coordinates, elevation coordinates, and positional displacement of the construction object at each acquisition time. After obtaining the location status information at each acquisition time, it is organized according to the construction object identifier and acquisition time to form the location status data corresponding to each construction object. The location status data serves as the component of the subsequent generation of multi-source status data.

[0029] Steps 1-3: Obtain attitude status information around the construction object identifier and form attitude status data. Specifically, attitude acquisition devices are deployed for each construction object. For lifting equipment, transportation equipment and temporary support components, attitude acquisition can be carried out using inertial measurement devices, tilt sensors, angle encoders or attitude feedback interfaces in the equipment's own controller. For the components being constructed, their installation orientation and tilt status can be collected using visual recognition devices or tilt sensors.

[0030] The attitude status information specifically refers to at least one of the following: spatial orientation, pitch angle, roll angle, slewing angle, and tilt angle of the construction object at each collection time. For lifting equipment, the attitude status information includes at least one of the slewing angle and the boom elevation angle. For the component being constructed, the attitude status information includes at least one of the component installation orientation and the component tilt angle. For transportation equipment, the attitude status information includes at least one of the vehicle body orientation and the vehicle body tilt angle.

[0031] After obtaining the attitude status information at each acquisition time, the information is organized according to the construction object identifier and acquisition time to form the attitude status data corresponding to each construction object; the attitude status data serves as the component of the subsequent generation of multi-source status data.

[0032] Steps 1-4: Obtain load status information, operational status information, and environmental status information based on the identified construction object, and form corresponding status data. Specifically, load status information is obtained for each construction object. For lifting equipment and temporary support components, load status information can be obtained through load cells, tension sensors, pressure sensors, strain sensors, or load feedback interfaces in the equipment controller. The load status information specifically refers to the lifting weight, support force, tension / compression, or force changes of the construction object at each acquisition time. After acquisition, the data is organized according to the construction object identification and acquisition time to form load status data.

[0033] Operational status information is acquired for each construction object. This operational status information can be obtained through equipment controller operation records, work instruction execution records, switch quantity feedback records, or manual terminal feedback records. Specifically, the operational status information refers to the start / stop status, action execution status, action switching status, operating gear, or current work action content of the construction object at each acquisition time. After acquisition, the information is organized according to the construction object identification and acquisition time to form operational status data.

[0034] Environmental status information is acquired for the work area where the construction object is located. This environmental status information can be obtained through wind speed sensors, temperature and humidity sensors, dust sensors, vibration sensors, or illuminance sensors. Specifically, the environmental status information refers to at least one of the following: wind speed, temperature, humidity, dust concentration, vibration intensity, and illuminance corresponding to the work area where the construction object is located. After acquisition, the information is organized according to the work area where the construction object is located and the time of acquisition, and further mapped to the construction object within that work area to form environmental status data. The load status data, the work status data, and the environmental status data serve as components for the subsequent generation of multi-source status data.

[0035] Steps 1-5: Perform time correction on the position status data, attitude status data, load status data, operation status data, and environmental status data to generate multi-source status data. Specifically, since different acquisition devices have different acquisition frequencies, time correction processing is performed on the aforementioned types of status data. The basis for the time correction processing is that multiple types of status data corresponding to the same construction object need to be read together at the same analysis time.

[0036] In specific processing, a preset settling time sequence is used as a unified time reference. For a certain settling time, the most recently collected result whose time deviation from the settling time is within the preset time deviation range is read and mapped to the settling time. When a certain type of status data is missing at the current settling time, the same type of status data corresponding to the construction object at the previous settling time is called as continuous status data.

[0037] After time realignment is completed, the location status data, attitude status data, load status data, operation status data, and environmental status data corresponding to the same construction object at the same realignment time are combined and organized to generate multi-source status data corresponding to the construction object. The multi-source status data includes at least the construction object identifier, realignment time, location status data, attitude status data, load status data, operation status data, and environmental status data. The multi-source status data is used in step 3 to extract temporal change features during the status reconstruction period.

[0038] To illustrate the time normalization process of multiple types of state data in this embodiment, a hoisting device is used as an example. At a certain normalization moment, the position state data has the most recently acquired result whose time deviation from the normalization moment is within a preset time deviation range, the attitude state data also has a corresponding most recently acquired result, the load state data has not returned a new acquisition result at the normalization moment, and the operation state data and environmental state data have corresponding acquisition results. At this time, the most recently acquired result in the position state data and the attitude state data whose time deviation from the normalization moment is within the preset time deviation range is first mapped to the normalization moment. For the load state data, the load state data corresponding to the hoisting device at the previous normalization moment is called as continuous state data. Then, the position state data, attitude state data, load state data, operation state data, and environmental state data corresponding to the normalization moment are combined and organized according to the same construction object identifier to generate multi-source state data corresponding to the hoisting device at the normalization moment. Through this example, it can be seen that after time normalization, multiple types of state data obtained at different acquisition frequencies can be jointly read at the same analysis moment and used for the extraction of temporal change features in the subsequent state reconstruction period.

[0039] Steps 1-6: Obtain the current process execution data based on the construction object identifier. Specifically, read the construction organization arrangement record, task scheduling record, equipment operation record, and work instruction record corresponding to each construction object. Among them, the construction organization arrangement record is used to represent the content of the planned process and the sequence of the planned processes; the task scheduling record is used to represent the currently assigned work tasks; the equipment operation record is used to represent the actions performed by the construction equipment and the timing of the actions; and the work instruction record is used to represent the action instructions issued on site and their execution feedback.

[0040] Subsequently, according to the correspondence between the construction object identifier and the time, the planned process content, current task content, equipment actions performed, and work instruction execution feedback are organized accordingly to determine the process identifier, preceding process identifier, subsequent process identifier, current action, action start time, action end time, and work area changes corresponding to each construction object at the current time, and generate current process execution data; the current process execution data is used in step S2 to identify the process switching stage and determine the state reconstruction period.

[0041] Steps 1-7: Generate construction object association data based on the site layout information, work sequence information, and collaborative work information corresponding to each construction object. Specifically, first, read the site layout information corresponding to each construction object; the site layout information can come from the construction site layout diagram, the object layout results in the digital twin scene, equipment installation location records, and work area division results; the site layout information includes at least the layout coordinates, the work area to which each construction object belongs, the work area boundary, and the distribution of adjacent construction objects.

[0042] Subsequently, based on the layout coordinates of each construction object and its corresponding work area, the spatial distance between any two construction objects is calculated, and it is determined whether the two objects are located in the same work area, adjacent work areas, or within a preset influence distance range. When any two construction objects are located in the same work area, adjacent work areas, or the spatial distance between them is not greater than the preset influence distance, the two construction objects are determined to have a spatial adjacency relationship. The preset influence distance is formed based on the action influence range of the corresponding construction object during the construction process, the safe operating distance of equipment, and the site area division requirements. When determining whether any two construction objects have a spatial adjacency relationship, the currently calculated spatial distance is compared with the preset influence distance.

[0043] Furthermore, the operation sequence information corresponding to each construction object is read; the operation sequence information can specifically come from the process arrangement results in the construction organization design, the construction network plan, the task decomposition results, and the on-site work assignment records; the operation sequence information includes at least the current process, the preceding process, the subsequent process, and the process succession order in which each construction object participates; then, based on the succession relationship between the processes in which each construction object participates, it is determined whether the completion of the process corresponding to one construction object constitutes the starting prerequisite for the process corresponding to another construction object; when the completion of the process corresponding to one construction object constitutes the starting prerequisite for the process corresponding to another construction object, or when two construction objects participate in the preceding and following processes in the same construction task, the two construction objects are determined to have an operation dependency relationship.

[0044] Further, the collaborative operation information corresponding to each construction object is read; the collaborative operation information may specifically come from hoisting plans, transportation scheduling records, equipment collaborative operation records, and on-site joint task arrangement results; the collaborative operation information includes at least the combination of construction objects participating in the same construction task, the start time of collaborative operation, the end time of collaborative operation, and the action coordination requirements in collaborative operation; then, based on the collaborative operation information, it is determined whether any two construction objects are participating in the same construction task within the same time range, and it is determined whether there are action coordination, position coordination, or load coordination requirements between the two; when any two construction objects are participating in the same construction task within the same time range, and there are at least one of action coordination, position coordination, or load coordination requirements between the two, the two construction objects are determined to have a collaborative relationship.

[0045] After determining spatial adjacency, operational dependency, and synergistic relationships, construction object identifier pairs are used as association units to organize the spatial adjacency, operational dependency, and synergistic relationships between any two construction objects, generating construction object association data. The construction object association data includes at least construction object identifier pairs, spatial adjacency determination results, operational dependency determination results, and synergistic relationship determination results. The construction object association data is used in subsequent steps to determine the coupling response relationship between construction objects.

[0046] Steps 1-8: Establish a unified correspondence and output the multi-source status data, the current process execution data, and the construction object association data. Specifically, establish an object correspondence between the multi-source status data, the current process execution data, and the construction object association data based on the construction object identifier; establish a time correspondence between the multi-source status data and the current process execution data based on the time sequence; after completing the unified correspondence, output the multi-source status data, the current process execution data, and the construction object association data.

[0047] This step first clarifies the identification of the construction object, and then generates multi-source status data, current process execution data, and construction object association data with clear collection sources and data content. On the one hand, this ensures that the process switching stage identification in step S2 is based on clear process execution facts; on the other hand, it ensures that the time sequence change feature extraction in step S3 is based on clear status collection results.

[0048] See Figure 2 As shown, step S2: Identify the process switching stage based on the current process execution data, and determine the state reconstruction period corresponding to the process switching stage.

[0049] In some implementations, the steps of identifying the process switching phase based on the current process execution data and determining the state reconstruction period corresponding to the process switching phase include: Step 2-1: Read the current process execution data and determine the process switching candidate time period. Specifically, read the process identifier, previous process identifier, subsequent process identifier, current execution action, action start time, action end time, equipment action code, and work area identifier corresponding to each construction object from the current process execution data.

[0050] When the process identifier changes within a continuous time range for the same construction object, or when the currently executed action changes from a previous type of operation to a later type of operation, or when the equipment action code changes and is accompanied by a change in the work area identifier, the corresponding continuous time range is determined as a candidate period for process switching; wherein, the equipment action code is used to characterize the specific action type performed by the construction equipment at the corresponding time, and the work area identifier is used to characterize the work area where the construction object is located at the corresponding time.

[0051] Step 2-2: Perform validity verification on the candidate time period for process switching and identify the process switching stage. Specifically, read the corresponding equipment action code sequence, work area identification sequence and candidate time period duration within the candidate time period for process switching. The equipment action code sequence is used to characterize the action switching process of the construction object within the candidate time period for process switching, and the work area identification sequence is used to characterize whether the construction object enters, exits or moves across work areas within the candidate time period for process switching.

[0052] The duration of the candidate time period is compared with the shortest confirmation time. The shortest confirmation time is formed based on the duration of the historical switching records of similar processes, the response time of equipment actions, and the adjustment time of on-site operations. When in use, the corresponding shortest confirmation time is read according to the current construction object category and the current process type, and compared with the duration of the current candidate time period.

[0053] When the candidate time period for process switching meets the following conditions simultaneously, the candidate time period for process switching is determined as the process switching stage: 1. The process identifier changes or the currently executed action changes; 2. There is a continuous action switching in the equipment action code sequence; 3. There is at least one of area entry, area exit, or cross-area movement in the work area identifier sequence; 4. The duration of the candidate time period is not less than the minimum confirmation duration.

[0054] Step 2-3: Determine the state reconstruction period corresponding to the process switching stage based on the process switching stage. Specifically, first determine the start time and end time of the process switching stage. The start time is determined based on the earliest of the following: the end of the corresponding action of the previous process, the start of the preparation action of the next process, the first switching of the equipment action code, or the first change of the work area identifier. The end time is determined based on the core action corresponding to the next process entering a stable execution state, and the position offset, attitude angle change, or load change corresponding to the construction object entering the corresponding preset stability judgment range. Here, the position offset refers to the coordinate change between adjacent adjustment times, the attitude angle change refers to the pitch angle, roll angle, rotation angle, or tilt angle change between adjacent adjustment times, and the load change refers to the load increase or decrease between adjacent adjustment times.

[0055] The preset stability judgment range is not set for the overall state data, but rather for the stability judgment intervals set separately for position offset, attitude angle change, and load change. It is formed based on the stability state records after the completion of similar historical processes, the stability state requirements of the corresponding processes in the construction plan, and the equipment operation requirements. When invoked, the corresponding position offset stability judgment interval, attitude angle stability judgment interval, and load stability judgment interval are read according to the current construction object category and the current process type, and the current position offset, current attitude angle change, and current load change are compared with the corresponding judgment intervals respectively.

[0056] When determining the end time of the process switching phase, the set of target stable state quantities participating in the end determination is first determined according to the current construction object category and the current process switching type; wherein, the set of target stable state quantities may include at least one of position offset, attitude angle change and load change.

[0057] The target stable state quantity set is formed based on: the stable state records after the completion of the switching of similar processes in the past, the stable state requirements of the corresponding process in the construction plan, and the equipment operation requirements; the calling method is: according to the current construction object category and the current process switching type, read the corresponding target stable state quantity set and the preset stability judgment interval corresponding to each state quantity; when each state quantity in the target stable state quantity set enters its corresponding preset stability judgment interval, and the equipment action code remains the core action code corresponding to the next process, the corresponding time is determined as the end time of the process switching stage.

[0058] After determining the start and end times of the process switching phase, the start time is used as the starting reference time of the state reconstruction period, and the end time is used as the ending reference time of the state reconstruction period. Then, based on the preparation time, stabilization time, and equipment inertia change time in similar historical process switching, the preparation time range is extended before the starting reference time, and the stabilization time range is extended after the ending reference time, to determine the state reconstruction period corresponding to the process switching phase.

[0059] This step establishes the identification of process switching stages based on changes in process identifiers, switching of action codes, changes in work area identifiers, and confirmation of duration. It also establishes the determination of the end of process switching based on comparable state quantities such as position offset, attitude angle change, and load change entering the corresponding stable determination range. This provides a clear quantitative basis for determining the process switching stages and their corresponding state reconstruction periods, thus providing a clear and feasible time range for subsequent extraction of temporal change features.

[0060] To illustrate the process of determining the process switching stage and the state reconstruction period in this embodiment, a scenario in which the hoisting equipment switches from the component lifting process to the component installation process is used as an example.

[0061] After reading the current process execution data, if it is found that the process identifier corresponding to the same hoisting equipment changes from hoisting to installation, the current execution action changes from lifting to slewing and luffing, and the work area identifier changes from material storage area to installation work area, then the corresponding continuous time range is first determined as a process switching candidate period. Further, if there are continuous action changes in the equipment action code sequence within this candidate period, cross-area changes in the work area identifier sequence, and the duration of the candidate period is not less than the shortest confirmation time corresponding to the current construction object category and the current process type, then this candidate period is determined as the process switching stage. Afterwards, the actions corresponding to the previous process are... The earliest moment among the following—the end of the process, the start of preparation for the next process, the first switch of equipment action code, or the first change of work area identifier—is determined as the start time. The moment when the core action corresponding to the next process enters a stable execution state, and the target stable state quantities involved in the end determination—position offset, attitude angle change, and load change—all enter their respective preset stable determination intervals, is determined as the end time. Finally, the state reconstruction period is determined by combining the preceding and following time ranges. This example illustrates that the state reconstruction period is not simply extracted based on changes in process name, but is determined by combining action switching, area switching, and the stability results of state quantities.

[0062] See Figure 2 As shown, in step S3, the multi-source state data within the state reconstruction period corresponding to the process switching stage is segmented and analyzed to extract time-series change features.

[0063] In some implementations, the steps for extracting temporal variation features include: Step 3-1: Read the multi-source state data within the state reconstruction period and form a state time series. Specifically, according to the state reconstruction period, read the multi-source state data corresponding to the state reconstruction period from the multi-source state data; then, arrange the read position state data, attitude state data, load state data, operation state data, and environmental state data according to the construction object identification and the order of the settling time to form a state time series corresponding to each construction object within the state reconstruction period.

[0064] Step 3-2: Generate a state change sequence based on the state time series. Specifically, read the state time series corresponding to two adjacent settling times for the same construction object, and use two adjacent settling times as a state change calculation unit to generate a state change sequence for each state change calculation unit.

[0065] For each state change calculation unit, the calculation is performed as follows: First, for the position status data, read the plane coordinates and elevation coordinates corresponding to the previous adjustment time, and the plane coordinates and elevation coordinates corresponding to the next adjustment time.

[0066] The planar coordinates corresponding to the next settling time are subtracted from the planar coordinates corresponding to the previous settling time to obtain the planar displacement component; the elevation coordinates corresponding to the next settling time are subtracted from the elevation coordinates corresponding to the previous settling time to obtain the elevation displacement component; then, based on the planar displacement component and the elevation displacement component, the position offset is determined; wherein, the position offset is used to characterize the degree of position change of the construction object between two adjacent settling times; the position offset direction is determined according to the direction of the planar displacement component; the position change rate is obtained by calculating the position offset in correspondence with the time interval between the previous settling time and the next settling time; thus, the position offset, position offset direction, and position change rate of the construction object in the state change calculation unit are generated.

[0067] Secondly, for the attitude state data, read the attitude angle data corresponding to the previous adjustment moment and the attitude angle data corresponding to the next adjustment moment.

[0068] The attitude angle data may include at least one of pitch angle, roll angle, yaw angle, or tilt angle. For the actual attitude angle type configured for the current construction object, the attitude angle values ​​at the next settling moment and the previous settling moment are differentially processed to obtain the corresponding attitude angle change. The direction of attitude angle change is then determined based on the positive or negative direction of the differential result. Subsequently, the attitude angle change is calculated in relation to the time interval between the previous and next settling moments to obtain the attitude angle change rate. If the current construction object has multiple attitude angle data, the change amount and change rate corresponding to each attitude angle are generated separately, and the attitude angle change result with the largest change amplitude is determined as the main attitude change result in the current state change calculation unit. Alternatively, all attitude angle change results are retained together as the attitude change result in the current state change calculation unit. Thus, the attitude angle change amount, attitude angle change direction, and attitude angle change rate of the construction object in the state change calculation unit are generated.

[0069] Furthermore, for the load status data, the load value corresponding to the previous settling time and the load value corresponding to the next settling time are read.

[0070] The load change is obtained by subtracting the load value corresponding to the next settling time from the load value corresponding to the previous settling time. The direction of load change is determined based on the positive or negative direction of the subtracting result. The load change rate is then calculated by subtracting the load change from the previous settling time to the next settling time. The load change is used to characterize the degree of force change of the construction object between two adjacent settling times, the load change direction is used to characterize whether the load is increasing or decreasing, and the load change rate is used to characterize the speed of load change. Thus, the load change, load change direction, and load change rate of the construction object in the state change calculation unit are generated.

[0071] Furthermore, for the job status data, the action code corresponding to the previous settling moment and the action code corresponding to the next settling moment are read.

[0072] The action code corresponding to the next settling moment is compared with the action code corresponding to the previous settling moment. When they match, the current state change calculation unit is determined as the action retention result. When they do not match, the action code corresponding to the previous settling moment is used as the cut-out action code, and the action code corresponding to the next settling moment is used as the cut-in action code. The change result from the cut-out action code to the cut-in action code is determined as the action code switching result. Thus, the action code switching result of the construction object in the state change calculation unit is generated.

[0073] Furthermore, for the work area data, the work area identifier corresponding to the previous settling time and the work area identifier corresponding to the next settling time are read.

[0074] The work area identifier corresponding to the next reorganization time is compared with the work area identifier corresponding to the previous reorganization time. When they match, the current state change calculation unit is determined as a result of no area switching. When they do not match, the coordinates of the construction object corresponding to the previous reorganization time, the coordinates of the construction object corresponding to the next reorganization time, and the corresponding area boundary in the site layout diagram are used to determine whether the construction object exits from the original work area and enters the target work area, or crosses from the original work area to an adjacent work area, and an area switching result is generated. Thus, the area switching result of the construction object in the state change calculation unit is generated.

[0075] After completing the aforementioned calculations, the position offset, position offset direction, position change rate, attitude angle change, attitude angle change direction, attitude angle change rate, load change, load change direction, load change rate, action code switching result, and area switching result of the same construction object in each state change calculation unit are arranged in chronological order from the previous settling time to the next settling time to generate the state change sequence corresponding to the construction object.

[0076] The state change sequence includes at least: construction object identifier, previous resetting time, next resetting time, position offset, position offset direction, position change rate, attitude angle change, attitude angle change direction, attitude angle change rate, load change, load change direction, load change rate, action code switching result, and area switching result.

[0077] Step 3-3: Determine the segment boundaries based on the state change sequence and form a segmented state sequence. Specifically, read the state change sequence corresponding to the same construction object, and use each state change calculation unit in the state change sequence as the judgment object to determine the segment boundaries corresponding to the pre-switching preparation segment, the switching execution segment, and the post-switching stable segment in sequence.

[0078] First, determine the starting boundary of the switching execution segment; read the position change rate, attitude angle change rate, load change rate, motion code switching result, and area switching result corresponding to each state change calculation unit in chronological order; compare the position change rate with the position change starting judgment range, compare the attitude angle change rate with the attitude change starting judgment range, and compare the load change rate with the load change starting judgment range.

[0079] The determination ranges for the starting of position change, attitude change, and load change are formed based on the following: the starting characteristics of position change, attitude change, and load change corresponding to the start stage of switching between similar processes, as well as the upper limit of sensor acquisition fluctuations. When invoked, the corresponding starting determination ranges are read according to the current construction object category and the current process switching type, and compared with the position change rate, attitude angle change rate, and load change rate in the current state change calculation unit.

[0080] When at least one of the position change rate, attitude angle change rate, and load change rate in a series of consecutive state change calculation units enters outside the corresponding initial judgment range, and an action code switching result or a region switching result appears in the current state change calculation unit or in a series of consecutive state change calculation units thereafter, the first state change calculation unit at the start of the continuous change is determined as the starting boundary of the switching execution segment; the formation of the series of consecutive state change calculation units is based on the shortest continuous change duration required for the transition from the preparation state to the execution state in historical similar process switching.

[0081] Furthermore, after determining the starting boundary of the switching execution segment, the ending boundary of the switching execution segment is determined.

[0082] Continue reading each state change calculation unit located after the starting boundary of the switching execution segment in chronological order, and compare the current position change rate with the position change stability judgment range, the current attitude angle change rate with the attitude change stability judgment range, and the current load change rate with the load change stability judgment range; at the same time, read the action code switching result and the region switching result corresponding to the current state change calculation unit.

[0083] The stability determination ranges for position change, attitude change, and load change are formed based on the following: the distribution ranges of position change rate, attitude change rate, and load change rate in the corresponding stable stage after the completion of the same type of process switching, as well as the equipment stable operation requirements; when called, the corresponding stability determination ranges are read according to the current construction object category and the current process switching type, and compared with the position change rate, attitude angle change rate, and load change rate in the current state change calculation unit.

[0084] When the position change rate, attitude angle change rate, and load change rate all enter their respective stable determination ranges in several consecutive state change calculation units, and the action code switching result indicates that the current construction object has maintained the core action code corresponding to the next process, and the area switching result indicates that the current construction object has stopped cross-area changes or has entered the target operation area and kept the area identifier unchanged, the first state change calculation unit that starts the continuous stability is determined as the starting boundary of the stable segment after the switch, and the previous state change calculation unit is determined as the ending boundary of the switching execution segment.

[0085] After determining the start and end boundaries of the switching execution segment, the state change calculation units before the start boundary are determined as the pre-switching preparation segment, the state change calculation units between the start boundary and the end boundary are determined as the switching execution segment, and the start boundary and the state change calculation units after the switching stable segment are determined as the switching stable segment.

[0086] Subsequently, the segment identifiers of each state change calculation unit are written into the corresponding state time series and state change quantity series to generate a segmented state sequence. The segmented state sequence includes at least: construction object identifier, the previous and next settling times corresponding to the state change calculation unit, segment identifier, position offset, position offset direction, position change rate, attitude angle change, attitude angle change direction, attitude angle change rate, load change, load change direction, load change rate, action code switching result, and area switching result.

[0087] Steps 3-4: Extract time-series change features based on the segmented state sequence. Specifically, read the segmented state sequence corresponding to the same construction object, and extract time-series change features for the pre-switching preparation segment, the switching execution segment, and the post-switching stable segment, respectively.

[0088] First, for each segmented state sequence under each segment identifier, the feature extraction interval is determined according to the state type. The state types include position state type, attitude state type, load state type, action state type, and region state type. For position state type, attitude state type, and load state type, the change amount, change direction, and change rate corresponding to each state change calculation unit in the same segment are read in sequence, and the continuity determination is performed in chronological order.

[0089] The criteria for determining continuity include: whether the change direction corresponding to adjacent state change calculation units is consistent, whether the change amount is outside the corresponding micro-fluctuation exclusion threshold, and whether the change rate remains continuously changing; wherein, the micro-fluctuation exclusion threshold is formed based on the resolution error range of the corresponding acquisition device, the natural fluctuation range under historical stable conditions, and the minimum effective change amount corresponding to the current construction object's state type; when invoked, the corresponding micro-fluctuation exclusion threshold is read according to the current construction object category and current state type, and the change amount corresponding to the current state change calculation unit is compared with the micro-fluctuation exclusion threshold.

[0090] When the change amount corresponding to a certain state change calculation unit exceeds the micro-fluctuation exclusion threshold, and its change direction is consistent with the change direction corresponding to several subsequent consecutive state change calculation units, the state change calculation unit is determined as the starting calculation unit of a feature extraction interval; when the change direction corresponding to a subsequent state change calculation unit changes, or the change amount returns to within the micro-fluctuation exclusion threshold several times consecutively, the last state change calculation unit before the change direction changes or before the change amount returns to within the micro-fluctuation exclusion threshold consecutively is determined as the ending calculation unit of the feature extraction interval; thus, one or more feature extraction intervals for position state type, attitude state type, or load state type are formed within the same segment.

[0091] For each action state type, the action code switching results corresponding to each state change calculation unit within the same segment are read sequentially. When the action code switching result corresponding to a certain state change calculation unit is not an action hold result, the state change calculation unit is determined as the starting calculation unit of the action feature extraction interval. Subsequent state change calculation units are read sequentially. When the cut-in action code corresponding to the subsequent state change calculation unit is continuously connected with the cut-in action code corresponding to the previous state change calculation unit, it is included in the same action feature extraction interval. When the action code switching result corresponding to the subsequent state change calculation unit changes back to an action hold result, or the cut-in action code is no longer continuously connected with the cut-in action code corresponding to the previous state change calculation unit, the previous state change calculation unit is determined as the ending calculation unit of the action feature extraction interval. Thus, one or more feature extraction intervals for action state types are formed within the same segment.

[0092] For each region state type, the region switching results corresponding to each state change calculation unit within the same segment are read sequentially. When the region switching result corresponding to a certain state change calculation unit is not a region not switched result, the state change calculation unit is determined as the starting calculation unit of the region feature extraction interval. Subsequent state change calculation units are read sequentially. When the region switching result corresponding to a subsequent state change calculation unit has a connection relationship with the target region corresponding to the previous state change calculation unit, it is included in the same region feature extraction interval. When the region switching result corresponding to a subsequent state change calculation unit changes back to a region not switched result, or the region switching target is no longer connected with the previous state change calculation unit, the previous state change calculation unit is determined as the ending calculation unit of the region feature extraction interval. Thus, one or more feature extraction intervals for the same region state type are formed within the same segment.

[0093] After determining the feature extraction intervals corresponding to each state type, the temporal change features corresponding to each feature extraction interval are determined.

[0094] For position state types, the position offset, position offset direction, and position change rate corresponding to each state change calculation unit within a certain position feature extraction interval are read, and the position temporal change characteristics are determined as follows: the previous settling time corresponding to the first state change calculation unit in the position feature extraction interval is determined as the position change start time; the next settling time corresponding to the last state change calculation unit in the position feature extraction interval is determined as the position change end time; the time difference between the position change end time and the position change start time is determined as the position change duration; the position change calculation unit within the position feature extraction interval is further analyzed. The position offsets corresponding to each element are accumulated sequentially over time to determine the cumulative position offset. The sum of the cumulative position offsets corresponding to different position offset directions for each state change calculation unit within the position feature extraction interval is calculated, and the direction with the largest sum of cumulative position offsets is determined as the primary position offset direction. The position change rate corresponding to each state change calculation unit within the position feature extraction interval is read, and the maximum value is determined as the maximum position change rate. The position state data of the last state change calculation unit in the position feature extraction interval at the next settling time in the segmented state sequence is read and determined as the position state corresponding to the end time. Thus, the temporal change features corresponding to the position state type are generated.

[0095] For attitude state types, read the attitude angle change, attitude angle change direction, and attitude angle change rate corresponding to each state change calculation unit within a certain attitude feature extraction interval, and determine the attitude temporal change characteristics according to the following method: The attitude change start time is determined by the preceding settling time corresponding to the first state change calculation unit in the attitude feature extraction interval; the attitude change end time is determined by the following settling time corresponding to the last state change calculation unit in the attitude feature extraction interval; the time difference between the attitude change end time and the attitude change start time is determined as the attitude change duration; the attitude angle changes corresponding to each state change calculation unit in the attitude feature extraction interval are accumulated in chronological order to determine the cumulative attitude angle changes; the sum of the cumulative attitude angle changes corresponding to different attitude change directions for each state change calculation unit in the attitude feature extraction interval is calculated, and the direction with the largest sum of cumulative attitude angle changes is determined as the main attitude change direction; the attitude angle change rate corresponding to each state change calculation unit in the attitude feature extraction interval is read, and the maximum value is determined as the maximum attitude angle change rate; the attitude state data in the segmented state sequence corresponding to the following settling time of the last state change calculation unit in the attitude feature extraction interval is read and determined as the attitude state corresponding to the end time; thus, the temporal change features corresponding to the attitude state type are generated.

[0096] For each load state type, the load change amount, load change direction, and load change rate corresponding to each state change calculation unit within a certain load feature extraction interval are read, and the load time series change characteristics are determined according to the following method: The load change start time is determined by the preceding settling time corresponding to the first state change calculation unit in the load feature extraction interval; the load change end time is determined by the following settling time corresponding to the last state change calculation unit in the load feature extraction interval; the time difference between the load change end time and the load change start time is determined as the load change duration; the load change amounts corresponding to each state change calculation unit in the load feature extraction interval are accumulated in chronological order to determine the cumulative load change amount; the sum of the cumulative load change amounts corresponding to different load change directions for each state change calculation unit in the load feature extraction interval is calculated, and the direction with the largest sum of cumulative load change amounts is determined as the main load change direction; the load change rate corresponding to each state change calculation unit in the load feature extraction interval is read, and the maximum value among them is determined as the maximum load change rate; the load state data in the segmented state sequence corresponding to the following settling time of the last state change calculation unit in the load feature extraction interval is read, and it is determined as the load state corresponding to the end time; thus, the temporal change features corresponding to the load state type are generated.

[0097] For action state types, read the action encoding switching results corresponding to each state change calculation unit within a certain action feature extraction interval, and determine the action temporal change characteristics according to the following method: The action switching start time is determined by the preceding settling time corresponding to the first state change calculation unit in the action feature extraction interval; the action switching end time is determined by the following settling time corresponding to the last state change calculation unit in the action feature extraction interval; the time difference between the action switching end time and the action switching start time is determined as the action switching duration; the cut-out action code corresponding to the first state change calculation unit in the action feature extraction interval is read and determined as the action switching start code; the cut-in action code corresponding to the last state change calculation unit in the action feature extraction interval is read and determined as the action switching end code; the action code switching results corresponding to each state change calculation unit in the action feature extraction interval are read in chronological order, and the same cut-in action codes that appear repeatedly are merged into one action switching, and the merged action code switching results are arranged in chronological order to determine the action switching sequence feature; thus, the temporal change feature corresponding to the action state type is generated.

[0098] For the region state type, read the region switching results corresponding to each state change calculation unit within a certain region feature extraction interval, and determine the region temporal change characteristics according to the following method: The region switching start time is determined by the preceding settling time corresponding to the first state change calculation unit in the region feature extraction interval; the region switching end time is determined by the following settling time corresponding to the last state change calculation unit in the region feature extraction interval; the time difference between the region switching end time and the region switching start time is determined as the region switching duration; the original work area identifier corresponding to the first state change calculation unit in the region feature extraction interval is read and determined as the region switching start area; the target work area identifier corresponding to the last state change calculation unit in the region feature extraction interval is read and determined as the region switching end area; the region switching results corresponding to each state change calculation unit in the region feature extraction interval are read in chronological order, and the original work area identifier, intermediate area identifier, and target work area identifier involved are arranged in sequence to determine the region switching path feature; thus, the temporal change feature corresponding to the region state type is generated.

[0099] After determining the temporal change characteristics corresponding to each of the aforementioned state types, the temporal change characteristics of the position state type, attitude state type, load state type, action state type, and area state type of the same construction object within the same segment are organized in chronological order of the start time of the change, thereby generating the temporal change characteristics of the construction object within that segment.

[0100] The temporal change features include general fields and state type-specific fields. The general fields include at least: construction object identifier, process switching stage identifier, segment identifier, state type, change start time, change end time, change duration, and the state result corresponding to the end time. When the temporal change feature corresponds to a position state type, attitude state type, or load state type, its specific fields include at least the cumulative change amount, main change direction, and maximum change rate. When the temporal change feature corresponds to an action state type, its specific fields include at least the action switching sequence feature. When the temporal change feature corresponds to a region state type, its specific fields include at least the region switching path feature.

[0101] Then, according to the construction object identification and the order of segmentation, the time-series change features corresponding to the pre-switching preparation segment, the time-series change features corresponding to the switching execution segment, and the time-series change features corresponding to the stable segment after switching are arranged to generate the time-series change features corresponding to the construction object; the time-series change features are used in step S4 to construct the preset stage response law, perform corresponding comparison and classify and generate staged abnormal features.

[0102] See Figure 2 As shown, in step S4, based on the preset stage response pattern corresponding to the process switching stage, the time-series change characteristics are classified to determine the switching response characteristics and continuous abnormal characteristics corresponding to the process switching stage, and staged abnormal characteristics are generated. In some implementations, the steps for generating staged anomaly features include: Step 4-1: Generate a preset stage response pattern based on the historical results of similar process switching and construction requirements. Specifically, first determine the process switching type and construction object category corresponding to the current construction object.

[0103] The process switching type is determined by the preceding process identifier and the subsequent process identifier identified in step S2, and the construction object category is determined by the construction object category determined in step S1. Subsequently, the historical process switching records of the same type corresponding to the current process switching type and the current construction object category, the process conversion requirements in the construction plan, the equipment action execution requirements, and the on-site operation constraint information are read.

[0104] Among them, historical switching records of similar processes are used to provide historical time-series change characteristics, process conversion requirements in the construction plan are used to provide the sequence of process switching steps, equipment action execution requirements are used to provide the action code switching sequence and action continuity requirements, and on-site operation constraint information is used to provide area switching requirements, load change restrictions, and attitude change restrictions.

[0105] Subsequently, historical temporal change characteristics were extracted from historical records of similar process switching, and categorized and organized according to segment identifiers and status types. Specifically, for multiple historical samples under the same process switching type and the same construction object category, the order of occurrence, main direction of change, cumulative change distribution, maximum change rate distribution, duration distribution, and end status distribution of position status type, attitude status type, load status type, action status type, and area status type were statistically analyzed in the pre-switching preparation segment, switching execution segment, and post-switching stabilization segment. The aforementioned statistical results were then cross-checked with the process conversion requirements, equipment action execution requirements, and on-site operation constraint information in the construction plan, retaining common temporal change results that appeared simultaneously in the historical samples and construction requirements.

[0106] After cross-checking, the standard occurrence segments of position status types during the process switching phase are determined, and the position change sequence requirements, position change direction requirements, position offset duration intervals, and position offset fallback judgment intervals corresponding to the position status types are generated; the standard occurrence segments of attitude status types during the process switching phase are determined, and the attitude change sequence requirements, attitude change direction requirements, attitude angle change duration intervals, and attitude angle change fallback judgment intervals corresponding to the attitude status types are generated. Determine the standard occurrence segments of load state types during the process changeover phase, and generate the load change sequence requirements, load change direction requirements, load change duration range, and load change fall-off judgment range corresponding to the load state type; determine the standard occurrence segments of action state types during the process changeover phase, and generate the action code switching sequence requirements and action duration range corresponding to the action state type; determine the standard occurrence segments of regional state types during the process changeover phase, and generate the regional switching sequence requirements and regional result requirements corresponding to the regional state type.

[0107] The standard occurrence segment is formed based on the high frequency occurrence segment of the corresponding state type in the historical records of similar process switching and the stage requirements of the corresponding process conversion steps for the state type in the construction plan; when called, the corresponding standard occurrence segment is read according to the current construction object category, the current process switching type and the current state type.

[0108] The aforementioned continuous interval and fallback judgment interval are formed based on the distribution range of changes in the historical samples of similar process switching that reach the preset sample ratio and the allowable change range in the construction requirements. When calling, the continuous interval and fallback judgment interval of the corresponding status type are read according to the current process switching type and the current construction object category.

[0109] After the data collection is completed, the common temporal change results retained after the aforementioned cross-checking are processed according to the current process switching type, current construction object category, segment identifier, and status type.

[0110] Specifically, for each position status type, the corresponding standard occurrence segment, position change sequence requirements, position change direction requirements, position offset duration interval, and position offset fallback judgment interval are read, and the position change results corresponding to the pre-switch preparation segment, the switch execution segment, and the post-switch stable segment are arranged in the order of segmentation. For each attitude state type, the corresponding standard occurrence segment, attitude change sequence requirements, attitude change direction requirements, attitude angle change duration range, and attitude angle change fallback judgment range are read, and the attitude change results corresponding to the pre-switch preparation segment, the switch execution segment, and the post-switch stabilization segment are arranged in the order of segmentation. For each load state type, the corresponding standard occurrence segment, load change sequence requirements, load change direction requirements, load change duration range, and load change fall-off judgment range are read, and the load change results corresponding to the pre-switch preparation segment, switch execution segment, and post-switch stabilization segment are arranged in the order of segmentation. For each action state type, the corresponding standard occurrence segment, action code switching sequence requirements, and action duration range are read, and the action switching results corresponding to the pre-switching preparation segment, switching execution segment, and post-switching stabilization segment are arranged in the order of segment sequence. For each region status type, the corresponding standard occurrence segment, region handover sequence requirements, and region result requirements are read, and the region handover results corresponding to the pre-handover preparation segment, handover execution segment, and post-handover stabilization segment are arranged in the order of segmentation.

[0111] When there are multiple candidate order requirements, multiple candidate direction requirements, multiple candidate duration intervals, multiple candidate fallback judgment intervals, or multiple candidate result requirements for the same state type in the same segment, retain the candidate requirements that simultaneously meet the historical high-frequency distribution range and the construction requirement restriction range; when there are multiple candidate requirements that are simultaneously met, select the candidate requirement with the largest number of corresponding historical samples.

[0112] Subsequently, in the order of pre-switching preparation segment, switching execution segment, and post-switching stabilization segment, the standard occurrence segment, sequence requirements, direction requirements, duration interval, fallback judgment interval, and result requirements corresponding to the position state type, attitude state type, load state type, action state type, and area state type are combined and organized to generate the preset stage response law corresponding to the process switching stage.

[0113] The preset stage response rules include at least the following: standard occurrence segmentation, change sequence requirements, change direction requirements, duration interval, fallback judgment interval, action switching sequence requirements, and regional result requirements corresponding to each state type; the preset stage response rules are used by step 4-2 to perform segment consistency judgment, sequence consistency judgment, direction consistency judgment, duration interval consistency judgment, fallback judgment consistency judgment, and result consistency judgment.

[0114] Step 4-2: Perform item-by-item comparison of the temporal change features according to the preset stage response rules, and generate feature comparison results. Specifically, read the temporal change features output in Step 3-4, and read the corresponding temporal change features item by item according to the construction object identifier, segment identifier, and state type. At the same time, read the stage sequence requirements, change direction requirements, duration interval, fallback judgment interval, action switching sequence requirements, and regional result requirements corresponding to the current construction object category, current process switching type, current segment identifier, and current state type in the preset stage response rules generated in Step 4-1.

[0115] Subsequently, for each time-series change feature, the following sub-item comparisons were performed: First, perform a segment consistency check.

[0116] Read the segment identifier corresponding to the current time-series change feature, and read the standard occurrence segment corresponding to the current state type in the preset stage response law; if the segment identifier corresponding to the current time-series change feature is consistent with the standard occurrence segment, then the segment consistency of the current time-series change feature is determined to be valid; if the segment identifier corresponding to the current time-series change feature is inconsistent with the standard occurrence segment, then the segment consistency of the current time-series change feature is determined to be invalid.

[0117] Secondly, the consistency of execution order is determined.

[0118] If the current time-series change feature corresponds to a position state type, attitude state type, or load state type, then the start time of the change of the current time-series change feature is read and sorted with the start times of the changes of other state types under the same construction object and the same process switching stage; then the corresponding state change sequence requirements in the preset stage response law are read; if the position of the current time-series change feature in the actual sorting is consistent with the state change sequence requirements, then the sequence consistency is determined to be valid; otherwise, the sequence consistency is determined to be invalid.

[0119] If the current time sequence change feature corresponds to the action state type, then the action switching sequence feature is read and compared with the action switching sequence requirement. If the action switching sequence feature is consistent with the action switching sequence requirement, then the sequence consistency is determined to be valid; otherwise, the sequence consistency is determined to be invalid.

[0120] If the current time-series change characteristics correspond to the regional state type, then the regional switching path characteristics are read and compared with the regional switching order requirements in the preset stage response rules. If the regional switching path characteristics are consistent with the regional switching order requirements, then the order consistency is determined to be valid; otherwise, the order consistency is determined to be invalid.

[0121] Secondly, the consistency of execution direction is determined.

[0122] If the current temporal change feature corresponds to the position state type, then the main position offset direction is read and compared with the position change direction requirement in the preset stage response law; if the main position offset direction is consistent with the position change direction requirement, then the direction consistency is determined to be valid; otherwise, the direction consistency is determined to be invalid.

[0123] If the current temporal change feature corresponds to the attitude state type, then the main attitude change direction is read and compared with the attitude change direction requirement in the preset stage response law; if the main attitude change direction is consistent with the attitude change direction requirement, then the direction consistency is determined to be valid; otherwise, the direction consistency is determined to be invalid.

[0124] If the current time-series change characteristics correspond to the load state type, then the main load change direction is read and compared with the load change direction requirement in the preset stage response law; if the main load change direction is consistent with the load change direction requirement, then the direction consistency is determined to be valid; otherwise, the direction consistency is determined to be invalid.

[0125] If the current temporal change feature corresponds to an action state type or a region state type, then the direction consistency determination will not be performed separately, but the direction-related content will be incorporated into the sequence consistency determination or the result consistency determination.

[0126] Furthermore, a consistency determination of the continuous interval is performed.

[0127] If the current temporal change feature corresponds to the position state type, then the cumulative position offset and the duration of position change are read and compared with the position offset duration interval and the position change duration interval in the preset stage response rule, respectively. If the cumulative position offset falls into the position offset duration interval and the duration of position change falls into the position change duration interval, then the consistency of the duration interval is determined to be valid; otherwise, the consistency of the duration interval is determined to be invalid.

[0128] If the current temporal change feature corresponds to the attitude state type, then the cumulative attitude angle change and the duration of attitude change are read and compared with the attitude angle change duration interval and the attitude change duration interval in the preset stage response law, respectively. If the cumulative attitude angle change falls into the attitude angle change duration interval and the duration of attitude change falls into the attitude change duration interval, then the consistency of the duration interval is determined to be valid; otherwise, the consistency of the duration interval is determined to be invalid.

[0129] If the current time-series change characteristic corresponds to the load state type, then the cumulative load change amount and load change duration are read and compared with the load change duration interval and load change duration interval in the preset stage response law, respectively; if the cumulative load change amount falls into the load change duration interval and the load change duration falls into the load change duration interval, then the consistency of the duration interval is determined to be valid; otherwise, the consistency of the duration interval is determined to be invalid.

[0130] If the current time-series change feature corresponds to the action state type, then the duration of the action switch is read and compared with the action duration interval; if the duration of the action switch falls within the action duration interval, then the consistency of the duration interval is determined to be valid; otherwise, the consistency of the duration interval is determined to be invalid.

[0131] If the current time-series change feature corresponds to the region status type, then the region switching duration is read and compared with the region switching duration interval; if the region switching duration falls within the region switching duration interval, then the consistency of the duration interval is determined to be valid; otherwise, the consistency of the duration interval is determined to be invalid; wherein, the method for calling each duration interval is as follows: based on the current construction object category, the current process switching type, the current segment identifier, and the current status type, the corresponding duration interval is read and compared with the cumulative change amount and duration corresponding to the current time-series change feature respectively.

[0132] Furthermore, a consistency determination of the pullback judgment is performed.

[0133] If the current time-series change feature corresponds to the position state type, then read the position state corresponding to the end time and calculate the position offset result corresponding to the end time; then compare the position offset result with the position offset amount fall-back judgment interval in the preset stage response law; if the position offset result falls into the position offset amount fall-back judgment interval, then the fall-back judgment consistency is determined to be valid; otherwise, the fall-back judgment consistency is determined to be invalid.

[0134] If the current temporal change feature corresponds to the attitude state type, then read the attitude state corresponding to the end time and calculate the attitude angle offset result corresponding to the end time; then compare the attitude angle offset result with the attitude angle change amount fallback judgment interval in the preset stage response law; if the attitude angle offset result falls into the attitude angle change amount fallback judgment interval, then the fallback judgment consistency is determined to be valid; otherwise, the fallback judgment consistency is determined to be invalid.

[0135] If the current time-series change characteristics correspond to the load state type, then read the load state corresponding to the end time and calculate the load offset result corresponding to the end time; then compare the load offset result with the load change amount fall-off judgment interval in the preset stage response law; if the load offset result falls into the load change amount fall-off judgment interval, then the fall-off judgment consistency is determined to be valid; otherwise, the fall-off judgment consistency is determined to be invalid.

[0136] If the current temporal change characteristics correspond to an action state type or a region state type, then a separate fallback consistency determination will not be performed.

[0137] The position offset result, attitude angle offset result, and load offset result are formed by performing differential calculation between the state result corresponding to the end time and the reference state corresponding to the stable stage of the next process. The reference state is retrieved by reading the reference state result of the stable stage of the next process according to the current construction object category and the current process switching type.

[0138] Finally, the consistency of the execution results is determined.

[0139] If the current time-series change characteristics correspond to the action state type, then read the action switching end code and compare it with the core action code of the next process in the preset stage response law; if the two are consistent, then the result consistency is determined to be valid; otherwise, the result consistency is determined to be invalid.

[0140] If the current time-series change characteristics correspond to the regional state type, then read the region where the regional switch ended and compare it with the target operation region identifier in the preset stage response rule; if the two are consistent, then the result consistency is determined to be valid; otherwise, the result consistency is determined to be invalid.

[0141] If the current time-series change characteristics correspond to the position state type, attitude state type, or load state type, then the result consistency determination is not performed separately, but the fallback determination consistency determination reflects whether the final result meets the requirements.

[0142] After completing the aforementioned sub-item comparisons, the segment consistency judgment results, sequence consistency judgment results, direction consistency judgment results, continuous interval consistency judgment results, fallback judgment consistency judgment results, and result consistency judgment results corresponding to the current time series change characteristics are organized accordingly to generate feature comparison results.

[0143] The feature comparison results include at least: construction object identifier, process switching stage identifier, segment identifier, status type, segment consistency judgment result, sequence consistency judgment result, direction consistency judgment result, continuous interval consistency judgment result, fallback judgment consistency judgment result, and result consistency judgment result.

[0144] Step 4-3: Determine the switching response features and continuous abnormal features based on the feature comparison results. Specifically, read the feature comparison results and perform classification judgment item by item according to the construction object identifier, process switching stage identifier, segment identifier and status type.

[0145] First, determine the criteria for judging the switching response characteristics.

[0146] If a feature comparison result meets the following conditions, then the time-series change feature corresponding to that feature comparison result will be determined as the switching response feature: The segment consistency determination result is true; The result of the sequence consistency determination is true; When a direction consistency determination result exists, the direction consistency determination result is true; When there is a persistent interval consistency determination result, the persistent interval consistency determination result is true; When there is a consistent determination result for the fallback determination, the consistent determination result for the fallback determination is valid; When there is a result consistency determination result, the result consistency determination result is valid.

[0147] When all of the above conditions are met, the current time-series change characteristics are determined as the switching response characteristics.

[0148] Secondly, determine the criteria for identifying persistent abnormal characteristics.

[0149] When a feature comparison result shows at least one of the following conditions, the time-series change feature corresponding to that feature comparison result will be identified as a persistent anomalous feature: The segment consistency determination result is invalid; The result of the sequence consistency determination is that it is not true; There is a direction consistency determination result, and the direction consistency determination result is not valid; There is a persistent interval consistency determination result, and the persistent interval consistency determination result is not valid; There is a consistent determination result for the fallback determination, and the consistent determination result for the fallback determination is not valid; There is a result consistency determination result, and the result consistency determination result is not valid.

[0150] If the current time series change characteristics meet any of the aforementioned abnormal conditions, the current time series change characteristics will be determined as persistent abnormal characteristics.

[0151] Furthermore, in order to distinguish the source of anomalies in persistent anomalies, anomaly source marking processing is performed on the current persistent anomaly.

[0152] Specifically, when the segment consistency determination result is invalid, the current persistent abnormal feature is marked as a segmented abnormal persistent abnormal feature; when the sequence consistency determination result is invalid, the current persistent abnormal feature is marked as a sequence abnormal persistent abnormal feature; when the direction consistency determination result is invalid, the current persistent abnormal feature is marked as a direction abnormal persistent abnormal feature; when the persistent interval consistency determination result is invalid, the current persistent abnormal feature is marked as a persistent over-limit persistent abnormal feature; when the fallback determination consistency determination result is invalid, the current persistent abnormal feature is marked as a non-fallback persistent abnormal feature. When the consistency determination result is not met, the current persistent abnormal feature is marked as a persistent abnormal feature of the result abnormality type. When multiple items of the same time-series change feature are not met at the same time, all corresponding abnormal source tags are written into the current persistent abnormal feature.

[0153] After completing the classification and anomaly source marking, the switching response feature and the continuous anomaly feature are output respectively. The switching response feature is used to characterize the state change result that conforms to the normal response law of process switching. The continuous anomaly feature is used to characterize the state change result that does not change according to the preset stage response law or does not enter the corresponding judgment interval after the process switching is completed.

[0154] To illustrate the item comparison and feature classification process in this embodiment, the temporal change characteristics of the position status type of a certain hoisting equipment during the process switching stage are used as an example.

[0155] If the temporal change characteristic of the position state type corresponds to a switching execution segment, its segment identifier is consistent with the standard occurrence segment in the preset stage response rule, the sorting position corresponding to the start time of the change is consistent with the position change order requirement in the preset stage response rule, the main position offset direction is consistent with the position change direction requirement, the cumulative position offset and the position change duration fall into the corresponding position offset duration interval and position change duration interval respectively, and at the end of the stable segment after the switch, the position offset result corresponding to the end time falls back to the corresponding position offset fallback judgment interval, then the temporal change characteristic of the position state type is determined as a switching response characteristic; otherwise... If, although the temporal change characteristic of the position state type occurs within the process switching phase, there are at least one of the following situations: inconsistent segment identification, inconsistent main position offset direction, cumulative position offset exceeding the corresponding continuous interval, or failure to fall back to the corresponding fallback judgment interval after the stable segment ends following the switch, then the temporal change characteristic of the position state type is determined as a continuous abnormal characteristic, and the corresponding invalid item is written into the abnormal source marker. This example illustrates that the state change that occurs within the process switching phase does not automatically constitute an abnormality. Instead, a sub-item comparison is first performed based on the preset phase response rules, and then it is distinguished into switching response characteristics and continuous abnormal characteristics.

[0156] Step 4-4: Generate staged anomaly features based on the continuous anomaly features. Specifically, read the continuous anomaly features and classify and organize them according to the construction object identifier, process switching stage identifier, segment identifier, and status type to generate staged anomaly features.

[0157] Specifically, first determine the anomaly classification unit.

[0158] Continuous abnormal features with the same construction object identifier, the same process switching stage identifier, the same segment identifier, and the same state type are identified as continuous abnormal features in the same abnormal classification unit; continuous abnormal features with different construction object identifiers, different process switching stage identifiers, different segment identifiers, or different state types are identified as continuous abnormal features in different abnormal classification units; thus, each abnormal classification unit corresponds to "the continuous abnormal change result of the same construction object in the same segment and in the same state type within the same process switching stage".

[0159] Subsequently, anomaly feature processing is performed for each anomaly classification unit.

[0160] First, read the start time of change, end time of change, duration of change, cumulative change amount, main direction of change, maximum rate of change, status result corresponding to the end time, and anomaly source marker for each persistent anomaly feature in the current anomaly classification unit.

[0161] The anomaly source marker is used to characterize the anomaly source type corresponding to the current persistent anomaly feature. The anomaly source type includes at least one of the following: segmented anomaly, sequential anomaly, directional anomaly, persistent over-limit, non-fallback, and result anomaly.

[0162] Secondly, determine the start and end times of the anomaly occurrence corresponding to the current anomaly classification unit.

[0163] The start times of changes corresponding to each persistent abnormal feature in the current anomaly classification unit are compared, and the earliest start time of change is taken as the start time of anomaly occurrence corresponding to the current anomaly classification unit. The end times of changes corresponding to each persistent abnormal feature in the current anomaly classification unit are compared, and the latest end time of change is taken as the end time of anomaly occurrence corresponding to the current anomaly classification unit. The time difference between the end time of anomaly occurrence and the start time of anomaly occurrence is then determined as the total duration of anomaly occurrence corresponding to the current anomaly classification unit.

[0164] Furthermore, determine the abnormal change amount corresponding to the current abnormal classification unit.

[0165] If the current anomaly classification unit contains only one persistent anomaly feature, then the cumulative change corresponding to that persistent anomaly feature is directly determined as the anomaly change result corresponding to the current anomaly classification unit. If the current anomaly classification unit contains multiple persistent anomaly features, the cumulative change amount corresponding to each persistent anomaly feature is read, and the changes in the same direction are accumulated in chronological order, while the changes in opposite directions are canceled out to obtain the anomaly change amount result corresponding to the current anomaly classification unit; where the changes in the same direction refer to changes with the same main direction of change, and the changes in opposite directions refer to changes with the opposite main direction of change.

[0166] If the current anomaly classification unit corresponds to a position state type, then the anomaly change result is an abnormal position offset; if the current anomaly classification unit corresponds to an attitude state type, then the anomaly change result is an abnormal attitude angle change; if the current anomaly classification unit corresponds to a load state type, then the anomaly change result is an abnormal load change; if the current anomaly classification unit corresponds to an action state type, then the anomaly change result is an abnormal action switching result; if the current anomaly classification unit corresponds to a region state type, then the anomaly change result is an abnormal region switching result.

[0167] Furthermore, determine the main direction of the anomaly corresponding to the current anomaly classification unit.

[0168] The cumulative change of each persistent abnormal feature in the current anomaly classification unit is calculated, and the direction with the largest cumulative change is determined as the main anomaly direction corresponding to the current anomaly classification unit.

[0169] If the current anomaly classification unit corresponds to an action state type, then the action switching direction that is inconsistent with the preset stage response rule in the action switching sequence is determined as the main anomaly direction; if the current anomaly classification unit corresponds to a region state type, then the region switching direction that is inconsistent with the preset region switching sequence requirement in the region switching path is determined as the main anomaly direction.

[0170] Furthermore, determine the maximum rate of change of the anomaly corresponding to the current anomaly classification unit.

[0171] Read the maximum rate of change corresponding to each persistent abnormal feature in the current abnormal classification unit, and determine the maximum value among them as the maximum rate of change of the abnormality corresponding to the current abnormal classification unit; if the current abnormal classification unit corresponds to an action state type or a region state type and there is no rate of change value, then determine the corresponding action switching frequency or region switching frequency as the abnormal change rate result.

[0172] Furthermore, determine the exception termination status corresponding to the current exception classification unit.

[0173] Read the status result corresponding to the end time of the exception in the current exception classification unit, and determine the status result as the exception end status corresponding to the current exception classification unit.

[0174] Specifically, if the current anomaly classification unit corresponds to a position state type, then the anomaly termination state is the position state corresponding to the time when the anomaly occurred ended; if the current anomaly classification unit corresponds to an attitude state type, then the anomaly termination state is the attitude state corresponding to the time when the anomaly occurred ended; if the current anomaly classification unit corresponds to a load state type, then the anomaly termination state is the load state corresponding to the time when the anomaly occurred ended; if the current anomaly classification unit corresponds to an action state type, then the anomaly termination state is the action state corresponding to the time when the anomaly occurred ended; if the current anomaly classification unit corresponds to a region state type, then the anomaly termination state is the region state corresponding to the time when the anomaly occurred ended.

[0175] Furthermore, determine the combination of anomaly sources corresponding to the current anomaly classification unit.

[0176] Read the anomaly source markers corresponding to each persistent anomaly feature in the current anomaly classification unit, and perform deduplication and sorting of the anomaly source markers to generate the anomaly source combination result corresponding to the current anomaly classification unit.

[0177] If multiple anomaly source tags exist simultaneously in the current anomaly classification unit, then all multiple anomaly source tags are retained in the anomaly source combination result; if only one anomaly source tag exists in the current anomaly classification unit, then that anomaly source tag is determined as the anomaly source combination result.

[0178] After completing the aforementioned processing, the construction object identifier, process switching stage identifier, segment identifier, state type, anomaly start time, anomaly end time, total anomaly duration, anomaly change result, anomaly main direction, anomaly maximum change rate, anomaly termination state, and anomaly source combination result corresponding to the current anomaly classification unit are combined and processed to generate the staged anomaly features corresponding to the current anomaly classification unit. The staged anomaly features are used to characterize the continuous anomaly change results and anomaly sources of a certain construction object in a certain segment and a certain state type during a certain process switching stage.

[0179] When the same construction object has staged abnormal features corresponding to multiple segments or multiple state types in the same process switching stage, they are arranged in the order of segment sequence and state type sequence to generate a staged abnormal feature set of the construction object in the process switching stage; each staged abnormal feature in the staged abnormal feature set can be read independently, so that the next step can perform consistency verification and transient anomaly screening based on the coupling response relationship between construction objects.

[0180] The staged anomaly features include at least: construction object identifier, process switching stage identifier, segment identifier, state type, anomaly start time, anomaly end time, total anomaly duration, anomaly change result, anomaly main direction, anomaly maximum change rate, anomaly end state, and anomaly source combination result; the staged anomaly features are used in step S5 to perform consistency verification and transient anomaly screening based on the coupling response relationship between construction objects.

[0181] Step 5: Determine the coupling response relationship between construction objects based on the associated data of the construction objects, and perform consistency verification on the staged abnormal features based on the coupling response relationship between the construction objects, filter out transient abnormalities caused by the process switching stage, and generate real abnormal results.

[0182] In some implementations, the steps for generating realistic anomaly results include: Step 5-1: Read the construction object association data and the staged anomaly features, and generate a coupling relationship analysis unit. Specifically, read the construction object association data output in step S1 and the staged anomaly features output in step S4.

[0183] The construction object association data includes at least construction object identifier pairs, spatial adjacency relationship determination results, operation dependency relationship determination results, and collaborative effect relationship determination results; the staged anomaly features include at least construction object identifier, process switching stage identifier, segment identifier, status type, anomaly start time, anomaly end time, anomaly duration, anomaly change result, anomaly main direction, anomaly maximum change rate, anomaly end status, and anomaly source combination result.

[0184] Subsequently, based on the construction object identifier pairs and process switching stage identifiers, construction object identifier pairs that have spatial adjacency, operational dependency, or synergistic relationships are identified as coupling relationship analysis units, and the staged abnormal features corresponding to the construction object identifier pairs and process switching stage identifiers are assigned to the corresponding coupling relationship analysis units; the coupling relationship analysis units are used in the next step to determine the coupling response relationship.

[0185] Step 5-2: Determine the coupling response relationship between construction objects based on the associated data of the construction objects. Specifically, for each coupling relationship analysis unit, read the corresponding spatial adjacency relationship determination result, operation dependency relationship determination result, and synergy relationship determination result.

[0186] If the determination result of the spatial adjacency relationship between the construction object identifier pairs is true, then it is determined that there is a spatial propagation type coupling response relationship between the construction object identifier pairs; if the determination result of the operation dependency relationship between the construction object identifier pairs is true, then it is determined that there is an operation undertaking type coupling response relationship between the construction object identifier pairs; if the determination result of the synergistic relationship between the construction object identifier pairs is true, then it is determined that there is a synergistic linkage type coupling response relationship between the construction object identifier pairs.

[0187] If the same construction object identifier has multiple relationships as mentioned above, then all relationships will be written into the same coupled response relationship.

[0188] Subsequently, based on the current combination of construction object categories, the current process switching type, and the current coupling relationship type, the corresponding response status type, response direction requirements, response time interval, and response release interval are read.

[0189] The response status type, response direction requirement, response time interval, and response release interval are formed based on: the linkage change record of similar construction object combinations in the historical process of switching similar procedures, the provisions of the succession relationship between previous and subsequent operations in the construction plan, the equipment action execution requirements, and on-site safety operation constraints; when invoked, the corresponding results are read according to the current construction object category combination, the current process switching type, and the current coupling relationship type.

[0190] After completing the aforementioned determination, the construction object identifier is used as the associated unit, and the coupling relationship type is used as the classification basis to organize the response state type, response direction requirements, response time interval, and response release interval accordingly.

[0191] Specifically, when the current construction object identifier pair corresponds to only one type of coupling relationship, the coupling relationship type and its corresponding response state type, response direction requirement, response time interval and response release interval are directly combined and organized to generate the coupling response relationship corresponding to the construction object identifier pair; When the current construction object identifier corresponds to two or more of the following: spatial propagation type coupling response relationship, operation acceptance type coupling response relationship, and collaborative linkage type coupling response relationship, the response state type, response direction requirement, response time interval, and response release interval corresponding to each coupling relationship type are read respectively, and then merged into the corresponding coupling response relationship of the same construction object identifier according to the coupling relationship type.

[0192] When there are multiple candidate response time intervals or multiple candidate response release intervals for the same construction object identifier under the same coupling relationship type, the candidate interval that simultaneously satisfies the historical linkage change records and on-site operation constraint information is retained; when there are multiple candidate intervals that simultaneously satisfy the conditions, the candidate interval with the largest number of corresponding historical samples is selected.

[0193] The coupled response relationship includes at least the construction object identification pair, the coupling relationship type, the response state type, the response direction requirement, the response time interval, and the response release interval; the coupled response relationship is used by the next step to generate the verification subunit and perform the consistency verification.

[0194] Step 5-3: Generate a verification subunit based on the coupling response relationship and the staged anomaly features. Specifically, read the coupling response relationship corresponding to the current coupling relationship analysis unit and read the staged anomaly features classified into the current coupling relationship analysis unit. Subsequently, use the construction object identifier pair, process switching stage identifier, segment identifier, and state type as classification criteria to filter the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction objects.

[0195] Specifically, firstly, a set of candidate current anomalies that are consistent with the current process switching stage identifier, current segment identifier, and current state type are selected from the phased anomalies corresponding to the current construction object; then, a set of candidate associated anomalies that are consistent with the current process switching stage identifier, current segment identifier, and current state type are selected from the phased anomalies corresponding to the associated construction object. When both the current abnormal feature candidate set and the associated abnormal feature candidate set contain only one staged abnormal feature, the two are directly paired and organized into a verification sub-unit. When there are multiple staged abnormal features in the current abnormal feature candidate set or the associated abnormal feature candidate set, the time difference between the start time of the abnormal occurrence of each candidate staged abnormal feature is calculated respectively, and the candidate staged abnormal feature whose time difference falls into the response time interval corresponding to the coupling response relationship and has the smallest time difference is determined as the pairing object. Then, the determined current staged abnormal feature and associated staged abnormal feature are paired and organized into a verification subunit. The verification subunit includes at least the current construction object identifier, associated construction object identifier, process switching stage identifier, segment identifier, status type, staged abnormal features corresponding to the current construction object, and staged abnormal features corresponding to the associated construction object.

[0196] When there are no candidate staged anomalies in the candidate set of associated anomalies whose time difference falls within the response time interval, the staged anomaly corresponding to the current construction object is retained separately and identified as an unpaired anomaly.

[0197] Step 5-4: Perform consistency verification on the staged anomaly features in the verification subunit based on the coupling response relationship, and generate transient anomaly judgment results and retained anomaly judgment results. Specifically, for each verification subunit, read the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction object, and perform the following verification: First, perform a state type consistency check; read the state type corresponding to the current staged anomaly feature and compare it with the response state type in the coupled response relationship; if the state type corresponding to the current staged anomaly feature belongs to the response state type, then the state type consistency check result is determined to be valid; otherwise, it is determined to be invalid.

[0198] Secondly, perform a direction consistency check; read the main direction of the anomaly corresponding to the current staged anomaly feature and compare it with the response direction requirement in the coupled response relationship; if the two are consistent, the direction consistency check result is determined to be valid; otherwise, it is determined to be invalid.

[0199] Next, perform a time consistency check; read the start time of the anomaly corresponding to the current staged anomaly feature and the start time of the anomaly corresponding to the staged anomaly feature of the associated construction object, and subtract the two to obtain the response time difference; then compare the response time difference with the response time interval in the coupled response relationship; if the response time difference falls into the response time interval, the time consistency check result is determined to be valid; otherwise, it is determined to be invalid.

[0200] Finally, a consistency check is performed; the end time of the anomaly occurrence corresponding to the current staged anomaly feature and the end time of the anomaly occurrence corresponding to the staged anomaly feature of the associated construction object are read, and the difference between the two is used to obtain the anomaly removal time difference; then the anomaly removal time difference is compared with the response removal interval in the coupled response relationship; if the anomaly removal time difference falls into the response removal interval, the consistency check result is determined to be valid; otherwise, it is determined to be invalid.

[0201] After completing the aforementioned verifications, when the consistency verification results of state type, direction, time, and deconsistency verification are all valid, the current stage-specific anomaly characteristics are determined to be transient anomalies caused by the process switching stage, and a transient anomaly judgment result is generated. If at least one of the above verification results is not true, the current staged abnormal feature is determined as an abnormal feature that has failed the coupling consistency verification, and a retained abnormal judgment result is generated; the retained abnormal judgment result is used for the next step to generate the real abnormal result.

[0202] Step 5-5: Based on the retained anomaly determination results, filter out transient anomalies and generate true anomaly results. Specifically, read the transient anomaly determination results and retained anomaly determination results output in Step 5-4; first, remove the staged anomaly features corresponding to the transient anomaly determination results from the anomaly feature set of the current process switching stage; then, classify and organize the staged anomaly features corresponding to the retained anomaly determination results according to the construction object identifier, process switching stage identifier, segment identifier, and status type.

[0203] Subsequently, for each retained anomaly judgment result, its corresponding construction object identifier, process switching stage identifier, segment identifier, status type, anomaly start time, anomaly end time, anomaly duration, anomaly change result, anomaly main direction, anomaly maximum change rate, anomaly termination status, and anomaly source combination result are read, and the aforementioned contents are combined and organized to generate the real anomaly result.

[0204] When the same construction object corresponds to multiple real abnormal results in the same process switching stage, they are arranged according to the order of segmentation and state type to generate a set of real abnormal results of the current construction object in the current process switching stage; the real abnormal results are used in step 6 to map the construction object to the digital twin scene and determine the linkage monitoring object and linkage action sequence.

[0205] This step determines the coupling response relationship between construction objects based on the associated data of the construction objects, and performs consistency checks on the staged abnormal features in four aspects: state type, direction, time, and resolution based on the verification subunit. It filters out transient abnormalities corresponding to linkage changes that occur synchronously in multiple associated construction objects during the process switching stage and are resolved within a preset time range, and retains abnormal changes that do not occur according to the coupling linkage rules, have abnormal duration, or are not resolved within the specified time range and generates real abnormal results, thereby improving the pertinence and accuracy of subsequent linkage monitoring object determination and linkage action output.

[0206] To illustrate the generation of the verification subunit and the process of filtering out transient anomalies in this embodiment, a scenario in which there is a collaborative relationship between the hoisting equipment and the constructed component is used as an example.

[0207] For the same segment and state type within the same process switching phase, firstly, select the current abnormal feature candidate set from the phased abnormal features corresponding to the hoisting equipment, and then select the associated abnormal feature candidate set from the phased abnormal features corresponding to the constructed component. If both the current abnormal feature candidate set and the associated abnormal feature candidate set contain only one phased abnormal feature, then directly combine them into a single verification sub-unit. If any candidate set contains multiple phased abnormal features, calculate the time difference between the abnormal occurrence start times of each candidate phased abnormal feature, and determine the candidate phased abnormal feature with the smallest time difference falling within the response time interval corresponding to the coupled response relationship as the pairing object. Then, integrate the paired phased abnormal features... This is considered a verification subunit. Further, for this verification subunit, if the staged anomaly characteristics corresponding to the current construction object are consistent with the response state type, response direction requirements, response time interval, and response release interval in the coupled response relationship in terms of state type, main direction of the anomaly, time difference of the anomaly occurrence start time, and time difference of the anomaly release, then the staged anomaly characteristic is determined to be a transient anomaly caused by the process switching stage. If at least one of the aforementioned four items is inconsistent, then the staged anomaly characteristic is determined to be a retained anomaly and used for subsequent generation of real anomaly results. This example illustrates that not all anomaly changes occurring during the process switching stage are directly retained; instead, they are first reviewed in conjunction with the coupled response relationship between the associated construction objects.

[0208] Step 6: Map the real anomaly results to the construction objects in the digital twin scene, and determine the linkage monitoring objects and corresponding linkage action sequences based on the mapping results, and output the linkage monitoring results.

[0209] In some implementation methods, the steps for outputting the linkage monitoring results include: Step 6-1: Read the actual anomaly results and object mapping information to generate a mapping result. Specifically, read the actual anomaly results and the pre-established object mapping information in the digital twin scene. The object mapping information is used to characterize the correspondence between the construction object identifier and the scene object identifier in the digital twin scene. The object mapping information is formed based on the object registration results, scene modeling results, and the correspondence registration results between the construction object identifier and the scene object identifier established when the construction object enters the digital twin scene. When called, the corresponding scene object identifier is read according to the construction object identifier in the actual anomaly results.

[0210] Subsequently, for each real anomaly result, the construction object identifier is read, and the scene object identifier corresponding to the construction object identifier is retrieved from the object mapping information; then, the construction object identifier, the scene object identifier, the process switching stage identifier, segment identifier, status type, and anomaly termination status in the real anomaly result are combined and organized to generate a mapping result; the mapping result is used by the next step to determine the linkage monitoring object.

[0211] Step 6-2: Determine the linked monitoring object based on the mapping result. Specifically, read the mapping result and read the scene correspondence relationship between the spatial adjacency object, the task undertaking object, and the collaborative operation object corresponding to the scene object identifier in the digital twin scene.

[0212] First, identify the monitoring objects of the anomaly center; read the scene object identifiers corresponding to each real anomaly result, and determine the scene object identifiers directly corresponding to each real anomaly result as the monitoring objects of the anomaly center.

[0213] When multiple real anomaly results correspond to the same scene object identifier, duplicate scene object identifiers are deduplicated, and only one anomaly center monitoring object is retained; when the same real anomaly result corresponds to multiple scene object identifiers, multiple corresponding anomaly center monitoring objects are retained according to the one-to-one correspondence in the object mapping information; thus, an anomaly center monitoring object set is generated.

[0214] Secondly, identify the regional linkage monitoring objects; read the work area identifier, scene coordinates, area boundary, and spatial adjacent objects corresponding to each anomaly center monitoring object; wherein, the spatial adjacent objects are used to represent scene objects that are spatially adjacent to the anomaly center monitoring object in the digital twin scenario; then, determine the preset influence range corresponding to the anomaly center monitoring object based on the object category, current process switching type, and anomaly change results of the anomaly center monitoring object.

[0215] The preset influence range is formed based on: the action influence radius corresponding to the current object category, the safe operating distance of the equipment, the area division requirements of the construction site, and the spatial range of the impact of similar historical anomalies on surrounding objects; when invoked, the corresponding preset influence range is read according to the object category, process switching type, and anomaly change result corresponding to the current anomaly center monitoring object.

[0216] Next, the object distance between each spatially adjacent object and the corresponding anomaly center monitoring object is calculated, and it is determined whether each spatially adjacent object meets at least one of the following conditions: 1. Located within the work area of ​​the monitoring object at the anomaly center; II. Located within the adjacent work area of ​​the work area where the monitoring object of the anomaly center is located; 3. The distance between objects shall not exceed the preset influence range.

[0217] When a spatially adjacent object meets at least one of the aforementioned conditions, the spatially adjacent object is identified as a regional linkage monitoring object; thereby, a set of regional linkage monitoring objects is generated.

[0218] Further, the monitoring targets for joint operation are determined; the operation targets corresponding to each anomaly center monitoring target are read; wherein, the operation targets are used to represent the scenario objects that have a preceding process or a subsequent process relationship with the anomaly center monitoring target in the current process switching stage; then, the operation targets are filtered for relationship based on the preceding process identifier and the subsequent process identifier corresponding to the current process switching stage.

[0219] When the completion result of the preceding process corresponding to a certain task assignment object constitutes the execution prerequisite for the current process corresponding to the abnormality center monitoring object, or when the completion result of the current process corresponding to the abnormality center monitoring object constitutes the execution prerequisite for the subsequent process corresponding to the task assignment object, the task assignment object is identified as the task assignment linkage monitoring object; thereby, a set of task assignment linkage monitoring objects is generated.

[0220] Further, the collaborative monitoring objects are identified; the collaborative operation objects corresponding to each anomaly center monitoring object are read; wherein, the collaborative operation objects are used to represent scene objects that have synchronous operation relationships, action coordination relationships, position coordination relationships or load coordination relationships with the anomaly center monitoring objects in the current construction task; then, the collaborative relationship filtering is performed on the collaborative operation objects according to the current construction task identifier, the current process switching stage identifier and the collaborative operation relationship type.

[0221] When a collaborative operation object and an anomaly center monitoring object are both in the same execution state in the same construction task, and there is at least one relationship between them in terms of action coordination, position coordination, or load coordination, the collaborative operation object is identified as a collaborative linkage monitoring object; thereby, a set of collaborative linkage monitoring objects is generated.

[0222] After completing the aforementioned determination, the sets of monitoring objects at the anomaly center, the sets of regional linkage monitoring objects, the sets of receiving linkage monitoring objects, and the sets of collaborative linkage monitoring objects are deduplicated and reorganized to generate linkage monitoring objects.

[0223] Specifically, when the same scene object belongs to multiple object types, including anomaly center monitoring object, regional linkage monitoring object, receiving linkage monitoring object, and collaborative linkage monitoring object, the scene object is deduplicated and retained, and its various linkage sources are written together into the linkage source tag corresponding to the scene object; the linkage source tag includes at least one of anomaly center source tag, regional linkage source tag, receiving linkage source tag, and collaborative linkage source tag; then, the deduplicated scene object identifier and its corresponding linkage source tag are combined and organized to generate a linkage monitoring object; the linkage monitoring object includes at least: scene object identifier, linkage source tag, corresponding anomaly center object identifier, and the work area identifier to which the object belongs; the linkage monitoring object is used in the next step to determine the linkage action sequence.

[0224] Step 6-3: Determine the linkage action sequence based on the actual anomaly results, the linkage monitoring objects, and the linkage action rules. Specifically, read the actual anomaly results, the linkage monitoring objects, and the pre-established linkage action rules. The linkage action rules are used to characterize the linkage action content, action execution objects, and action execution order corresponding to different state types, different combinations of anomaly sources, different anomaly change amounts, different construction object categories, and different linkage source markers. The linkage action rules are generated according to the correspondence between state types, anomaly source combinations, anomaly change amounts, construction object categories, and linkage source markers.

[0225] Specifically, based on the construction site monitoring requirements, first determine the scene marking action corresponding to the anomaly center source marker, the range display action corresponding to the regional linkage source marker, the acceptance prompt action corresponding to the acceptance linkage source marker, and the collaborative alarm action corresponding to the collaborative linkage source marker. Based on the abnormal handling requirements, determine the abnormal display actions, abnormal alarm actions, and restriction prompt actions corresponding to the position status type, attitude status type, load status type, action status type, and area status type, respectively. Then, based on the equipment's motion restriction requirements and safe operation constraints, determine the motion execution objects and preset motion execution order corresponding to each linkage action; Then, the status type, abnormal source combination result, abnormal change result, construction object category, and linkage source marker are matched with the corresponding linkage action content, action execution object, and action execution order to generate the linkage action rule; when called, the corresponding linkage action rule is read according to the status type, abnormal source combination result, abnormal change result, construction object category, and linkage source marker in the linkage monitoring object in the actual abnormal result.

[0226] Subsequently, for each real abnormal result and its corresponding linked monitoring object, the linked action content was determined.

[0227] First, determine the source-related actions based on the linked source markers.

[0228] When the monitored object corresponds to the source marker of the anomaly center, the scene marking action corresponding to the anomaly center object is determined; the scene marking action includes at least one of highlighting, color changing, bounding box display, or anomaly status label display on the anomaly center object.

[0229] When the linked monitoring object corresponds to the region linkage source mark, the range display action corresponding to the region linkage object is determined; the range display action includes at least one of the following: displaying the region boundary, displaying the impact path, or displaying the adjacent object mark for the abnormal impact range.

[0230] When the monitored object corresponds to the source marker of the linkage, the corresponding linkage prompt action is determined; the linkage prompt action includes at least one of the following: displaying the linkage relationship, prompting the linkage status, or prompting the process connection risk to the preceding or subsequent linkage object.

[0231] When the monitored object corresponds to the collaborative linkage source mark, the collaborative alarm action corresponding to the collaborative linkage object is determined; the collaborative alarm action includes at least one of the following: displaying the collaborative relationship, prompting abnormal collaborative operation, or issuing a collaborative risk alarm to the collaborative object.

[0232] Furthermore, based on the status type of the actual abnormal result, determine the corresponding abnormal display action, abnormal alarm action, and restriction prompt action.

[0233] When the actual abnormal result corresponds to the location status type, determine the location abnormality display action and the location abnormality alarm action; wherein, the location abnormality display action includes at least one of displaying the abnormal location offset direction, the abnormal location offset amount, and the abnormal impact range, and the location abnormality alarm action includes at least one of outputting location abnormality prompt information or regional risk prompt information.

[0234] When a real abnormal result corresponds to an attitude state type, an attitude abnormality display action and an attitude abnormality alarm action are determined; wherein, the attitude abnormality display action includes at least one of displaying abnormal attitude direction, abnormal attitude angle change amount and abnormal attitude state, and the attitude abnormality alarm action includes at least one of outputting attitude abnormality prompt information or attitude risk prompt information.

[0235] When the actual abnormal result corresponds to the load status type, determine the load abnormality display action and the load abnormality alarm action; wherein, the load abnormality display action includes at least one of displaying the abnormal load change amount, the abnormal load change direction and the affected collaborative object, and the load abnormality alarm action includes at least one of outputting load abnormality prompt information or load risk prompt information.

[0236] When a real abnormal result corresponds to an action state type, determine the abnormal action display action and the restriction prompt action; wherein, the abnormal action display action includes at least one of displaying the abnormal action switching result, the abnormal action persistence state, or the abnormal action path, and the restriction prompt action includes at least one of action restriction prompt action, pause prompt action, or manual review prompt action.

[0237] When a real anomaly result corresponds to a region status type, determine the region anomaly display action and the region anomaly alarm action; wherein, the region anomaly display action includes at least one of displaying an abnormal region switching path, an abnormal entry region, or an abnormal stay region, and the region anomaly alarm action includes at least one of outputting a region anomaly prompt message or a region intrusion risk prompt message.

[0238] After determining the content of each of the aforementioned linked actions, the linked actions are sorted according to the preset action execution order.

[0239] The preset action execution order is used to characterize the sequential execution order of various linkage actions among multiple linkage actions corresponding to the same real abnormal result. The preset action execution order is formed based on the processing requirements of the construction site abnormal monitoring process: "first locate the abnormal object, then display the scope of influence, then output the abnormal alarm, and finally output the handling prompt", as well as equipment action restriction requirements and safe operation constraints. The calling method is: read the corresponding action execution order according to the status type and linkage source mark corresponding to the current real abnormal result.

[0240] Specifically, the scene marking action corresponding to the exception center object is executed first; Then execute the range display action corresponding to the area linkage object, the acceptance prompt action corresponding to the receiving linkage object, and the collaborative alarm action corresponding to the collaborative linkage object; Then, execute the exception display action and the exception alarm action; Finally, a restriction prompt action is executed; wherein, the restriction prompt action includes an action restriction prompt action, a pause prompt action, or a manual review prompt action.

[0241] When there are duplicate actions in the linked action content corresponding to multiple real abnormal results, duplicate actions with the same execution object and the same action content are deduplicated and retained.

[0242] When there are conflicting actions among the linked actions corresponding to multiple real abnormal results, the conflicting actions are filtered out according to the pre-set action conflict handling rules.

[0243] The action conflict handling rules are generated by establishing conflict correspondences based on the combination of linked actions with the same execution object but different action content.

[0244] Specifically, first select the linked action combinations that have the same execution object and overlapping execution sequence as candidate conflict action combinations; Based on the safety requirements at the construction site, it was determined that when both pause prompts and action restriction prompts exist in the candidate conflict action combination, the pause prompt action should be retained; when both action restriction prompts and display actions exist, the action restriction prompt action should be retained; and when both manual review prompts and repetition prompts exist, the manual review prompt action should be retained. Based on the equipment control restrictions, determine the combination of actions that cannot be executed simultaneously on the same execution object, and identify the actions that cause the execution object to enter a stopped state, a restricted state, or a verification state as reserved actions; For multiple display actions that only show different content but do not affect device control, they are identified as merged display actions. Then, the candidate conflict action combinations are organized with the corresponding retained actions, merged actions, and deleted actions to generate the action conflict handling rules. When calling, when generating the linkage action sequence, multiple linkage actions with the same execution object but inconsistent action content are screened for conflict, and the corresponding conflict handling results are retained.

[0245] Specifically, when multiple linked actions contain both restriction prompts and display actions, the restriction prompts are retained, and the corresponding display actions are merged into the same prompt result; when multiple linked actions contain multiple display actions, display actions with the same content are merged and displayed.

[0246] After completing the aforementioned sorting, deduplication, and conflict filtering, the action execution object, action content, action source marker, and action execution order corresponding to each linkage action are combined and organized to generate a linkage action sequence; the linkage action sequence includes at least: action execution object, action content, action source marker, and action execution order; the linkage action sequence is used to output linkage monitoring results in the next step.

[0247] Step 6-4: Execute the linked action sequence and output the linked monitoring results. Specifically, read the linked action sequence output in step 6-3, and read the action execution object, action content and action source marker corresponding to each linked action in the order of action execution.

[0248] Specifically, for each linked action in the linked action sequence, the corresponding linked action processing is executed.

[0249] If the current linked action corresponds to a scene-marked action, then the action execution object corresponding to the current linked action is read, and at least one of the following is performed on the action execution object in the digital twin scene: highlighting, color change display, bounding box display, or abnormal status label display, to generate a scene marking result; wherein, the scene marking result is used to characterize the abnormal location result of the abnormal center object in the digital twin scene.

[0250] If the current linkage action corresponds to the range display action, then the action execution object corresponding to the current linkage action is read, and at least one of the following is performed on the action execution object in the digital twin scene: region boundary display, influence path display, adjacent object mark display, or abnormal influence range filling display, to generate a range display result; wherein, the range display result is used to characterize the scene display result of the abnormal center object affecting the surrounding area linkage objects, receiving linkage objects, or collaborative linkage objects.

[0251] If the current linkage action corresponds to a receiving prompt action, then the action execution object corresponding to the current linkage action is read, and at least one of the following is performed on the action execution object in the digital twin scenario: display of receiving relationship connection, receiving status prompt, or process connection risk prompt, to generate a receiving prompt result; wherein, the receiving prompt result is used to characterize the process receiving risk result between the abnormal center object and the preceding receiving object or the subsequent receiving object.

[0252] If the current linkage action corresponds to a collaborative alarm action, then the action execution object corresponding to the current linkage action is read, and at least one of the following is performed on the action execution object in the digital twin scenario: display of collaborative relationship, collaborative operation anomaly prompt, or collaborative risk alarm, to generate a collaborative alarm result; wherein, the collaborative alarm result is used to characterize the collaborative linkage risk result between the anomaly center object and the collaborative object.

[0253] If the current linked action corresponds to an abnormal display action, then the action execution object and action content corresponding to the current linked action are read, and according to the state type corresponding to the real abnormal result, at least one of abnormal direction display, abnormal change amount display, abnormal path display, or abnormal state display is executed in the digital twin scene to generate an abnormal display result; wherein, the abnormal display result is used to characterize the visualization display result of the real abnormal result in the digital twin scene.

[0254] If the current linkage action corresponds to an abnormal alarm action, then the action execution object and action content corresponding to the current linkage action are read, and the abnormal type, abnormal object, abnormal segment, abnormal change result, abnormal source combination result and corresponding alarm prompt content are output in the monitoring interface or field terminal to generate an abnormal alarm result; wherein, the abnormal alarm result is used to characterize the alarm output result corresponding to the actual abnormal result.

[0255] If the current linkage action corresponds to a restriction prompt action, then the action execution object and action content corresponding to the current linkage action are read, and at least one of the following is output: action restriction prompt, pause prompt, or manual review prompt, generating a restriction prompt result; wherein, the restriction prompt result is used to characterize the subsequent handling prompt result output for the real abnormal result.

[0256] After completing the aforementioned linkage actions, the scene marking results, range display results, acceptance prompt results, collaborative alarm results, anomaly display results, anomaly alarm results, and restriction prompt results are organized according to the action execution object and action source marking to generate linkage monitoring results; wherein, the linkage monitoring results include at least: anomaly center object scene display results, linkage object range display results, process acceptance prompt results, collaborative alarm results, anomaly display results, anomaly alarm results, and restriction prompt results.

[0257] When the same action execution object corresponds to multiple linked action execution results, the multiple linked action execution results are merged and organized according to the action execution order; when the display content of multiple linked action execution results is duplicated, the duplicate content is deduplicated and retained; when multiple linked action execution results correspond to scene display, alarm output and restriction prompt respectively, their respective contents are retained and written together into the linked monitoring result; the linked monitoring result is used to characterize the abnormal object display result, linked object display result, alarm output result and handling prompt result of the real abnormal result in the digital twin scenario.

[0258] As another example, in order to illustrate the process of determining the linkage monitoring object and linkage action sequence in this embodiment, it is assumed that a certain real abnormal result corresponds to an abnormal attitude state type of the hoisting equipment in the installation operation area.

[0259] First, based on the object mapping information, the construction object identifier corresponding to the hoisting equipment is mapped to the scene object identifier in the digital twin scene, and this is identified as the anomaly center monitoring object. Then, the spatially adjacent objects, work-receiving objects, and collaborative work objects corresponding to this scene object are read, and the preset influence range is determined by combining the object category corresponding to the hoisting equipment, the current process switching type, and the anomaly change results. Scene objects located in the same work area, adjacent work areas, or with an object distance not exceeding the preset influence range are identified as area-linked monitoring objects. Scene objects with preceding and following process relationships with the hoisting equipment are identified as receiving-linked monitoring objects. Scene objects with action coordination, position coordination, or load coordination relationships with the hoisting equipment in the same construction task are identified as... The objects to be monitored in a collaborative manner are identified. Then, based on the state type corresponding to the actual anomaly and the linkage source markers corresponding to each monitored object, scene marker actions, range display actions, receiving prompt actions, collaborative alarm actions, posture anomaly display actions, posture anomaly alarm actions, and manual review prompt actions are determined and sorted according to a preset action execution order. When multiple linkage actions are duplicated, duplicate actions are removed and retained. When multiple linkage actions conflict, conflict filtering is performed according to action conflict handling rules to generate a linkage action sequence. This example illustrates that after the actual anomaly is mapped into the digital twin scene, it not only corresponds to the display and processing of the anomaly object itself but also further drives the linkage monitoring and processing of surrounding objects, receiving objects, and collaborative objects. Example 2

[0260] See Figure 3 As shown, this embodiment provides a digital twin on-site perception and linkage monitoring system for engineering construction. Since this system uses a digital twin on-site perception and linkage monitoring method for engineering construction from Embodiment 1, it has the same effect, which will not be repeated here. The system includes: The data acquisition module is used to acquire multi-source status data, current process execution data, and construction object related data from the construction site. The process switching identification module is used to identify the process switching stage based on the current process execution data, and determine the state reconstruction period corresponding to the process switching stage; The temporal feature extraction module is used to perform segmented analysis on the multi-source state data within the state reconstruction period corresponding to the process switching stage, and extract temporal change features. The stage anomaly identification module is used to classify the time-series change features according to the preset stage response rules corresponding to the process switching stage, determine the switching response features and continuous anomaly features corresponding to the process switching stage, and generate staged anomaly features. The coupling verification module is used to determine the coupling response relationship between construction objects based on the associated data of the construction objects, and to perform consistency verification on the staged abnormal features based on the coupling response relationship between the construction objects, thereby filtering out transient abnormalities caused by the process switching stage and generating real abnormal results. The linkage monitoring module is used to map the real abnormal results to the construction objects in the digital twin scene, and determine the linkage monitoring objects and corresponding linkage action sequences based on the mapping results, and output the linkage monitoring results.

[0261] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope defined in the claims.

Claims

1. A digital twin-based on-site sensing and linkage monitoring method for engineering construction, characterized in that, include: Acquire multi-source status data of the construction site, current process execution data, and construction object related data; Identify the process switching phase based on the current process execution data, and determine the state reconstruction period corresponding to the process switching phase; The multi-source state data within the state reconstruction period corresponding to the process switching stage is segmented and analyzed to extract temporal change features. Based on the preset stage response rules corresponding to the process switching stage, the time-series change features are classified to determine the switching response features and continuous abnormal features corresponding to the process switching stage, and staged abnormal features are generated. The coupling response relationship between construction objects is determined based on the associated data of the construction objects, and the consistency of the staged anomaly features is verified based on the coupling response relationship between the construction objects to screen out transient anomalies caused by the process switching stage and generate real anomaly results. The real anomaly results are mapped to construction objects in the digital twin scene, and the linked monitoring objects and corresponding linkage action sequences are determined based on the mapping results, and the linked monitoring results are output.

2. The digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1, characterized in that, The method for determining the state reconstruction period corresponding to the process switching stage includes: Based on the current process execution data, a process switching candidate time period is determined. When the process switching candidate time period meets the following conditions: process identifier change or current execution action change is established, equipment action code sequence has continuous action switching, work area identifier sequence has area entry, area exit or cross-area movement, and the duration of the candidate time period is not less than the shortest confirmation time, the process switching candidate time period is determined as the process switching stage. The state reconstruction period is determined based on the start and end times of the process switching phase, combined with the preceding and following time ranges.

3. The digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1, characterized in that, Methods for extracting temporal variation features include: Read multi-source state data within the state reconstruction period to form a state time series; Generate a sequence of state changes based on the state time series corresponding to adjacent settling times. Based on the state change sequence, determine the pre-switching preparation segment, the switching execution segment, and the post-switching stabilization segment, and generate a segmented state sequence; Based on the segmented state sequence corresponding to each segment, extract the temporal change features corresponding to the position state type, attitude state type, load state type, action state type, and region state type.

4. The digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1, characterized in that, Methods for generating staged anomaly features include: Based on historical records of similar process switching, process conversion requirements in the construction plan, equipment action execution requirements, and on-site operation constraints, a preset stage response pattern is generated for the process switching stage. Based on the preset stage response pattern, the time-series change features are compared item by item to generate feature comparison results; Based on the feature comparison results, the switching response features and persistent anomaly features are determined, and staged anomaly features are generated based on the persistent anomaly features.

5. The digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1, characterized in that, The method for determining the coupling response relationship between construction objects based on the associated data of the construction objects includes: Based on the spatial adjacency determination results, operation dependency determination results, and synergy determination results in the construction object association data, determine the coupling relationship type between construction objects; Then, based on the current combination of construction object categories, the current process switching type, and the current coupling relationship type, determine the response state type, response direction requirements, response time interval, and response release interval to generate the coupling response relationship.

6. A digital twin-based on-site perception and linkage monitoring method for engineering construction according to any one of claims 1-5, characterized in that, The method for generating verification sub-units based on the coupling response relationship and the staged anomaly features includes: Filter from the phased anomaly features corresponding to the current construction object and the phased anomaly features corresponding to related construction objects; When the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction object conform to the response time interval corresponding to the coupling response relationship, the staged anomaly features corresponding to the current construction object and the staged anomaly features corresponding to the associated construction object are paired, and the paired staged anomaly features are organized into verification sub-units.

7. A digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 6, characterized in that, Methods for performing consistency checks include: For the staged anomaly characteristics in the verification subunit, perform state type consistency verification, direction consistency verification, time consistency verification, and deconsistency verification. When all consistency checks are successful, the corresponding staged anomaly characteristics are identified as transient anomalies caused by the process switching stage. If any consistency check fails, the corresponding staged anomaly characteristic will be identified as a retained anomaly.

8. The digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1, characterized in that, Methods for determining the objects to be monitored in a coordinated manner include: The monitoring objects of the anomaly center are determined based on the mapping results; Based on the work area identifier, scene coordinates, area boundary, and preset impact range corresponding to the monitoring object in the anomaly center, determine the area linkage monitoring object; Based on the task receiving object corresponding to the monitoring object in the anomaly center, determine the task receiving and linkage monitoring object; Based on the collaborative operation objects corresponding to the monitoring objects in the anomaly center, determine the collaborative monitoring objects; The abnormality center monitoring objects, regional linkage monitoring objects, receiving linkage monitoring objects, and collaborative linkage monitoring objects are deduplicated and sorted to generate the linkage monitoring objects.

9. A digital twin-based on-site perception and linkage monitoring method for engineering construction according to claim 1 or 8, characterized in that, Methods for determining the sequence of linked actions include: Based on the actual abnormal results and the monitored objects, determine the linkage actions; sort the linkage actions, remove duplicate actions and retain duplicates, and filter conflicting actions to generate the linkage action sequence.

10. A digital twin-based on-site perception and linkage monitoring system for engineering construction, characterized in that, include: The data acquisition module is used to acquire multi-source status data, current process execution data, and construction object related data from the construction site. The process switching identification module is used to identify the process switching stage based on the current process execution data, and determine the state reconstruction period corresponding to the process switching stage; The temporal feature extraction module is used to perform segmented analysis on the multi-source state data within the state reconstruction period corresponding to the process switching stage, and extract temporal change features. The stage anomaly identification module is used to classify the time-series change features according to the preset stage response rules corresponding to the process switching stage, determine the switching response features and continuous anomaly features corresponding to the process switching stage, and generate staged anomaly features. The coupling verification module is used to determine the coupling response relationship between construction objects based on the associated data of the construction objects, and to perform consistency verification on the staged abnormal features based on the coupling response relationship between the construction objects, thereby filtering out transient abnormalities caused by the process switching stage and generating real abnormal results. The linkage monitoring module is used to map the real abnormal results to the construction objects in the digital twin scene, and determine the linkage monitoring objects and corresponding linkage action sequences based on the mapping results, and output the linkage monitoring results.

Images