Project progress quality space-time tracing system based on digital twin sand table

Abstract

The invention relates to the technical field of engineering management, in particular to an engineering progress quality space-time tracing system based on a digital twin sand table, which comprises a behavior factor extraction module for acquiring equipment operation records, process instruction streams, personnel tracks and quality detection data; constructing a construction behavior factor including an operation type, a trigger source identifier, an execution node and an execution duration; a behavior driving chain reconstruction module uses the behavior factors as nodes, identifies causal paths between adjacent or cross-processes, constructs a behavior driving chain, and maps the behavior driving chain with a construction quality result to generate a behavior quality association chain; the space-time twinborn projection and inversion module maps the association chain to a digital twinborn sand table, generates a tracing view based on a time axis and space coordinates, and supports inversion of an operation chain and responsibility factor information forward from a quality abnormal point. According to the method, high-precision, verifiable and responsibility-traceable time-space visual tracing of the abnormal quality events in the whole construction process is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering management, and particularly relates to an engineering progress and quality space-time tracing system based on a digital twin sand table. BACKGROUND

[0002] With the continuous expansion of the scale of engineering projects and the increasing complexity of the construction process, digital and fine management has gradually become a mainstream trend in the field of engineering construction. In order to improve the construction efficiency and quality control level, more and more large-scale engineering projects begin to introduce visual management systems based on BIM, GIS or digital twin technology to realize dynamic monitoring, data integration and information tracing of the construction site. Among them, the digital twin sand table is widely used in construction progress simulation, equipment scheduling and quality visualization display due to its space modeling and real-time data fusion capabilities, and has become one of the key technical means to support intelligent construction.

[0003] However, the existing construction management systems based on digital twins are mostly still in the static state labeling or result presentation stage, lacking the ability of structured modeling of construction behaviors, and unable to establish time logic or causal path from the behavior level. In the face of quality abnormalities, progress delays and other problems, the system can only provide result position and state display, and cannot automatically trace the relevant behavior chain and locate the trigger source, resulting in difficulty in determining responsibility, relying on manual experience for problem analysis, and insufficient data credibility. In addition, the multi-source construction data (such as equipment action, process instruction, personnel trajectory and detection record) in the existing system lacks a unified time reference and causal modeling framework, making it difficult to form a verifiable behavior-quality correlation. SUMMARY

[0004] The present application provides an engineering progress and quality space-time tracing system based on a digital twin sand table, which realizes structured analysis, causal path identification and responsibility information visual tracing of multi-source construction data, establishes an interpretable and verifiable behavior-quality correlation chain taking construction behavior as the basic unit, and finally projects it to the digital twin sand table to realize the reconstruction of the whole process responsibility chain from the quality abnormality to the forward tracing, thereby improving the accuracy, efficiency and credibility of engineering event tracing.

[0005] An engineering progress and quality space-time tracing system based on a digital twin sand table, comprising an original behavior factor extraction module, a behavior driving chain reconstruction module and a space-time twin projection and inversion module, wherein; The behavior factor extraction module collects multi-source data of the construction site, including equipment operation records, process instruction flow, personnel position information and quality detection events, and constructs construction behavior factors at the engineering unit level according to time sequence, including operation type, trigger source identification, execution node and execution time information; The behavior-driven chain reconstruction module takes construction behavior factors as nodes, identifies adjacent or cross-process behavior trigger paths with causal relationships, constructs a chain behavior-driven model, and logically maps key nodes in the chain to corresponding construction quality results to generate a behavior-quality association chain; The space-time twin projection and inversion module maps the behavior-quality association chain into a digital twin sand table model, generates an interactive traceability view based on a time axis and spatial coordinates, and supports the forward inversion of dependent operation behavior chains and related responsibility factors based on any quality abnormal result point.

[0006] Optionally, the behavior factor extraction module includes: Multi-source data time alignment: uniform time format conversion is performed on the collected device operation records, process instruction flow, personnel position information, and quality detection events, and alignment processing is performed according to timestamps to generate a data stream set with consistent time sequence reference; Job event identification and grouping: according to the trigger conditions of device start, stop, and state switching in the device operation records, combined with the trajectory changes of personnel position information and the start and end identifiers of action segments in the process instruction flow, job events are identified from the time sequence data, and events are grouped according to construction units; Factor extraction and normalization: construction behavior factors are extracted from each job event, including operation type, trigger source identifier (such as instruction ID or task ID), execution node (such as construction unit number or device code), and operation start and end time, execution duration information is calculated and generated, and the extracted construction behavior factors are normalized.

[0007] Optionally, the multi-source data time alignment includes: Time format standardization: uniform format conversion is performed on the time fields in the multi-source data to convert them into a unified standard time format (such as ISO 8601 standard timestamp); Timestamp precision normalization: the precision of the timestamps in the multi-source data is unified to the target precision , and the original timestamps are normalized by discretization alignment; Uniform time sequence alignment and sorting: all multi-source data records with normalized timestamps are merged and sorted in chronological order to construct a unified time sequence data stream set .

[0008] Optionally, the job event identification and grouping includes: Device event interval extraction: based on the start event, stop event, and state switching event marked in the device operation records, device job intervals are identified in chronological order to form a device event segment set , wherein, is the start time of a device, Each interval represents an independent work segment corresponding to a stop time or state switching time; Personnel trajectory linkage analysis: the position information of the construction personnel is processed in time sequence, and in each equipment event segment interval , personnel trajectory points with a distance less than a threshold from the equipment position coordinates are identified to form a personnel linkage subset , satisfying , wherein, is the spatial coordinates of the equipment, is the position coordinates of the personnel at the time point ; Process action segment matching: according to the start and end times of the action segments recorded in the process instruction flow , time intersection judgment is performed with the equipment event interval , and if is satisfied, it is considered that the equipment event and the process action segment have a logical corresponding relationship; Construction event construction and grouping: events that satisfy the equipment-personnel-process ternary coupling relationship are identified as construction events, the corresponding construction unit number is extracted , and the events are grouped according to to form a construction unit-level construction event set .

[0009] Optionally, the factor extraction and normalization includes: Field analysis and mapping extraction: extracting construction behavior factors from the identified construction events, including operation type (equipment behavior category such as start, stop, switch, run), trigger source identification (record the control identification that triggers the event, such as task ID, scheduling instruction ID), execution node identification (indicating the position or object of the event, such as construction unit number or equipment code), operation start and end time (the start and end time of the event); Execution duration calculation: calculating the execution duration of the construction event based on the start and end time ; Normalization processing: standard format conversion and normalization processing are performed on the extracted construction behavior factors.

[0010] Optionally, the behavior-driven chain reconstruction module includes: Behavior factor graph construction: the extracted construction behavior factors are taken as graph nodes, and a directed initial graph is constructed according to the occurrence time sequence and execution node distribution , wherein, is the node set composed of all construction behavior factors, a trigger edge between events; Causal relationship identification: trigger edges of the directed initial graph Causal screening, retaining the edge set with logical driving characteristics , according to the time difference between the construction behavior factors, calculate the causal driving confidence, used for screening the behavior trigger edges that meet the causal conditions; Chain behavior driving path generation: based on the screened edge set , using depth-first search (DFS), recursively build behavior driving chain from source node , each path represents a sequence of construction behaviors that exist potential operational dependency relationship; Quality event mapping and label construction: map the quality abnormal events in the construction quality detection result set According to the time and space correspondence, map to one or more key nodes in the behavior driving chain , establish the association label structure of behavior factors Quality results, generate a set of behavior quality association chains , wherein, An effective behavior driving chain path, Quality detection results.

[0011] Optionally, the space-time twin projection and inversion module comprises: Behavior chain coordinate analysis and layer binding: for each construction behavior factor node in the behavior quality association chain , extract its execution node space coordinates And time stamp , and bind it to the corresponding construction unit three-dimensional structure in the digital twin sand table model , construct the time-space double-axis mapping relationship ; Time axis and space level visualization modeling: build an interactive time-space axis mechanism for the twin view, set the time window length And the space display granularity level , and dynamically activate the state change animation or color identification of the corresponding area and time period in the sand table model driven by the behavior chain time sequence, realize the visual expression of behavior-quality state in the twin model; Quality abnormal driving inversion path generation: when selecting a quality abnormal event , automatically locate its corresponding node in the behavior quality association chain , and perform depth backtracking along its forward causal path, extract its dependent link , and synchronously activate the historical state playback and responsibility factor display of the related construction unit in the sand table; Inverted chain visual rendering and interactive output: based on path Each node in the In the form of a connected line or time jump point in the twin sand table, a visual path of operation behavior chain is generated, and at the same time, the responsibility unit, behavior type and time length of each behavior node are synchronously displayed on the interface interaction layer, outputting an interactive traceability view .

[0012] Optionally, the timeline and spatial level visual modeling includes: Time window and spatial granularity setting: setting the global time window length And the spatial display granularity level , used to control the time span and construction unit resolution degree of behavior state change displayed in the twin sand table; Behavior sequence driven space-time mapping activation: based on the behavior quality association chain set All behavior nodes in the When its timestamp , activate its bound construction unit layer , and jointly encode its behavior type state And quality state , generate visual display code; Sand table view dynamic rendering output: all activated construction behavior nodes within the current time window , according to the generated visual display code, drive its bound construction unit layer Dynamic rendering, including color filling, flashing frequency or transparency change, and generating an interactive space-time animation layer evolving over time in the sand table view.

[0013] Optionally, the quality anomaly driven inversion path generation includes: Quality event node positioning: when the user selects a quality anomaly event in the twin sand table interface , automatically find the behavior node corresponding to the event in the behavior quality association chain , satisfy , wherein The recording time of the quality event is , and the timestamp of the node ; Causal path depth backtracking extraction: based on the causal edge set in the behavior driven graph From node Forward depth-first backtracking is performed to extract all the predecessor nodes that have an impact on it in logic, forming an inversion path ; ; ​Sand table layer activation and responsibility factor display: for each behavior node in the inversion path , extract its bound construction unit layer , and activate the historical operation type, historical execution time and historical quality state visualization display, while extracting the responsibility factor information recorded by the node, including the responsible unit, the operation personnel and the execution task ID, and displaying them in a superimposed manner on the sand table interactive interface.

[0014] Advantages of the present application: The present application can extract structured and standardized construction behavior factors from multi-source data of the construction site by constructing an original behavior factor extraction module, covering key fields such as operation type, execution node, execution time and trigger source, and normalizing based on a unified time axis and space structure, which solves the problems of construction data fragmentation, time desynchronization and difficulty in defining responsibility subjects in the prior art.

[0015] The present application, by constructing a behavior-driven chain reconstruction module, proposes a cause and effect confidence calculation method based on the time difference and trigger logic between behavior factors, effectively identifies the construction behavior dependency path between adjacent or cross-processes, and then constructs a chain behavior-driven model with logical traceability, and one-to-one maps with construction quality abnormal results to generate an explainable behavior-quality association chain, which can reverse the behavior sequence induced by the quality problem, and enhances the logical verifiability and responsibility traceability of the construction process.

[0016] The present application, by constructing a time-space twin projection and inversion module, dynamically maps the behavior-quality association chain to the digital twin sand table model by combining the timestamp and spatial coordinates of the behavior chain nodes, generates an interactive traceability view by setting the time window length and spatial granularity level, and automatically traces the forward causal path of the selected quality abnormal event by the user, activates the historical operation type, execution time and quality state of the corresponding construction unit layer, and superimposes the responsibility factor information for display, which significantly improves the intuitiveness, efficiency and credibility of engineering abnormal source tracing. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only illustrate the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Fig. 1 The system function module schematic diagram of the embodiment of the present application; Fig. 2 ​​A behavior-driven chain reconstruction module of an embodiment of the present application is shown in the schematic diagram. DETAILED DESCRIPTION

[0019] The present application will be described in detail below with reference to the accompanying drawings and specific embodiments. For some known technologies, other alternative ways can also be implemented by those skilled in the art; and the accompanying drawings are only for a more specific description of the embodiments, and are not intended to specifically limit the present application.

[0020] As shown in the schematic diagram, a digital-twin sand table-based engineering progress quality space-time tracing system includes an original behavior factor extraction module, a behavior-driven chain reconstruction module, and a space-time twin projection and inversion module, wherein: Figs. 1-2 The behavior factor extraction module collects multi-source data of the construction site, including equipment operation records, process instruction flow, personnel location information, and quality detection events, and constructs construction behavior factors at the engineering unit level in chronological order, including operation type, trigger source identification, execution node, and execution duration information; The behavior-driven chain reconstruction module takes the construction behavior factors as nodes, identifies behavior trigger paths with causal relationships between adjacent or cross-processes, constructs a chain behavior-driven model, and logically maps key nodes in the chain to corresponding construction quality results to generate a behavior-quality association chain. The space-time twin projection and inversion module maps the behavior-quality association chain to the digital-twin sand table model, generates an interactive tracing view based on the time axis and spatial coordinates, and supports the inversion of the dependent operation behavior chain and related responsibility factors based on any quality abnormal result point. The behavior factor extraction module includes:

[0021] Multi-source data time alignment: uniform time format conversion is performed on the collected equipment operation records, process instruction flow, personnel location information, and quality detection events, and alignment processing is performed according to the time stamp to generate a data stream set with consistent time sequence reference; Job event identification and grouping: according to the trigger conditions of equipment start, stop, and state switching in the equipment operation records, combined with the trajectory changes of personnel location information and the start and end identifiers of action segments in the process instruction flow, job event identification is performed on the time sequence data, and event grouping is performed according to the construction unit; Factor extraction and normalization: construction behavior factors are extracted from each job event, including operation type, trigger source identification (such as instruction ID or task ID), execution node (such as construction unit number or equipment code), and operation start and end time, execution duration information is calculated and generated, and the extracted construction behavior factors are normalized. Multi-source data time alignment includes:

[0022] ​Time format standardization: unify the format of time fields in multi-source data and convert them into a unified standard time format (such as ISO 8601 standard timestamp); Timestamp precision normalization: unify the timestamp precision in multi-source data to the target precision , using discrete alignment to normalize the original timestamp, represented as: ; Where, is the normalized timestamp, is the set unified time axis start time, is the timestamp of the original data record; Unified time sequence alignment and sorting: merge and sort all multi-source data records with normalized timestamps in chronological order to construct a unified time series data stream set .

[0023] Job event identification and grouping includes: Device event interval extraction: based on the start event, stop event and state switching event marked in the device operation record, identify the device job interval in chronological order to form a device event segment set , where, is the start time of the device, is the corresponding stop time or state switching time, each interval represents an independent job segment; Personnel trajectory linkage analysis: time series processing of construction personnel location information, in each device event segment interval , identify personnel trajectory points with a distance less than a threshold from the device location coordinates, form a personnel linkage subset , satisfying , where, is the spatial coordinates of the device, is the location coordinates of the personnel at time point ; Process action segment matching: according to the start and end time of the action segment recorded in the process instruction flow , time intersection judgment is performed with the device event interval , if is satisfied, it is considered that the device event and the process action segment have a logical corresponding relationship; Job event construction and grouping: identify the event segment that satisfies the device-personnel-process ternary coupling relationship as a job event, extract its corresponding construction unit number , and group the events according to to form a construction unit level job event set , where, are the sets of operation events of the th construction unit, respectively.

[0024] threshold is represented as: ; wherein, is the equipment entity space influence radius, which is preset according to the equipment type (such as tower crane, excavator, concrete pump), representing the influence range of the equipment operation on the periphery, is the average moving speed of personnel in the construction site, which is calculated based on historical trajectories, is the allowed maximum personnel-equipment time interaction tolerance, that is, the personnel is considered to be effective cooperation within the time range, is the equipment space weight coefficient, is the personnel dynamic compensation coefficient.

[0025] Factor extraction and normalization includes: Field analysis and mapping extraction: extracting construction behavior factors from the identified operation events, including operation type (equipment behavior category, such as start, stop, switch, run), trigger source identification (record the control identification that triggers the event, such as task ID, scheduling instruction ID), execution node identification (indicate the location or object of the event, such as construction unit number or equipment code), operation start and end time (the start and end time of the event); Execution duration calculation: calculating the execution duration of the operation event based on the start and end time , represented as: ; wherein, , are the start and end time stamps of the operation behavior, respectively; Normalization processing: standard format conversion and normalization processing of the extracted construction behavior factors, specifically including: Mapping the operation type to a predefined coding label, including start , stop , switch ; Hash encoding the trigger source identification and the execution node identification ; Standard interval scaling normalization processing is performed on the execution duration , represented as: ; wherein, is the normalized standardization execution duration, , are the minimum and maximum values of the operation duration in the historical operation data, respectively.

[0026] The behavior-driven chain reconstruction module comprises: Behavior factor graph construction: taking the extracted construction behavior factors as graph nodes, constructing a directed initial graph according to the time sequence of their occurrence and the distribution of execution nodes wherein, is a node set composed of all construction behavior factors, is a trigger edge existing between events; Causal relationship identification: performing causality screening on the trigger edges of the directed initial graph , retaining the edge set with logical driving characteristics , calculating the causal driving confidence according to the time difference between the construction behavior factors, which is used to screen the behavior trigger edges that meet the causal conditions, and is represented as: ; wherein, is the probability of constituting a causal drive, when the edge is retained, is a confidence threshold, , is a confidence weighting coefficient, , are the minimum and maximum time differences between the operation behavior factors in the historical data, respectively, , are the occurrence times of the nodes , is a preset minimum causal reaction time threshold, is a decay coefficient; Chain behavior-driven path generation: based on the screened edge set , using depth-first search (DFS) to recursively construct behavior-driven chains from the source node, each path represents a sequence of construction behaviors with potential operational dependency relationships, specifically including: (1) Source node identification and initialization: identifying the node without an incoming edge as the starting point of the path according to the construction behavior factor set in the behavior factor graph , i.e., satisfying , initializing , path set ; (2) DFS-based causal path recursive construction: starting from the source node , a depth-first search recursive procedure is performed to construct the behavior-driven chain , denoted as: ; wherein is the successor node set of the current node, if the node has no successor, i.e., the end of the path, the current path is added to the result set , all possible paths are traversed to extract all construction behavior sequences with potential causal dependency; (3) Path structured output: each valid path is structured as an ordered sequence, recording the operation type, execution node, execution time and trigger source information of each node; Quality event mapping and label construction: the quality abnormal events in the construction quality detection result set are mapped to one or more key nodes in the behavior-driven chain according to the time and space correspondence , establishing the associated label structure of the behavior factor quality result, generating a set of behavior quality association chains , wherein is an effective behavior-driven chain path, is the quality detection result.

[0027] The space-time twin projection and inversion module includes: Behavior chain coordinate analysis and layer binding: for each construction behavior factor node in the behavior quality association chain , its execution node spatial coordinates and time stamp are extracted and bound to the corresponding construction unit three-dimensional structure in the digital twin sandbox model, constructing a time-space two-axis mapping relationship ; Time axis and space level visualization modeling: an interactive time-space axis mechanism of the twin view is constructed, the time window length and the space display granularity level are set, and the corresponding area and time period state change animation or color identification in the sandbox model is dynamically activated driven by the behavior chain time sequence, realizing the visual expression of behavior-quality state in the twin model; Quality abnormality driven inversion path generation: when a quality abnormal event is selected, its corresponding node in the behavior quality association chain is automatically located and perform deep backtracking along its forward-causal path to extract its dependency chain and activate the historical state playback and responsibility factor display of the relevant construction unit in the sand table synchronously Inversion chain visual rendering and interactive output: based on path of each node in the path In the twin sand table, the visual path of the operation behavior chain is generated in the form of a connected line or a time jump point, and at the same time, the responsibility unit, behavior type, and duration of each behavior node are synchronously displayed on the interface interaction layer to output an interactive trace view .

[0028] The timeline and space level visual modeling includes: Time window and space granularity setting: set the global time window length and the space display granularity level , which are used to control the time span and construction unit resolution degree of behavior state change in the twin sand table, represented as: ; ; wherein, , are the start and end time stamps of the current visualization window, is the total number of construction units in the model, is the number of construction units participating in the behavior event in the current time window; Behavior sequence driven space-time mapping activation: based on the behavior quality association chain set All behavior nodes when their time stamps activate the construction unit layer bound to them and jointly encode their behavior type state and quality state to generate a visual display code, represented as: ; wherein, is the visual display code, is the state mapping function, is the behavior type, is the quality state (qualified, warning, unqualified), is the basic color coding value corresponding to the behavior type (such as start = 1, abnormal = 3), is the warning weight of the quality state (such as qualified = 0, warning = 1, unqualified = 2), , are the weight coefficients; Sand table view dynamic rendering output: all activated construction behavior nodes within the current time window According to the generated visualization display code, drive its bound construction unit layer Dynamic rendering, including color filling, flashing frequency or transparency change, and generating interactive space-time animation layer evolving over time in the sand table view.

[0029] Quality anomaly driven inversion path generation includes: Quality event node positioning: when the user selects a quality anomaly event in the twin sand table interface , automatically find the behavior node corresponding to the event in the behavior quality association chain , satisfying , where is the recording time of the quality event, is the timestamp of the node ; Causal path depth backtracking extraction: based on the causal edge set in the behavior driven graph , perform a depth-first backtracking from node , extract all the predecessor nodes that logically affect it, form the inversion path , represented as: ; Where is the path from to exists; Sand table layer activation and responsibility factor display: for each behavior node in the inversion path , extract its bound construction unit layer , and activate its historical operation type, historical execution time and historical quality state visualization display, while extracting the responsibility factor information recorded by the node , including the responsible unit, the operator, the execution task ID, to be displayed in the sand table interactive interface in an overlay manner. The present application covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present application. In order to make the public have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be fully understood without the description of these details to those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits, etc. are not described in detail. The present application covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present application. In order to make the public have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be fully understood without the description of these details to those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits, etc. are not described in detail.

[0030]

[0031] ​​The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A spatiotemporal tracing system for engineering progress quality based on digital twin sandbox, characterized by: It includes a behavior factor extraction module, a behavior drive chain reconstruction module, and a spatiotemporal twin projection and inversion module, among which; The behavior factor extraction module collects multi-source data from the construction site, including equipment operation records, process instruction flow, personnel location information, and quality inspection events, and constructs construction behavior factors at the engineering unit level in chronological order, including operation type, trigger source identification, execution node, and execution duration information; The behavior-driven chain reconstruction module uses construction behavior factors as nodes to identify behavior trigger paths with causal relationships between adjacent or cross-processes, builds a chain behavior-driven model, and logically maps key nodes in the chain with corresponding construction quality results to generate a behavior-quality association chain. The space-time twin projection and inversion module maps the behavior quality association chain into the digital twin sandbox model, generates an interactive traceability view based on the time axis and spatial coordinates, and supports forward inversion of the operation behavior chain and related responsibility factors based on any quality abnormality result point.

2. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 1 is characterized in that: The behavior factor extraction module includes: Multi-source data time alignment: The collected equipment operation records, process instruction flows, personnel location information, and quality inspection events are converted to a unified time format, aligned based on timestamps, and generated into a data stream set with a consistent timing benchmark. Operation event identification and grouping: Based on the trigger conditions of equipment start, stop, and state switching in the equipment operation record, combined with the changes in personnel location information trajectory and the start and end identifiers of the action segments in the process instruction flow, operation events are identified in the time series data and grouped by construction unit; Factor extraction and normalization: Extract construction behavior factors from each operation event, including operation type, trigger source identifier, execution node, and operation start and end time, calculate and generate corresponding execution duration information, and normalize the extracted construction behavior factors.

3. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 2 is characterized in that: The multi-source data time alignment includes: Time format standardization: convert the time fields in multi-source data into a unified standard time format; Timestamp accuracy normalization: unify the timestamp accuracy in multi-source data to the target accuracy ,The original timestamps are normalized using discretization alignment; Unified time series alignment and sorting: merge and sort all multi-source data records with normalized timestamps in chronological order to build a unified time series data stream set .

4. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 3 is characterized in that: The job event identification and grouping includes: Equipment event interval extraction: Based on the start events, stop events, and state switching events marked in the equipment operation records, the equipment operation interval is identified in chronological order to form a collection of equipment event fragments ,in, The time when the device was started. is the corresponding stop time or state switching time, and each interval represents an independent operation segment; Personnel trajectory linkage analysis: Time series processing of construction personnel location information, in each equipment event segment interval Within, the distance between the identification and the device location coordinates is less than the threshold Personnel trajectory points form a personnel linkage subset ,satisfy ,in, is the spatial coordinate of the device, For personnel at the time point The location coordinates of Process action segment matching: based on the start and end time of the action segment recorded in the process instruction flow , and the device event interval Perform time intersection judgment, if it meets , then it is considered that there is a logical correspondence between the equipment event and the process action segment; Construction and grouping of operation events: Identify the event segments that satisfy the equipment-personnel-process ternary coupling relationship as operation events and extract their corresponding construction unit numbers , and press Grouping to form a construction unit level operation event set .

5. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 4 is characterized in that: The factor extraction and normalization include: Field parsing and mapping extraction: Extract construction behavior factors, including operation types, from identified work events , trigger source identification , Execution node identification , Operation start and end time ; Execution duration calculation: Calculates the execution duration of a job event based on the start and end time. ; Normalization processing: The extracted construction behavior factors are converted into standard format and normalized.

6. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 1 is characterized in that: The behavior-driven chain reconstruction module includes: Behavior factor graph construction: The extracted construction behavior factors are used as graph nodes, and a directed initial graph is constructed according to their occurrence time sequence and execution node distribution. ,in, is the node set consisting of all construction behavior factors, is the trigger edge between events; Causality Identification: Triggering Edges in a Directed Initial Graph Perform causal screening to retain edge sets with logically driven features ,According to the time difference between the construction behavior factors, the causal driving confidence is calculated,which is used to screen the behavior triggering edges that meet the causal conditions; Chain behavior driven path generation: based on filtered edge sets , using depth-first search, recursively build the behavior-driven chain from the source node backward ,Each path represents a set of construction behavior sequences with potential operation dependencies; Quality event mapping and label construction: collect construction quality inspection results The abnormal quality events in the process are mapped to one or more key nodes in the behavior-driven chain according to the time and space correspondence. , establish behavioral factors The associated label structure of the quality results generates a set of behavior quality association chains ,in, For an effective behavior-driven chain path, For quality inspection results.

7. The engineering progress quality spatiotemporal tracing system based on digital twin sandbox according to claim 6 is characterized in that: The space-time twin projection and inversion module includes: Behavior chain coordinate analysis and layer binding: for each construction behavior factor node in the behavior quality association chain , extract the spatial coordinates of its execution node and timestamp and bind it to the corresponding 3D structure of the construction unit in the digital twin sandbox model On the other hand, we construct a time-space dual-axis mapping relationship ; Timeline and spatial hierarchical visualization modeling: building an interactive time-space axis mechanism for twin views and setting the time window length and spatial display granularity levels Driven by the behavior chain time series, it dynamically activates the state change animation or color identification of the corresponding area and time period in the sandbox model, realizing the visual expression of behavior-quality status in the twin model; Mass anomaly driven inversion path generation: When a mass anomaly event is selected , automatically locate the corresponding node in the behavior quality association chain , and perform deep backtracking along its forward causal path to extract its dependency links , and simultaneously activate the historical status playback and responsibility factor display of relevant construction units in the sandbox; Inversion chain visual rendering and interactive output: path-based Each node in , a visual path of the operation behavior chain is generated in the form of lines or time jump points in the twin sandbox. At the same time, the responsible unit, behavior type and duration of each behavior node are synchronously displayed on the interface interaction layer, and an interactive traceability view is output. .

8. The digital twin sandbox-based spatiotemporal tracing system for project progress quality according to claim 7 is characterized in that: The timeline and space level visualization modeling includes: Time window and spatial granularity settings: set the global time window length and spatial display granularity level , used to control the time span of behavioral state changes and the degree of construction unit resolution in the twin sandbox; Behavior sequence drives spatiotemporal mapping activation: based on the collection of behavioral quality association chains All behavior nodes in , when its timestamp Activate the construction unit layer to which it is bound , and its behavior type status and quality status Perform joint coding and generate visual display coding; Dynamic rendering output of sand table view: all construction behavior nodes activated in the current time window are rendered , based on the generated visual display code, drive the bound construction unit layer Perform dynamic rendering, including color fill, flickering frequency, or transparency changes, and generate interactive spatiotemporal animation layers that evolve over time in the sandbox view.

9. The digital twin sandbox-based spatiotemporal tracing system for project progress quality according to claim 8 is characterized in that: The mass anomaly driven inversion path generation includes: Quality event node positioning: When the user selects a quality abnormality event in the twin sandbox interface , automatically in the behavior quality association chain Find the behavior node corresponding to the event ,satisfy ,in, The recording time of the quality event, For nodes timestamp; Deep backtracking extraction of causal paths: based on behavior-driven graphs The set of causal edges in , from the node Perform depth-first backtracking forward to extract all the previous nodes that logically affect it and form an inversion path ; Sand table layer activation and responsibility factor display: for inversion path Each behavior node in , extract its bound construction unit layer , and activate its historical operation type, historical execution time and historical quality status visualization, and extract the responsibility factor information recorded by the node , including responsible units, operating personnel, and task execution ID, are displayed in an overlaid manner on the sandbox interactive interface.

Images