A method and system for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring

Abstract

This invention discloses a method and system for risk assessment in building engineering based on multi-source anomaly monitoring, specifically relating to the field of building engineering risk assessment. The method includes analyzing the mapping and backhaul latency characteristics of equipment-to-component nodes based on multi-source monitoring data received during idle task periods, establishing a cross-device time alignment model and a pseudo-synchronization segment time offset trajectory, identifying potential anomaly path sets through latency drift trends, and providing a temporal feature basis for subsequent structural anomaly reasoning and risk assessment. Based on the final pseudo-synchronization segment index matrix and channel path mapping, the method derives the node triggering order and temporal causal chain. By receiving multi-source monitoring data during idle task periods and establishing a cross-device time alignment model, the method effectively identifies timing misalignment problems caused by mechanisms such as cache latency and packet loss retransmission, preventing historical anomaly data from being misjudged as future risks from the source.

Description

Technical Field

[0001] This invention relates to the field of building engineering risk assessment technology, and more specifically, to a building engineering risk assessment method and system for multi-source anomaly monitoring. Background Technology

[0002] In risk assessment of building engineering projects oriented towards multi-source anomaly monitoring, the system often needs to integrate monitoring data from different types of sensors, such as structural displacement sensors, environmental image modules, stress gauges, etc., to make time-series risk judgments.

[0003] However, when a project is in a non-operational period or construction is suspended, due to the caching delay, intermittent backhaul, or packet loss retransmission mechanism of some equipment, its historical data may be delayed and incorrectly mapped to the current time period by the system, thus causing a time sequence misalignment with other real-time data. In this case, the system is very likely to misjudge "historical abnormal data" as "abnormal events that have already occurred in the future", forming the so-called "abnormality advance retrospective illusion", thereby triggering incorrect structural warnings and affecting the correctness of the overall risk reasoning path. Summary of the Invention

[0004] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a method and system for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring. By synchronously verifying the cross-device time consistency of multi-source monitoring data during idle task cycles, dynamically extracting pseudo-synchronous offset trajectories and performing structural causal chain inference and response time correction, the method eliminates anomaly misjudgments caused by historical data lag, ensuring the continuity and accuracy of the risk assessment inference chain, thereby solving the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring, comprising:

[0006] Based on the multi-source monitoring data received during the idle task cycle, the mapping and backhaul latency characteristics from the device to the component node are analyzed, a cross-device time alignment model and pseudo-synchronization segment time offset trajectory are established, and potential abnormal path sets are identified through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment.

[0007] Based on the final pseudo-synchronous segment index matrix and channel path mapping, the node triggering order and time causal chain are derived. Combined with delay statistical analysis, time misalignment events are identified, and structural segments with causal residual mismatch in the response chain are located, providing support for pseudo-synchronous segment time correction and modified causal chain graph reconstruction.

[0008] Based on the modified causal chain graph, the response time and order of pseudo-synchronous segment nodes are retrieved, effective time correction functions are derived and screened, time labels and node triggering order are corrected, the synchronization dataset is reconstructed and the path index and delay characteristics are updated, and the continuity closure of the structural response chain and cross-cycle synchronization consistency are achieved.

[0009] Archive synchronization instability indicators for each cycle and construct a time series risk indicator set. Predict the evolution of synchronization instability risk based on trend-driven dynamics. Adjust the detection parameters of high-risk path groups according to the prediction results. Continuously optimize the screening process for cross-cycle pseudo-synchronous segments and the correction process for structural chains to form an iterative risk feedback system.

[0010] In a preferred embodiment, the original monitoring records uploaded by each heterogeneous monitoring device when the risk assessment and monitoring platform is in an idle state are collected, and the device identifier, timestamp and channel protocol fields of the records are extracted to form an original data tuple set; channel protocol classification and path identifier grouping operations are performed on the original data tuple set respectively to generate a device channel table and record the return start time and periodic interval of each channel.

[0011] The data of the start time and periodic interval of each channel recorded in the device channel table are aggregated according to the device identifier and timestamp to generate a reporting periodic sequence, and the three statistical indicators of average delay, range and frequency hopping rate are calculated based on this to generate a delay statistical parameter group.

[0012] Project the set of delay statistics parameters onto the global time axis, solve the time alignment difference map between devices, record the response drift trend, and extract the set of paths that form a stable offset trend.

[0013] In a preferred embodiment, an offset trajectory map is constructed based on the time alignment difference map between the delay statistical parameter set and the device, and a set of paths with continuous drift gradients exceeding a stable threshold is identified. This set of paths is defined as a statistically unstable path set, and a cross-analysis is performed between it and the alignment difference map to screen out paths that simultaneously satisfy both trend anomaly and offset amplitude exceeding the threshold to form a potential abnormal path set.

[0014] Extract the time period response records corresponding to each path in the potential abnormal path set, and aggregate them according to the time window to generate a pseudo-synchronization segment candidate set; perform drift trend consistency screening on the pseudo-synchronization segment candidate set, and output the final pseudo-synchronization segment index matrix. The final pseudo-synchronization segment index matrix records the device number, start timestamp and path identification information corresponding to each pseudo-synchronization interval, which serves as the basic data for subsequent path structure mapping and causal inference.

[0015] For each pseudo-synchronization interval in the final pseudo-synchronization segment index matrix, the consistency of the response sequence distribution is checked. The amplification characteristics of the drift residual change trend within the time segment are evaluated. Candidate synchronization instability segments with drift amplitude exceeding a set threshold are selected. The candidate synchronization instability segments selected by drift residual are further labeled as key drift segments, and the corresponding time offset gradient and sensor node mapping relationship are recorded.

[0016] In a preferred embodiment, the device number recorded in the final pseudo-synchronization segment index matrix is ​​used as the index entry for the potential abnormal path response record. Through the path rules defined by the channel protocol classification and path identifier grouping, it is mapped to the corresponding component node to construct the component-level node sequence under the device channel path mapping.

[0017] The original timestamps of each node in the component-level node sequence are sorted sequentially to determine the triggering order of each node in the device path structure, which serves as the basis for constructing the time axis of the structural response chain. The device path mapping relationship is derived according to the path rules, the connection method between component nodes is parsed, the ordered structure of node connections is extracted, and a set of directional time-driven edges is constructed.

[0018] For each node pair in the time-driven edge set, calculate the trigger interval, and combine it with the device channel mapped by the node pair. In the delay statistics parameter group, define the average delay and range parameters, call the propagation delay fitting model of the corresponding channel, and generate the time residual vector.

[0019] The time residual vector is segmented according to a fixed time window. Within each segment, the slope change value is continuously calculated based on a sliding window to form a slope change data sequence. In this data sequence, slope abrupt change points and slope reversal points are identified, and a set of marker points with abnormal change trends is extracted. The time intervals covered between adjacent marker points are extracted to form a preliminary set of fluctuation candidate segments.

[0020] In a preferred embodiment, the fluctuation amplitude and slope change rate are calculated for each time segment in the preliminary fluctuation candidate segment set to construct a joint fluctuation evaluation index set; the index vector corresponding to each time segment in the joint fluctuation evaluation index set is compared with the set evaluation threshold segment by segment to screen out the time segments that meet the fracture drift triggering characteristics and mark them as candidate drift fracture segments.

[0021] Using the time segment covered by the candidate drift break segment as an index window, filter the set of edge segments in the time-driven edge set whose trigger time falls into the segment, and aggregate them into a set of causal mismatch paths.

[0022] Perform response path elimination operation on the edge segments in the causal mismatch path set, strip away logical conflict relationships, reconstruct a modified causal chain graph that satisfies structural temporal consistency, and synchronously update the path index mapping table corresponding to the response path structure; feed the modified causal chain graph back to the drift segment data correction process.

[0023] In a preferred embodiment, a modified causal chain graph structure is retrieved, the set of edge segments corresponding to the key drift segments is extracted, and the original timestamp, component node number and response time trajectory in the path structure contained in each pseudo-synchronization segment are parsed. The parsed response time trajectory and the original timestamp are jointly used to construct a node response residual table, and a node lag probability evaluation model is introduced as a constraint term for time offset correction.

[0024] Based on the node response residual table, a recursive error inversion algorithm is executed to derive the time correction function node by node, and the time correction functions derived from all nodes are collected into a time correction function group to generate a path consistency index.

[0025] Perform a topological logic consistency check on the time correction function set, remove illegal function subsets with time reversal or path intersection, and retain the set of valid time adjustment functions under topological constraints.

[0026] In a preferred embodiment, the set of valid time adjustment functions selected by the topological logic consistency check is applied to the pseudo-synchronous segment record. Based on the time correction function corresponding to each node in the node response residual table, the response time label of the node in the structural path is updated. The new node triggering order is deduced synchronously based on the updated response time label, and a node sequence mapping record is generated to reconstruct the temporal causal chain relationship between nodes, forming a synchronous reconstruction dataset that conforms to the coherence of the corrected structural path.

[0027] The synchronous reconstructed dataset is inserted into the time segment position of the corresponding pseudo-synchronous segment in the original monitoring data stream, and the response path index mapping table is updated synchronously to form a new structural response timing link.

[0028] The corrected response time labels and node sequence mapping records of the synchronous reconstruction dataset are sent back to the latency statistics parameter group. The latency characteristic records between devices are dynamically updated based on the adjustment range of each node time and the continuity of the triggering order. The overall timing consistency baseline is further superimposed and corrected, and a basis for cross-cycle time synchronization evaluation is formed.

[0029] In a preferred embodiment, after each idle task execution cycle ends, based on the pseudo-synchronous segment screening results and the structural path reconstruction data recorded in the modified causal chain graph generation process, the frequency of synchronous instability segment triggering, the number of associated abnormal paths, and the stability change index of the modified structural response chain that occurred in this cycle are statistically analyzed.

[0030] The frequency of synchronous instability segment triggers, the number of associated abnormal paths, and the stability change indicators of the corrected structural chain obtained in each idle period are archived and organized to form a time series risk indicator set, which is used to record the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability change.

[0031] Based on the time series risk indicator set obtained from archiving and collation, the trend extrapolation analysis is performed. By fitting the evolution curves of the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability, the trajectory of synchronous instability risk in future cycles is predicted, and high-risk path groups are dynamically marked in advance according to the risk threshold.

[0032] In a preferred embodiment, for each device involved in the high-risk path group, the corresponding device item is located in the delay statistics parameter group, and its delay detection threshold and drift detection sensitivity configuration are dynamically adjusted to improve the abnormal drift capture capability and timing instability identification accuracy of the high-risk path segment.

[0033] The dynamically adjusted delay detection parameters and drift detection sensitivity configuration of high-risk path groups are synchronously applied to each process step of pseudo-synchronous segment screening, structural chain reconstruction and causal mismatch testing during the idle period of the next cycle, forming a continuous iterative cross-cycle risk accumulation detection and response calibration feedback mechanism.

[0034] In a preferred embodiment, a building engineering risk assessment system for multi-source anomaly monitoring includes: an offset extraction module, a mismatch inference module, a correction synchronization module, and an early warning iteration module;

[0035] The offset extraction module extracts multi-source monitoring data received during idle task cycles, analyzes the mapping and backhaul latency characteristics from devices to component nodes, establishes a cross-device time alignment model and pseudo-synchronous segment time offset trajectory, and identifies potential abnormal path sets through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment.

[0036] The mismatch inference module is based on the final pseudo-synchronization segment index matrix and channel path mapping to deduce the node triggering order and time causal chain. Combined with delay statistical analysis, it identifies time misalignment events and locates structural segments in the response chain that have causal residual mismatches, providing support for pseudo-synchronization segment time correction and modified causal chain graph reconstruction.

[0037] The corrected synchronization module is based on a corrected causal chain graph. It retrieves the response time and order of pseudo-synchronous segment nodes, derives and filters effective time correction functions, completes the correction of time tags and node triggering order, reconstructs the synchronization dataset and updates the path index and delay characteristics, and achieves continuous closure of the structural response chain and cross-cycle synchronization consistency.

[0038] The early warning iteration module archives synchronization instability indicators for each cycle and constructs a time series risk indicator set. Based on the trend-driven dynamics, it predicts the evolution of synchronization instability risk, adjusts the detection parameters of high-risk path groups according to the prediction results, and continuously optimizes the process of screening pseudo-synchronization segments and correcting structural chains across cycles, thus forming an iterative risk feedback system.

[0039] The technical effects and advantages of this invention are as follows:

[0040] 1. By receiving multi-source monitoring data during the idle task cycle and establishing a cross-device time alignment model, this invention can effectively identify timing misalignment problems caused by mechanisms such as cache latency and packet loss retransmission, avoid misjudging historical abnormal data as future risks from the source, correct the illusion of abnormal advance retrospection, and ensure the timeliness of structural early warning.

[0041] 2. This invention constructs a pseudo-synchronous segment time offset trajectory, extracts potential abnormal paths based on drift trends, and combines node-level causal chain reasoning and time residual analysis to locate structural segments in the response chain that have temporal breaks or causal mismatches, thereby improving the integrity and continuity of the structural anomaly reasoning chain.

[0042] 3. This invention constructs a pseudo-synchronous segment time offset trajectory, extracts potential abnormal paths based on drift trends, and combines node-level causal chain reasoning and time residual analysis to locate structural segments in the response chain that have temporal breaks or causal mismatches, thereby improving the integrity and continuity of the structural anomaly reasoning chain.

[0043] 4. By archiving synchronous instability indicators of each period, this invention establishes a time series risk indicator set and performs trend extrapolation prediction. It can dynamically adjust the detection parameters of high-risk path groups and construct an adaptive and iteratively optimized cross-period risk accumulation detection and feedback calibration system, thereby continuously improving the foresight of overall risk identification. Attached Figure Description

[0044] Figure 1 This is a flowchart of the method steps of the present invention.

[0045] Figure 2 This is a schematic diagram of the system modules of the present invention.

[0046] Figure 3 This is a flowchart of the multi-source data time-series offset extraction process of the present invention.

[0047] Figure 4 This is a flowchart illustrating the pseudo-synchronous segment causal chain mismatch inference process of the present invention.

[0048] Figure 5 This is a flowchart of the pseudo-synchronization period time correction and synchronization reconstruction of the present invention.

[0049] Figure 6This is a flowchart of the cross-cycle synchronous instability risk prediction and feedback process of the present invention. Detailed Implementation

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0051] Refer to the instruction manual appendix Figure 1-6 An embodiment of the present invention provides a risk assessment method for building engineering projects oriented towards multi-source anomaly monitoring, comprising:

[0052] Based on the multi-source monitoring data received during the idle task cycle, the mapping and backhaul latency characteristics from the device to the component node are analyzed, a cross-device time alignment model and pseudo-synchronization segment time offset trajectory are established, and potential abnormal path sets are identified through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment.

[0053] Based on the final pseudo-synchronous segment index matrix and channel path mapping, the node triggering order and time causal chain are derived. Combined with delay statistical analysis, time misalignment events are identified, and structural segments with causal residual mismatch in the response chain are located, providing support for pseudo-synchronous segment time correction and modified causal chain graph reconstruction.

[0054] Based on the modified causal chain graph, the response time and order of pseudo-synchronous segment nodes are retrieved, effective time correction functions are derived and screened, time labels and node triggering order are corrected, the synchronization dataset is reconstructed and the path index and delay characteristics are updated, and the continuity closure of the structural response chain and cross-cycle synchronization consistency are achieved.

[0055] Archive synchronization instability indicators for each cycle and construct a time series risk indicator set. Predict the evolution of synchronization instability risk based on trend-driven dynamics. Adjust the detection parameters of high-risk path groups according to the prediction results. Continuously optimize the screening process for cross-cycle pseudo-synchronous segments and the correction process for structural chains to form an iterative risk feedback system.

[0056] The system collects raw monitoring records uploaded by various heterogeneous monitoring devices when the risk assessment and monitoring platform is idle, and extracts the device identifier, timestamp, and channel protocol fields of the records to form a raw data tuple set. The system then performs channel protocol classification and path identifier grouping operations on the raw data tuple set to generate a device channel table and records the return start time and periodic interval of each channel.

[0057] The data of the start time and periodic interval of each channel recorded in the device channel table are aggregated according to the device identifier and timestamp to generate a reporting periodic sequence, and the three statistical indicators of average delay, range and frequency hopping rate are calculated based on this to generate a delay statistical parameter group.

[0058] Project the set of delay statistics parameters onto the global time axis, solve the time alignment difference map between devices, record the response drift trend, and extract the set of paths that form a stable offset trend.

[0059] An offset trajectory map is constructed based on the time alignment difference map between the delay statistical parameter set and the device, and a set of paths with continuous drift gradients exceeding the stability threshold is identified. This set of paths is defined as a statistically unstable path set, and a cross-analysis is performed between it and the alignment difference map. Paths that simultaneously satisfy the abnormal trend and the offset magnitude exceeding the threshold are selected to form a potential abnormal path set.

[0060] Extract the time period response records corresponding to each path in the potential abnormal path set, and aggregate them according to the time window to generate a pseudo-synchronization segment candidate set; perform drift trend consistency screening on the pseudo-synchronization segment candidate set, and output the final pseudo-synchronization segment index matrix. The final pseudo-synchronization segment index matrix records the device number, start timestamp and path identification information corresponding to each pseudo-synchronization interval, which serves as the basic data for subsequent path structure mapping and causal inference.

[0061] For each pseudo-synchronization interval in the final pseudo-synchronization segment index matrix, the consistency of the response sequence distribution is checked. The amplification characteristics of the drift residual change trend within the time segment are evaluated. Candidate synchronization instability segments with drift amplitude exceeding the set threshold are selected. The candidate synchronization instability segments selected by drift residual are further marked as key drift segments. The corresponding time offset gradient and sensor node mapping relationship are recorded and used as the input basis for subsequent causal structure continuity verification.

[0062]

[0063] in:

[0064]

[0065] in Represents the pseudo-synchronization segment index matrix; τ i (t),τ j (t) represents the timestamps of the i-th and j-th devices at time t, respectively; Δt i,j (t) represents the time offset between devices i and j; A matrix representing the channel connection relationships between devices; Indicates the rate of change of the time offset; The acceleration represents the change in time offset; λ and κ represent the first-order and second-order feature enhancement indices, respectively. θ represents the dynamic modulation factor for channel stability between i and j; sync This represents the threshold for determining pseudo-synchronization drift anomalies; Seg(·) represents the set of time periods that satisfy the drift intensity determination condition extracted from the joint drift discriminant expression, which is used as the extraction operation for pseudo-synchronization offset segments; This represents the operation of joint traversal and feature extraction for all device pairs (i,j) forming sensor path combinations; This indicates that an element-wise multiplication operation is performed on two vectors or function results of the same dimension.

[0066] The above formula uses a joint computing device to analyze the first derivative (drift rate) and second derivative (drift acceleration) characteristics of time offset, and combines them with channel stability weights to form a comprehensive drift anomaly intensity index. This allows the identification of drift anomalies exceeding a set threshold on a continuous time axis, enabling the extraction of pseudo-synchronization segments. The overall structure aims to highlight the abrupt and aggravating characteristics in the drift evolution trend, supporting subsequent pseudo-synchronization time window screening and anomaly path inference.

[0067] The device number recorded in the final pseudo-synchronization segment index matrix is ​​used as the index entry for potential abnormal path response records. Through the path rules defined by the channel protocol classification and path identifier grouping, it is mapped to the corresponding component node to construct the component-level node sequence under the device channel path mapping.

[0068] The original timestamps of each node in the component-level node sequence are sorted sequentially to determine the triggering order of each node in the device path structure, which serves as the basis for constructing the time axis of the structural response chain. The device path mapping relationship is derived according to the path rules, the connection method between component nodes is analyzed, the ordered structure of node connections is extracted, and a directional time-driven edge set is constructed as the basic edge set for causal reasoning.

[0069] For each node pair in the time-driven edge set, calculate the trigger interval, and combine it with the device channel mapped by the node pair. In the delay statistics parameter group, define the average delay and range parameters, call the propagation delay fitting model of the corresponding channel, and generate the time residual vector.

[0070] The time residual vector is segmented according to a fixed time window. Within each segment, the slope change value is continuously calculated based on a sliding window to form a slope change data sequence. In this data sequence, slope abrupt change points and slope reversal points are identified, and a set of marker points with abnormal change trends is extracted. The time intervals covered between adjacent marker points are extracted to form a preliminary set of fluctuation candidate segments.

[0071] For each time segment in the preliminary candidate fluctuation segment set, calculate the fluctuation amplitude and slope change rate to construct a joint fluctuation assessment index set; compare the index vector corresponding to each time segment in the joint fluctuation assessment index set with the set assessment threshold segment by segment to screen out the time segments that meet the fracture drift triggering characteristics and mark them as candidate drift fracture segments.

[0072] Using the time segment covered by the candidate drift fracture segment as an index window, filter the set of edge segments in the time-driven edge set whose trigger time falls into the segment, and aggregate them into a causal mismatch path set to identify the structural path fracture region caused by response misalignment.

[0073] The process involves removing response paths from the edges in the set of causal mismatched paths, stripping away logical conflicts, reconstructing a modified causal chain graph that satisfies structural temporal consistency, and synchronously updating the path index mapping table corresponding to the response path structure. This modified causal chain graph is then fed back to the drift segment data correction process to correct the response record time stamps, close the causal inference path, and establish a new definition. Represents the set of paths to nodes with causal mismatch;

[0074]

[0075] in,

[0076] Ψ u,v (t)=(τ u (t)-τ v (t));

[0077] Where τ u (t), τ v (t) represents the timestamps of nodes u and v at time t, respectively; Ψ u,v (t) represents the difference in triggering time between node u and node v; ω represents the rate of change of node u with respect to time difference; ζ and η represent the drift rate feature enhancement index and the drift amplitude feature enhancement index, respectively; u,v (t) represents the dynamic weight function of the edge weight of node pair (u,v) at time t, used to measure the significance or importance of the causal relationship of this edge segment in the structural response path at the current time. The unit is "dimensionless coefficient" or "weight coefficient / time" depending on the specific physical quantity; θ mismatch This represents the comprehensive threshold for mismatch determination; This represents a structural segment extraction function used to filter out substructures that satisfy the causal chain mismatch condition from the global residual index set; This means performing a set and operation on the causal connection edges between all node pairs (u,v). In the entire structural response graph, all node pairs (u,v) with time-driven relationships are traversed, and the judgment expressions on each node pair are collected into a set to construct an overall causal mismatch index.

[0078] exist In the formula, the first-order rate of change (drift rate anomaly) of the difference in trigger time between nodes is calculated and the absolute time offset amplitude is superimposed with connectivity weighting to screen out the trigger chain misalignment region caused by drift instability; the formula design structure aims to dynamically capture the causal relationship breakpoints hidden inside the pseudo-synchronization segment, ensure the sensitivity and accuracy of the causal chain integrity test, and provide basic support for subsequent response correction and causal chain correction.

[0079] The modified causal chain graph structure is retrieved, the set of edge segments corresponding to the key drift segments is extracted, and the original timestamp, component node number and response time trajectory in the path structure contained in each pseudo-synchronization segment are analyzed. The analyzed response time trajectory and the original timestamp are combined to construct the node response residual table, and the node lag probability evaluation model is introduced as a constraint term for time offset correction.

[0080] A recursive error inversion algorithm is executed based on the node response residual table. The time correction function is derived node by node, and the time correction functions derived from all nodes are collected into a time correction function group to generate a path consistency index for verifying the integrity of the structural correction.

[0081] Perform a topological logic consistency check on the time correction function set, remove illegal function subsets with time reversal or path intersection, and retain the set of valid time adjustment functions under topological constraints.

[0082] The set of valid time adjustment functions selected by the topological logic consistency check is applied to the pseudo-synchronous segment record. Based on the time correction function corresponding to each node in the node response residual table, the response time label of the node in the structural path is updated. The new node triggering order is deduced synchronously based on the updated response time label, and a node sequence mapping record is generated to reconstruct the temporal causal chain relationship between nodes, forming a synchronous reconstruction dataset that conforms to the coherence of the corrected structural path.

[0083] The synchronous reconstructed dataset is inserted into the time segment position of the corresponding pseudo-synchronous segment in the original monitoring data stream, and the response path index mapping table is updated synchronously to form a new structural response timing link.

[0084] The corrected response time labels and node sequence mapping records of the synchronous reconstruction dataset are sent back to the latency statistics parameter group. The latency characteristic records between devices are dynamically updated according to the adjustment range of each node time and the continuity of the triggering order. The overall timing consistency baseline is further superimposed and corrected, and a cross-cycle time synchronization evaluation basis is formed.

[0085]

[0086] in,

[0087] Φ u,v (t)=(τ u (t)+δ u )-(τ v (t)+δ v );

[0088] τ represents the objective of synchronous dataset correction and optimization; u (t), τ v (t) represents the original timestamps of node u and node v at time t, respectively; δ u δ v Φ represents the time correction offset between node u and node v, respectively; u,v (t) represents the corrected residual between node u and node v; κ u,v (t) represents the connectivity weighting factor, used to adjust the synchronization importance between nodes; t s t e These represent the start and end times of the correction time window, respectively; Represents the set of all optional time correction offsets {δ u In the expression, find the set of δ that minimizes the sum of the weighted terms in the subsequent integral. u Value combination; additionally, the fourth power in the formula is used to amplify the difference in timing offset error, its meaning being related to Φ u,v (t) is processed by higher-order powers so that the larger the offset value, the more significant its impact on the overall optimization objective function, thereby strengthening the penalty effect on large offset paths and improving the sensitivity and selectivity of synchronous adjustment.

[0089] This formula minimizes the synchronization error of the structural chain by accumulating the corrected time residuals of node pairs in the form of a fourth-power penalty and then using connectivity-weighted integration to globally optimize the set of corrections within the overall time window. The formula structure aims to strengthen the strong suppression effect on nodes with large synchronization offsets, while ensuring the closed consistency of the causal chain logic, and supporting the reconstruction of the synchronized dataset and the updating of the path index.

[0090] After each idle task execution cycle ends, based on the pseudo-synchronous segment screening results and the structural path reconstruction data recorded in the modified causal chain graph generation process, the frequency of synchronous instability segment triggering, the number of associated abnormal paths, and the stability change index of the modified structural response chain are statistically analyzed.

[0091] The frequency of synchronous instability segment triggers, the number of associated abnormal paths, and the stability change indicators of the corrected structural chain obtained in each idle period are archived and organized to form a time series risk indicator set, which is used to record the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability change.

[0092] Based on the time series risk indicator set obtained from archiving and collation, the trend extrapolation analysis is performed. By fitting the evolution curves of the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability, the trajectory of synchronous instability risk in future cycles is predicted, and high-risk path groups are dynamically marked in advance according to the risk threshold.

[0093] For each device involved in the high-risk path group, locate the corresponding device item in the delay statistics parameter group, and dynamically adjust its delay detection threshold and drift detection sensitivity configuration to improve the abnormal drift capture capability and timing instability identification accuracy of the high-risk path segment.

[0094] The dynamically adjusted delay detection parameters and drift detection sensitivity configuration of high-risk path groups are synchronously applied to each process step of pseudo-synchronous segment screening, structural chain reconstruction and causal mismatch testing during the idle period of the next cycle, forming a continuous iterative cross-cycle risk accumulation detection and response calibration feedback mechanism.

[0095]

[0096] This represents the predicted trend value of synchronous instability risk; Σ instability (t) represents the cumulative number of synchronous instability events up to time t; The third derivative represents the rate of change of acceleration, indicating the trend of synchronous instability; χ(t) represents the synchronous instability sensitivity modulation function, which is used to adjust the risk response sensitivity; Υ(·) represents the higher-order dynamic extrapolation function, which is used to predict the synchronous instability trend in the future cycle.

[0097] This formula, after extracting the third derivative of the cumulative number of synchronous instability curves and adjusting the weighted sensitivity, predicts the evolution trend of synchronous instability risk in future cycles. The purpose of the formula structure design is to quickly respond to the acceleration and abrupt changes in instability trend, dynamically deduce the cross-cycle risk evolution situation, and guide the adaptive adjustment of early warning and detection strategies for synchronous instability risk.

[0098] A risk assessment system for building engineering with multi-source anomaly monitoring includes: an offset extraction module, a mismatch inference module, a correction synchronization module, and an early warning iteration module;

[0099] The offset extraction module extracts multi-source monitoring data received during idle task cycles, analyzes the mapping and backhaul latency characteristics from devices to component nodes, establishes a cross-device time alignment model and pseudo-synchronous segment time offset trajectory, and identifies potential abnormal path sets through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment.

[0100] The mismatch inference module is based on the final pseudo-synchronization segment index matrix and channel path mapping to deduce the node triggering order and time causal chain. Combined with delay statistical analysis, it identifies time misalignment events and locates structural segments in the response chain that have causal residual mismatches, providing support for pseudo-synchronization segment time correction and modified causal chain graph reconstruction.

[0101] The corrected synchronization module is based on a corrected causal chain graph. It retrieves the response time and order of pseudo-synchronous segment nodes, derives and filters effective time correction functions, completes the correction of time tags and node triggering order, reconstructs the synchronization dataset and updates the path index and delay characteristics, and achieves continuous closure of the structural response chain and cross-cycle synchronization consistency.

[0102] The early warning iteration module archives synchronization instability indicators for each cycle and constructs a time series risk indicator set. Based on the trend-driven dynamics, it predicts the evolution of synchronization instability risk, adjusts the detection parameters of high-risk path groups according to the prediction results, and continuously optimizes the process of screening pseudo-synchronization segments and correcting structural chains across cycles, thus forming an iterative risk feedback system.

[0103] In the formula structure involved in this scheme, dimensionless terms can be used as proportional or structural adjustment factors. When combined with quantities with units, they only play a role in numerical scaling and do not introduce new physical dimensions. Therefore, they will not change or confuse the overall unit system. This combination of "dimensionless terms and terms with units" can be understood as a composite structural expression commonly used in mathematical physics modeling. It conforms to the principle of dimensional consistency and has a clear physical interpretation basis.

[0104] Secondly, in the formula structure of this scheme, if multiple variables with different physical units are involved, including but not limited to time, mass or energy variables, their joint appearance is to express the collaborative modeling relationship of multiple physical mechanisms. Each variable forms a unified structure through function mapping, ratio combination or normalization adjustment, with clear units and clear meaning. The overall expression conforms to the principle of dimensional consistency and the conventional formula of engineering modeling.

[0105] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A risk assessment method for building engineering projects oriented towards multi-source anomaly monitoring, characterized in that, include: Based on the multi-source monitoring data received during the idle task cycle, the mapping and backhaul latency characteristics from the device to the component node are analyzed, a cross-device time alignment model and pseudo-synchronization segment time offset trajectory are established, and potential abnormal path sets are identified through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment. Based on the final pseudo-synchronous segment index matrix and channel path mapping, the node triggering order and time causal chain are derived. Combined with delay statistical analysis, time misalignment events are identified, and structural segments with causal residual mismatch in the response chain are located, providing support for pseudo-synchronous segment time correction and modified causal chain graph reconstruction. Based on the modified causal chain graph, the response time and order of pseudo-synchronous segment nodes are retrieved, effective time correction functions are derived and screened, time labels and node triggering order are corrected, the synchronization dataset is reconstructed and the path index and delay characteristics are updated, and the continuity closure of the structural response chain and cross-cycle synchronization consistency are achieved. Archive synchronization instability indicators for each cycle and construct a time series risk indicator set. Predict the evolution of synchronization instability risk based on trend-driven dynamics. Adjust the detection parameters of high-risk path groups according to the prediction results. Continuously optimize the screening process for cross-cycle pseudo-synchronous segments and the correction process for structural chains to form an iterative risk feedback system.

2. The method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring as described in claim 1, characterized in that: The system collects raw monitoring records uploaded by various heterogeneous monitoring devices when the risk assessment and monitoring platform is idle, and extracts the device identifier, timestamp, and channel protocol fields of the records to form a raw data tuple set. The system then performs channel protocol classification and path identifier grouping operations on the raw data tuple set to generate a device channel table and records the return start time and periodic interval of each channel. The data of the start time and periodic interval of each channel recorded in the device channel table are aggregated according to the device identifier and timestamp to generate a reporting periodic sequence, and the three statistical indicators of average delay, range and frequency hopping rate are calculated based on this to generate a delay statistical parameter group. Project the set of delay statistics parameters onto the global time axis, solve the time alignment difference map between devices, record the response drift trend, and extract the set of paths that form a stable offset trend.

3. The method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring according to claim 2, characterized in that: An offset trajectory map is constructed based on the time alignment difference map between the delay statistical parameter set and the device, and a set of paths with continuous drift gradients exceeding the stability threshold is identified. This set of paths is defined as a statistically unstable path set, and a cross-analysis is performed between it and the alignment difference map. Paths that simultaneously satisfy the abnormal trend and the offset magnitude exceeding the threshold are selected to form a potential abnormal path set. Extract the time period response records corresponding to each path in the potential abnormal path set, and aggregate them according to the time window to generate a pseudo-synchronization segment candidate set; perform drift trend consistency screening on the pseudo-synchronization segment candidate set, and output the final pseudo-synchronization segment index matrix. The final pseudo-synchronization segment index matrix records the device number, start timestamp and path identification information corresponding to each pseudo-synchronization interval, which serves as the basic data for subsequent path structure mapping and causal inference. For each pseudo-synchronization interval in the final pseudo-synchronization segment index matrix, the consistency of the response sequence distribution is checked. The amplification characteristics of the drift residual change trend within the time segment are evaluated. Candidate synchronization instability segments with drift amplitude exceeding a set threshold are selected. The candidate synchronization instability segments selected by drift residual are further labeled as key drift segments, and the corresponding time offset gradient and sensor node mapping relationship are recorded.

4. The method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring according to claim 3, characterized in that: The device number recorded in the final pseudo-synchronization segment index matrix is ​​used as the index entry for potential abnormal path response records. Through the path rules defined by the channel protocol classification and path identifier grouping, it is mapped to the corresponding component node to construct the component-level node sequence under the device channel path mapping. The original timestamps of each node in the component-level node sequence are sorted sequentially to determine the triggering order of each node in the device path structure, which serves as the basis for constructing the structural response chain timeline. Based on the path rules, the device path mapping relationship is derived, the connection method between component nodes is analyzed, the ordered structure of node connections is extracted, and a set of directional time-driven edges is constructed. For each node pair in the time-driven edge set, calculate the trigger interval, and combine it with the device channel mapped by the node pair. In the delay statistics parameter group, define the average delay and range parameters, call the propagation delay fitting model of the corresponding channel, and generate the time residual vector. The time residual vector is segmented according to a fixed time window. Within each segment, the slope change value is continuously calculated based on a sliding window to form a slope change data sequence. In this data sequence, slope abrupt change points and slope reversal points are identified, and a set of marker points with abnormal change trends is extracted. The time intervals covered between adjacent marker points are extracted to form a preliminary set of fluctuation candidate segments.

5. A method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring according to claim 4, characterized in that: For each time segment in the preliminary candidate fluctuation segment set, calculate the fluctuation amplitude and slope change rate to construct a joint fluctuation assessment index set; compare the index vector corresponding to each time segment in the joint fluctuation assessment index set with the set assessment threshold segment by segment to screen out the time segments that meet the fracture drift triggering characteristics and mark them as candidate drift fracture segments. Using the time segment covered by the candidate drift break segment as an index window, filter the set of edge segments in the time-driven edge set whose trigger time falls into the segment, and aggregate them into a set of causal mismatch paths. Perform response path elimination operation on the edge segments in the causal mismatch path set, strip away logical conflict relationships, reconstruct a modified causal chain graph that satisfies structural temporal consistency, and synchronously update the path index mapping table corresponding to the response path structure; feed the modified causal chain graph back to the drift segment data correction process.

6. The method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring according to claim 5, characterized in that: The modified causal chain graph structure is retrieved, the set of edge segments corresponding to the key drift segments is extracted, and the original timestamp, component node number and response time trajectory in the path structure contained in each pseudo-synchronization segment are analyzed. The analyzed response time trajectory and the original timestamp are combined to construct the node response residual table, and the node lag probability evaluation model is introduced as a constraint term for time offset correction. Based on the node response residual table, a recursive error inversion algorithm is executed to derive the time correction function node by node, and the time correction functions derived from all nodes are collected into a time correction function group to generate a path consistency index. Perform a topological logic consistency check on the time correction function set, remove illegal function subsets with time reversal or path intersection, and retain the set of valid time adjustment functions under topological constraints.

7. A method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring as described in claim 6, characterized in that: The set of valid time adjustment functions selected by the topological logic consistency check is applied to the pseudo-synchronous segment record. Based on the time correction function corresponding to each node in the node response residual table, the response time label of the node in the structural path is updated. The new node triggering order is deduced synchronously based on the updated response time label, and a node sequence mapping record is generated to reconstruct the temporal causal chain relationship between nodes, forming a synchronous reconstruction dataset that conforms to the coherence of the corrected structural path. The synchronous reconstructed dataset is inserted into the time segment position of the corresponding pseudo-synchronous segment in the original monitoring data stream, and the response path index mapping table is updated synchronously to form a new structural response timing link. The corrected response time labels and node sequence mapping records of the synchronous reconstruction dataset are sent back to the latency statistics parameter group. The latency characteristic records between devices are dynamically updated based on the adjustment range of each node time and the continuity of the triggering order. The overall timing consistency baseline is further superimposed and corrected, and a basis for cross-cycle time synchronization evaluation is formed.

8. A method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring as described in claim 7, characterized in that: After each idle task execution cycle ends, based on the pseudo-synchronous segment screening results and the structural path reconstruction data recorded in the modified causal chain graph generation process, the frequency of synchronous instability segment triggering, the number of associated abnormal paths, and the stability change index of the modified structural response chain are statistically analyzed. The frequency of synchronous instability segment triggers, the number of associated abnormal paths, and the stability change indicators of the corrected structural chain obtained in each idle period are archived and organized to form a time series risk indicator set, which is used to record the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability change. Based on the time series risk indicator set obtained from archiving and collation, the trend extrapolation analysis is performed. By fitting the evolution curves of the cumulative number of synchronous instability, the success rate of correction, and the trend of path stability, the trajectory of synchronous instability risk in future cycles is predicted, and high-risk path groups are dynamically marked in advance according to the risk threshold.

9. A method for risk assessment of building engineering projects oriented towards multi-source anomaly monitoring as described in claim 8, characterized in that: For each device involved in the high-risk path group, locate the corresponding device item in the delay statistics parameter group, and dynamically adjust its delay detection threshold and drift detection sensitivity configuration to improve the abnormal drift capture capability and timing instability identification accuracy of the high-risk path segment. The dynamically adjusted delay detection parameters and drift detection sensitivity configuration of high-risk path groups are synchronously applied to each process step of pseudo-synchronous segment screening, structural chain reconstruction and causal mismatch testing during the idle period of the next cycle, forming a continuous iterative cross-cycle risk accumulation detection and response calibration feedback mechanism.

10. A risk assessment system for building engineering projects oriented towards multi-source anomaly monitoring, comprising an offset extraction module, a mismatch inference module, a correction synchronization module, and an early warning iteration module, characterized in that: The offset extraction module extracts multi-source monitoring data received during idle task cycles, analyzes the mapping and backhaul latency characteristics from devices to component nodes, establishes a cross-device time alignment model and pseudo-synchronous segment time offset trajectory, and identifies potential abnormal path sets through latency drift trends, providing a time feature basis for subsequent structural anomaly reasoning and risk assessment. The mismatch inference module is based on the final pseudo-synchronization segment index matrix and channel path mapping to deduce the node triggering order and time causal chain. Combined with delay statistical analysis, it identifies time misalignment events and locates structural segments in the response chain that have causal residual mismatches, providing support for pseudo-synchronization segment time correction and modified causal chain graph reconstruction. The corrected synchronization module is based on a corrected causal chain graph. It retrieves the response time and order of pseudo-synchronous segment nodes, derives and filters effective time correction functions, completes the correction of time tags and node triggering order, reconstructs the synchronization dataset and updates the path index and delay characteristics, and achieves continuous closure of the structural response chain and cross-cycle synchronization consistency. The early warning iteration module archives synchronization instability indicators for each cycle and constructs a time series risk indicator set. Based on the trend-driven dynamics, it predicts the evolution of synchronization instability risk, adjusts the detection parameters of high-risk path groups according to the prediction results, and continuously optimizes the process of screening pseudo-synchronization segments and correcting structural chains across cycles, thus forming an iterative risk feedback system.

Images