A building construction process risk assessment method and system based on a fault-tolerant mechanism

Abstract

The application discloses a kind of based on fault-tolerant mechanism's building construction process risk assessment method and system, specifically related to building construction process risk assessment field, including by establishing the index of measuring point level of building construction process and synchronous arrangement time series displacement data, the spatial gradient evolution sequence is obtained by executing difference calculation, it is extracted from the trend change interval set mapped as path node set by sliding window processing, and the topological connection relationship is constructed, and the initial risk path atlas is formed;By abnormal identification to the data state of each path node in risk path atlas, the access control effect of abnormal node in path is analyzed, whether it constitutes path fracture risk is determined in combination with sensitive monitoring area. By constructing the risk atlas structure based on trend path, in combination with path break node identification and alternative path dynamic deduction mechanism, the continuous evaluation and atlas self-correction of risk trend are realized under the condition of measuring point data failure.

Description

Technical Field

[0001] This invention relates to the field of risk assessment technology for building construction processes, and more specifically, to a method and system for risk assessment of building construction processes based on a fault-tolerant mechanism. Background Technology

[0002] In deep foundation pit support engineering during building construction, inclinometers serve as key slope deformation monitoring equipment, and their multi-layered distributed measuring point data constitute an important input basis for risk assessment.

[0003] However, in actual construction, inclinometers are easily affected by mud erosion, structural disturbance or sensor aging, which can cause sudden failures or data drift at a certain layer or local measuring point. Once the measuring point data fails, traditional risk assessment methods often cannot respond effectively to the missing data due to the lack of fault tolerance mechanism, which can lead to abnormal model fitting and distorted trend judgment, ultimately causing deviations or delays in slope risk identification.

[0004] Therefore, there is currently a lack of a risk assessment mechanism with fault tolerance capabilities that can identify and correct structural data gaps caused by the failure of measuring points during the construction process. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a risk assessment method and system for building construction processes based on a fault-tolerant mechanism. By constructing a risk map structure based on trend paths and combining a path interruption node identification and alternative path dynamic deduction mechanism, the method achieves continuous assessment of risk trends and self-correction of the map under the condition of measurement point data failure, thereby solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a risk assessment method for building construction process based on a fault-tolerant mechanism, comprising:

[0007] By establishing a hierarchical index of measuring points during the building construction process and simultaneously organizing time-series displacement data, a spatial gradient evolution sequence is obtained by performing differential calculations. The set of trend change intervals extracted by the sliding window processing is mapped to a set of path nodes, and their topological connection relationships are constructed to form an initial risk path map.

[0008] By identifying anomalies in the data status of each path node in the risk path map, analyzing the access control role of the abnormal nodes in the path, and combining the sensitive monitoring area to determine whether a path breakage risk is constituted, candidate nodes for path breakage are screened out.

[0009] Based on the candidate nodes of path breakage, candidate paths with structural consistency exceeding a set threshold are retrieved in the risk path map, and alternative paths are determined after judging their trend output and error offset.

[0010] By extracting offset feedback from the corrected path and recording path evolution behavior, the label confidence and node usage frequency in the risk path graph are updated, and the path activation strategy is adjusted.

[0011] In a preferred embodiment, construction monitoring data is acquired during the building construction process, and structural monitoring data related to the foundation support stage is extracted from the construction monitoring data. The structural monitoring data includes the measuring point number of each monitoring layer, structural depth information, time-series displacement data continuously recorded during the construction process and its timestamp, and a measuring point hierarchical index sequence is established through the measuring point layout scheme and the hierarchical positioning relationship in the construction drawings.

[0012] The time-series displacement data is synchronously organized according to the measurement point hierarchical index sequence to obtain the displacement record sequence corresponding to each level of measurement point. Then, the measurement point hierarchical index sequence and the displacement record sequence are structurally matched, and gradient difference operation between adjacent levels is performed to solve the initial spatial gradient evolution sequence. The directional change is synchronously marked in each time step.

[0013] In a preferred embodiment, the spatial gradient evolution sequence is divided into sliding window segments in the time dimension, and within each window, trend direction aggregation, perturbation slope statistics and continuity judgment are performed to extract the set of trend change intervals.

[0014] The set of trend change intervals is compared with the set of risk evolution samples marked during historical construction periods. Trend label mapping sequences are generated based on the similarity of trend shape and time rhythm.

[0015] The trend label mapping sequence is processed into nodes according to hierarchical order to generate a path node set. The path nodes in the path node set include the measurement point hierarchical position, trend status label, direction offset mark and time window positioning information. Based on predefined connection rules, dependency directionality and topological constraints between nodes, an ordered connection relationship between the path node sets is constructed to form an initial risk path map. The path in the risk path map is then verified for node accessibility and loop structure is removed.

[0016] In a preferred embodiment, the sequence of path nodes contained in each path in the established risk path map is read sequentially, and the real-time data update status of the corresponding measurement points of the path nodes is retrieved.

[0017] In the real-time data update state, if there is an abnormal behavior such as interruption, sudden value or long-term stable value of the measurement point data, the node of the measurement point data is marked as a data status abnormal node, and its corresponding time window and hierarchical position are added to the abnormal node set.

[0018] Perform path dependency resolution on the set of abnormal nodes to determine whether each abnormal node in the set is at the target connection position in the current path. The target connection position refers to the access control node that cannot be bypassed in the path.

[0019] In a preferred embodiment, if the path dependency analysis results indicate that the abnormal node is in a position of access control in the path, that is, its absence will cause the nodes before and after to be disconnected, then the abnormal node is determined to meet the necessary conditions for path accessibility, and then it is regarded as a candidate node for path breakage, and the node position structure hierarchy analysis is performed to confirm whether it is located in the preset sensitive monitoring area.

[0020] The sensitive monitoring area consists of prior geological conditions and areas of active historical deformation. Prior geological conditions include anomalous geological strata such as soft interlayers and structural turning points, while areas of active historical deformation include strata with high frequency of historical cumulative deformation and areas frequently disturbed by construction.

[0021] If a candidate node for path breakage meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, then the node is upgraded to a priority replacement evaluation node, and its position code and path number in the path structure are recorded.

[0022] In a preferred embodiment, for each priority alternative evaluation node, the risk path map is searched for a set of paths that do not contain the node but whose path node sequences have a structural consistency higher than a preset similarity threshold in the start-end level feature dimension, trend state distribution feature dimension, and connection density feature dimension, to obtain an initial path migration candidate set.

[0023] Perform path node sequence alignment operation on the path migration candidate set, and calculate its structural consistency index with the original path in terms of trend state sequence, hierarchical span and connection density;

[0024] If there are candidate paths with a structural consistency index higher than the set threshold, they will be included in the equivalent path set and their data integrity will be checked to verify whether there are any missing data points in the current monitoring period.

[0025] In a preferred embodiment, a trend output comparison operation is performed on all complete equivalent paths of data to analyze whether the trend output sequence of the target time period is consistent with the output trend of the original path.

[0026] If the output trend is consistent, the equivalent path is marked as an alternative path. If the output trend deviates, it is determined whether the deviation of the output trend is within the error tolerance range of the preset risk level. If it meets the error tolerance range, it is marked as a correction path and an offset correction factor is added as a candidate risk evolution path.

[0027] In a preferred embodiment, the offset correction factor of each correction path is read and compared with the weight offset threshold set in the risk assessment mechanism to determine whether the offset degree of the correction path is within the weight offset threshold limit.

[0028] If the offset correction factor is less than the preset weight offset threshold, the correction path will be included in the set of path sequences currently participating in risk trend judgment, and the corresponding paragraph nodes that have failed in the original path will be replaced with it.

[0029] Input the main assessment path after the modified path is replaced into the risk trend judgment process to calculate the continuous risk level assessment results for the corresponding time period.

[0030] The path output trend before and after correction are compared to find the offset vector caused by the path replacement. The path number, replacement position and time window associated with the offset vector are recorded.

[0031] In a preferred embodiment, the offset vector and the evaluation error result are jointly written into the map evolution record set to generate an evaluation offset trajectory file that reflects the dynamics of path replacement.

[0032] The confidence level of trend labels and the frequency of node use in the risk path map are updated by using the assessment offset trajectory archive to obtain the map evolution status in the current period;

[0033] The updated graph evolution state is input into the path selection rule set to dynamically adjust the path activation strategy in the evaluation model, thereby achieving a continuous optimization closed loop under path structure fault tolerance.

[0034] A risk assessment system for building construction process based on fault tolerance mechanism includes a path construction module, a fracture identification module, an alternative inference module, and a map update module;

[0035] The path construction module establishes a hierarchical index of measuring points during the construction process and simultaneously organizes temporal displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence, maps the set of trend change intervals extracted by sliding window processing to a set of path nodes, and constructs their topological connection relationships to form an initial risk path map.

[0036] The fracture identification module identifies anomalies in the data status of each path node in the risk path map, analyzes the access control role of the abnormal nodes in the path, and determines whether a path fracture risk is constituted by combining sensitive monitoring areas, thus screening out candidate nodes for path fracture.

[0037] The alternative inference module retrieves candidate paths with structural consistency exceeding a set threshold in the risk path map based on path break candidate nodes, and determines the alternative path after judging its trend output and error offset.

[0038] The map update module extracts offset feedback from corrected paths and records path evolution behavior to update label confidence and node usage frequency in the risk path map, and adjusts path activation strategies.

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

[0040] 1. By constructing a risk path map and introducing a fault node identification and path migration mechanism, an equivalent trend path can still be dynamically generated when the inclinometer monitoring data fails. This effectively makes up for the shortcomings of traditional models that cannot continuously assess risks under the condition of missing measurement point data, thereby solving the problem of delayed slope risk identification caused by lack of fault tolerance.

[0041] 2. By constructing a hierarchical index of measuring points and organizing time-series displacements from structural monitoring data, performing gradient difference and combining it with time window trend aggregation, a high-density trend representation of spatial deformation of the foundation support structure is achieved, improving the real-time performance of trend identification.

[0042] 3. By performing abnormal pattern identification on the data update status of path nodes, and combining path dependency analysis and sensitive area location mechanism, candidate nodes for break are screened to ensure that only key locations with the potential to block risk propagation enter the fault-tolerant path replacement process.

[0043] 4. By constructing a structural consistency index and trend output comparison mechanism, alternative paths with matching structural features and controllable trend deviations are screened, thereby realizing the substitution of risk trend paths and making the assessment results continuous. Attached Figure Description

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

[0045] Figure 2 This is a flowchart for constructing the risk path map of the present invention.

[0046] Figure 3 This is a flowchart of the path break node identification process of the present invention.

[0047] Figure 4 This is a flowchart illustrating the path substitution deduction process of the present invention.

[0048] Figure 5 This is a flowchart of the map update and risk reassessment process of the present invention.

[0049] Figure 6 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Refer to the instruction manual appendix Figure 1-6 An embodiment of the present invention provides a risk assessment method for building construction processes based on a fault-tolerant mechanism, comprising:

[0052] By establishing a hierarchical index of measuring points during the building construction process and simultaneously organizing time-series displacement data, a spatial gradient evolution sequence is obtained by performing differential calculations. The set of trend change intervals extracted by the sliding window processing is mapped to a set of path nodes, and their topological connection relationships are constructed to form an initial risk path map.

[0053] By identifying anomalies in the data status of each path node in the risk path map, analyzing the access control role of the abnormal nodes in the path, and combining the sensitive monitoring area to determine whether it constitutes a path breakage risk, candidate nodes for path breakage are screened out for subsequent fault tolerance simulation process.

[0054] Based on the candidate nodes of path breakage, candidate paths with structural consistency exceeding a set threshold are retrieved in the risk path map. After judging their trend output and error offset, alternative paths are determined to restore the complete chain of risk assessment.

[0055] By extracting the offset feedback of the corrected path and recording the path evolution behavior, the label confidence and node usage frequency in the risk path graph are updated, the path activation strategy is adjusted, and a closed loop of graph update based on fault tolerance feedback is completed.

[0056] During the construction process, construction monitoring data is acquired, and structural monitoring data related to the foundation support stage is extracted from the construction monitoring data. The structural monitoring data includes the measuring point number of each monitoring layer, structural depth information, time-series displacement data and its timestamps continuously recorded during the construction process, and a measuring point hierarchical index sequence is established through the measuring point layout plan and the hierarchical positioning relationship in the construction drawings. The construction monitoring data includes various types of continuous monitoring information that are dynamically related to the structural response, such as inclination data, settlement data, stress and strain data, and construction progress information.

[0057] The time-series displacement data is synchronously organized according to the measurement point hierarchical index sequence to obtain the displacement record sequence corresponding to each level of measurement point. Then, the measurement point hierarchical index sequence and the displacement record sequence are structurally matched, and gradient difference operation between adjacent levels is performed to solve the initial spatial gradient evolution sequence. The directional change is synchronously marked in each time step.

[0058] The spatial gradient evolution sequence is divided into sliding window segments in the time dimension, and within each window, trend direction aggregation, perturbation slope statistics and continuity judgment are performed to extract the set of trend change intervals;

[0059] The set of trend change intervals is compared with the set of risk evolution samples marked during historical construction periods. Trend label mapping sequences are generated based on the similarity of trend shape and time rhythm.

[0060] The trend label mapping sequence is processed into nodes according to hierarchical order to generate a path node set. The path nodes in the path node set include the measurement point hierarchical position, trend status label, direction offset mark and time window positioning information. Based on predefined connection rules, dependency directionality and topological constraints between nodes, an ordered connection relationship between the path node sets is constructed to form an initial risk path map. The path in the risk path map is then verified for node accessibility and loop structure is removed.

[0061] For each path in the established risk path map, read the sequence of path nodes it contains in sequence, and retrieve the real-time data update status of the corresponding measurement points of the path nodes.

[0062] In the real-time data update state, if there is an abnormal behavior such as interruption, sudden value or long-term stable value of the measurement point data, the node of the measurement point data is marked as a data status abnormal node, and its corresponding time window and hierarchical position are added to the abnormal node set.

[0063] Perform path dependency resolution on the set of abnormal nodes to determine whether each abnormal node in the set is at the target connection position in the current path. The target connection position refers to the access control node that cannot be bypassed in the path.

[0064] If the path dependency analysis results indicate that the abnormal node is in a position of access control in the path, that is, its absence will cause the nodes before and after to be disconnected, then the abnormal node is determined to meet the necessary conditions for path accessibility. Then it is regarded as a candidate node for path breakage, and the node position structure hierarchy analysis is performed to confirm whether it is located in the preset sensitive monitoring area.

[0065] The sensitive monitoring area consists of prior geological conditions and areas of active historical deformation. Prior geological conditions include anomalous geological strata such as soft interlayers and structural turning points, while areas of active historical deformation include strata with high frequency of historical cumulative deformation and areas frequently disturbed by construction.

[0066] If a candidate node for path breakage meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, then the node is upgraded to a priority replacement evaluation node, and its position code and path number in the path structure are recorded.

[0067] For each priority alternative evaluation node, the risk path map is searched for a set of paths that do not contain that node but whose path node sequences have a structural consistency higher than a preset similarity threshold in the start-end level feature dimension, trend state distribution feature dimension, and connection density feature dimension, thus obtaining an initial path migration candidate set.

[0068] Perform path node sequence alignment operation on the path migration candidate set, and calculate its structural consistency index with the original path in terms of trend state sequence, hierarchical span and connection density;

[0069] If there are candidate paths with a structural consistency index higher than the set threshold, they will be included in the equivalent path set and their data integrity will be checked to verify whether there are any missing data points in the current monitoring period.

[0070] Perform a trend output comparison operation on all complete equivalent paths of data to analyze whether the trend output sequence within the target time period is consistent with the output trend of the original path.

[0071] If the output trend is consistent, the equivalent path is marked as an alternative path. If the output trend deviates, it is determined whether the deviation of the output trend is within the error tolerance range of the preset risk level. If it meets the error tolerance range, it is marked as a correction path and an offset correction factor is added as a candidate risk evolution path.

[0072] For each correction path, its offset correction factor is read and compared with the weight offset threshold set in the risk assessment mechanism to determine whether the degree of offset of the correction path is within the weight offset threshold limit.

[0073] If the offset correction factor is less than the preset weight offset threshold, the correction path will be included in the set of path sequences currently participating in risk trend judgment, and the corresponding paragraph nodes that have failed in the original path will be replaced with it.

[0074] Input the main assessment path after the modified path is replaced into the risk trend judgment process to calculate the continuous risk level assessment results for the corresponding time period.

[0075] The path output trend before and after correction are compared to find the offset vector caused by the path replacement. The path number, replacement position and time window associated with the offset vector are recorded.

[0076] The offset vector and the evaluation error results are jointly written into the map evolution record set to generate an evaluation offset trajectory file that reflects the dynamics of path replacement.

[0077] The confidence level of trend labels and the frequency of node use in the risk path map are updated by using the assessment offset trajectory archive to obtain the map evolution status in the current period;

[0078] The updated graph evolution state is input into the path selection rule set to dynamically adjust the path activation strategy in the evaluation model, thereby achieving a continuous optimization closed loop under path structure fault tolerance.

[0079] A risk assessment system for building construction process based on fault tolerance mechanism includes a path construction module, a fracture identification module, an alternative inference module, and a map update module;

[0080] The path construction module establishes a hierarchical index of measuring points during the construction process and simultaneously organizes temporal displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence, maps the set of trend change intervals extracted by sliding window processing to a set of path nodes, and constructs their topological connection relationships to form an initial risk path map.

[0081] The fracture identification module identifies anomalies in the data status of each path node in the risk path map, analyzes the access control role of the abnormal nodes in the path, and determines whether a path fracture risk is constituted by combining the sensitive monitoring area. It then selects candidate nodes for path fracture for use in the subsequent fault tolerance simulation process.

[0082] The alternative inference module searches for candidate paths with structural consistency exceeding a set threshold in the risk path map based on the candidate nodes of path breakage, and determines the alternative paths after judging their trend output and error offset, so as to restore the complete chain of risk assessment.

[0083] The graph update module extracts the offset feedback of the correction path and records the path evolution behavior, updates the label confidence and node usage frequency in the risk path graph, adjusts the path activation strategy, and completes the graph update closed loop based on fault tolerance feedback.

[0084] The following points need further clarification regarding the above plan:

[0085] Firstly, 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.

[0086] 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 can form 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.

[0087] In this scheme, constants, weights, adjustment factors, threshold parameters, proportional coefficients, etc., are all adjustable control parameters for different application environments. Their values ​​depend on the target equipment configuration, data input characteristics, and performance optimization goals. During the implementation phase, they are converged within a reasonable range through model verification, performance constraints, or engineering calibration. Although these parameters do not have a unique preset value, they have clear adjustment logic and calculation paths. They belong to the deterministic setting process in engineering implementation. The purpose of this setting is to ensure that the scheme is both universally adaptable and reproducible and operable, without affecting its technical clarity and feasibility.

[0088] Based on structural monitoring data during the construction phase, trend evolution nodes are constructed according to the measurement point hierarchy and time window. Trend nodes are generated through nonlinear gradient and disturbance aggregation, and structural connections are executed to form an initial risk path map, which is then formulated. This is the initial risk path map, which is a connected graph structure composed of trend nodes;

[0089]

[0090] Where i is the index of the measuring point in the vertical hierarchy, ranging from 1 ≤ i ≤ N, and N is the total number of measuring point hierarchies; j is the time window index, ranging from 1 ≤ j ≤ T, and T is the total number of time steps in all construction stages; D i,j The displacement record sequence of the i-th layer measuring point within the time window j; t j h is the center time of the j-th time window; i M represents the spatial depth value of the i-th layer measurement point; i,j Let α be the perturbation direction tensor of the i-th layer measurement point at time j. The perturbation direction tensor includes indices such as the number of inflection points, the direction of fluctuation, and the rate of change of slope; i,jThis indicates the directional aggregation index of the node within its time window; the directional aggregation index reflects the directional concentration. Here, the perturbation label mapping function represents the trend state after clustering from the input perturbation tensor. This is a trend gradient aggregation operator under a sliding time window, which is used to test the stability of clustering features within the trend window. This involves mapping the trend node state encoding to the label state combination. This represents the node connection operation in the path graph structure, which is used to construct the relationship between graph edges and nodes;

[0091] Additionally, in the formula Partially representing the second derivative of the displacement of the i-th layer measuring point at time j with respect to time, used to reflect the acceleration of trend changes and capture sudden changes in trend; This represents the rate of change of displacement of the measuring point with respect to depth, reflecting the strain gradient under the structural level; log(1+·) indicates that the nonlinear response with high gradient changes is enhanced by using a logarithmic function, while compressing low-frequency background disturbances. This indicates that the perturbation tensor is mapped to node labels to classify trend types (such as "rising unstable" / "falling stable"). This indicates calculating the time gradient of the directional aggregation indicator within the window to ensure the stability of trend nodes, with the outer two-layer union symbol. and This indicates that trend nodes across all levels and time periods are integrated to generate a graph.

[0092] By analyzing the temporal abrupt changes in the node displacement function in the path map, the path interruption connectivity analysis, and the sensitive area discrimination, candidate nodes for path fractures that constitute risky structural fractures are screened.

[0093]

[0094] in v is the set of candidate nodes for path breakage. k D represents the k-th trend node in the risk path graph. k (t) represents node v k The continuous-time displacement function of the represented measuring point; v k The infinitesimal neighborhood before and after the corresponding time point; ζ k The accessibility control factor is set to 1 if a node failure causes a path to be disconnected, and 0 otherwise; Θ k This is a sensitive indicator; a positive value is given if the node is located in an anomalous stratum or a historically highly variable area; δ bThis represents the boundary threshold that constitutes the fracture determination, and in practical applications, it is set as a tolerance index jointly determined by the site and the structure; t represents a continuous time variable.

[0095] for Explanation of the components in the formula:

[0096] in This represents the degree of abrupt change in the trend time series at that node, i.e., the instantaneous discontinuity of the time slope; ζ k ·Θ k This represents the fracture contribution weight after the structural control strength and the risk area are superimposed; it is then summed with δ after all terms are weighted. b Comparison serves as a criterion for determining whether a node constitutes a break;

[0097] By matching and evaluating candidate paths in the structural map that have similar topological structures to the fracture path and whose risk trend output error is tolerable, an alternative set of risk paths is constructed.

[0098]

[0099] in p is the set of alternative paths derived from path substitution analysis. m Candidate paths refer to path segments in the path graph that retain structural integrity after removing path segments containing candidate fracture nodes; φ m,d (τ) is p m Risk trend response function over time τ; ψ d (τ) is the trend function of the original fracture path over time τ; τ s ,τ e These represent the start and end times of the trend comparison period, respectively; κ l The risk trend cosine similarity threshold; ∈ d This is the tolerance for the integral error of the third-order trend offset;

[0100] exist The numerator of the first term in the formula represents the risk trend dot product of the candidate path and the original path; in The denominator of the first term in the formula represents the energy modulus of the risk trends of the two factors (a complete expression of cosine similarity); in The second term in the formula is the third-order time integral of the trend difference, which is used to measure the cumulative sensitivity of nonlinear differences.

[0101] Based on the sensitivity response of node confidence and activation density calculated by alternative path offset feedback, weight updates are implemented in the path graph structure to form a closed loop for the optimization of the evaluation strategy.

[0102]

[0103] in The updated risk path map; For the initial risk path map; p m To correct the path; v k To correct the nodes in the path; ω k Use frequency weights for the original nodes in the graph; π k (Δ k ) is node v k The risk level output function depends on the trend offset; Δ k The degree of trend deviation of this node between the alternative path and the original path; ρ k (t) represents the temporal activity density of node risk output; θ ρ For trend deviation tolerance; This is a logical judgment function. When the condition within the parentheses of the logical judgment function is met, the value is 1.

[0104] The risk assessment method and system for building construction process based on fault tolerance mechanism proposed in this invention is designed to solve key problems in current deep foundation pit support engineering, such as interruption of risk assessment, misjudgment of trend model, and lag in overall structural safety identification caused by the failure of measuring point data.

[0105] In actual construction processes, especially in complex scenarios such as deep foundation pits, diaphragm walls, and reverse construction methods, construction monitoring systems typically deploy a large number of inclinometers, settlement gauges, and stress-strain sensing devices to build a dynamic cognitive mechanism for structural responses. However, due to the extreme complexity of the underground environment, monitoring points are prone to single-point failures, abnormal drifts, or sudden data gaps under construction disturbances, mud erosion, and sensor aging. Such failures are sudden and highly concealed, often not detectable in a timely manner through hardware malfunctions. Traditional risk assessment methods mostly rely on data integrity and lack dynamic compensation mechanisms to address gaps in structural data. Therefore, once monitoring point data is missing, the model will be unable to accurately reconstruct risk trends, ultimately leading to failed safety warnings or even misjudgments of working conditions, resulting in decision-making risks.

[0106] This solution is designed based on this technical bottleneck, proposing to use a "path graph" as the core modeling object, integrating a four-level process of "trend state expression, fault-tolerant reconstruction, structural equivalent transfer, and feedback closed-loop update" to form a complete fault-tolerant risk assessment chain. The method is divided into four functional parts, each of which is closely connected and uses consistent terminology to ensure a closed-loop structure and strong dynamic adaptability.

[0107] 1. Constructing a Risk Path Map: In this stage, structural monitoring data is extracted from construction monitoring data, and a hierarchical index of monitoring points is established based on the hierarchical and temporal relationships of the monitoring points. Then, gradient difference operations are performed on the temporal displacement data corresponding to each level to capture its spatial evolution trend. Trend change behavior is extracted within the sliding time window to form a "set of trend change intervals". These trend segments are then represented as nodes, and path connection relationships are constructed based on structural topology and dependency constraints to finally form an "initial risk path map". This map abstracts and encapsulates all identifiable risk trend evolution chains in a graph structure, providing a trend-level structural foundation for subsequent analysis.

[0108] 2. Identifying Path Disruption Nodes: This section traverses each path and its associated trend nodes from the constructed risk path map, detecting the real-time data status of corresponding measurement points, identifying any data interruptions, abrupt changes, or abnormally stable behavior patterns, and marking them as "abnormal data status nodes." Subsequently, based on the path dependency structure, it determines whether the abnormal node is located in an unavoidable "access control location." If the failure of this node would cause the path to become disconnected, it is identified as a "path break candidate node." Further, based on the definition of "sensitive monitoring areas" (including soft interlayers, structural turning zones, and historically active deformation layers), break nodes "with risk significance" are selected as target points for subsequent alternative simulations. This process ensures that not all faults enter the compensation process, but only "core interruptions in a structural sense" are addressed.

[0109] 3. Execution Path Substitution Inference: For each candidate node of path breakage, the system will search for feasible paths that are "structurally similar and have complete data" in the risk path map. The judgment criteria cover three-dimensional structural features such as the hierarchical distribution of path nodes, trend state sequence, and connection density, forming a "path migration candidate set". Then, the candidate paths and the original paths are matched and analyzed on the risk trend output sequence, and their trend consistency (such as cosine similarity) and nonlinear error offset (such as the third integral of the trend difference) are calculated. Finally, the path that is "structurally substitutable and has tolerable output error" is selected as the "corrected path". This process is not a forced fitting of the original path data, but rather the establishment of a new risk evolution channel through path migration, thereby ensuring that the assessment chain is not interrupted.

[0110] 4. Update the evaluation structure and graph weights: After the corrected path is enabled, the system replaces the failed segments in the original path with it, and uses its risk output results as part of the formal trend judgment results. It calculates the output offset vector before and after the replacement, and writes this vector, along with the path segments, time windows, and node positions, into the "graph evolution record". Based on this, the usage frequency of all path nodes is updated, the trend label confidence is adjusted, and the graph structure state is reconstructed accordingly, thereby realizing the dynamic update of the "path enabling strategy". This part is the system's learning loop, making fault tolerance not a one-time "replacement" but a continuous "self-optimization evolution".

[0111] This design not only solves the problem of "risk trend path breakage" from the perspective of structural graph, but also innovatively introduces the mechanisms of "structural equivalent transfer" and "trend evolution feedback", which enable the risk assessment process to dynamically replace failure paths, adaptively select the optimal risk link, and continuously learn the evolution of graph state.

[0112] In practical applications, this method is applicable to construction environments that require in-depth monitoring, such as underground engineering, bridge foundations, and subway enclosures. It can also be seamlessly integrated with BIM platforms and digital twin monitoring systems to achieve fault-tolerant assessment and real-time adjustment of multi-path risk trends. It is particularly suitable for project scenarios where "node failure is inevitable and data sampling is frequently interrupted." Its multi-path recovery capability is significantly better than that of traditional linear prediction models, and it can significantly improve the continuity of overall risk perception and the ability to avoid misjudgments.

[0113] 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 construction process based on fault tolerance mechanism, characterized in that, include: By establishing a hierarchical index of measuring points during the building construction process and simultaneously organizing time-series displacement data, a spatial gradient evolution sequence is obtained by performing differential calculations. The set of trend change intervals extracted by the sliding window processing is mapped to a set of path nodes, and their topological connection relationships are constructed to form an initial risk path map. By identifying anomalies in the data status of each path node in the risk path map, analyzing the access control role of the abnormal nodes in the path, and combining the sensitive monitoring area to determine whether a path breakage risk is constituted, candidate nodes for path breakage are screened out. Based on the candidate nodes of path breakage, candidate paths with structural consistency exceeding a set threshold are retrieved in the risk path map, and alternative paths are determined after judging their trend output and error offset. By extracting offset feedback from the corrected path and recording path evolution behavior, the label confidence and node usage frequency in the risk path graph are updated, and the path activation strategy is adjusted.

2. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 1, characterized in that: During the construction process, construction monitoring data is acquired, and structural monitoring data related to the foundation support stage is extracted from the construction monitoring data. The structural monitoring data includes the measuring point number of each monitoring layer, structural depth information, time-series displacement data continuously recorded during the construction process and its timestamp. A measuring point hierarchical index sequence is established through the measuring point layout scheme and the hierarchical positioning relationship in the construction drawings. The time-series displacement data is synchronously organized according to the measurement point hierarchical index sequence to obtain the displacement record sequence corresponding to each level of measurement point. Then, the measurement point hierarchical index sequence and the displacement record sequence are structurally matched, and gradient difference operation between adjacent levels is performed to solve the initial spatial gradient evolution sequence. The directional change is synchronously marked in each time step.

3. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 2, characterized in that: The spatial gradient evolution sequence is divided into sliding window segments in the time dimension, and within each window, trend direction aggregation, perturbation slope statistics and continuity judgment are performed to extract the set of trend change intervals; The set of trend change intervals is compared with the set of risk evolution samples marked during historical construction periods. Trend label mapping sequences are generated based on the similarity of trend shape and time rhythm. The trend label mapping sequence is processed into nodes according to hierarchical order to generate a path node set. The path nodes in the path node set include the measurement point hierarchical position, trend status label, direction offset mark and time window positioning information. Based on predefined connection rules, dependency directionality and topological constraints between nodes, an ordered connection relationship between the path node sets is constructed to form an initial risk path map. The path in the risk path map is then verified for node accessibility and loop structure is removed.

4. The risk assessment method for building construction process based on a fault-tolerant mechanism according to claim 3, characterized in that: For each path in the established risk path map, read the sequence of path nodes it contains in sequence, and retrieve the real-time data update status of the corresponding measurement points of the path nodes. In the real-time data update state, if there is an abnormal behavior such as interruption, sudden value or long-term stable value of the measurement point data, the node of the measurement point data is marked as a data status abnormal node, and its corresponding time window and hierarchical position are added to the abnormal node set. Perform path dependency resolution on the set of abnormal nodes to determine whether each abnormal node in the set is at the target connection position in the current path. The target connection position refers to the access control node that cannot be bypassed in the path.

5. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 4, characterized in that: If the path dependency analysis results indicate that the abnormal node is in a position of access control in the path, that is, its absence will cause the nodes before and after to be disconnected, then the abnormal node is determined to meet the necessary conditions for path accessibility. Then it is regarded as a candidate node for path breakage, and the node position structure hierarchy analysis is performed to confirm whether it is located in the preset sensitive monitoring area. The sensitive monitoring area consists of prior geological conditions and areas of active historical deformation. Prior geological conditions include anomalous geological strata such as soft interlayers and structural turning points, while areas of active historical deformation include strata with high frequency of historical cumulative deformation and areas frequently disturbed by construction. If a candidate node for path breakage meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, then the node is upgraded to a priority replacement evaluation node, and its position code and path number in the path structure are recorded.

6. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 5, characterized in that: For each priority alternative evaluation node, the risk path map is searched for a set of paths that do not contain that node but whose path node sequences have a structural consistency higher than a preset similarity threshold in the start-end level feature dimension, trend state distribution feature dimension, and connection density feature dimension, thus obtaining an initial path migration candidate set. Perform path node sequence alignment operation on the path migration candidate set, and calculate its structural consistency index with the original path in terms of trend state sequence, hierarchical span and connection density; If there are candidate paths with a structural consistency index higher than the set threshold, they will be included in the equivalent path set and their data integrity will be checked to verify whether there are any missing data points in the current monitoring period.

7. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 6, characterized in that: Perform a trend output comparison operation on all complete equivalent paths of data to analyze whether the trend output sequence within the target time period is consistent with the output trend of the original path. If the output trend is consistent, the equivalent path is marked as an alternative path. If the output trend deviates, it is determined whether the deviation of the output trend is within the error tolerance range of the preset risk level. If it meets the error tolerance range, it is marked as a correction path and an offset correction factor is added as a candidate risk evolution path.

8. The risk assessment method for building construction process based on fault tolerance mechanism according to claim 7, characterized in that: For each correction path, its offset correction factor is read and compared with the weight offset threshold set in the risk assessment mechanism to determine whether the degree of offset of the correction path is within the weight offset threshold limit. If the offset correction factor is less than the preset weight offset threshold, the correction path will be included in the set of path sequences currently participating in risk trend judgment, and the corresponding paragraph nodes that have failed in the original path will be replaced with it. Input the main assessment path after the modified path is replaced into the risk trend judgment process to calculate the continuous risk level assessment results for the corresponding time period. The path output trend before and after correction are compared to find the offset vector caused by the path replacement. The path number, replacement position and time window associated with the offset vector are recorded.

9. A risk assessment method for building construction process based on a fault-tolerant mechanism according to claim 8, characterized in that: The offset vector and the evaluation error results are jointly written into the map evolution record set to generate an evaluation offset trajectory file that reflects the dynamics of path replacement. The confidence level of trend labels and the frequency of node use in the risk path map are updated by using the assessment offset trajectory archive to obtain the map evolution status in the current period; The updated graph evolution state is input into the path selection rule set to dynamically adjust the path activation strategy in the evaluation model, thereby achieving a continuous optimization closed loop under path structure fault tolerance.

10. A risk assessment system for building construction process based on a fault-tolerant mechanism, comprising a path construction module, a fracture identification module, an alternative deduction module, and a map update module, characterized in that: The path construction module establishes a hierarchical index of measuring points during the construction process and simultaneously organizes temporal displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence, maps the set of trend change intervals extracted by sliding window processing to a set of path nodes, and constructs their topological connection relationships to form an initial risk path map. The fracture identification module identifies anomalies in the data status of each path node in the risk path map, analyzes the access control role of the abnormal nodes in the path, and determines whether a path fracture risk is constituted by combining sensitive monitoring areas, thus screening out candidate nodes for path fracture. The alternative inference module retrieves candidate paths with structural consistency exceeding a set threshold in the risk path map based on path break candidate nodes, and determines the alternative path after judging its trend output and error offset. The map update module extracts offset feedback from corrected paths and records path evolution behavior to update label confidence and node usage frequency in the risk path map, and adjusts path activation strategies.

Images