Building construction process risk assessment method and system based on fault tolerance mechanism

Abstract

The invention discloses a building construction process risk assessment method and system based on a fault tolerance mechanism, and particularly relates to the field of building construction process risk assessment, and the method comprises the steps: building a measurement point hierarchical index of a building construction process, synchronously arranging time sequence displacement data, and executing difference calculation to obtain a space gradient evolution sequence; mapping a trend change interval set extracted by sliding window processing into a path node set, and constructing a topological connection relation of the path node set to form an initial risk path map; the data state of each path node in the risk path map is subjected to anomaly identification, the access control effect of the abnormal node in the path is analyzed, and whether the path fracture risk is formed or not is judged in combination with the sensitive monitoring area. By constructing a risk map structure based on a trend path and combining a path interruption node identification and alternative path dynamic deduction mechanism, continuous assessment and map self-correction of the risk trend under the condition of measuring point data failure are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of risk assessment in a construction process, and more particularly to a method and system for risk assessment in a construction process based on a fault tolerance mechanism. Background Art

[0002] In deep foundation pit support projects during construction, inclinometers are key slope deformation monitoring equipment. The data from their multi-layered distribution of measurement points form an important input basis for risk assessment.

[0003] However, during actual construction, inclinometers are extremely susceptible to mud erosion, structural disturbances, or sensor aging, which can lead to sudden failures or data drift at certain layers or local measuring points. Once the measuring point data fails, traditional risk assessment methods, lacking a fault-tolerance mechanism, often fail to effectively respond to the missing data, leading to model fitting anomalies and distorted trend judgments, ultimately causing deviations or delays in slope risk identification.

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

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a construction process risk assessment method and system based on a fault-tolerant mechanism. By constructing a risk map structure based on trend paths, combined with path interruption node identification and alternative path dynamic deduction mechanism, continuous assessment of risk trends and self-correction of the map are achieved under the condition of measurement point data failure, so as to solve the problems raised in the above-mentioned background technology.

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

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

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

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

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

[0011] In a preferred embodiment, construction monitoring data is acquired during the 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 measurement point number of each monitoring layer, structural depth information, time-series displacement data continuously recorded during the construction process and its timestamp. A measurement point hierarchical index sequence is established based on the hierarchical positioning relationship in the measurement point layout plan and the construction drawings.

[0012] The time-series displacement data are synchronously sorted according to the measuring point level index sequence to obtain the displacement record sequence corresponding to the measuring points at each level. The measuring point level index sequence is then structured matched with the displacement record sequence, and gradient difference operations are performed between adjacent levels to solve the initial spatial gradient evolution sequence. Its directional changes are synchronously marked in each time step.

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

[0014] Perform a morphological comparison between the trend change interval set and the risk evolution sample set that has been annotated during the historical construction period, and generate a trend label mapping sequence based on the trend shape similarity and time rhythm similarity matching;

[0015] The trend label mapping sequence is node-based in hierarchical order to generate a path node set. The path nodes in the path node set include the hierarchical position of the measuring point, trend status label, direction offset mark and time window positioning information. According to the predefined connection rules, dependency directionality and topological constraints between nodes, the ordered connection relationship between the path node sets is constructed to form an initial risk path map. The paths in the risk path map are then verified for node accessibility and ring structure is eliminated.

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

[0017] In the real-time data update state, if there is an interruption in the measurement point data or abnormal behavior of sudden changes or long-term stable values, the node of the measurement point data will be marked as a data state abnormal node, and its corresponding time window and hierarchical position will be added to the abnormal node set;

[0018] A path dependency parsing operation is performed on the abnormal node set to determine whether each abnormal node in the abnormal node set is at a target connection position in the current path, where the target connection position refers to a reachable control node that cannot be bypassed in the path.

[0019] In a preferred embodiment, if the results of the path dependency analysis indicate that the abnormal node is in an accessibility control position in the path, that is, its absence will cause the previous and next nodes 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. A node position structure hierarchical analysis is performed to confirm whether it is located in a preset sensitive monitoring area.

[0020] The sensitive monitoring area is composed of prior geological conditions and historical deformation active areas. Prior geological conditions include soft interlayers and geological abnormal layers in structural turning sections. Historical deformation active areas include historical cumulative deformation high-frequency layers and areas with frequent construction disturbances.

[0021] If a path-broken candidate node meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, the node will be upgraded to a priority alternative evaluation node, and its position code and path sequence number in the path structure will be recorded.

[0022] In a preferred embodiment, for each priority replacement assessment node, a set of paths that does not contain the node but whose path node sequence has a structural consistency higher than a preset similarity threshold in the start and end level feature dimension, the trend state distribution feature dimension, and the connection density feature dimension is searched in the risk path map to obtain an initial path migration candidate set;

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

[0024] If there is a candidate path whose structural consistency index is higher than the set threshold, it will be included in the equivalent path set and the monitoring data integrity check will be performed on it to verify whether there are any data missing points in the current monitoring cycle.

[0025] In a preferred embodiment, a trend output comparison operation is performed on all equivalent paths with complete data to analyze whether their trend output sequences within the target time period are consistent with the output trends of the original paths;

[0026] If the output trends are consistent, the equivalent path will be marked as an alternative path. If there is an offset in the output trend, it will be determined whether the offset of the output trend is within the error tolerance range of the preset risk level. If the error tolerance range is met, it will be marked as a correction path and an offset correction factor will be added as a candidate risk evolution path.

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

[0028] If the offset correction factor is less than the preset weight offset threshold, the corrected path is included in the path sequence set currently participating in the risk trend judgment, and is used to replace the corresponding invalid segment node in the original path;

[0029] The main assessment path after the correction path is replaced is input into the risk trend judgment process to calculate the continuous risk level assessment results within the corresponding period;

[0030] Compare the path output trends before and after correction to find the offset vector caused by the path replacement, and record the path number, replacement position, and time window associated with the offset vector.

[0031] In a preferred embodiment, the offset vector and the evaluation error result are jointly written into a graph evolution record set to generate an evaluation offset trajectory archive reflecting the path replacement dynamics;

[0032] Use the assessment deviation trajectory archive to update the trend label confidence and node usage frequency in the risk path map to obtain the map evolution status in the current cycle;

[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, realizing a continuous tuning closed loop under path structure fault tolerance.

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

[0035] The path construction module establishes a hierarchical index of measurement points during the construction process and simultaneously organizes time-series displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence. The set of trend change intervals extracted by sliding window processing is mapped into a set of path nodes, and their topological connections are constructed to form an initial risk path map.

[0036] The fracture identification module identifies abnormalities in the data status of each path node in the risk path map, analyzes the accessibility control role of abnormal nodes in the path, and determines whether there is a path fracture risk in combination with sensitive monitoring areas, and screens out candidate path fracture nodes;

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

[0038] 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, and adjusts the path activation strategy.

[0039] Technical effects and advantages of the present invention:

[0040] 1. By constructing a risk path map and introducing a mechanism for identifying broken nodes and migrating paths, an equivalent trend path can be dynamically generated even when inclinometer monitoring data fails. This effectively addresses the inability of traditional models to continuously assess risks when there are gaps in measurement point data, thereby resolving the issue of delayed slope risk identification caused by a lack of fault tolerance.

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

[0042] 3. By performing abnormal pattern recognition on the data update status of path nodes, combined with path dependency analysis and sensitive area positioning mechanism, candidate fracture nodes 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 building a structural consistency index and trend output comparison mechanism, we screen alternative paths with matching structural features and controllable trend deviations, thereby achieving the replacement of risk trend pathways and ensuring the continuity of assessment results. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0045] Figure 2 Construct a flow chart for the risk path map of the present invention.

[0046] Figure 3 This is a flow chart of path-broken node identification according to the present invention.

[0047] Figure 4 This is a flow chart of the path replacement deduction of the present invention.

[0048] Figure 5 This is the map update and risk reassessment flow chart of the present invention.

[0049] Figure 6 Schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] Refer to the instruction manual Figure 1-6 According to an embodiment of the present invention, a construction process risk assessment method based on a fault tolerance mechanism includes:

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

[0053] By identifying abnormalities in the data status of each path node in the risk path map, analyzing the access control role of abnormal nodes in the path, and combining sensitive monitoring areas to determine whether there is a path breakage risk, candidate path breakage nodes are screened for subsequent fault tolerance deduction processes;

[0054] Based on the candidate nodes of path breakage, candidate paths with structural consistency exceeding the 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 correction 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 the graph update loop based on fault-tolerant feedback is completed.

[0056] Acquire construction monitoring data during the construction process, and extract structural monitoring data related to the foundation support stage from the construction monitoring data. The structural monitoring data includes the measurement point number of each monitoring layer, structural depth information, time-series displacement data continuously recorded during the construction process, and its timestamp. The measurement point hierarchical index sequence is established based on the hierarchical positioning relationship in the measurement point layout plan and the construction drawings. Construction monitoring data includes various types of continuous monitoring information dynamically associated with structural response, such as inclination data, settlement data, stress and strain data, and construction progress information.

[0057] The time-series displacement data are synchronously sorted according to the measuring point level index sequence to obtain the displacement record sequence corresponding to the measuring points at each level. The measuring point level index sequence is then structured matched with the displacement record sequence, and gradient difference operations are performed between adjacent levels to solve the initial spatial gradient evolution sequence. Its directional changes are synchronously marked in each time step.

[0058] The spatial gradient evolution sequence is segmented into sliding windows in the time dimension, and trend direction aggregation, disturbance slope statistics and continuity judgment are performed in each window to extract the trend change interval set;

[0059] Perform a morphological comparison between the trend change interval set and the risk evolution sample set that has been annotated during the historical construction period, and generate a trend label mapping sequence based on the trend shape similarity and time rhythm similarity matching;

[0060] The trend label mapping sequence is node-based in hierarchical order to generate a path node set. The path nodes in the path node set include the hierarchical position of the measuring point, trend status label, direction offset mark and time window positioning information. According to the predefined connection rules, dependency directionality and topological constraints between nodes, the ordered connection relationship between the path node sets is constructed to form an initial risk path map. The paths in the risk path map are then verified for node accessibility and ring structure is eliminated.

[0061] For each path in the established risk path map, the path node sequence contained therein is read in turn, and the real-time data update status of the measurement points corresponding to the path nodes is retrieved;

[0062] In the real-time data update state, if there is an interruption in the measurement point data or abnormal behavior of sudden changes or long-term stable values, the node of the measurement point data will be marked as a data state abnormal node, and its corresponding time window and hierarchical position will be added to the abnormal node set;

[0063] A path dependency parsing operation is performed on the abnormal node set to determine whether each abnormal node in the abnormal node set is at a target connection position in the current path, where the target connection position refers to a reachable control node that cannot be bypassed in the path.

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

[0065] The sensitive monitoring area is composed of prior geological conditions and historical deformation active areas. Prior geological conditions include soft interlayers and geological abnormal layers in structural turning sections. Historical deformation active areas include historical cumulative deformation high-frequency layers and areas with frequent construction disturbances.

[0066] If a path-broken candidate node meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, the node will be upgraded to a priority alternative evaluation node, and its position code and path sequence number in the path structure will be recorded.

[0067] For each priority alternative assessment node, a set of paths that does not contain this node but whose path node sequence has a structural consistency higher than a preset similarity threshold in the start and end level feature dimension, trend state distribution feature dimension, and connection density feature dimension is searched in the risk path map to obtain the initial path migration candidate set;

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

[0069] If there is a candidate path whose structural consistency index is higher than the set threshold, it will be included in the equivalent path set and the monitoring data integrity check will be performed on it to verify whether there are any data missing points in the current monitoring cycle.

[0070] Perform trend output comparison operations on all equivalent paths with complete data to analyze whether their trend output sequences within the target time period are consistent with the output trends of the original paths;

[0071] If the output trends are consistent, the equivalent path will be marked as an alternative path. If there is an offset in the output trend, it will be determined whether the offset of the output trend is within the error tolerance range of the preset risk level. If the error tolerance range is met, it will be marked as a correction path and an offset correction factor will be added as a candidate risk evolution path.

[0072] Read the offset correction factor for each corrected path and compare it with the weight offset threshold set in the risk assessment mechanism to determine whether the offset degree of the corrected path is within the weight offset threshold limit;

[0073] If the offset correction factor is less than the preset weight offset threshold, the corrected path is included in the path sequence set currently participating in the risk trend judgment, and is used to replace the corresponding invalid segment node in the original path;

[0074] The main assessment path after the correction path is replaced is input into the risk trend judgment process to calculate the continuous risk level assessment results within the corresponding period;

[0075] Compare the path output trends before and after correction to find the offset vector caused by the path replacement, and record the path number, replacement position, and time window associated with the offset vector.

[0076] The offset vector and the evaluation error result are jointly written into the graph evolution record set to generate an evaluation offset trajectory archive reflecting the path replacement dynamics;

[0077] Use the assessment deviation trajectory archive to update the trend label confidence and node usage frequency in the risk path map to obtain the map evolution status in the current cycle;

[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, realizing a continuous tuning closed loop under path structure fault tolerance.

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

[0080] The path construction module establishes a hierarchical index of measurement points during the construction process and simultaneously organizes time-series displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence. The set of trend change intervals extracted by sliding window processing is mapped into a set of path nodes, and their topological connections are constructed to form an initial risk path map.

[0081] The fracture identification module identifies abnormalities in the data status of each path node in the risk path map, analyzes the accessibility control role of abnormal nodes in the path, and determines whether there is a path fracture risk based on sensitive monitoring areas. It then selects candidate path fracture nodes for subsequent fault-tolerance deduction processes.

[0082] The alternative deduction module searches for candidate paths with structural consistency exceeding a set threshold in the risk path map based on the candidate nodes of the path break, determines the alternative path after judging its trend output and error offset, and restores 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 loop based on fault-tolerant feedback.

[0084] The following further explanations are needed for the above scheme:

[0085] First, in the formula structure involved in this solution, dimensionless terms can serve as proportionality or structural adjustment factors. When combined with quantities with units, they only play a numerical scaling role and do not introduce new physical dimensions. Therefore, they do not change or confuse the overall unit system of expression. This combination of "dimensionless terms and units" can be understood as a composite structural expression commonly used in mathematical and physical 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 be formed into a unified structure through function mapping, ratio combination or normalization adjustment. The units and meanings are clear, and the overall expression conforms to the principle of dimensional consistency and the common formula of engineering modeling.

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

[0088] Based on the structural monitoring data of the construction phase, the trend evolution nodes are constructed according to the measurement point level and time window. The trend nodes are generated through nonlinear gradient and disturbance aggregation, and the structural connection is performed to form the initial risk path map and formulate is the initial risk path map, which is a connection 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, where N is the total number of measuring point levels; j is the time window index, ranging from 1≤j≤T, where T is the number of time steps in all construction phases; D i,j is the displacement record sequence of the i-th layer measuring point in the time window j; t j is the central moment of the jth time window; h i is the spatial depth value of the i-th layer measurement point; M i,j is the disturbance direction tensor of the i-th layer measurement point at time j, which includes indicators such as the number of turning points, fluctuation direction, and slope change rate; α i,jIt represents the directional aggregation index of the node in its time window. The directional aggregation index reflects the directional concentration. is the perturbation label mapping function, which represents the trend state of the input perturbation tensor output after clustering; It is the trend gradient aggregation operator under the sliding time window, which is used to test the stability of clustering features within the trend window; Combine the trend node state encoding and label state mapping operation; Represents the node connection operation in the path graph structure, which is used to construct the relationship between graph edges and nodes;

[0091] In addition, in the formula The part represents the second-order derivative of the displacement of the i-th layer measuring point at time j with respect to time, which is used to reflect the acceleration of trend change and capture the sudden change of trend turning point; It represents the rate of change of the displacement of the measuring point in this layer with respect to depth, reflecting the strain gradient at the structural level; log(1+·) indicates the use of a logarithmic function to enhance the nonlinear response of high gradient changes while compressing low-frequency background disturbances; Indicates mapping the disturbance tensor to node labels and classifying trend types (such as "rising unstable" / "falling stable", etc.); Indicates that the time gradient of the directional aggregation index is calculated within the window to ensure the stability of the trend node. The external two-layer union symbol and Indicates that trend nodes at all levels and all time periods are integrated to generate a graph;

[0092] Through the time series mutation characteristics of node displacement functions in the path graph, path interruption connectivity analysis and sensitive area identification, the candidate path break nodes that constitute the risk structure break are screened;

[0093]

[0094] in is the set of candidate nodes for path breaking; v k is the kth trend node in the risk path graph; D k (t) is the 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 is the accessibility control factor. If the node fails and the path is disconnected, the accessibility control factor is set to 1, otherwise it is 0; Θ k is a sensitive area indicator, which is positive if the node is located in an abnormal stratum or a historically high variation area; δ bIt represents the boundary threshold for fracture determination, which is set as the tolerance index jointly determined by the site and structure in practical application; t represents the continuous time variable;

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

[0096] in It represents the degree of mutation of the node in the trend time series, that is, the instantaneous discontinuity of the time slope; k ·Θ k Represents the fracture contribution weight after the superposition of structural control strength and risk area; after all items are weighted and δ b Comparison, as the criterion for whether a node constitutes a fracture;

[0097] By matching and evaluating candidate paths in the structural map that are similar to the fracture path topology and have tolerable risk trend output errors, a set of alternative risk paths is constructed;

[0098]

[0099] in is the set of alternative paths for path substitution deduction; p m Candidate paths are the path segments that remain structurally intact after the path segments containing the candidate nodes for the break are removed from the path graph; m,d (τ) is p m Risk trend response function at time τ; ψ d (τ) is the trend function of the original fault path at time τ; τ s ,τ e are the start and end time of the trend comparison period respectively; κ l is the risk trend cosine similarity threshold; ∈ d is the third-order trend shift integral error tolerance;

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

[0101] Based on the alternative path deviation feedback, the sensitivity response of node confidence and activation density is calculated, and the weight update is implemented in the path graph structure to form a closed loop of tuning of the evaluation strategy;

[0102]

[0103] in This is the updated risk path map; is the initial risk path map; p m is the correction path; v k is the node in the correction path; ω k The frequency weight of the original node in the graph; π k (Δ k ) is the node v k The risk level output function depends on the trend shift; Δ k is the trend deviation degree of the node between the alternative path and the original path; ρ k (t) is the active density of node risk output in time; θ ρ is the trend deviation tolerance; It is a logic judgment function. When the conditions in the brackets of the logic judgment function are met, the value is 1.

[0104] The proposed method and system for risk assessment of construction processes based on a fault-tolerant mechanism is designed to address key issues in current deep foundation pit support projects, such as interrupted risk assessment, misjudgment of trend models, and delayed identification of overall structural safety, caused by failed measurement point data.

[0105] During actual construction, especially in complex scenarios like deep foundation pits, underground continuous walls, and reverse construction, construction monitoring systems commonly deploy a large number of inclinometers, settlement gauges, and stress-strain sensing devices to establish a dynamic cognitive mechanism for structural response. However, due to the extremely complex underground environment, measurement points are prone to single-point failures, abnormal drift, or sudden data interruptions due to construction disturbances, mud erosion, and sensor aging. Such failures are sudden and highly hidden, often not detectable in a timely manner through hardware failures. Traditional risk assessment methods mostly rely on data integrity and lack dynamic compensation mechanisms for structural data gaps. Therefore, if measurement point data is missing, the model will be unable to accurately reconstruct risk trends, ultimately resulting in failed safety warnings or even misjudgment of working conditions, creating decision-making risks.

[0106] The design concept of this solution is precisely based on this technical bottleneck. It proposes to use the "path map" as the core modeling object and integrate the four-level process of "trend state expression, fault-tolerant reconstruction, structural equivalent migration, 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 has consistent terminology, ensuring the method's closed structure and strong dynamic adaptability:

[0107] 1. Constructing a risk path map: In this phase, structural monitoring data is extracted from the construction monitoring data, and a measurement point hierarchy index is established based on the relationship between the measurement point hierarchy and the time series. Gradient differential operations are then performed on the time series displacement data corresponding to each hierarchy to capture its spatial evolution trend. Trend change behavior is extracted within the sliding time window to form a "trend change interval set." These trend segments are then represented as nodes, and path connection relationships are constructed based on structural topology and dependency constraints, ultimately forming an "initial risk path map." This map abstractly encapsulates all identifiable risk trend evolution chains in the form of a graph structure, providing a trend-level structural foundation for subsequent analysis.

[0108] 2. Identify path interruption nodes: This section traverses each path and its contained trend nodes from the constructed risk path map, detects the real-time data status of the corresponding measuring points, identifies whether there are data interruptions, abnormal mutations, or abnormally stable behavior patterns, and marks them as "data status abnormal nodes." Subsequently, based on the path dependency structure, it is determined whether the abnormal node is located in an "accessible control position" that cannot be circumvented. If the failure of this node will cause the path to be disconnected, it is identified as a "path interruption candidate node." Further, based on the definition of "sensitive monitoring areas" (including soft interlayers, structural turning zones, and historically active deformation layers), "risk-significant" fault nodes are screened as target points for subsequent replacement deduction. This process ensures that not all faults enter the compensation process, but only "core interruptions in a structural sense" are processed.

[0109] 3. Execute path substitution deduction: For each candidate node of path breakage, the system searches the risk path map for feasible paths that are "structurally similar and have complete data." Its judgment criteria include three-dimensional structural characteristics such as the hierarchical distribution of path nodes, trend state sequence, and connection density, forming a "path migration candidate set." The candidate paths are then matched and analyzed with the original path on the risk trend output sequence, calculating their trend consistency (such as cosine similarity) and nonlinear error offset (such as the third-order integral of the trend difference). Ultimately, paths with "structural substitution and tolerable output errors" are selected as "corrected paths." This process does not involve a forced fit of the original path data, but rather establishes a new risk evolution channel through path migration, thereby ensuring an uninterrupted assessment chain.

[0110] 4. Update the evaluation structure and graph weights: After the corrected path is activated, the system replaces the invalid segment in the original path with it and uses its risk output as part of the formal trend judgment result. The system calculates the output offset vector before and after the replacement and writes this vector, along with the path segment, time window, and node position, into the "graph evolution record." Based on this, the usage frequency of all path nodes is updated, the confidence of the trend label is adjusted, and the graph structure state is reconstructed accordingly, thereby dynamically updating the "path activation strategy." This part forms the system's learning closed loop, ensuring that fault tolerance is not a one-time "replacement" but a continuous "self-optimizing evolution."

[0111] This solution not only solves the problem of "broken risk trend paths" from a structural graph perspective, but also innovatively introduces "structural equivalent migration" and "trend evolution feedback" mechanisms, enabling the risk assessment process to dynamically replace failed paths, adaptively select the optimal risk link, and continuously learn from the evolution of the graph state.

[0112] In practical applications, this method can be applied to construction environments that require in-depth monitoring, such as underground projects, bridge foundations, and subway enclosures. It can also be seamlessly connected with BIM platforms and digital twin monitoring systems to achieve fault-tolerant assessment and real-time adjustment of multi-path risk trends. It is especially 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 can significantly improve the overall risk perception continuity and misjudgment avoidance capabilities.

[0113] The above description is only 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 in the scope of protection of the present invention.

Claims

1. A construction process risk assessment method based on fault tolerance mechanism, characterized in that: include: By establishing a hierarchical index of measurement points during the construction process and synchronously organizing time-series displacement data, differential calculations are performed to obtain a spatial gradient evolution sequence. The set of trend change intervals extracted by sliding window processing is mapped into a set of path nodes, and their topological connection relationships are constructed to form an initial risk path map. By identifying abnormalities in the data status of each path node in the risk path map, analyzing the access control role of abnormal nodes in the path, and combining sensitive monitoring areas to determine whether there is a path breakage risk, candidate nodes for path breakage are screened; Based on the candidate nodes of path breakage, the candidate paths with structural consistency exceeding the set threshold are retrieved in the risk path map, and the alternative paths are determined after judging their trend output and error offset; By extracting the offset feedback of the correction path and recording the path evolution behavior, the label confidence and node usage frequency in the risk path map are updated, and the path activation strategy is adjusted.

2. A construction process risk assessment method based on a fault tolerance mechanism according to claim 1, characterized in that: Acquire construction monitoring data during the construction process and extract structural monitoring data related to the foundation support stage from the construction monitoring data. The structural monitoring data includes the measurement point number of each monitoring layer, structural depth information, time-series displacement data continuously recorded during the construction process and its timestamp. Establish a hierarchical index sequence of measurement points based on the hierarchical positioning relationship in the measurement point layout plan and the construction drawings. The time-series displacement data are synchronously sorted according to the measuring point level index sequence to obtain the displacement record sequence corresponding to the measuring points at each level. The measuring point level index sequence is then structured matched with the displacement record sequence, and gradient difference operations are performed between adjacent levels to solve the initial spatial gradient evolution sequence. Its directional changes are synchronously marked in each time step.

3. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 2, characterized in that: The spatial gradient evolution sequence is segmented into sliding windows in the time dimension, and trend direction aggregation, disturbance slope statistics and continuity judgment are performed in each window to extract the trend change interval set; Perform a morphological comparison between the trend change interval set and the risk evolution sample set that has been annotated during the historical construction period, and generate a trend label mapping sequence based on the trend shape similarity and time rhythm similarity matching; The trend label mapping sequence is node-based in hierarchical order to generate a path node set. The path nodes in the path node set include the hierarchical position of the measuring point, trend status label, direction offset mark and time window positioning information. According to the predefined connection rules, dependency directionality and topological constraints between nodes, the ordered connection relationship between the path node sets is constructed to form an initial risk path map. The paths in the risk path map are then verified for node accessibility and ring structure is eliminated.

4. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 3 is characterized by: For each path in the established risk path map, the path node sequence contained therein is read in turn, and the real-time data update status of the measurement points corresponding to the path nodes is retrieved; In the real-time data update state, if there is an interruption in the measurement point data or abnormal behavior of sudden changes or long-term stable values, the node of the measurement point data will be marked as a data state abnormal node, and its corresponding time window and hierarchical position will be added to the abnormal node set; A path dependency parsing operation is performed on the abnormal node set to determine whether each abnormal node in the abnormal node set is at a target connection position in the current path, where the target connection position refers to a reachable control node that cannot be bypassed in the path.

5. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 4 is characterized in that: If the results of the path dependency analysis indicate that the abnormal node is in an accessibility-control position in the path, that is, its absence will cause the previous and next nodes to be disconnected, then the abnormal node is determined to meet the necessary conditions for path accessibility. It is then regarded as a candidate node for path breakage and a node position structure hierarchical analysis is performed to confirm whether it is located in the preset sensitive monitoring area. The sensitive monitoring area is composed of prior geological conditions and historical deformation active areas. Prior geological conditions include soft interlayers and geological abnormal layers in structural turning sections. Historical deformation active areas include historical cumulative deformation high-frequency layers and areas with frequent construction disturbances. If a path-broken candidate node meets the necessary conditions for path accessibility and belongs to a sensitive monitoring area, the node will be upgraded to a priority alternative evaluation node, and its position code and path sequence number in the path structure will be recorded.

6. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 5, characterized in that: For each priority alternative assessment node, a set of paths that does not contain this node but whose path node sequence has a structural consistency higher than a preset similarity threshold in the start and end level feature dimension, trend state distribution feature dimension, and connection density feature dimension is searched in the risk path map to obtain the initial path migration candidate set; Perform path node sequence alignment on the path migration candidate set and calculate the structural consistency index with the original path in terms of trend state sequence, hierarchical span and connection density; If there is a candidate path whose structural consistency index is higher than the set threshold, it will be included in the equivalent path set and the monitoring data integrity check will be performed on it to verify whether there are any data missing points in the current monitoring cycle.

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

8. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 7 is characterized by: Read the offset correction factor for each corrected path and compare it with the weight offset threshold set in the risk assessment mechanism to determine whether the offset degree of the corrected path is within the weight offset threshold limit; If the offset correction factor is less than the preset weight offset threshold, the corrected path is included in the path sequence set currently participating in the risk trend judgment, and is used to replace the corresponding invalid segment node in the original path; The main assessment path after the correction path is replaced is input into the risk trend judgment process to calculate the continuous risk level assessment results within the corresponding period; Compare the path output trends before and after correction to find the offset vector caused by the path replacement, and record the path number, replacement position, and time window associated with the offset vector.

9. The method for risk assessment of a construction process based on a fault tolerance mechanism according to claim 8, characterized in that: The offset vector and the evaluation error result are jointly written into the graph evolution record set to generate an evaluation offset trajectory archive reflecting the path replacement dynamics; Use the assessment deviation trajectory archive to update the trend label confidence and node usage frequency in the risk path map to obtain the map evolution status in the current cycle; The updated graph evolution state is input into the path selection rule set to dynamically adjust the path activation strategy in the evaluation model, realizing a continuous tuning closed loop under path structure fault tolerance.

10. A construction process risk assessment system based on a fault tolerance mechanism, comprising a path construction module, a fracture identification module, an alternative deduction module, and a map update module, characterized by: The path construction module establishes a hierarchical index of measurement points during the construction process and simultaneously organizes time-series displacement data. It then performs differential calculations to obtain a spatial gradient evolution sequence. The set of trend change intervals extracted by sliding window processing is mapped into a set of path nodes, and their topological connections are constructed to form an initial risk path map. The fracture identification module identifies abnormalities in the data status of each path node in the risk path map, analyzes the accessibility control role of abnormal nodes in the path, and determines whether there is a path fracture risk in combination with sensitive monitoring areas, and screens out candidate path fracture nodes; The alternative deduction module searches for candidate paths with structural consistency exceeding a set threshold in the risk path map based on the candidate nodes of the path break, and determines the alternative path after judging its trend output and error offset; 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, and adjusts the path activation strategy.

Images