A tunnel boring machine tunneling efficiency uncertainty tracing diagnosis method and device

Abstract

This application relates to a diagnostic method and apparatus for tracing the source of uncertainty in tunnel boring machine (TBM) tunneling efficiency. It collects multi-source operational data of the TBM, using tunneling efficiency as the target variable. Based on information theory, it performs uncertainty contribution analysis on candidate influencing variables to obtain a set of observed variables for constructing a diagnostic network. Under mutual information constraints, it constructs a tunneling efficiency diagnostic network structure, performs parameter learning on the conditional probability relationships within the network to form a tunneling efficiency diagnostic network model. Based on the diagnostic network, it performs probabilistic diagnostic inference on tunneling efficiency and calculates the prediction residuals. When the prediction residuals are abnormal, it analyzes the dependency relationship between the residuals and observed variables to identify potential unmodeled influencing factors, thus tracing the source of tunneling efficiency uncertainty. The method explicitly characterizes the structural dependencies between tunneling parameters, improving the accuracy and interpretability of tunneling efficiency diagnosis, and is applicable to the diagnosis and decision support of TBM operating status under complex working conditions.

Description

Technical Field

[0001] This application relates to the field of engineering machinery operation status diagnosis technology, and in particular to a diagnostic method and device for tracing the source of uncertainty in the tunneling efficiency of a tunnel boring machine. Background Technology

[0002] During tunnel construction, the tunneling efficiency of a tunnel boring machine (TBM) is influenced by a combination of factors, including geological conditions, equipment operating status, construction parameters, and environmental factors, exhibiting characteristics of high coupling and significant uncertainty. In actual construction, abnormal fluctuations in tunneling efficiency can not only affect the construction progress but also potentially induce equipment overload, structural damage, and safety risks.

[0003] Existing analytical methods for addressing the uncertainty of tunneling efficiency mainly include statistical analysis methods, empirical rule methods, and information entropy analysis methods based on single indicators. These methods typically identify factors that significantly impact tunneling efficiency by statistically modeling or ranking historical data.

[0004] However, the above methods generally have the following shortcomings: most methods fail to explicitly characterize the structural dependencies between tunneling parameters and are difficult to distinguish between direct and indirect influencing factors; when there are many variables and the working conditions are complex, variable redundancy and collinearity problems can easily lead to unstable analysis results; when tunneling efficiency anomalies are caused by unobserved or unmodeled potential factors, existing methods are difficult to effectively trace the source; and there is a lack of dynamic updating and iterative analysis mechanisms for diagnosing tunneling efficiency failures.

[0005] Therefore, there is an urgent need for a technical solution that can characterize the relationship between tunneling efficiency and other factors in a structured manner under multi-source data conditions, and further locate the source of uncertainty when diagnostic errors occur, so as to improve the operational diagnosis and decision support capabilities of the tunneling process. Summary of the Invention

[0006] To address or partially address the problems existing in related technologies, this application provides a diagnostic method and apparatus for tracing the source of uncertainty in tunnel boring machine efficiency. By constructing a tunneling efficiency diagnostic network with engineering constraints, it realizes probabilistic diagnostic reasoning of tunneling efficiency and identifies potential unmodeled factors when diagnostic deviations occur, thereby achieving the tracing and iterative updating of tunneling efficiency uncertainty.

[0007] The first aspect of this application provides a diagnostic method for tracing the source of uncertainty in the tunnel boring machine's tunneling efficiency, comprising the following steps: Acquire multi-source operating data of the tunnel boring machine and determine the target variable. The multi-source operating data includes at least tunneling parameters, equipment status parameters and working environment parameters. The target variable is used to characterize the tunneling efficiency. Uncertainty contribution analysis is performed on candidate variables of multi-source operational data, mutual information between each candidate variable and the target variable is calculated, and the set of observed variables is obtained by screening based on the relative uncertainty contribution rate and the redundancy information constraint between variables. Using the set of observed variables as the set of nodes in the diagnostic network, candidate dependencies are generated based on the mutual information strength between nodes. Under the conditions of mutual information constraints and no directed loop constraints, a diagnostic network structure for tunneling efficiency diagnosis is constructed based on the structure search method. The conditional probability relationships of each node in the diagnostic network structure are learned by parameters to form a tunneling efficiency diagnostic network model and perform probabilistic diagnostic inference on tunneling efficiency. The prediction residual between the diagnostic results and the actual tunneling efficiency is calculated. If the predicted residuals meet the preset anomaly criteria, then based on the conditional dependency relationship between the predicted residuals and the observed variables, the residual dependency patterns of potential unmodeled influencing factors are identified, thereby achieving source diagnosis of tunneling efficiency uncertainty.

[0008] The selection of the set of observed variables includes: Using the information entropy of the target variable as a benchmark, the relative uncertainty contribution rate of each candidate variable is calculated; Candidate variables whose relative uncertainty contribution rate is lower than the preset statistical threshold are removed; When the mutual information between two candidate variables is significantly higher than their mutual information with the target variable, the candidate variable that contributes more to the uncertainty of the target variable is retained to reduce the redundancy of the diagnostic network structure.

[0009] The process of constructing the diagnostic network structure includes the following steps: Generate a set of candidate dependencies based on the mutual information between candidate variables; Candidate dependencies are sorted according to mutual information strength and the diagnostic network structure is initialized. Without introducing directed cycles, structural search is performed by adding, deleting, or reversing dependencies, where the reversal of dependencies must satisfy the mutual information rationality constraint.

[0010] Among them, when constructing the diagnostic network structure based on the structure search method, the evaluation of the candidate diagnostic network structure adopts a scoring mechanism that combines model fit and structural complexity, and introduces mutual information as a structural prior constraint to suppress dependencies that are inconsistent with the engineering mechanism; when multiple consecutive structural adjustments fail to improve the score value, the diagnostic network structure is determined to have converged.

[0011] Among them, the anomaly detection criteria for the predicted residuals shall include at least one of the following: The negative log-likelihood residual calculated based on the tunneling efficiency diagnostic probability exceeds a preset threshold. The information entropy of the residual sequence is higher than the preset uncertainty level.

[0012] After identifying the residual dependency pattern, the process also includes the following steps: Identifying the direction of potential unmodeled influencing factors based on residual dependency patterns; Potential influencing factors can be observed by adding new sensors or constructing proxy variables; Newly observed variables are introduced into the diagnostic network, and the variable selection, network construction, and residual analysis steps are repeated to achieve iterative source tracing of tunneling efficiency uncertainty.

[0013] The second aspect of this application provides a diagnostic network device for tracing the uncertainty of tunnel boring machine efficiency, applicable to the diagnostic method for tracing the uncertainty of tunnel boring machine efficiency as described in the first aspect of this application. The device includes: The data acquisition unit is used to collect multi-source operational data during the tunnel boring machine's excavation process; The variable screening unit is used to perform uncertainty contribution analysis on multi-source operational data based on information theory and to screen diagnostic network observation variables. Diagnostic network building unit, used to construct a tunneling efficiency diagnostic network structure under mutual information constraints; The diagnostic reasoning unit is used to perform probabilistic diagnostic reasoning on tunneling efficiency based on the diagnostic network. Uncertainty tracing unit is used to identify potential sources of uncertainty in tunneling efficiency based on predictive residuals.

[0014] The diagnostic network construction unit includes a candidate dependency generation module, a mutual information constraint structure search module, and a diagnostic network structure evaluation and convergence determination module.

[0015] The uncertainty tracing unit includes a potential influencing factor indication module and a new variable introduction and diagnostic network update module. The new variable introduction and diagnostic network update module is used to realize dynamic tracing of tunneling efficiency uncertainty and adaptive model update.

[0016] A third aspect of this application provides a computer-readable storage medium storing a control program for a diagnostic network device for tracing the uncertainty of tunnel boring machine efficiency. When the control program for the diagnostic network device for tracing the uncertainty of tunnel boring machine efficiency is executed by a processor, it implements the diagnostic method for tracing the uncertainty of tunnel boring machine efficiency as provided in the first aspect of this application.

[0017] The technical solution provided in this application may include the following beneficial effects: A diagnostic method and apparatus for tracing the source of uncertainty in tunnel boring machine (TBM) efficiency are disclosed. By constructing a TBM efficiency diagnostic network, the structural dependencies between TBM parameters are explicitly characterized, improving the interpretability of the diagnostic results. An information-theoretic-based variable selection mechanism effectively reduces variable redundancy and enhances the stability of the diagnostic network structure. An uncertainty tracing mechanism driven by predictive residuals enables the directional positioning of potential unmodeled factors. The method supports dynamic updating and iterative optimization of the diagnostic network, enhancing its adaptability to complex working conditions and making it suitable for online diagnostic and decision support scenarios in tunnel boring operations.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0019] The above and other objects, features and advantages of this application will become more apparent from the more detailed description of exemplary embodiments thereof in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments thereof.

[0020] Figure 1 This is a schematic diagram illustrating the construction and diagnosis method of the tunnel boring machine tunneling efficiency uncertainty tracing and diagnosis network in an embodiment of this application; Figure 2 This is a schematic diagram of the diagnostic network device structure shown in an embodiment of this application. Detailed Implementation

[0021] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While embodiments of this application are shown in the drawings, it should be understood that this application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to make this application more thorough and complete, and to fully convey the scope of this application to those skilled in the art.

[0022] Example 1: Diagnostic Network Construction and Diagnostic Methods like Figure 1 As shown, the diagnostic network construction and diagnostic method of this embodiment includes the following steps.

[0023] Step S101: Multi-source operation data acquisition Collect multi-source operational data during the tunnel boring machine's excavation process, including but not limited to thrust, cutterhead rotation speed, cutterhead torque, thrust speed, equipment vibration status, and geological parameters, and determine excavation efficiency as the target variable.

[0024] Step S102: Variable selection based on information theory The mutual information between each candidate variable and tunneling efficiency is calculated, and the relative uncertainty contribution rate is calculated based on the information entropy of tunneling efficiency. Variables with low contributions to the uncertainty of tunneling efficiency are removed, and highly redundant variables are further eliminated to obtain the set of observed variables used to construct the diagnostic network.

[0025] (1) Mutual information calculation formula: Candidate variables With target variable Mutual information between (tunneling efficiency) Defined as: in, It is a joint probability distribution. and These are the marginal probability distributions.

[0026] (2) Information entropy of tunneling efficiency tunneling efficiency Information entropy for: Information entropy characterizes the uncertainty of tunneling efficiency and provides a benchmark for subsequent calculation of relative contribution rate.

[0027] (3) Contribution rate of relative uncertainty: The relative contribution rate of candidate variable Xi to the uncertainty of tunneling efficiency is defined as: And remove low-contribution variables that meet the criteria: Removed in This is a preset threshold.

[0028] 4. Redundant variable removal criteria If two candidate variables and The mutual information between them is significantly higher than the mutual information between them and the target variable: Then retain those that contribute more to the target variable. or Eliminate one to reduce redundancy.

[0029] Step S103: Diagnostic Network Structure Construction Using observed variables as nodes in the diagnostic network, candidate dependencies are generated based on the mutual information strength between nodes. Without introducing the constraint of directed loops, a tunneling efficiency diagnostic network structure is constructed through a structure search method. This diagnostic network is used to characterize the structured influence relationship between tunneling parameters.

[0030] 1. Quantitative criteria for constructing dependency relationships based on mutual information In the diagnostic network, candidate nodes and Establish a directed dependency relationship between them Conditions: Maximize while satisfying structural constraints.

[0031] in for The set of parent nodes.

[0032] 2. Objective function for structure search without directed cycle constraints The diagnostic network structure scoring function can be defined as: in, The log-likelihood function of the network model. Network structure complexity (number of nodes, number of edges): λ: Complexity penalty coefficient. During the structure search process, ensure... For a directed acyclic graph 3. Formal Representation of Diagnostic Network Structure A network is defined as a directed acyclic graph. : : Set of observed variables (set of nodes) Dependencies between nodes (edge ​​set) Each node Conditional probability table Used for probabilistic reasoning.

[0033] Step S104: Diagnose network parameter learning After the diagnostic network structure is determined, the conditional probability relationships of each node in the network are learned as parameters. In one implementation, parameter learning can be achieved using a probability-statistic-based parameter estimation method to obtain the diagnostic network model.

[0034] Step S105: Tunneling efficiency diagnosis reasoning Based on the completed tunneling efficiency diagnostic network model, probabilistic diagnostic reasoning is performed on the tunneling efficiency under the current tunneling conditions, and the prediction residual between the diagnostic results and the actual tunneling efficiency is calculated.

[0035] Assume the predicted tunneling efficiency is The actual observed value is Then predict the residual: Probabilistic diagnostic reasoning can be based on Bayes' theorem: Residuals are used to measure diagnostic errors and drive uncertainty attribution.

[0036] Step S106: Uncertainty Origin Tracing and Iterative Update When the prediction residual exceeds the preset threshold, the conditional dependency between the prediction residual and the observed variable is analyzed, potential influencing factors that have not been explicitly modeled are identified, and new observed variables are introduced by adding new sensor data or constructing proxy variables to update the diagnostic network structure and parameters, thereby achieving iterative source tracing of uncertainty.

[0037] 1. Conditional Dependence of Predicted Residuals on Observed Variables residual With observed variables Conditional mutual information is defined as: Used to identify potential unmodeled influencing factors.

[0038] 2. Identification and Judgment of Unmodeled Influencing Factors If conditions are mutually informative If so, it is determined that there are potential unmodeled factors, and new observation variables need to be introduced.

[0039] 3. Network structure and parameter iterative update Introducing new variables Then, update the network: Dependency relationship: E←E∪ Re-execute variable selection, network structure search, and parameter learning to achieve iterative tracing.

[0040] Example 2: Diagnostic Network Device like Figure 2 As shown, the diagnostic network device provided in this embodiment includes: The data acquisition unit is used to collect multi-source operational data during the tunnel boring machine's excavation process; The variable screening unit is used to perform uncertainty contribution analysis on multi-source operational data based on information theory and to screen diagnostic network observation variables. Diagnostic network building unit, used to construct a tunneling efficiency diagnostic network structure under mutual information constraints; The diagnostic reasoning unit is used to perform probabilistic diagnostic reasoning on tunneling efficiency based on the diagnostic network. Uncertainty tracing unit is used to identify potential sources of uncertainty in tunneling efficiency based on predictive residuals.

[0041] The diagnostic network construction unit includes a candidate dependency generation module, a mutual information constraint structure search module, and a diagnostic network structure evaluation and convergence determination module.

[0042] The uncertainty tracing unit includes a potential influencing factor indication module and a new variable introduction and diagnostic network update module. The new variable introduction and diagnostic network update module is used to realize dynamic tracing of tunneling efficiency uncertainty and adaptive model update.

[0043] Example 3: A computer-readable storage medium storing a control program for a diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency. When the control program for the diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency is executed by a processor, it implements the diagnostic method for tracing the uncertainty of tunnel boring machine excavation efficiency as provided in Example 1.

[0044] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0045] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0046] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0047] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency, characterized in that, Includes the following steps: Acquire multi-source operating data of the tunnel boring machine and determine the target variable. The multi-source operating data includes at least tunneling parameters, equipment status parameters and working environment parameters. The target variable is used to characterize the tunneling efficiency. Uncertainty contribution analysis is performed on the candidate variables of the multi-source operational data, the mutual information between each candidate variable and the target variable is calculated, and the set of observed variables is obtained by screening based on the relative uncertainty contribution rate and the redundancy information constraint between variables. Using the set of observed variables as the node set of the diagnostic network, candidate dependencies are generated based on the mutual information strength between nodes. Under the conditions of mutual information constraints and no directed loop constraints, a diagnostic network structure for tunneling efficiency diagnosis is constructed based on the structure search method. The conditional probability relationships of each node in the diagnostic network structure are learned by parameters to form a tunneling efficiency diagnostic network model and perform probabilistic diagnostic reasoning on tunneling efficiency. The prediction residual between the diagnostic results and the actual tunneling efficiency is calculated. If the predicted residuals meet the preset anomaly judgment conditions, then based on the conditional dependency relationship between the predicted residuals and the observed variables, the residual dependency patterns of potential unmodeled influencing factors are identified, thereby achieving source tracing and diagnosis of tunneling efficiency uncertainty.

2. The diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency according to claim 1, characterized in that, The selection of the set of observed variables includes: Based on the information entropy of the target variable, the relative uncertainty contribution rate of each candidate variable is calculated; Candidate variables whose relative uncertainty contribution rate is lower than the preset statistical threshold are removed; When the mutual information between two candidate variables is significantly higher than their mutual information with the target variable, the candidate variable that contributes more to the uncertainty of the target variable is retained to reduce the redundancy of the diagnostic network structure.

3. The diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency according to claim 1, characterized in that, The process of constructing the diagnostic network structure includes the following steps: Generate a set of candidate dependencies based on the mutual information between candidate variables; Candidate dependencies are sorted according to mutual information strength and the diagnostic network structure is initialized. Without introducing directed cycles, structural search is performed by adding, deleting, or reversing dependencies, where the reversal of dependencies must satisfy the mutual information rationality constraint.

4. The diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency according to claim 1, characterized in that, When constructing the diagnostic network structure based on the structure search method, the evaluation of the candidate diagnostic network structure adopts a scoring mechanism that combines model fit and structural complexity, and introduces mutual information as a structural prior constraint to suppress dependencies that are inconsistent with engineering mechanisms. When multiple structural adjustments fail to improve the score, the diagnostic network structure is deemed to have converged.

5. The diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency according to claim 1, characterized in that, The anomaly detection criteria for the predicted residuals include at least one of the following: The negative log-likelihood residual calculated based on the tunneling efficiency diagnostic probability exceeds a preset threshold. The information entropy of the residual sequence is higher than the preset uncertainty level.

6. The diagnostic method for tracing the source of uncertainty in the tunnel boring machine's excavation efficiency according to claim 1, characterized in that, After identifying the residual dependency pattern, the following steps are also included: Identifying the direction of potential unmodeled influencing factors based on residual dependency patterns; The potential influencing factors can be observed by adding new sensors or constructing proxy variables; Newly observed variables are introduced into the diagnostic network, and the variable selection, network construction, and residual analysis steps are repeated to achieve iterative source tracing of tunneling efficiency uncertainty.

7. A diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency, applicable to the diagnostic method for tracing the uncertainty of tunnel boring machine excavation efficiency as described in any one of claims 1-6, characterized in that, The device includes: The data acquisition unit is used to collect multi-source operational data during the tunnel boring machine's excavation process; The variable screening unit is used to perform uncertainty contribution analysis on the multi-source operational data based on information theory and to screen diagnostic network observation variables. Diagnostic network building unit, used to construct a tunneling efficiency diagnostic network structure under mutual information constraints; A diagnostic reasoning unit is used to perform probabilistic diagnostic reasoning on tunneling efficiency based on the diagnostic network. Uncertainty tracing unit is used to identify potential sources of uncertainty in tunneling efficiency based on predictive residuals.

8. The diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency according to claim 7, characterized in that, The diagnostic network construction unit includes a candidate dependency generation module, a mutual information constraint structure search module, and a diagnostic network structure evaluation and convergence determination module.

9. The diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency according to claim 1, characterized in that, The uncertainty tracing unit includes a potential influencing factor indication module and a new variable introduction and diagnostic network update module. The new variable introduction and diagnostic network update module is used to realize dynamic tracing of tunneling efficiency uncertainty and adaptive model update.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a control program for a diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency. When the control program for the diagnostic network device for tracing the uncertainty of tunnel boring machine excavation efficiency is executed by a processor, it implements the diagnostic method for tracing the uncertainty of tunnel boring machine excavation efficiency as described in any one of claims 1 to 6.

Images