A data analysis-based shield machine operation and maintenance management system

Abstract

This invention relates to the field of tunnel boring machine (TBM) operation and maintenance management technology, and discloses a TBM operation and maintenance management system based on data analysis. The system includes: a state modeling module for acquiring multi-source construction data to form a unified state dataset and merging it to determine health and risk indicators; a quota allocation module for generating risk budgets, performance quotas, and lifespan quotas; a decision constraint module for forming a set of feasible decisions based on a prediction model; a cost optimization module for weighted calculation of multi-dimensional costs and determining the final operation and maintenance decision; a task scheduling module for generating a standardized task set and freezing quotas; a deviation assessment module for comparing actual and predicted data and identifying deviations; and a settlement ledger module for settling quotas and updating the ledger based on actual risks and lifespan consumption. This invention achieves quantitative management and intelligent decision optimization throughout the entire TBM operation and maintenance process.

Description

Technical Field

[0001] This invention belongs to the field of tunnel boring machine operation and maintenance management technology, specifically relating to a tunnel boring machine operation and maintenance management system based on data analysis. Background Technology

[0002] Tunnel boring machines (TBMs) are the most widely used full-face tunneling equipment in underground engineering construction. Their operation is affected by a combination of factors, including geological conditions, equipment health, spoil behavior, silo pressure control, and cutter wear. With the continuous expansion of subway, municipal tunnel, and underground utility tunnel projects, the TBM construction environment is characterized by strong geological heterogeneity, frequent disturbances to construction parameters, and rapid equipment lifespan depletion. Traditional operation and maintenance methods that rely on manual experience or static rules are no longer sufficient to meet the requirements of safety, stability, and refined management.

[0003] In existing technologies, some systems can collect data and analyze trends during tunnel boring machine (TBM) construction, but they generally suffer from the following problems: First, the dimensions and sampling frequencies of multi-source data are inconsistent, and there is a lack of a unified way to express the status, resulting in inconsistent standards for risk assessment and health judgment. Second, it is impossible to establish a dynamic risk budget and lifespan management mechanism based on the actual status and lifespan consumption of the equipment, making it difficult to quantify and constrain operation and maintenance strategies. Third, most existing decision-making systems only focus on propulsion efficiency or energy consumption indicators, lacking comprehensive optimization methods that unify and quantify multiple factors such as risk, lifespan, and energy consumption. Fourth, traditional operation and maintenance models often lack systematic feedback on the actual effects after execution and do not have the ability to identify deviations, settle accounts, and manage multi-cycle closed loops.

[0004] With the development of digital twin technology, predictive assessments of tunnel boring machine (TBM) operating status can be achieved by constructing a mapping model between equipment operating characteristics, geological conditions, and construction behavior, providing simulation and simulation capabilities for operation and maintenance decisions. However, existing digital twin systems still suffer from problems such as inconsistencies between the model and on-site data, lack of resource allocation mechanisms, and insufficient decision-making feasibility when applied in actual engineering projects. Summary of the Invention

[0005] This invention provides a data analysis-based operation and maintenance management system for tunnel boring machines, which solves the technical problems in related technologies such as the difficulty in uniformly expressing multi-source construction data, the inability to quantify and allocate risk and lifespan resources, the lack of predictive evaluation in operation and maintenance decisions, and the lack of closed-loop verification in the construction process.

[0006] This invention provides a data analysis-based operation and maintenance management system for tunnel boring machines, comprising: The state modeling module is used to acquire geological conditions, shield equipment operating status, slag and silo pressure information, cutter wear data and remaining life data, form a unified state data set, and determine equipment health and construction risk indicators. The quota allocation module is used to generate risk budgets, performance quotas, and life quotas based on construction risk indicators, equipment health, and remaining lifespan, and allocate them to future tunneling cycles according to the allocation ratio. The decision constraint module is used to determine the operation and maintenance decision variables based on the target advance speed, target inventory pressure, maintenance operation interval and tool replacement strategy parameters, and input them with the unified state data set into the risk increment prediction model and life consumption prediction model to obtain the risk increment and life consumption, forming a feasible decision set. The cost optimization module is used to calculate the total cost by weighting the various decision combinations in the feasible decision set based on capacity loss cost, energy consumption cost, risk overrun cost and lifetime consumption cost, and to determine the final operation and maintenance decision. The task orchestration module is used to transform the final operation and maintenance decision into a standardized set of tasks, which are then distributed to the execution layer, freezing the corresponding risk budget and lifetime allowance. The deviation assessment module is used to collect actual construction data to form actual unified status data, actual equipment health and actual construction risk indicators, compare them with the corresponding predicted values ​​to obtain deviation data, and identify abnormal deviation events. The settlement ledger module is used to settle and update the remaining amount ledger based on actual construction risk indicators and actual life consumption, as well as the risk budget and life limit.

[0007] The beneficial effects of this invention are as follows: By constructing a unified state data set, this invention achieves standardized expression of geological conditions, equipment operating status, slag and silo pressure information, and tool wear and lifespan data. By introducing a risk budget, performance allowance, and lifespan allowance generation mechanism based on equipment health, construction risk indicators, and remaining lifespan, it quantifies risk and lifespan resources. Through a digital twin model, it extrapolates the risk increment and lifespan consumption of candidate operation and maintenance decisions, and combines multi-dimensional cost weighting to select the final operation and maintenance decision, thus achieving systematic operation and maintenance optimization under the constraints of risk, safety, and efficiency. Through a task-based structure, frozen list, and concurrent flow limiting mechanism, it ensures the feasibility and traceability of operation and maintenance decisions at the execution layer. Through deviation identification, allowance settlement, and ledger updates, it achieves closed-loop verification between predicted status and actual execution results, enabling dynamic correction of operation and maintenance strategies across cycles. Overall, it improves the digital management capabilities of the tunnel boring machine construction process and the level of precision in operation and maintenance decisions. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of a data analysis-based tunnel boring machine operation and maintenance management system according to the present invention. Detailed Implementation

[0009] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0010] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0011] like Figure 1 As shown, a data analysis-based tunnel boring machine (TBM) operation and maintenance management system includes: State modeling module 1 is used to acquire geological conditions, shield equipment operating status, slag and silo pressure information, cutter wear data and remaining life data, form a unified state data set, and determine equipment health and construction risk indicators. The quota allocation module 2 is used to generate risk budgets, performance quotas, and life quotas based on construction risk indicators, equipment health, and remaining lifespan, and allocate them to future tunneling cycles according to the allocation ratio. Decision constraint module 3 is used to determine the operation and maintenance decision variables based on the target advance speed, target inventory pressure, maintenance operation interval and tool replacement strategy parameters, and input them with the unified state data set into the risk increment prediction model and life consumption prediction model to obtain the risk increment and life consumption, forming a feasible decision set. Cost optimization module 4 is used to calculate the total cost of each decision combination in the feasible decision set based on the weighted average of capacity loss cost, energy consumption cost, risk overrun cost and lifetime consumption cost, and to determine the final operation and maintenance decision. Task orchestration module 5 is used to transform the final operation and maintenance decision into a standardized set of tasks, distribute them to the execution layer, and freeze the corresponding risk budget and lifetime limit. Deviation assessment module 6 is used to collect actual construction data to form actual unified status data, actual equipment health and actual construction risk indicators, compare them with the corresponding predicted values ​​to obtain deviation data, and identify abnormal deviation events. The settlement ledger module 7 is used to settle the risk budget and life limit based on the actual construction risk indicators and actual life consumption, and update the remaining limit ledger.

[0012] In one embodiment of the present invention, geological conditions, shield tunneling equipment operating status, excavated soil and silo pressure information, cutter wear data, and remaining life data are acquired to form a unified status data set, and equipment health and construction risk indicators are determined, including: Step 11: Within the decision-making cycle, acquire geological conditions, shield machine operating status, excavated soil and silo pressure information, cutter wear data, and remaining life data, and organize them according to a unified timestamp and shield ring number to form a multi-source raw data set. The decision-making cycle refers to a fixed interval for updating operation and maintenance decisions, which can be divided by time window or by shield ring number. The shield ring number refers to the unique number of the current tunneling ring. Geological conditions include soil type, moisture content, and representative value of ground stress. Shield machine operating status includes propulsion force, cutterhead torque, propulsion speed, and screw speed. Excavated soil and silo pressure information includes excavated soil moisture content, slag pressure, and mud silo pressure. Cutter wear data includes wear amount, missing blade mark, and chipped blade mark. Remaining life data represents the life estimate obtained based on historical wear and replacement records. Perform linear interpolation or removal on missing records and retain the missing identifier in the records. Truncate values ​​that obviously exceed the limits according to the upper or lower limit of engineering experience to finally form a multi-source raw data set. Step 12: Assign each original data item in the multi-source original data set to a state variable, perform normalization processing to obtain normalized state data, and combine them in a fixed order to form a unified state data set. Step 13: According to the preset health weight vector, the normalized state data in the unified state data set is weighted and summed, and the summation result is subtracted from 1 to determine the equipment health, so that the value is between 0 and 1, and the larger the value, the healthier the equipment is. According to the preset risk index weight vector, the normalized state data is weighted and summed to determine the construction risk index, so that the value is between 0 and 1, and the larger the value, the higher the construction risk.

[0013] This implementation process uses a unified timestamp and shield ring number alignment strategy and a unified status data set with a fixed field order to stably map heterogeneous and multi-source raw data into standardized status vectors and two types of indicators that can be directly consumed by the digital twin side. This process provides guarantees in terms of data consistency, dimensional comparability, and cross-cycle traceability, and reduces the interference of data noise and inconsistent standards on digital twin calculations and subsequent operation and maintenance decisions.

[0014] In one embodiment of the present invention, a risk budget, performance allowance, and life allowance are generated based on construction risk indicators, equipment health status, and remaining lifespan, and then allocated to future tunneling cycles according to a specified ratio, including: Step 21: Within the decision-making cycle, calculate the difference between the preset nominal risk upper limit and the construction risk index. If the difference is positive, it is directly used as the risk budget for the current cycle. If the difference is negative, it means that the current construction risk has exceeded the safety threshold, and the risk budget is set to 0. The nominal risk upper limit refers to the safety threshold of the acceptable construction risk index determined by engineering experience, which is used to construct the upper limit of the risk budget. Step 22: Calculate the product of the equipment health status and the preset health status conversion coefficient to obtain the performance limit; calculate the product of the remaining lifespan and the preset lifespan conversion coefficient to obtain the lifespan limit; the health status conversion coefficient and the lifespan conversion coefficient are used to proportionally map the equipment health status and remaining lifespan to the performance limit and lifespan limit, respectively; the higher the equipment health status, the larger the performance limit, reflecting the performance margin that the equipment can undertake in the current cycle for operations such as propulsion and adjustment; the lifespan limit reflects the lifespan resources that the tool or key component can be consumed in the current cycle, and is the upper bound constraint when the digital twin lifespan prediction module performs lifespan consumption simulation; Step 23: According to the preset allocation ratio for each cycle, the current risk budget and performance allowance are divided into each future tunneling cycle to form the risk budget, performance allowance, and lifespan allowance for the future tunneling cycles. The allocation ratio refers to a preset non-negative proportional sequence for multiple future tunneling cycles, the sum of which does not exceed 1, and is used to distribute the budget and allowance values ​​generated in the current cycle in future cycles.

[0015] Through the above process, this embodiment transforms the original decision-making mechanism that relied solely on real-time data into a resource management mechanism based on risk capacity, performance capacity, and lifespan capacity. By quantifying, decomposing, and periodically managing the three types of resources—risk, performance, and lifespan—the operation and maintenance decisions not only depend on the current state but are also controlled by future trends under digital twin simulation, thereby achieving steady-state tunneling and controllable resource utilization of the tunnel boring machine under complex geological conditions.

[0016] In one embodiment of the present invention, operation and maintenance decision variables are determined based on target advance speed, target chamber pressure, maintenance operation interval, and tool replacement strategy parameters. These variables are then input into a risk increment prediction model and a lifespan consumption prediction model along with a unified state data set to obtain risk increment and lifespan consumption, forming a feasible decision set, including: Step 31: Read the target feed speed range and step size, target pressure range and step size, and maintenance interval value set; combine the replacement threshold, inspection window, and replacement order to form tool usage strategy parameters; combine and number them according to the first field order to form a candidate maintenance decision variable sequence; the target feed speed range and step size refer to the value sequence of the feed speed selectable values ​​after being discretized by a fixed step size; the target pressure range and step size refer to the value sequence of the pressure control target after being discretized by the upper and lower limits of the range and the step size; the maintenance interval value set refers to the discrete interval values ​​that can be used for maintenance tasks; the tool replacement strategy parameters consist of three fields: replacement threshold, which is used to determine whether the tool needs to be replaced based on the remaining life or wear threshold; inspection window, which is used for the fixed ring length of tool inspection; and replacement order, which is used to indicate the execution order of tool replacement. Step 32: Input the unified status data set and candidate operation and maintenance decision variables into the risk increment prediction model and the life consumption prediction model one by one according to the number. The risk increment prediction model and the life consumption prediction model are both prediction sub-models in the digital twin system. They are used to simulate the changes in construction risks and life consumption that may be caused by the current operation and maintenance decision variables in the actual tunneling process. Before the system is deployed, these two models are trained based on historical tunneling data, equipment operation status records, slag and silo pressure information, tool wear records and maintenance files, etc. After the training is completed, they are solidified in the form of fixed model parameters so that they can continuously receive input from the unified status data set and operation and maintenance decision variables during the operation period and output risk increment and life consumption respectively, and establish records corresponding to the number. Step 33: Compare the risk increment and risk budget in the records in sequence according to the number, and retain the records that are not greater than the risk budget; and on the basis of the retained records, compare the life consumption and life limit, and retain the records that are not greater than the life limit and meet the target advance speed, target warehousing pressure and maintenance operation interval boundaries, so as to ensure the physical executability of the decision variables, and collect them into a feasible decision set according to the number.

[0017] This embodiment discretizes and combines the operating parameters in space, and uses a digital twin model to simulate each candidate scheme one by one. This allows the risk increment and life consumption to be quantified before decision-making, thereby enabling predictive assessment of the safety and durability of different operating schemes. This ensures that the tunnel boring machine can still achieve stable and controllable operation and maintenance decisions under complex geological conditions and life-limited conditions.

[0018] In one embodiment of the present invention, a feasible domain repair mechanism is further included, comprising: Candidate maintenance decision variables that are outside the feasible decision set and whose excess ratio relative to the risk budget or lifespan limit does not exceed the preset repair limit are evaluated. The excess ratio relative to the risk budget and lifespan limit is calculated separately. The candidate maintenance decision variables consist of target advance speed, target inventory pressure, maintenance operation interval, and tool usage strategy parameters. The excess ratio is used to quantify the degree to which the model prediction output of a candidate maintenance decision variable exceeds the corresponding limit, and is defined as the difference between the predicted value and the limit divided by the limit. If the excess ratio of the predicted risk increment or lifespan consumption relative to the risk budget or lifespan limit does not exceed the repair limit, it is marked as a repairable object. If either excess ratio exceeds the repair limit, it indicates that the candidate solution has significantly deviated from the safety or lifespan boundary and is not considered as a repairable object. First, the target advance speed and target silo pressure are scaled according to the corresponding excess ratio in a fixed order. Then, the maintenance operation interval is adjusted by a fixed increment to obtain the candidate operation and maintenance decision variables after repair. The scaling means that the parameters are adjusted proportionally by subtracting the scaling factor corresponding to the excess ratio from 1. The corrected candidate operation and maintenance decision variables are checked against risk budget and lifetime allowance. Those satisfying both types of constraints are incorporated into the feasible decision set, and a correction marker is recorded. Specifically, the corrected candidate operation and maintenance decision variables are re-input into the risk increment prediction model and lifetime consumption prediction model to recalculate the corresponding risk increment and lifetime consumption. These are then checked against the risk budget and lifetime allowance, respectively. If the corrected variables simultaneously satisfy both types of constraints, the candidate solution is incorporated into the feasible decision set, and a correction marker is added to distinguish its origin; otherwise, it is removed.

[0019] Through the aforementioned feasible domain repair mechanism, this invention provides a limited opportunity to correct slightly out-of-bounds decision variables under the deduction conditions of the digital twin model, thereby enhancing the coverage of the search space and enabling the system to retain the potential executable schemes generated in a structured manner even when the parameters are discretized.

[0020] In one embodiment of the present invention, for each decision combination in the feasible decision set, a total cost is obtained by weighting the cost of capacity loss, energy consumption, risk overrun cost, and lifetime consumption cost, and the final operation and maintenance decision is determined, including: Step 41: For each decision combination in the feasible decision set, calculate the tunneling progress deviation to obtain the capacity loss cost based on the unified state data set and the decision combination. This cost represents the degree of deviation of a decision combination from the target tunneling capacity. Calculate the energy consumption to obtain the energy consumption cost, which represents the expected energy consumption of the decision combination under the control of propulsion force, cutterhead torque, and chamber pressure. Calculate the risk increment relative to the risk budget to obtain the risk overrun cost, which represents the risk increment prediction relative to the risk budget, i.e., how much risk resources the current decision combination uses. Calculate the lifetime consumption relative to the lifetime allowance to obtain the lifetime consumption cost, which represents the predicted lifetime consumption relative to the lifetime allowance, reflecting the degree of utilization of tool or key structural component lifetime resources. Step 42: The total cost of each decision combination is obtained by weighted summation of capacity loss cost, energy consumption cost, risk overrun cost and lifetime consumption cost. The total cost is used to select the best option within the set of feasible decisions. Step 43: Compare the total costs within the feasible decision set and select the decision combination with the minimum total cost as the final operation and maintenance decision; when there are multiple decision combinations with the same total cost, compare the capacity loss cost, risk overrun cost, lifetime consumption cost and energy consumption cost in sequence, and select a unique final operation and maintenance decision according to the preset priority to ensure that only a unique and optimal operation and maintenance decision is selected in the end.

[0021] Through the aforementioned multi-dimensional cost assessment mechanism, this invention integrates the risk increment and lifespan consumption information generated by the digital twin with the tunneling efficiency and energy consumption information into the operation and maintenance decision assessment system, realizing data-based global decision optimization. Under conditions of complex strata and severe variable coupling, the system can automatically select the executable and resource-optimal operation and maintenance strategy.

[0022] In one embodiment of the present invention, the final operation and maintenance decision is transformed into a standardized set of tasks, which are then distributed to the execution layer, freezing the corresponding risk budget and lifetime allowance, including: Step 51: Map the target advance speed and target inventory pressure in the final operation and maintenance decision to the execution parameters of the parameter setting task, map the maintenance operation interval to the execution parameters of the maintenance arrangement task, map the tool usage strategy parameters to the execution parameters of the tool strategy task, and generate a unique number. Combine them into a standardized task set according to the order of the second field to realize the transformation of decision variables into instruction units that can be directly triggered. Step 52: Set up a risk budget allocation coefficient set and a lifetime quota allocation coefficient set. For each task in the standardized task set, calculate the risk budget occupancy by multiplying the risk budget by the corresponding risk budget allocation coefficient, and calculate the lifetime quota occupancy by multiplying the lifetime quota by the corresponding lifetime quota allocation coefficient. When the sum of the occupancy of each task exceeds the total amount of the corresponding category, adjust the occupancy of the last task so that the total occupancy is equal to the total amount of the corresponding category. This ensures that the resource freeze amount is consistent with the resource budget and avoids the problem of resource pre-allocation overflow or insufficiency. Generate a freeze list corresponding to the task number. Step 53: Construct a distribution package containing a standardized task set, a freeze list, and version numbers, distribute it to the execution layer, and receive a receipt. The distribution package not only provides the parameter information required for task execution but also includes a freeze list, enabling the execution layer to strictly adhere to the risk budget and lifetime limits during task execution.

[0023] Through the aforementioned task-based transformation and freezing mechanism, this invention transforms the optimal operation and maintenance decisions derived from the digital twin system into structured, executable, and traceable task units. By binding risk budgets and lifetime limits to specific tasks in the form of a frozen list, risk control and lifetime resource management can be precisely executed at the task level. This embodiment achieves the unification of decision deduction, budget generation, and task-based structure on the digital twin side, and the unification of resource constraints, execution instructions, and task traceability on the execution side, thereby improving the intelligent management and process controllability of tunnel boring machine construction under complex geological conditions.

[0024] In one embodiment of the present invention, after the standardized task set is generated, task concurrency limiting and timing rearrangement are performed, including: The standardized task set is mapped to time intervals according to the triggering conditions, and the sum of the risk budget occupancy and lifespan quota occupancy of the frozen list is calculated in each overlapping interval; the triggering conditions refer to the rules defining the execution start point of each task, such as reaching a certain shield ring number, passing a specified time interval, or the trigger point obtained by mapping from the maintenance operation interval. Set risk concurrency limits and lifetime concurrency limits. If the sum of the occupancy in any overlapping interval exceeds the corresponding limit, it is determined that there is a potential resource concurrency conflict in that interval. The lower priority tasks are postponed to non-overlapping positions according to a fixed priority. The postponement operation only adjusts the execution sequence of the tasks and does not change the task number, task content, or resource occupancy in the frozen list, so as to ensure the consistency and traceability of task records. The process is repeated until all intervals meet the concurrency limit constraint, generating a rearranged task sequence and updating the delivery package and version number. Specifically, after each postponement operation, the system re-executes overlapping interval identification and resource consumption calculation, and continues to verify against the concurrency limit. If intervals still violate the concurrency limit, lower priority tasks are postponed. After repeated iterations, the system completes task sequence rearrangement until all time intervals meet the risk concurrency limit and lifetime concurrency limit. Finally, the rearranged task sequence is written to the delivery package, and the version number of the delivery package is updated to ensure that the task sequence received by the execution layer is completely consistent with the system record.

[0025] Through the above-mentioned task concurrency limiting and timing reordering mechanisms, this invention achieves smooth management of risk budget and lifetime allowance at the task level, avoiding the accumulation of hidden risks caused by short-term high load; at the same time, it ensures the consistency of task execution order between digital twin simulation, task issuance and on-site execution, improving the stability and controllability of operation and maintenance scheduling.

[0026] In one embodiment of the present invention, actual construction data is collected to form actual unified state data, actual equipment health status, and actual construction risk indicators. These are compared with corresponding predicted values ​​to obtain deviation data, and abnormal deviation events are identified, including: Step 61: Collect actual construction data within the decision-making cycle, including propulsion force, cutterhead torque, propulsion speed, silo pressure, soil parameters, and cutter wear data, and align them with the shield ring number according to a unified timestamp; linearly interpolate or remove missing records to form an actual unified state data set, and call the aforementioned health weight vector and risk index weight vector to calculate the actual equipment health and actual construction risk index. Step 62: Obtain the final operation and maintenance decision and version number, call the risk increment prediction model and the life consumption prediction model to obtain the risk increment prediction value and the life consumption prediction value, generate a unified state data set, and call the aforementioned health weight vector and risk index weight vector to calculate the predicted equipment health and the predicted construction risk index. Subtract the predicted unified state data set from the actual unified state data set item by item and calculate the state deviation according to the Euclidean norm to reflect the overall deviation of the multi-dimensional state. Calculate the absolute difference between the actual health and the predicted health as the health deviation, the absolute difference between the actual risk index and the predicted risk index as the risk deviation, and the absolute difference between the actual life consumption and the predicted life consumption as the life consumption deviation, and generate deviation records. Step 63: If any of the risk deviation, health deviation, lifespan consumption deviation, or state deviation exceeds the corresponding deviation threshold, an abnormal deviation event record is generated.

[0027] It should be noted that actual construction data refers to the multi-source raw data collected by on-site sensors, monitoring systems, and equipment control units during the tunnel boring machine (TBM) construction process. The unified state data set is a standardized set of state vectors formed by organizing and normalizing the multi-source raw data according to a preset unified timestamp and TBM ring number. The predicted unified state data set refers to the set of predicted state vectors derived by the system based on the unified state data set, given the final operation and maintenance decision, by calling the risk increment prediction model and the life consumption prediction model respectively. The actual unified state data set refers to the standardized set of state vectors formed by applying the same time alignment, normalization, and field sorting rules as the unified state data set to the actual construction data. It is strictly consistent with the predicted unified state data set in terms of field structure and is used for consistency analysis between the actual construction state and the digital twin predicted state.

[0028] Through the aforementioned actual monitoring and deviation identification mechanism, this invention ensures that the predicted behavior of the digital twin model remains consistent with the actual operation of the tunnel boring machine. It can also promptly trigger subsequent clearing, redistribution, and task rollback mechanisms when significant deviations are detected, enabling the system to have self-monitoring, self-verification, and self-repair capabilities, thereby enhancing the stability, controllability, and adaptability to complex geological disturbances in the tunnel boring process.

[0029] In one embodiment of the present invention, the risk budget and lifespan allowance are settled and the remaining allowance ledger is updated based on actual construction risk indicators and actual lifespan consumption, including: Step 71: Obtain construction risk indicators as risk baselines, calculate the difference between actual construction risk indicators and risk baselines as actual risk increments, and record the difference as zero if it is negative, and use the actual life consumption as the settlement base. Step 72: Calculate the total risk budget usage and lifespan quota usage in the frozen list. Calculate the ratio of actual risk increment to total risk budget usage as the first liquidation ratio coefficient, and the ratio of actual lifespan consumption to total lifespan quota usage as the second liquidation ratio coefficient. Ratios exceeding one are rounded down to one. Write off the two types of usage for each task according to the corresponding liquidation ratio, and calculate the difference between actual consumption and total usage as risk overrun and lifespan overrun. Negative values ​​are counted as zero. Step 73: Calculate the difference between the risk budget occupancy and the current risk consumption for each task as the task-level risk budget balance, and the difference between the lifetime quota occupancy and the current actual lifetime quota consumption as the task-level lifetime quota balance. Calculate the difference between the total risk budget occupancy for this period and the current total risk consumption as the period-level risk budget balance, and the difference between the total lifetime quota occupancy for this period and the current total actual lifetime quota consumption as the period-level lifetime quota balance. Generate a remaining quota ledger and establish a connection with the task number, frozen list, and package version number to provide input basis for budget generation, allocation ratio adjustment, and potential abnormal rollback in the next period.

[0030] Through the above-mentioned clearing and ledger update process, this embodiment realizes the verification of resource usage based on digital twin predicted values ​​and actual on-site values, making the use of risk budget and lifespan allowance transparent, auditable and data consistent, enabling the system to continuously maintain resource usage constraints in multi-cycle tunneling tasks, strengthening the safety of the construction process and the controllability of lifespan management, and realizing a data-driven dynamic operation and maintenance closed loop.

[0031] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0032] The embodiments of the present invention have been described above, but the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of the present embodiments, all of which are within the protection scope of the present embodiments.

Claims

1. A shield tunneling machine operation and maintenance management system based on data analysis, characterized in that, include: The state modeling module is used to acquire geological conditions, shield equipment operating status, slag and silo pressure information, cutter wear data and remaining life data, form a unified state data set, and determine equipment health and construction risk indicators. The quota allocation module is used to generate risk budgets, performance quotas, and life quotas based on construction risk indicators, equipment health, and remaining lifespan, and allocate them to future tunneling cycles according to the allocation ratio. The decision constraint module is used to determine the operation and maintenance decision variables based on the target advance speed, target inventory pressure, maintenance operation interval and tool replacement strategy parameters, and input them with the unified state data set into the risk increment prediction model and life consumption prediction model to obtain the risk increment and life consumption, forming a feasible decision set. The cost optimization module is used to calculate the total cost by weighting the various decision combinations in the feasible decision set based on capacity loss cost, energy consumption cost, risk overrun cost and lifetime consumption cost, and to determine the final operation and maintenance decision. The task orchestration module is used to transform the final operation and maintenance decision into a standardized set of tasks, which are then distributed to the execution layer, freezing the corresponding risk budget and lifetime allowance. The deviation assessment module is used to collect actual construction data to form actual unified status data, actual equipment health and actual construction risk indicators, compare them with the corresponding predicted values ​​to obtain deviation data, and identify abnormal deviation events. The settlement ledger module is used to settle and update the remaining amount ledger based on actual construction risk indicators and actual life consumption, as well as the risk budget and life limit.

2. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, Generate a unified status data set and determine equipment health and construction risk indicators, including: Step 11: Acquire geological conditions, shield tunneling equipment operating status, slag and silo pressure information, cutter wear data and remaining life data within the decision-making cycle, and organize them according to a unified timestamp and shield ring number to form a multi-source raw data set; Step 12: Assign each original data item in the multi-source original data set to a state variable, perform normalization processing to obtain normalized state data, and combine them in a fixed order to form a unified state data set. Step 13: According to the preset health weight vector, the normalized state data in the unified state data set is weighted and summed, and the summation result is subtracted by 1 to determine the equipment health. According to the preset risk index weight vector, the normalized state data is weighted and summed to determine the construction risk index.

3. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, Based on construction risk indicators, equipment health status, and remaining lifespan, a risk budget, performance allowance, and lifespan allowance are generated and allocated to future tunneling cycles according to the allocation ratio, including: Step 21: Within the decision-making cycle, calculate the difference between the preset nominal risk ceiling and the construction risk index. If the difference is positive, it is directly used as the risk budget for the current cycle; if the difference is negative, the risk budget is set to 0. Step 22: Calculate the product of the equipment health status and the preset health status conversion coefficient to obtain the performance limit; calculate the product of the remaining lifespan and the preset lifespan conversion coefficient to obtain the lifespan limit. Step 23: According to the preset allocation ratio for each cycle, the current risk budget and performance allowance are divided into each future tunneling cycle to form the risk budget, performance allowance and lifespan allowance for the future tunneling cycle.

4. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, Determine the operation and maintenance decision variables, and input them, along with a unified state data set, into the risk increment prediction model and the lifetime consumption prediction model to obtain the risk increment and lifetime consumption, forming a feasible decision set, including: Step 31: Read the target advance speed range and step size, target pressure range and step size, and maintenance operation interval value set; combine the replacement threshold, inspection window, and replacement order to form tool usage strategy parameters; combine and number them according to the first field order to form a candidate operation and maintenance decision variable sequence; Step 32: Input the unified status data set and candidate operation and maintenance decision variables into the risk increment prediction model and the life consumption prediction model one by one according to the number. The two models are digital twin sub-models trained and solidified based on historical tunneling data and maintenance records. The two models output risk increment and life consumption respectively, and establish records corresponding to the number. Step 33: Compare the risk increment and risk budget in the records in sequence according to the number, and retain the records that are not greater than the risk budget; and on the basis of the retained records, compare the life consumption and life limit, and retain the records that are not greater than the life limit and meet the target advance speed, target inventory pressure and maintenance operation interval boundaries, and collect them into a feasible decision set according to the number.

5. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 4, characterized in that, It also includes feasible domain repair mechanisms, including: For candidate operation and maintenance decision variables that are outside the feasible decision set and whose excess ratio of risk budget or lifetime limit does not exceed the preset repair limit, the excess ratio of risk budget and lifetime limit is calculated respectively. First, the target advance speed and target silo pressure are scaled according to the corresponding excess ratio in a fixed order. Then, the maintenance operation interval is adjusted by a fixed increment to obtain the candidate operation and maintenance decision variables after repair. Risk budget and lifetime limit verification are performed on the candidate operation and maintenance decision variables after repair. Those that meet the two types of constraints are incorporated into the feasible decision set and repair markers are recorded.

6. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, For each decision combination in the feasible decision set, the total cost is obtained by weighting the costs of capacity loss, energy consumption, risk overrun, and lifetime depletion, and the final operation and maintenance decision is determined, including: Step 41: For each decision combination in the feasible decision set, calculate the tunneling progress deviation to obtain the production capacity loss cost based on the unified state data set and the decision combination, calculate the energy consumption to obtain the energy consumption cost, calculate the risk increment relative to the risk budget to obtain the risk overrun cost, and calculate the lifetime consumption relative to the lifetime allowance to obtain the lifetime consumption cost. Step 42: Weighted summation of capacity loss cost, energy consumption cost, risk overrun cost, and lifetime consumption cost to obtain the total cost of each decision combination; Step 43: Compare the total costs within the feasible decision set and select the decision combination with the minimum total cost as the final operation and maintenance decision; when there are multiple decision combinations with the same total cost, compare the capacity loss cost, risk overrun cost, lifetime consumption cost and energy consumption cost in sequence, and select a unique final operation and maintenance decision according to the preset priority.

7. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, The final operational decisions are transformed into a standardized set of tasks, which are then distributed to the execution layer, freezing the corresponding risk budget and lifetime allowance, including: Step 51: Map the target advance speed and target inventory pressure in the final operation and maintenance decision to the execution parameters of the parameter setting task, map the maintenance operation interval to the execution parameters of the maintenance arrangement task, map the tool usage strategy parameters to the execution parameters of the tool strategy task, and generate a unique number, and combine them into a standardized task set according to the order of the second field. Step 52: Set up a risk budget allocation coefficient set and a lifetime quota allocation coefficient set. For each task in the standardized task set, calculate the risk budget occupancy by multiplying the risk budget by the corresponding risk budget allocation coefficient, and calculate the lifetime quota occupancy by multiplying the lifetime quota by the corresponding lifetime quota allocation coefficient. When the sum of the occupancy of each task exceeds the total amount of the corresponding category, adjust the occupancy of the last task so that the total occupancy is equal to the total amount of the corresponding category, and generate a freeze list corresponding to the task number. Step 53: Construct a distribution package containing a standardized task set, a freeze list, and a version number, distribute it to the execution layer, and receive a receipt.

8. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 7, characterized in that, After the standardized task set is generated, task concurrency limiting and timing reordering are performed, including: The standardized task set is mapped to time intervals according to the triggering conditions, and the sum of the risk budget occupancy and life limit occupancy of the frozen list is calculated in each overlapping interval; Set risk concurrency limits and lifetime concurrency limits. If the sum of the occupancy in any overlapping interval exceeds the corresponding limit, the lower priority tasks will be postponed to non-overlapping positions according to a fixed priority. Repeat the process until all intervals meet the concurrency limit constraint, generate the rearranged task sequence, and update the distribution package and version number.

9. The shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, Collect actual construction data to form actual unified status data, actual equipment health status, and actual construction risk indicators. Compare these with corresponding predicted values ​​to obtain deviation data and identify abnormal deviation events, including: Step 61: Collect actual construction data within the decision-making cycle, align it with the shield ring number according to the unified timestamp; linearly interpolate or remove missing records to form an actual unified status data set, and calculate the actual equipment health and actual construction risk indicators. Step 62: Obtain the final operation and maintenance decision and version number, call the risk increment prediction model and the life consumption prediction model to obtain the risk increment prediction value and the life consumption prediction value, generate a unified state data set, and calculate the predicted equipment health and the predicted construction risk index. Subtract the predicted unified state data set from the actual unified state data set item by item and calculate the state deviation according to the Euclidean norm. Calculate the absolute difference between the actual health and the predicted health as the health deviation, the absolute difference between the actual risk index and the predicted risk index as the risk deviation, and the absolute difference between the actual life consumption and the predicted life consumption as the life consumption deviation, and generate deviation records. Step 63: If any of the risk deviation, health deviation, lifespan consumption deviation, or state deviation exceeds the corresponding deviation threshold, an abnormal deviation event record is generated.

10. A shield tunneling machine operation and maintenance management system based on data analysis according to claim 1, characterized in that, Based on actual construction risk indicators and actual lifespan consumption, the risk budget and lifespan allowance are settled and the remaining allowance ledger is updated, including: Step 71: Obtain construction risk indicators as risk baselines, calculate the difference between actual construction risk indicators and risk baselines as actual risk increments, and record the difference as zero if it is negative, and use the actual life consumption as the settlement base. Step 72: Calculate the total risk budget usage and lifespan quota usage in the frozen list, and use the ratio of actual risk increment to total risk budget usage as the first liquidation ratio coefficient, and the ratio of actual lifespan consumption to total lifespan quota usage as the second liquidation ratio coefficient; write off the two types of usage for each task according to the corresponding liquidation ratio, and calculate the difference between actual consumption and total usage as risk overspending and lifespan overspending. Step 73: Calculate the difference between the risk budget occupancy and the current risk consumption for each task as the task-level risk budget balance, the difference between the lifetime quota occupancy and the current actual lifetime quota consumption as the task-level lifetime quota balance, calculate the difference between the total risk budget occupancy for this period and the current total risk consumption as the period-level risk budget balance, and the difference between the total lifetime quota occupancy for this period and the current total actual lifetime quota consumption as the period-level lifetime quota balance, and generate the remaining quota ledger.

Images