Intelligent tunnel supporting method and system based on data drift detection

Abstract

The invention discloses an intelligent tunnel supporting method and system based on data drift detection. The method comprises the steps that actual monitoring data in the tunnel supporting process are obtained; determining a preliminary support scheme according to the actual monitoring data and the support intelligent decision model; calculating a drift index between the actual monitoring data and the reference monitoring data; judging whether the actual monitoring data drifts or not according to the drifting index; when it is judged that the actual monitoring data drifts seriously, the conservative support emergency plan is used as a final support scheme, and the support intelligent decision model and the reference monitoring data are updated, so that the accuracy of the support scheme is ensured, and the safety of support engineering is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction technology for tunnel engineering, and in particular to an intelligent tunnel support method and system based on data drift detection. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Tunnel support is a core component ensuring the safety and stability of tunnels during construction and operation. In recent years, data-driven intelligent support methods have become a research hotspot. These methods involve deploying various sensors to collect real-time mechanical response data of the tunnel's surrounding rock and support structure, and using machine learning models to establish a mapping relationship between monitoring data and support status or required support parameters, thereby enabling intelligent recommendations for support solutions.

[0004] However, existing intelligent support methods generally suffer from a critical flaw: their implicit assumption is that the historical data used in the model training phase follows the same probability distribution as the current actual monitoring data. In actual tunnel engineering, this assumption often fails. Numerous factors can cause changes in data distribution (i.e., "data drift"), such as: tunnel excavation traversing different geological units, changes in construction techniques, the influence of seasonal environmental factors, and the performance degradation or failure of the sensors themselves. When data drift occurs, the predictive performance of machine learning models trained on historical data will significantly decrease, or even become completely ineffective. Continuing to use a degraded model for support schemes may lead to serious misjudgments or recommend inapplicable or unsafe support schemes, thereby triggering engineering risks. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a method and system for intelligent tunnel support based on data drift detection, which improves the accuracy of the final support scheme.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a tunnel intelligent support method based on data drift detection is proposed, including: Obtain actual monitoring data during the tunnel support process; Based on actual monitoring data and the intelligent support decision-making model, a preliminary support plan is determined; Calculate the drift index between actual monitoring data and baseline monitoring data; Determine whether there is drift in the actual monitoring data based on drift indicators; When it is determined that there is a serious drift in the actual monitoring data, the conservative support emergency plan will be used as the final support plan, and the support intelligent decision-making model and benchmark monitoring data will be updated.

[0007] Furthermore, when it is determined that there is no drift in the actual monitoring data, the preliminary support plan will be used as the final support plan. When it is determined that there is potential drift in the actual monitoring data, the preliminary support plan is marked with risks and a potential drift warning is issued.

[0008] Furthermore, when the drift index is less than or equal to the first threshold, it is determined that there is no drift in the actual monitoring data; When the drift index is less than or equal to the second threshold and greater than the first threshold, it is determined that there is potential drift in the actual monitoring data. When the drift index exceeds the second threshold, it is determined that there is serious drift in the actual monitoring data.

[0009] Furthermore, when the cumulative number of potential drift warnings issued exceeds the set number, the support intelligent decision-making model and benchmark monitoring data are updated.

[0010] Furthermore, the support intelligent decision-making model includes classifiers and regressors; The classifier takes monitoring data as input and outputs the status of the support project. The regression analyzer takes monitoring data as input and outputs a preliminary support plan.

[0011] Furthermore, the process of calculating the drift index between the actual monitoring data and the baseline monitoring data is as follows: The actual monitoring data is reconstructed by a trained data encoder to obtain reconstructed data. The trained data encoder is obtained by training on the benchmark monitoring data. The error between the actual monitoring data and the reconstructed data is calculated as an indicator of the drift between the actual monitoring data and the baseline monitoring data.

[0012] Secondly, a tunnel intelligent support system based on data drift detection is proposed, including: The data acquisition unit is used to acquire actual monitoring data during the tunnel support process. The preliminary support scheme determination unit is used to determine the preliminary support scheme based on actual monitoring data and the intelligent support decision-making model; The drift index calculation unit is used to calculate the drift index between the actual monitoring data and the benchmark monitoring data. The drift status determination unit is used to determine whether there is drift in the actual monitoring data based on the drift index. The decision control unit is used to adopt the conservative support emergency plan as the final support scheme when it is determined that there is a serious drift in the actual monitoring data, and to update the support intelligent decision model and the baseline monitoring data.

[0013] Thirdly, a computer device is proposed, the device comprising: A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program, which, when executed by the processor, implements the intelligent tunnel support method based on data drift detection proposed in the first aspect.

[0014] Fourthly, a computer-readable storage medium is proposed, wherein the computer-readable storage medium stores a computer program adapted to be loaded and executed by a processor, the first aspect proposing a method for intelligent tunnel support based on data drift detection.

[0015] Fifthly, a computer program product is proposed, which includes a computer program that, when executed by a processor, implements the intelligent tunnel support method based on data drift detection proposed in the first aspect.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes a tunnel intelligent support method and system based on data drift detection. After determining the preliminary support scheme according to the intelligent support decision model, the method also calculates the drift index between the actual monitoring data and the benchmark monitoring data. Then, it determines whether the actual monitoring data has drifted based on the drift index. When it is determined that the actual monitoring data has serious drift, it indicates that the preliminary support scheme is inaccurate. Therefore, a conservative support emergency plan is adopted as the final support scheme to ensure the safety of the support project. At the same time, the intelligent support decision model and the benchmark monitoring data are updated to improve the accuracy of subsequent preliminary support schemes and drift index calculations.

[0017] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0019] Figure 1 This is a flowchart of a tunnel intelligent support method based on data drift detection proposed in an embodiment of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0022] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0023] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0024] The present invention proposes a tunnel intelligent support method based on data drift detection, which is applied to the application scenario of determining tunnel support scheme.

[0025] Tunnel support is a core component ensuring the safety and stability of tunnels during construction and operation. Traditional tunnel support design primarily relies on engineering analogy, theoretical calculation, and feedback design based on on-site monitoring and measurement information. However, these methods have significant limitations when facing complex and variable geological conditions. Engineering analogy is highly subjective and struggles to accurately address specific geological formations; theoretical calculation relies on numerous idealized assumptions, leading to discrepancies with actual engineering responses; while feedback design offers the advantage of dynamic adjustment, its decision-making heavily depends on engineers' experience, and it typically only takes action after significant adverse deformations have occurred in the surrounding rock or support structure, exhibiting a lag and failing to provide proactive risk warnings and control.

[0026] In recent years, with the development of the Internet of Things, sensor technology, and artificial intelligence, data-driven intelligent support methods have become a research hotspot. These methods involve deploying various sensors to collect real-time mechanical response data of the tunnel surrounding rock and support structure, and using machine learning models to establish a mapping relationship between monitoring data and support status or required support parameters, thereby enabling intelligent recommendations for support schemes. This type of method reduces reliance on experience to some extent and improves the objectivity of decision-making.

[0027] A key flaw in existing intelligent support methods is their implicit assumption that the historical data used in model training follows the same probability distribution as the current actual monitoring data. In practical tunnel engineering, this assumption often fails. Numerous factors can cause changes in data distribution (i.e., "data drift"), such as tunnel excavation traversing different geological units, changes in construction techniques, the influence of seasonal environmental factors, and the performance degradation or failure of the sensors themselves. When data drift occurs, the predictive performance of machine learning models trained on historical data will significantly decrease, or even become completely ineffective. Continuing to use a degraded model for support schemes may lead to serious misjudgments or recommend unsuitable or unsafe support solutions, thereby triggering engineering risks.

[0028] Therefore, ensuring the continuous reliability and adaptability of intelligent support models throughout the entire tunnel lifecycle, especially under dynamic changes in geological and construction conditions, is a core challenge that urgently needs to be addressed in the field of intelligent tunnel construction. Data drift detection technology offers a new solution to this problem. By monitoring the distribution changes of input data in real time and triggering corresponding model updates or alarm mechanisms, the long-term effectiveness of the intelligent support system can theoretically be maintained. However, there is currently no mature solution for how to organically and efficiently embed data drift detection technology into the complete decision-making process of intelligent tunnel support and design a methodology that balances accuracy, timeliness, and engineering feasibility.

[0029] To accurately determine the support scheme and ensure the safety of the support project, this invention proposes a tunnel intelligent support method based on data drift detection. This method deeply integrates the data drift detection mechanism into the closed-loop process of support scheme determination. By monitoring changes in the statistical characteristics of the data stream in real time, it automatically identifies the failure risk of the intelligent support decision-making model and intelligently triggers updates, drift warnings, or conservative decision-making processes. This ensures that the support scheme maintains high reliability and adaptability throughout the entire tunnel construction process, achieving a leap from "static intelligence" to "dynamic adaptive intelligence," guaranteeing the accuracy of the final support decision, and thus ensuring the safety of the support project.

[0030] like Figure 1 As shown in the figure, an intelligent tunnel support method based on data drift detection proposed in this embodiment of the invention includes: Obtain actual monitoring data during the tunnel support process; Based on actual monitoring data and the intelligent support decision-making model, a preliminary support plan was determined. ; Calculate the drift index between actual monitoring data and baseline monitoring data; Determine whether there is drift in the actual monitoring data based on drift indicators; When it is determined that there is a serious drift in the actual monitoring data, the conservative support emergency plan will be used as the final support plan, and the support intelligent decision-making model and benchmark monitoring data will be updated.

[0031] In addition, when it is determined that there is no drift in the actual monitoring data, the preliminary support plan will be used as the final support plan. When it is determined that there is potential drift in the actual monitoring data, the preliminary support plan is marked with risks and a potential drift warning is issued.

[0032] This invention proposes a tunnel intelligent support method based on data drift detection. After determining the preliminary support scheme according to the intelligent support decision model, it also calculates the drift index between the actual monitoring data and the benchmark monitoring data. Then, it determines whether the actual monitoring data has drifted based on the drift index. When it is determined that the actual monitoring data has serious drift, it indicates that the preliminary support scheme is inaccurate. Therefore, a conservative support emergency plan is adopted as the final support scheme to ensure the safety of the support project. At the same time, the intelligent support decision model and the benchmark monitoring data are updated using verified monitoring data to improve the accuracy of subsequent preliminary support schemes and drift index calculations.

[0033] In this embodiment of the invention, a sensor network is systematically deployed at a typical monitoring section at a certain distance behind the tunnel excavation face to acquire multi-dimensional monitoring time-series data that comprehensively reflects the mechanical state of the surrounding rock-support system.

[0034] Multi-source monitoring time series data includes, but is not limited to: (1) Convergence deformation data: The settlement and horizontal convergence data of the tunnel arch are obtained by laser rangefinder or total station.

[0035] (2) Surrounding rock stress / strain data: The stress changes inside the surrounding rock are obtained by using embedded stress gauges or strain gauges.

[0036] (3) Internal force data of support structure: The stress and strain of the lining or steel arch frame are obtained by concrete strain gauge and steel reinforcement gauge.

[0037] (4) Contact pressure data: The contact pressure between the surrounding rock and the initial support is obtained through the earth pressure cell.

[0038] (5) Environmental factors data: such as groundwater level, temperature, etc.

[0039] In this embodiment of the invention, after obtaining multi-dimensional monitoring time-series data, the multi-dimensional monitoring time-series data is preprocessed to obtain preprocessed data, and data features are extracted from the preprocessed data to obtain actual monitoring data.

[0040] The preprocessing of multivariate monitoring time-series data includes: (1) Cleaning: Process missing values ​​(fill with interpolation or average of previous and next time points), identify and remove obvious abnormal values ​​caused by instantaneous sensor failure.

[0041] (2) Noise reduction: Use methods such as moving average, wavelet transform or Kalman filtering to smooth random noise and retain the real trend signal.

[0042] (3) Normalization: Scale data of different dimensions and magnitudes to [0,1] or standard normal distribution to eliminate the influence of feature scale on the model.

[0043] The actual monitoring data obtained by extracting data features from the preprocessed data includes the data deformation rate and acceleration, the spatial gradient of stress, the logarithmic function relationship between data deformation and time, and the data statistical features (mean, variance, extreme values, trend slope, etc.) within a specific time window.

[0044] Finally, the multivariate monitoring time series data at each monitoring time t is transformed into a d-dimensional standardized feature vector. and will As actual monitoring data.

[0045] In this embodiment of the invention, actual monitoring data is input into the support intelligent decision-making model. In the process, the initial support decision is output, and the intelligent support decision model includes a classifier and a regressor; The classifier takes monitoring data as input and outputs the status of the support project. The regression analyzer takes monitoring data as input and outputs a preliminary support plan. Based on benchmark monitoring data The constructed classifier and regressor are trained, and once training is complete, a support intelligent decision-making model is obtained.

[0046] The classifier can be a random forest or XGBoost, and the output of the support engineering status is the probability of "stable", "warning", or "dangerous".

[0047] The regressor uses a neural network model or a GBDT regression model, and the output preliminary support scheme is a suggested support parameter, such as the number of anchor bolts to be added in the next cycle.

[0048] Benchmark monitoring data Feature data is obtained by preprocessing historical monitoring time-series data of known support engineering conditions and corresponding effective support schemes, and then extracting data features from the preprocessed data. Historical monitoring time-series data can be monitoring time-series data acquired in the early stages of tunnel construction, or monitoring time-series data from the historical stable phases of similar projects.

[0049] In this embodiment of the invention, monitoring data from sections with relatively uniform and stable geological conditions and normal support response during the initial stage of tunnel construction, or corresponding monitoring data from completed tunnels with similar geological conditions, are selected as benchmark monitoring data. This data should ensure that the corresponding support engineering state is "stable," and this data distribution serves as the benchmark for "normal working conditions." All benchmark monitoring data form a benchmark dataset. , .

[0050] Whether the support structure is stable can be determined by experienced experts.

[0051] The process of calculating the drift index between actual monitoring data and benchmark monitoring data in this embodiment of the invention is as follows: The actual monitoring data is reconstructed by a trained data encoder to obtain reconstructed data. The trained data encoder is obtained by training on the benchmark monitoring data. The error between the actual monitoring data and the reconstructed data is calculated as an indicator of the drift between the actual monitoring data and the baseline monitoring data. .

[0052] In this embodiment of the invention, a univariate drift detector is constructed for each data feature in the benchmark monitoring data (such as crown settlement rate, maximum contact pressure, etc.), and a data encoder is constructed to capture the relationship between the monitoring data features. This data encoder is an autoencoder (AE).

[0053] The univariate drift detector uses the KS test to calculate... The cumulative distribution function (CDF) of each feature is used as the baseline CDF.

[0054] use Train the data encoder and learn The data is efficiently compressed and represented (encoded), and reconstructed (decoded) with low error. Once training is complete, a well-trained data encoder is obtained.

[0055] Through the trained data encoder All benchmark monitoring data samples are reconstructed to obtain the benchmark reconstructed data for each benchmark monitoring data. Calculate the reconstruction error (e.g., mean square error MSE) between each baseline monitoring data and its baseline reconstructed data. Calculate the mean μ and standard deviation σ of the reconstruction errors between all baseline monitoring data and their baseline reconstructed data. Based on the mean μ and standard deviation σ, calculate and determine the first threshold T1 and the second threshold T2.

[0056] T1 = μ + 2σ; T2 = μ + 4σ.

[0057] Because AE is It is trained on, it is The reconstruction error of the data is relatively small and concentrated. When the distribution of the input data changes, the reconstruction error increases significantly. Therefore, the reconstruction error can be used as an indicator to monitor data drift.

[0058] T1 = μ + 2σ; T2 = μ + 4σ.

[0059] This invention calculates actual monitoring data using a univariate drift detector. The KS statistic for each eigenvalue relative to its baseline CDF, etc.

[0060] and actual monitoring data Input the trained data into the encoder for encoding reconstruction, obtain the reconstructed data, and calculate... The error between the reconstructed data and the original data is used to obtain the reconstruction error. and the reconstruction error As a drift indicator of actual monitoring data, the drift indicator is used to determine whether there is drift in the actual monitoring data.

[0061] When drift indicator When the actual monitoring data is less than or equal to the first threshold T1, it is determined that there is no drift in the actual monitoring data, indicating that there is no significant difference between the actual monitoring data and the benchmark monitoring data, and the model operating environment is stable. When drift indicator If the actual monitoring data is less than or equal to the second threshold T2 and greater than the first threshold T1, it is determined that there is potential drift in the actual monitoring data, indicating that the distribution of the actual monitoring data has begun to deviate from the "normal operating conditions". The predictions may be biased, but they are not completely invalid. When drift indicator When the actual monitoring data exceeds the second threshold T2, it is determined that there is a serious drift in the actual monitoring data. The actual monitoring data differs greatly from the benchmark monitoring data, which may mean that the surrounding rock conditions and construction environment have undergone fundamental changes (such as sudden faulting or large-scale water inrush). The model is no longer reliable.

[0062] Subsequently, based on whether there is any drift in the actual monitoring data, a decision will be made on the preliminary support plan. (1) When it is determined that there is no drift in the actual monitoring data, the preliminary support plan is taken as the final support plan, and the final support plan is sent to the construction command system or display terminal, and the process enters the next monitoring cycle. (2) When it is determined that there is potential drift in the actual monitoring data, the preliminary support plan should be adjusted accordingly. Risks are flagged, and potential drift warnings are issued. Specifically: By means of By labeling the preliminary support plan with "Reliability to be verified," a risk marker is applied. Potential drift warnings are then issued via audible and visual alerts or graphic warnings, notifying the site engineer: "The monitoring data pattern has changed; please review the current support plan." , and drift index The message was also sent to the engineer's desktop, suggesting that they make a comprehensive judgment based on geological sketches, on-site observations, and other factors.

[0063] In addition, when issuing a potential drift warning, the current moment will also be considered. The data is stored in a "new data pool to be verified". The system is preparing to enter the incremental model update process, but the update is not executed immediately, waiting for more new data or manual confirmation.

[0064] (3) When it is determined that there is serious drift in the actual monitoring data, the conservative support emergency plan shall be used as the final support plan, and the support intelligent decision-making model and benchmark monitoring data shall be updated.

[0065] In this embodiment, when a severe drift is determined in the actual monitoring data, it is immediately ignored. Output The system automatically acquires a pre-set conservative support emergency plan and uses it as the final support solution. This conservative support emergency plan is based on the most unfavorable situation or is a plan limited by specifications. Simultaneously, it triggers the highest-level alarm, notifying that manual emergency handling is mandatory, and forcibly initiates the support intelligent decision-making model and baseline monitoring data update process.

[0066] After the support intelligent decision-making model and benchmark monitoring data update process are started, the model is urgently updated or retrained using the recently accumulated data in the "new data pool to be verified" and the data before and after the current moment in a state of severe drift, so as to adapt to the new working conditions as soon as possible.

[0067] Adopting conservative emergency plans until a new model is developed. After training is completed and validation is passed, the system switches back to the new model for decision-making, and generates a preliminary support plan using the retrained intelligent support decision-making model.

[0068] In addition to updating the support intelligent decision-making model and benchmark monitoring data when there is serious drift in the actual monitoring data, the support intelligent decision-making model and benchmark monitoring data are also updated when the cumulative number of potential drift warnings issued exceeds a set number; updates can also be triggered periodically or manually.

[0069] Among them, periodic triggering updates refer to the automatic updating of the support intelligent decision-making model and benchmark monitoring data after each certain distance of tunneling or at regular intervals.

[0070] Manually triggered updates refer to engineers manually triggering updates to the support intelligent decision-making model and baseline monitoring data based on requirements.

[0071] When updating the support intelligent decision-making model and baseline monitoring data, all monitoring data since the last update that have been manually verified as correct or subsequently monitored to prove that the support is effective are collected. This ensures that the data used for updates is high-quality, representative, and based on new operating conditions.

[0072] The methods for updating support intelligent decision-making models include incremental learning and local retraining. For support intelligent decision-making models that support incremental learning, the data can be directly updated. Incremental updates are performed using new samples to quickly adjust model parameters.

[0073] For other types of support intelligent decision-making models, (or partial data from the previous round) and Merge the datasets and retrain the model on the new mixed dataset. To avoid forgetting old knowledge, the old data samples can be weighted.

[0074] The baseline for drift detection should not be fixed indefinitely. After model updates, or when the project clearly enters a new, stable geological phase, the baseline can be updated using stability data from the initial stage of the new phase. The AE detector is retrained, and thresholds T1 and T2 are recalculated. This is equivalent to adjusting the system's "normal" baseline to the current new environment, making drift detection still sensitive to subsequent changes.

[0075] This embodiment records each support plan and its effect over a subsequent period. Regularly analyzing these "plan-effect" chains can be used for: 1) The rationality of calibrating drift thresholds T1 and T2.

[0076] 2) Evaluate the effectiveness of different update strategies.

[0077] 3) Identify shortcomings in the model or feature engineering to guide long-term optimization.

[0078] This embodiment proposes a tunnel intelligent support method based on data drift detection, which realizes the self-sensing and self-adaptation of the support scheme.

[0079] (1) By embedding data drift detection, the intelligent support system is equipped with the ability to sense changes in the external data environment. It automatically identifies working conditions where the model may fail and actively adjusts the support scheme strategy, fundamentally improving the robustness and long-term applicability in the dynamically changing tunnel engineering environment.

[0080] (2) A tiered early warning and decision-making support mechanism was established: By setting multi-level drift thresholds, a refined decision-making process was realized, from "normal adoption" to "risk warning" and then to "conservative risk avoidance". This not only avoids overreaction under slight changes, but also provides a safety net when major changes occur, effectively balancing decision-making efficiency and engineering safety.

[0081] (3) A complete closed loop of “monitoring-decision-detection-update” has been formed: This invention regards data drift detection and model update as an indispensable link in the intelligent support closed loop, so that the system can continuously evolve with new data and form a life form that continuously learns and optimizes, thus solving the drawback of the traditional data-driven model that is “one training determines life”.

[0082] (4) Improved tunnel construction safety and risk management level: This method can indirectly issue early warnings through data drift in the early stage when the surrounding rock conditions change adversely, prompting engineers to pay attention and take measures in advance, changing passive response to active prevention and control, and significantly reducing engineering risks caused by model misjudgment or decision lag.

[0083] (5) Excellent scalability and engineering applicability: The method framework described in this invention does not rely on specific machine learning models or drift detection algorithms, and can be flexibly configured according to the data characteristics, computing resources, and accuracy requirements of specific projects. Its process is clear, easy to integrate with existing tunnel monitoring and measurement systems, and highly practical for engineering applications. This invention also proposes a tunnel intelligent support system based on data drift detection, comprising: The data acquisition unit is used to acquire actual monitoring data during the tunnel support process. The preliminary support scheme determination unit is used to determine the preliminary support scheme based on actual monitoring data and the intelligent support decision-making model; The drift index calculation unit is used to calculate the drift index between the actual monitoring data and the benchmark monitoring data. The drift status determination unit is used to determine whether there is drift in the actual monitoring data based on the drift index. The decision control unit is used to adopt the conservative support emergency plan as the final support scheme when it is determined that there is a serious drift in the actual monitoring data, and to update the support intelligent decision model and the baseline monitoring data.

[0084] It should be noted that the above-described intelligent tunnel support system based on data drift detection is only illustrated by the division of the functional modules described above when determining the support scheme. In practical applications, the functions described above can be assigned to different functional modules as needed, that is, the internal structure of the equipment can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the intelligent tunnel support system based on data drift detection provided in the above-described embodiments and the intelligent tunnel support method based on data drift detection belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0085] The present invention also discloses a computer device, the device comprising: A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program, which, when executed by the processor, implements a tunnel intelligent support method based on data drift detection disclosed in the embodiments of the present invention.

[0086] The present invention also discloses a computer-readable storage medium storing a computer program adapted for loading and execution by a processor of a tunnel intelligent support method based on data drift detection disclosed in the embodiments of the present invention.

[0087] The present invention also discloses a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements a tunnel intelligent support method based on data drift detection disclosed in the embodiments of the present invention.

[0088] The method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.

[0089] Those skilled in the art will recognize that the units and algorithm steps described in conjunction with the embodiments herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0090] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for intelligent tunnel support based on data drift detection, characterized in that, The method comprises: acquiring actual monitoring data in a tunnel support process; determining a preliminary support scheme according to the actual monitoring data and a support intelligent decision-making model; calculating a drift index between the actual monitoring data and benchmark monitoring data; judging whether the actual monitoring data has drift according to the drift index; when it is determined that the actual monitoring data has serious drift, taking a conservative support emergency plan as a final support scheme, and updating the support intelligent decision-making model and the benchmark monitoring data.

2. A method of intelligent tunnel support based on data drift detection as claimed in claim 1, wherein, when it is determined that the actual monitoring data has no drift, taking the preliminary support scheme as a final support scheme; when it is determined that the actual monitoring data has potential drift, marking the preliminary support scheme as risky and issuing a potential drift warning.

3. A method of intelligent tunneling based on data drift detection as claimed in claim 2, wherein, when the drift index is less than or equal to a first threshold value, it is determined that the actual monitoring data has no drift; when the drift index is less than or equal to a second threshold value and greater than the first threshold value, it is determined that the actual monitoring data has potential drift; when the drift index is greater than the second threshold value, it is determined that the actual monitoring data has serious drift.

4. The method of intelligent tunnel support based on data drift detection as claimed in claim 1, wherein, when the number of times of issuing the potential drift warning reaches a set number of times, the support intelligent decision-making model and the benchmark monitoring data are updated.

5. A method of intelligent tunnel support based on data drift detection as claimed in claim 1, wherein, The support intelligent decision-making model comprises a classifier and a regressor; the classifier takes the monitoring data as input and outputs a support engineering state; the regressor takes the monitoring data as input and outputs the preliminary support scheme.

6. A method of intelligent tunnel support based on data drift detection as claimed in claim 1, wherein, The process of calculating the drift index between the actual monitoring data and the benchmark monitoring data comprises: reconstructing the actual monitoring data by using a trained data encoder to obtain reconstructed data, wherein the trained data encoder is obtained by training the benchmark monitoring data; calculating an error between the actual monitoring data and the reconstructed data as the drift index between the actual monitoring data and the benchmark monitoring data.

7. A tunnel intelligent support system based on data drift detection, characterized in that, The method comprises: a data acquisition unit configured to acquire actual monitoring data in a tunnel support process; a preliminary support scheme determination unit configured to determine a preliminary support scheme according to the actual monitoring data and a support intelligent decision-making model; a drift index calculation unit configured to calculate a drift index between the actual monitoring data and benchmark monitoring data; a drift state judgment unit configured to judge whether the actual monitoring data has drift according to the drift index; a decision-making regulation unit configured to, when it is determined that the actual monitoring data has serious drift, take a conservative support emergency plan as a final support scheme, and update the support intelligent decision-making model and the benchmark monitoring data.

8. An electronic device, comprising: The device comprises: a processor adapted to execute a computer program; a computer readable storage medium having a computer program stored therein, wherein the computer program is executed by the processor to implement the tunnel intelligent support method based on data drift detection according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is adapted to be loaded and executed by the processor to implement the tunnel intelligent support method based on data drift detection according to any one of claims 1-6.

10. A computer program product, characterised in that, The computer program product comprises a computer program, and the computer program is executed by the processor to implement the tunnel intelligent support method based on data drift detection according to any one of claims 1-6.

Images