Dual-mode shield blocking machine identification method based on data driving

Abstract

The invention discloses a dual-mode shield blocking machine identification method based on data driving. The method comprises the following steps: collecting tunneling parameters of a plurality of normal tunneling sections and blocking machine tunneling sections of dual-mode shield engineering; carrying out data preprocessing on the collected original data to remove invalid data; tunneling parameters capable of effectively describing the tunneling state of the dual-mode shield tunneling machine are optimized based on a mathematical statistics method; calculating the weight of each tunneling parameter based on an entropy weight method; calculating to obtain a quantitative comprehensive index based on the weight of each tunneling parameter and the cumulative probability density function; and calculating through a mathematical statistical method to obtain a quantitative comprehensive index to judge a threshold value of the jamming machine, and further obtaining an identification mechanism capable of identifying the tunneling state of the dual-mode shield jamming machine. According to the method, the tunneling state of the dual-mode shield can be well described, and the tunneling state of the stuck machine can be well evaluated and divided.

Description

Technical Field

[0001] This invention relates to the field of evaluation technology for dual-mode shield tunnel construction, and is particularly applicable to a data-driven method for identifying dual-mode shield tunnel shields. Background Technology

[0002] The tunneling process of a tunnel boring machine is essentially a dynamic process of interaction between the rock mass and the machine. The tunneling parameters, as key feedback indicators of this process, are not only closely related to the physical and mechanical properties of the surrounding rock mass and the geological structure, but can also reflect the working status of the tunneling equipment in real time.

[0003] Currently, tunneling parameters are widely used for identifying adverse geological conditions. However, a more crucial aspect is how to achieve proactive early warning and prevention of construction disasters based on tunneling parameters. Since construction disasters frequently occur in adverse geological conditions or high-risk strata, in-depth analysis of the response patterns of tunneling parameters under such geological conditions, extraction of indicative data features, and the subsequent construction of a disaster early warning mechanism based on tunneling parameters are of paramount importance for ensuring construction safety.

[0004] In shield tunneling, common hazards include surface subsidence, collapse, sudden water inrush, large deformation of surrounding rock, and machine jamming. Among these, machine jamming holds a special place: it is not only a potential final result of hazards such as collapse and large deformation of surrounding rock, but also a direct manifestation of the imbalance between the rock mass and the shield machine. Therefore, from the perspective of this typical hazard, systematically analyzing the evolution of tunneling parameters before, during, and after its occurrence, deeply extracting characteristic parameters that can characterize the risk of machine jamming, and constructing corresponding risk identification and early warning models are of great value for improving the safety, efficiency, and intelligence level of shield tunneling and represent a core technological direction that urgently needs in-depth research. Summary of the Invention

[0005] The purpose of this invention is to provide a data-driven dual-mode shield tunneling machine jamming identification method, which is used to solve the problems of delayed early warning of machine jamming disasters and reliance on manual experience in existing shield tunneling construction based on real-time data of tunneling parameters, and to build a method that can identify machine jamming risks in advance.

[0006] To achieve the above objectives, the data-driven dual-mode shield tunneling machine identification method of the present invention includes the following steps: S1 collects the original tunneling parameters of the normal tunneling section and the machine-operated tunneling section of multiple dual-mode shield tunneling projects; S2, based on the binary discriminant function method and the 2σ principle method, preprocesses the original tunneling parameters to remove invalid data; S3, based on mathematical statistics, select effective tunneling parameters from the preprocessed tunneling parameters to effectively describe the tunneling state of the dual-mode shield machine; S4, calculate the weights of each effective tunneling parameter based on the entropy weight method; S5, Calculates a quantitative comprehensive index based on the weights of each effective tunneling parameter and the cumulative probability density function; S6, calculate the judgment threshold of the quantitative comprehensive index of the card machine, and verify the judgment threshold through mathematical statistics of multiple engineering card machine cases; S7 calculates the quantitative comprehensive index of the tunneling parameters to be evaluated, and identifies the tunneling status of the dual-mode shield tunneling machine based on the judgment threshold.

[0007] Furthermore, the tunneling parameters include thrust F, penetration depth p, cutterhead torque T, and cutterhead rotation speed n; the binary discriminant function formula is: , , That is, when S When the value is 0, it is considered invalid data and is removed.

[0008] Furthermore, the 2σ principle, following a normal distribution, eliminates […]. μ-2σ , μ+2σ Beyond the excavation data, μ The mean, σ The standard deviation is denoted as .

[0009] Furthermore, the specific steps of step S3 are as follows: the preprocessed tunneling parameters are classified into normal tunneling sections and jack-up tunneling sections, and the tunneling parameters for each section are calculated. x overlap The formula is: ,in, x These are the tunneling parameters; , These are the tunneling parameters for the normal tunneling section. x Average value and tunneling parameters of the jacking section x The average value; , Normal tunneling parameters x The variance and tunneling parameters of the card machine tunneling section x The variance; when the overlap At that time, tunneling parameters x For effective tunneling parameters.

[0010] Furthermore, in step S4, an evaluation matrix of the measured values ​​of each effective tunneling parameter is constructed based on the entropy weight method, and the entropy value and difference coefficient of each effective tunneling parameter are calculated; thus, the weight of each effective tunneling parameter is obtained.

[0011] Furthermore, the formula for calculating the quantitative comprehensive index is as follows: ,in, For quantitative comprehensive indicators; Let x be the cumulative probability density function of the effective tunneling parameter x in the machine section; Let x be the weight of the i-th effective tunneling parameter.

[0012] Furthermore, the determination threshold In the formula, For quantitative comprehensive indicators The degree of overlap; , These are quantitative comprehensive indicators. The average values ​​for the normal tunneling section and the jammed tunneling section; , For quantitative comprehensive indicators The variance values ​​of the normal tunneling section and the jacking tunneling section.

[0013] Furthermore, the determination threshold For quantitative comprehensive indicators The distinguishing value between the normal tunneling section and the jammed tunneling section.

[0014] Compared with the prior art, the beneficial effects of the present invention are mainly reflected in the following aspects: 1. More targeted: Existing TBM jamming identification methods do not fully consider the structural and working characteristics of dual-mode shield tunnels. This invention focuses on the shield jamming problem for the first time on the special model of EPB / TBM dual-mode shield tunnels, and studies the shield-surrounding rock interaction mechanism in the process of mode transformation and adaptation to composite strata, filling the gap in identification methods in this field.

[0015] 2. More systematic and in-depth identification mechanism: By systematically collecting multi-source tunneling parameters of dual-mode shield tunneling under different tunneling conditions, and combining mathematical statistics and multivariate analysis methods, key parameter combinations that are sensitive to shield jamming were screened out. Furthermore, entropy weight method and statistical testing were introduced to establish a new indicator that can quantify the risk of jamming, realizing a leap from "experience-based judgment" to "parameterized and indexed" identification.

[0016] 3. The method is innovative and practical: For the first time, a complete quantitative identification method system for dual-mode shield tunneling machine chucks is proposed, including key parameter screening, risk indicator construction, threshold determination and status discrimination criteria. It provides an operable and reusable technical tool for engineering practice and significantly improves the objectivity and accuracy of chuck identification.

[0017] 4. Significant engineering guidance significance: This method can effectively identify shield jamming machines by relying only on conventional tunneling parameters, without the need for additional sensors. It has good field applicability and economy, and provides a new technical approach for real-time monitoring and early warning of jamming risks in dual-mode shield tunneling, which helps to improve construction safety and efficiency. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method described in this invention.

[0019] Figure 2 This is a normal distribution diagram of tunneling parameters under normal tunneling conditions obtained in an embodiment of the present invention.

[0020] Figure 3 This is a normal distribution diagram of tunneling parameters under the tunneling state obtained in an embodiment of the present invention.

[0021] Figure 4 This is a feature map of the preprocessed tunneling parameters obtained in an embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of the probability density distribution and threshold of the tunneling parameters of the present invention.

[0023] Figure 6 This is a graph showing the verification results of the quantitative comprehensive index of the dual-mode shield tunneling machine obtained in an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1 like Figure 1 As shown, the data-driven dual-mode shield machine identification method of the present invention includes the following steps: S1 collects the original tunneling parameters of the normal tunneling section and the machine-operated tunneling section of multiple dual-mode shield tunneling projects.

[0026] S2, based on the collected raw tunneling parameters, preprocess the raw tunneling parameters to remove invalid data using the binary discriminant function method and the 2σ principle method.

[0027] The original tunneling parameters, including thrust F, penetration depth p, cutterhead torque T, and cutterhead rotation speed n, are used to determine whether the EPB / TBM dual-mode shield tunnel is in normal working condition.

[0028] The formula for the binary discriminant function is: , , That is, when S When the value is 0, it is considered invalid data and is removed.

[0029] Statistical analysis of the distribution of key tunneling parameters revealed that the tunneling parameters for normal tunneling and when the rig is stuck generally follow a normal distribution, such as... Figure 2 and 3 The image shows a statistical distribution of subway tunneling parameters in a certain city. Figure 2 a, 2b, and 2c are statistical distribution diagrams of thrust F, cutterhead torque T, and tunneling speed v during normal tunneling, respectively. Figure 3 a, 3b, and 3c are statistical distribution diagrams of thrust F, cutterhead torque T, and tunneling speed v when the machine is stuck.

[0030] To ensure the accuracy of the test samples, this application selects data points with relatively concentrated normal distributions for subsequent analysis. Specifically, based on the 2σ principle of normal distribution, data points are removed according to the normal distribution. μ-2σ , μ+2σ Beyond the excavation data, μ The mean, σ The value is the standard deviation. This is because, according to the definition of probability density, when a sample follows a normal distribution, data points that are more than twice the standard deviation from the mean can be considered low-probability events, i.e., P(|x-μ|>2σ) is a low-probability event. Data points outside the range [μ-2σ, μ+2σ] are considered outliers.

[0031] S3. Based on mathematical statistics, effective tunneling parameters describing the tunneling state of the dual-mode shield tunneling machine are selected from the preprocessed tunneling parameters. Specifically, the preprocessed tunneling parameters are categorized into normal tunneling sections and machine-locked tunneling sections, and the tunneling parameters for each section are calculated. x overlap The formula is: in, x These are the tunneling parameters; , These are the tunneling parameters for the normal tunneling section. x Average value and tunneling parameters of the jacking section x The average value; , Normal tunneling parameters x The variance and tunneling parameters of the card machine tunneling section x The variance; when the overlap At that time, tunneling parameters x There are significant differences between the normal tunneling section and the machine-operated tunneling section. These are effective tunneling parameters that can better identify the tunneling status.

[0032] S4, calculate the weights of each effective tunneling parameter based on the entropy weight method. The specific steps are as follows: ① Obtain n Effective tunneling parameters mBased on the samples, the evaluation matrix is ​​constructed as follows: The effective tunneling parameters are normalized according to positive and negative indices using the following formula: Positive indicators: Negative indicators: ② Calculate the entropy value of each effective tunneling parameter according to the following formula: when season ; ③ Calculate the difference coefficient based on the entropy value: ④ Calculate the weights of each effective tunneling parameter: The objective weighting coefficients corresponding to each effective tunneling parameter can be calculated using the above methods.

[0033] S5 calculates a quantitative comprehensive index based on the weights of each effective tunneling parameter and the cumulative probability density function.

[0034] The formula for calculating the quantitative comprehensive index is as follows: in, For quantitative comprehensive indicators; Let x be the cumulative probability density function of the effective tunneling parameter x in the machine section; This represents the weight of the i-th effective tunneling parameter x. Effective tunneling parameters x include thrust F, penetration depth p, thrust penetration index FPI, torque penetration index TPI, etc.

[0035] S6, Calculate the threshold for quantitative comprehensive index determination of the card machine. Determine the threshold For quantitative comprehensive indicators The distinction value between normal tunneling sections and jammed tunneling sections was determined, and the threshold was verified through statistical cases of jammed tunneling in multiple projects.

[0036] The threshold for judgment is calculated as follows: ; In the formula, For quantitative comprehensive indicators The degree of overlap; , These are quantitative comprehensive indicators. The average values ​​for the normal tunneling section and the jammed tunneling section; , For quantitative comprehensive indicators The variance values ​​of the normal tunneling section and the jacking tunneling section.

[0037] This invention also addresses the distribution of effective tunneling parameters x, such as thrust F, penetration depth p, thrust penetration index FPI, and torque penetration index TPI, in normal tunneling sections and jammed tunneling sections. Based on the threshold calculation method, an independent threshold for each effective tunneling parameter x is determined, and the calculation formula is expressed as follows: ; In the formula, η x For effective tunneling parameters, the overlap ratio is used. , These represent the effective tunneling parameters multiplied by the average values ​​of the normal tunneling section and the jammed tunneling section, respectively. , The effective tunneling parameter x represents the variance of the normal tunneling section and the jacking tunneling section.

[0038] like Figure 5 As shown, taking the tunneling parameter torque T as an example, x is calculated. c So that when the torque T is at the threshold x c When the torque T is on the left, the tunnel boring machine is determined to be in the shield jamming tunneling state; when the torque T is at the threshold value... x c When the threshold is on the right, the tunnel boring machine (TBM) is determined to be in normal tunneling mode. Mathematical statistical verification shows that this threshold can serve as a distinguishing value between normal TBM tunneling and a jammed state. In other words, the method in this invention is effective.

[0039] S7 calculates the quantitative comprehensive index of the tunneling parameters to be evaluated, and identifies the tunneling status of the dual-mode shield tunneling machine based on the judgment threshold.

[0040] Example 2 Excavation parameters from multiple dual-mode shield tunneling projects were collected and divided into normal excavation sections and jammed excavation sections according to the actual excavation status. Data preprocessing was performed on the collected raw data using the binary discriminant function method and the 2σ principle to remove invalid data. The preprocessed data is as follows: Figure 4 As shown. Figure 4 af represents the tunneling thrust F, cutterhead torque T, tunneling speed v, penetration depth p, thrust penetration index FPI, and torque penetration index TPI, respectively.

[0041] Based on mathematical statistics, the overlap degree of the tunneling parameters was calculated. Based on the overlap degree values, the tunneling parameters that effectively describe the tunneling state of the dual-mode tunnel boring machine were selected. The selected tunneling parameters and their overlap degrees are shown in Table 1. Table 1. Overlap of various tunneling parameters η x and threshold x c Calculation results As shown in Table 1, these five parameters... η x The absolute values ​​are all greater than 1, which proves that they all have good card recognition performance.

[0042] The entropy weight method was introduced to calculate the weight of each tunneling parameter for the optimized tunneling section of the gantry crane. The specific calculation results are shown in Table 2. Table 2 Weighting values ​​for each indicator A quantitative comprehensive index is calculated based on the weights of each tunneling parameter and the cumulative probability density function. The final formula for calculating the comprehensive index is as follows: ; In the formula, As a comprehensive indicator; Let be the cumulative probability density function of the card segment; The weights of each indicator are used. From this, a comprehensive indicator that can be quantitatively assessed for the shield card machine can be calculated. .

[0043] Quantitative comprehensive indicators were obtained through mathematical statistics methods. The threshold for the stuck machine was determined, and thus an identification mechanism capable of recognizing the stuck tunneling state of the dual-mode shield tunneling machine was obtained. The threshold calculation results are shown in Table 3. Table 3. Overlap of Comprehensive Indicators and Judgment Thresholds in, η x Since 1.31 > 1, it is reasonable and effective to use it to distinguish between normal tunneling and jammed conditions.

[0044] Quantitative comprehensive indicators Threshold formula for card detection In the formula, For quantitative comprehensive indicators The degree of overlap; , These are quantitative comprehensive indicators. The average values ​​for the normal tunneling section and the jammed tunneling section; , For quantitative comprehensive indicators The variance values ​​of the normal tunneling section and the jacking tunneling section.

[0045] Through multiple engineering card machine case studies, the quantitative comprehensive identification index and its judgment threshold of the present invention were verified, and the verification results are as follows: Figure 6 As shown. From Figure 6 a and Figure 6 As can be seen from b, when quantitative comprehensive indicators When the value exceeds the threshold, the tunneling process enters an abnormal state.

[0046] Compared with existing TBM jamming risk assessment methods, the method proposed in this invention can reasonably assess the jamming risk of dual-mode shield tunnels. On the one hand, there is limited research on jamming risk assessment for dual-mode shield tunnels; on the other hand, the state values ​​of factors in actual engineering projects change in real time. The dynamic adaptive weight calculation model established in this invention can better adapt to the changes in the state values ​​of risk factors, calculating more reasonable weights. This invention can effectively reflect the jamming risk level of dual-mode shield tunnels to a certain extent.

[0047] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data-driven dual-mode shield tunneling machine identification method, characterized in that, Includes the following steps: S1 collects the original tunneling parameters of the normal tunneling section and the machine-operated tunneling section of multiple dual-mode shield tunneling projects; S2, based on the binary discriminant function method and the 2σ principle method, preprocesses the original tunneling parameters to remove invalid data; S3, based on mathematical statistics, select effective tunneling parameters from the preprocessed tunneling parameters to effectively describe the tunneling state of the dual-mode shield machine; S4, calculate the weights of each effective tunneling parameter based on the entropy weight method; S5, Calculate a quantitative comprehensive index based on the weights of each effective tunneling parameter and the cumulative probability density function; S6, calculate the judgment threshold of the quantitative comprehensive index of the card machine, and verify the judgment threshold through mathematical statistics and multiple engineering card machine cases; S7 calculates the quantitative comprehensive index of the tunneling parameters to be evaluated, and identifies the tunneling status of the dual-mode shield tunneling machine based on the judgment threshold.

2. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The tunneling parameters include thrust F, penetration depth p, cutterhead torque T, and cutterhead rotation speed n; the binary discriminant function formula is: , , That is, when S When the value is 0, it is considered invalid data and is removed.

3. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The 2σ principle, based on a normal distribution, eliminates […]. μ-2σ , μ+2σ Beyond the excavation data, μ The mean, σ The standard deviation is denoted as .

4. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The specific steps of step S3 are as follows: The preprocessed tunneling parameters are categorized into normal tunneling sections and jack-up tunneling sections; and the tunneling parameters for each section are calculated. x overlap The formula is: ,in, x These are the tunneling parameters; , These are the tunneling parameters for the normal tunneling section. x Average value and tunneling parameters of the jacking section x The average value; , Normal tunneling parameters x The variance and tunneling parameters of the card machine tunneling section x The variance; when the overlap At that time, tunneling parameters x For effective tunneling parameters.

5. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: Step S4 constructs an evaluation matrix of the measured values ​​of each effective tunneling parameter based on the entropy weight method, calculates the entropy value and difference coefficient of each effective tunneling parameter, and then obtains the weight of each effective tunneling parameter.

6. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The formula for calculating the quantitative comprehensive index is as follows: ,in, For quantitative comprehensive indicators; Let x be the cumulative probability density function of the effective tunneling parameter x in the machine section; Let x be the weight of the i-th effective tunneling parameter.

7. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The determination threshold In the formula, For quantitative comprehensive indicators The degree of overlap; , These are quantitative comprehensive indicators. The average values ​​for the normal tunneling section and the jammed tunneling section; , For quantitative comprehensive indicators The variance values ​​of the normal tunneling section and the jacking tunneling section.

8. The data-driven dual-mode shield tunneling machine identification method according to claim 1, characterized in that: The determination threshold For quantitative comprehensive indicators The distinguishing value between the normal tunneling section and the jammed tunneling section.

Images