Method and system for advanced prediction of toxic gases in shield tunnel based on multi-source information

Abstract

The disclosure provides a shield tunnel toxic gas advanced prediction method and system based on multi-source information, relating to the technical field of shield tunnel toxic gas prediction, comprising: acquiring multi-source geological information of the shield tunnel, and respectively preprocessing various types of information; analyzing the correlation between geological characteristics and toxic and harmful gas types and contents, and constructing a geological-gas matching identification library; using a multi-source information fusion mechanism based on hierarchical fusion and weight distribution, the multi-source geological information is fused into a stratum-gas feature vector; the stratum-gas feature vector is matched with the geological-gas matching identification library, and the advanced gas prediction result is output; based on the advanced gas prediction result, the minimum thickness of the stratum in the unexcavated area that needs to be retained is determined by using a toxic gas overflow minimum safety thickness calculation model, and toxic gas protection measures are taken. The disclosure effectively controls the gas content prediction error and risk level judgment error, and can significantly improve the prediction accuracy.

Description

Technical Field

[0001] This disclosure relates to the field of toxic gas prediction technology for shield tunnels, specifically to a method and system for advanced prediction of toxic gases in shield tunnels based on multi-source information. Background Technology

[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.

[0003] During shield tunnel construction, toxic and harmful gases such as hydrogen sulfide, carbon monoxide, and methane hidden in the strata are key hidden dangers threatening construction safety. On the one hand, if toxic gases leak and their concentration exceeds the safety threshold, they can easily lead to poisoning of construction workers, causing serious consequences such as difficulty breathing, coma, or even death. On the other hand, when flammable gases such as methane reach their explosive limits in the air, they can cause deflagration or explosion upon encountering a source of ignition, which can not only damage the shield equipment and tunnel structure but also cause casualties.

[0004] Currently, the methods for predicting and monitoring toxic and harmful gases in shield tunnels have significant limitations, mainly in terms of methodology and technical means: (1) Prediction based on data from a single geological survey report: Prediction before construction mainly relies on the previous geological borehole survey report. This method obtains information on the lithology, structure, and gas content of discrete points by intermittently drilling cores and logging along the planned route, and infers the area between boreholes based on the experience of geologists. Its technical drawbacks are: First, the number of boreholes is limited, and the data is discrete point data, resulting in low prediction accuracy for the spatial distribution of complex geological structures and gas-rich areas, with significant blind spots and uncertainties. Second, this is a static prediction, which cannot be dynamically updated and corrected based on new geological information revealed in real time during shield tunneling, resulting in delayed prediction results and insufficient reliability.

[0005] (2) Empirical analogy method based on geological information revealed at the excavation face: During the tunneling process, geological engineers will conduct real-time identification and recording of the rock debris excavated by the tunnel boring machine cutterhead. By analyzing the characteristics of the current excavation face, such as lithology, joints, and water content, they will empirically compare and infer the potential risks of the unexcavated strata ahead. The technical drawback is that this method is highly dependent on the engineer's personal experience and subjective judgment, lacks quantitative criteria and standards, has poor repeatability, and is prone to misjudgment.

[0006] (3) Lag feedback method for real-time sensor monitoring inside the tunnel: This is currently the most commonly used technical means, which involves deploying various gas concentration sensors on the shield machine body and in the area of ​​assembled tunnel segments to monitor the ambient air inside the tunnel in real time. Its technical drawback is that this method is essentially "post-event monitoring" rather than "advance prediction". The sensors can only trigger an alarm after toxic and harmful gases have already flowed from the ground into the tunnel interior space. At this time, the gas has already begun to diffuse, endangering the health and safety of construction personnel. Summary of the Invention

[0007] To address the aforementioned issues, this disclosure proposes a method and system for predicting toxic gases in shield tunnels based on multi-source information. By collecting multi-source geological information throughout the entire shield tunnel construction cycle, a geological-gas matching identification database is constructed. A multi-source information fusion mechanism is used to predict the types, contents, and leakage risks of toxic and harmful gases in unexcavated areas. Furthermore, a calculation model for the minimum safe thickness for toxic gas spillage is established to determine the minimum thickness of the strata that must be retained in unexcavated areas, providing a basis for the formulation of construction safety protection measures.

[0008] According to some embodiments, the present disclosure adopts the following technical solutions: A method for predicting toxic gases in shield tunnels based on multi-source information includes: Acquire multi-source geological information of the shield tunnel and preprocess each type of information separately; Analyze the correlation between geological features and the types and contents of toxic and harmful gases, and construct a geological-gas matching identification library; A multi-source information fusion mechanism based on hierarchical fusion and weight allocation is adopted to fuse the preprocessed multi-source geological information into a formation-gas feature vector; The formation-gas feature vector is matched with the geological-gas matching and identification library to determine the gas type, and the advanced gas prediction results are output. Based on the results of advanced gas prediction, the minimum thickness of the strata to be retained in the unexcavated area is determined by using the calculation model of the minimum safe thickness for toxic gas spillage, and protective measures against toxic gases are taken.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: A shield tunnel toxic gas prediction system based on multi-source information includes: The information acquisition module is used to acquire multi-source geological information of the shield tunnel and preprocess various types of information respectively; The database construction module is used to analyze the correlation between geological features and the types and contents of toxic and harmful gases, and to build a geological-gas matching and identification database. The information fusion module is used to fuse preprocessed multi-source geological information into a formation-gas feature vector by adopting a multi-source information fusion mechanism based on hierarchical fusion and weight allocation. The gas prediction module is used to match the formation-gas feature vector with the geological-gas matching and identification library to obtain the gas type and output the advanced gas prediction results. The risk quantification module is used to determine the minimum thickness of the strata to be retained in the unexcavated area based on the advanced gas prediction results and the minimum safe thickness calculation model for toxic gas spillage, and to take protective measures against toxic gases.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program that, when executed by a processor, implements the aforementioned method for predicting toxic gases in shield tunnels based on multi-source information.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is provided for storing computer instructions, which, when executed by a processor, implement the aforementioned method for predicting toxic gases in shield tunnels based on multi-source information.

[0012] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the aforementioned method for predicting toxic gases in shield tunnels based on multi-source information.

[0013] Compared with the prior art, the beneficial effects of this disclosure are as follows: This disclosed method for predicting toxic gases in shield tunnels based on multi-source information overcomes the limitations of single-source information by integrating three types of geological information: preliminary exploration, advance prediction, and excavation face identification. It enables advanced prediction of gases in the strata ahead of the tunnel, reducing the time for handling issues by at least 1-2 tunneling cycles compared to traditional real-time monitoring inside the tunnel. Furthermore, by establishing a geological-gas matching identification library and a dynamic update mechanism, combined with safe thickness calculation, it effectively controls the errors in gas content prediction and risk level judgment, significantly improving prediction accuracy.

[0014] This disclosed method for predicting toxic gases in shield tunnels based on multi-source information achieves data-level fusion, feature-level fusion, and decision-level fusion through a multi-layered information fusion mechanism. It not only predicts the type and content of gases but also quantifies the risk level of gas overflow by calculating the minimum safe thickness for toxic gas overflow and outputs targeted prevention and control measures. This avoids the low construction efficiency caused by the traditional "one-size-fits-all" prevention and control approach that suspends tunneling regardless of the risk level, and achieves a balance between safety and efficiency.

[0015] The method for predicting toxic gases in shield tunnels based on multi-source information disclosed herein uses multi-source geological information that is either existing or easily accessible during shield tunnel construction. The geological-gas matching identification library can be trained and dynamically updated through historical data, and is applicable to different lithological strata such as sandstone, mudstone, and limestone. It has good prospects for promotion and application in various types of shield tunnels. Attached Figure Description

[0016] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.

[0017] Figure 1 This is an overall flowchart of the shield tunnel toxic gas prediction method based on multi-source information according to an embodiment of the present disclosure; Figure 2 This is a block diagram for advanced prediction of toxic gases in shield tunnels based on multi-source information, according to an embodiment of this disclosure. Detailed Implementation

[0018] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.

[0019] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this disclosure. 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 disclosure pertains.

[0020] 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 disclosure. 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.

[0021] Example 1 One embodiment of this disclosure provides a method for predicting toxic gases in shield tunnels based on multi-source information. The method includes the following steps: Step 1: Obtain multi-source geological information of the shield tunnel and preprocess each type of information separately; Step 2: Analyze the correlation between geological features and the types and contents of toxic and harmful gases, and construct a geological-gas matching identification library; Step 3: Employ a multi-source information fusion mechanism based on hierarchical fusion and weight allocation to fuse the preprocessed multi-source geological information into a stratigraphic-gas feature vector; Step 4: Match the formation-gas feature vector with the geological-gas matching and identification library to determine the gas type, and output the advanced gas prediction results; Step 5: Based on the advanced gas prediction results, use the minimum safe thickness calculation model for toxic gas spillage to determine the minimum thickness of the strata to be retained in the unexcavated area, and take protective measures against toxic gases.

[0022] As one embodiment, the method for predicting toxic gases in shield tunnels based on multi-source information disclosed herein collects multi-source geological information throughout the entire shield tunnel construction cycle, constructs a geological-gas matching identification database, and utilizes a multi-source information fusion mechanism to predict the types, contents, and leakage risks of toxic and harmful gases in unexcavated areas. Furthermore, it establishes a calculation model for the minimum safe thickness for toxic gas spillage, providing a basis for formulating construction safety protection measures. The specific implementation process is as follows: Step 1: Obtain multi-source geological information of the shield tunnel and preprocess each type of information separately; Step 11: Obtain multi-source geological information of the shield tunnel; The multi-source geological information covers the entire process of shield tunnel construction, from pre-construction geological surveys and advanced geological forecasting during construction to real-time identification of the excavation face. This includes pre-construction geological survey information, advanced geological forecasting information during shield tunnel construction, excavation face stratum identification information, and tunnel interior gas environment monitoring data. Specific content includes: (1) Preliminary geological exploration information: obtained through geological exploration boreholes, including data such as the distribution range and thickness of strata lithology, strata porosity data, strata permeability data, and gas sampling and analysis results in the exploration boreholes.

[0023] (2) Advanced geological forecast information during shield tunneling: Seismic wave method is used to collect information. By deploying seismic wave transmitters and receivers in the excavated section of the shield tunnel, high-frequency seismic waves are transmitted to the unexcavated area and the seismic wave signals reflected by the strata are received. The seismic wave signals are filtered and processed to extract the reflection characteristics of the strata interface and the wave velocity change law, so as to determine the location and scale of gas storage structures such as lithological changes, karst caves, and fracture zones in the unexcavated area (for example, if there is a sudden decrease in the seismic wave velocity in the limestone strata, it usually indicates the presence of karst caves or fracture zones, which may provide space for the storage of toxic and harmful gases).

[0024] (3) Excavation face stratum identification information: The vibration wave feature identification method of the cutterhead is adopted. Vibration sensors are installed in the cutterhead drive system of the shield machine to collect vibration signals in real time during the excavation process of the cutterhead. According to the difference in resistance of different lithological strata to the cutterhead, the vibration signals are analyzed in time and frequency using wavelet transform and other methods to extract characteristic parameters such as vibration frequency, amplitude, and energy spectrum, so as to realize the real-time identification of the lithology of the excavation face and record the correspondence between the lithology of the excavation face and the gas concentration of the excavated area.

[0025] (4) Gas environment monitoring data inside the tunnel: By deploying multiple gas sensor arrays along the excavation direction inside the shield tunnel, real-time gas concentration data of the excavated area is collected, and the trend of gas concentration change with time and excavation mileage is recorded; at the same time, combined with ventilation parameters such as tunnel ventilation volume and wind speed, the gas concentration monitoring results are corrected to eliminate the interference of ventilation on the concentration. The correction formula is as follows:

[0026] in, This is the corrected gas concentration. For the actual concentration measured by the sensor, Standard ventilation volume This represents the actual ventilation volume.

[0027] Step 12: Preprocess each type of information separately; The collected multi-source geological information is preprocessed to eliminate noise interference and data redundancy, ensuring the validity of the information: (1) For the information of the previous geological exploration, if there are missing data of the exploration boreholes, the interpolation method is used to supplement the missing data; the lithological data are standardized and coded (e.g., sandstone is coded as 1, mudstone as 2, and limestone as 3) to facilitate subsequent data fusion.

[0028] (2) For advanced geological forecast information during the shield tunneling construction stage, the wavelet threshold denoising method is used to process the seismic wave signal to remove noise such as construction machinery vibration and electromagnetic interference; the seismic wave signal is converted into spatial distribution data of strata lithology and gas storage structure through the tomographic imaging wave velocity inversion algorithm.

[0029] The specific process was as follows: A commonly used tunnel seismic wave advanced geological prediction system was used to probe the strata at a certain distance in front of the tunnel boring machine cutterhead. The probe results showed that the strata in this area belonged to the medium-hard rock category, and a limestone cave was identified, clarifying the relative position and approximate size of the cave to the cutterhead, as well as the integrity of the surrounding rock mass. Subsequently, through professional data denoising processing technology, the boundary range of the cave and the positional relationship between the cave and the tunnel axis were further clarified, providing accurate structural information for subsequent analysis.

[0030] (3) For the stratum identification information of the excavation face, the vibration signal is smoothed by the sliding window filtering method and stable vibration characteristic parameters are extracted; a mapping relationship table between vibration characteristic parameters and lithology is established to verify the identified lithology. For example, if the error exceeds 5% when compared with the lithology data of the previous geological exploration borehole, the threshold of vibration characteristic parameters is adjusted.

[0031] The specific process is as follows: Vibration signals generated when the tunnel boring machine's cutterhead cuts through the strata are collected in real time. Signal conversion technology is used to transform these vibration signals into analyzable characteristic parameters, including vibration frequency and the torque magnitude during cutterhead operation. By comparing geological data from similar past projects, it is determined that the strata currently in contact with the excavation face are moderately weathered sandstone, consistent with the lithological information obtained through geological exploration boreholes before construction, thus verifying the reliability of the previous geological data.

[0032] (4) For the gas environment monitoring data inside the tunnel, outliers such as sudden data caused by sensor failure are removed (if the deviation of the data from the mean exceeds 3 times the standard deviation, it is considered an outlier and replaced with the mean of the adjacent time); the gas concentration is corrected by the ventilation volume correction formula to ensure that the data reflects the real situation of ground gas release.

[0033] Step 2: Analyze the correlation between geological features and the types and contents of toxic and harmful gases, and construct a geological-gas matching identification library; A geological-gas matching identification database is constructed to establish the correlation between geological characteristics such as stratigraphic lithology, porosity, and gas-reservoir structures and the types and contents of toxic and harmful gases. The process of constructing the geological-gas matching identification database includes the construction of a basic correlation database and a dynamic updating process, specifically including: Step 21: Building the basic relational database; This study collected engineering cases of shield tunnels covering different geological conditions and gas types. Geological parameters such as strata lithology, porosity, permeability, and gas-bearing structure type and scale were extracted from these cases, along with gas detection data such as gas types and content ranges in the corresponding areas. Correlation relationships were then established to create a basic correlation database, including multiple pairs of corresponding relationships as follows: (1) Lithology-gas type correlation: For example, hydrogen sulfide gas is easily generated in mudstone strata (due to the decomposition of sulfur-containing organic matter in mudstone), methane gas is easily stored in limestone caves (due to the reaction of carbonate rocks and organic matter in limestone strata to generate methane), and carbon monoxide gas may exist in sandstone fracture zones (due to the oxidation of coal interlayers in sandstone).

[0034] (2) Porosity-Gas Content Correlation: A fitting formula for porosity and gas content was established through statistical analysis, as follows:

[0035] in, Gas content (unit: ppm). Formation porosity (unit: %) Lithology coefficient, The porosity coefficient (determined based on core tests) indicates that the greater the porosity, the larger the space for gas storage in the rock mass, and the higher the gas content is usually.

[0036] (3) Gas storage structure-leakage risk association: The leakage risk level is determined based on the type and size of the gas storage structure, such as whether it is a karst cave or a fissure zone. For example, a limestone karst cave with a volume greater than 100m³ has a leakage risk level of "high"; a sandstone fissure zone with an extension length of less than 10m has a leakage risk level of "low".

[0037] Step 22: Dynamic update process; The preprocessed multi-source geological information collected during shield tunnel construction is input into the dynamic update module to update the basic associated database in real time, including: (1) When a certain lithology is identified at the excavation face and the corresponding gas is detected by the gas sensor in the tunnel, the correspondence between the lithology and the gas type is added to the basic association database.

[0038] (2) When advanced geological forecasting discovers a new gas storage structure and subsequent excavation verifies the gas release situation of the structure, the correlation between the scale, location and gas content and leakage risk of the gas storage structure is added to the database to optimize the leakage risk level classification standard.

[0039] Step 3: Employ a multi-source information fusion mechanism based on hierarchical fusion and weight allocation to fuse the preprocessed multi-source geological information into a formation-gas feature vector; Specifically, a multi-source information fusion mechanism based on hierarchical fusion and weight allocation is adopted to fuse the preprocessed multi-source geological information into a unified stratigraphic-gas feature vector. This mechanism includes three levels: data layer fusion, feature layer fusion, and decision layer fusion, as follows: Step 31: Data layer fusion; Data layer fusion involves spatial matching and fusion of information from previous geological exploration and advanced geological forecasting. A three-dimensional coordinate system was established using the tunnel axis as a reference. The X-axis represents the tunnel excavation direction, the Y-axis is horizontal and perpendicular to the tunnel axis, and the Z-axis is vertical. The lithology and porosity data from previous geological survey boreholes were converted into continuous three-dimensional stratigraphic data using three-dimensional interpolation. The spatial distribution data of gas-bearing structures obtained from seismic wave prediction was overlaid with the three-dimensional stratigraphic data to obtain a three-dimensional distribution model of "lithology-porosity-gas-bearing structures" in the unexcavated area. For example, at X=100m, Y=5m, Z=3m, the fused data shows that the location is a limestone stratum (lithology code 3), with a porosity of 15% and a karst cave with a volume of 50m³.

[0040] In the data fusion stage, spatial interpolation is used to transform the discrete data obtained from geological exploration boreholes before construction into a stratigraphic parameter map that reflects the continuous distribution along the tunnel, thus achieving continuous data presentation.

[0041] Step 32: Feature layer fusion; Feature layer fusion includes extracting key feature parameters from each information source and performing weight allocation and fusion, as detailed below: (1) Extracting characteristic parameters from each information source: The characteristic parameters of the early geological exploration information are lithology code (F1), porosity (F2), and gas content in the exploration borehole (F3); the characteristic parameters of the advanced geological prediction information are the scale of the gas-bearing structure (F4, such as the volume of the karst cave, unit: m). 3 The features of the gas storage structure and the excavation face are: F5 (unit: m); F6 (characteristic parameters of the strata identification information of the excavation face) and F7 (characteristic matching degree of cutterhead vibration feature, representing the credibility of the current lithology identification, with a value of 0 to 1, the closer to 1 the higher the credibility); F8 (characteristic parameters of the gas monitoring information in the tunnel) and F9 (unit: ppm / min).

[0042] (2) Determine the weights of each feature parameter: The weights are determined using the analytic hierarchy process (AHP). Experts are invited to score the importance of each feature parameter, a judgment matrix is ​​constructed, and the assigned weights w are calculated. i .

[0043] For example, the size of the gas-bearing structure (F4) has the greatest impact on gas prediction, with a weight of w4=0.25; porosity (F2) has a weight of w2=0.2; gas content in exploration wells (F3) has a weight of w3=0.15; gas concentration (F8) has a weight of w8=0.15; lithological coding (F1, F6) each has a weight of w1=w6=0.08; distance from the gas-bearing structure (F5) has a weight of w5=0.05; vibration characteristic matching degree (F7) has a weight of w7=0.03; and gas concentration change rate (F9) has a weight of w9=0.01 (the total weight is 1).

[0044] Using the weights calculated above as the base weights, a stage coefficient is further introduced. Based on the three stages of the tunnel boring machine (TBM) construction process (preliminary exploration period, initial stage of TBM tunneling, and mid-to-late stage of TBM tunneling), the base weights are dynamically adjusted to obtain the corrected weights: P=W×

[0045] Where W is the basic weight, This is the stage coefficient.

[0046] Furthermore, the stage coefficient is set based on information reliability and demand priority, for example, taking a value of 0.8 to 1.2. After all parameters are adjusted, the total weight is still 1. The specific rules are as follows: a. Preliminary exploration phase: Since tunneling has not yet begun and we are relying solely on exploration data, the stage coefficient for the preliminary geological exploration information is set to 1.2. For example, the correction weight for porosity F2 is 0.2 × 1.2 = 0.24. The stage coefficient for advanced forecasting and excavation face identification information is set to 0.8. For example, the correction weight for gas storage structure scale F4 is 0.25 × 0.8 = 0.2.

[0047] b. Initial stage of tunnel boring: Excavation has begun, and the advanced geological forecast data is becoming increasingly reliable. Therefore, the stage coefficient for the advanced geological forecast information can be set to 1.2, then the weight after F4 correction = 0.25 × 1.2 = 0.3, the stage coefficient for the preliminary exploration information is set to 1.0, and the stage coefficient for the excavation face identification information is set to 0.9.

[0048] c. Mid-to-late stages of tunnel boring: Since sufficient data has been accumulated at the excavation face, the stage coefficients for excavation face identification information and tunnel gas monitoring information are set to 1.2. For example, the weight after correction for F7 is 0.03×1.2=0.036, the weight after correction for F8 is 0.15×1.2=0.18, and the stage coefficient for preliminary exploration information is set to 0.9.

[0049] By incorporating stage coefficients, dynamic adjustments were made during the construction phase, reducing subjective bias in expert scoring and improving the rationality of weight allocation.

[0050] (3) Feature layer fusion calculation: Standardize each feature parameter (e.g., standardize the gas storage structure scale F4 to...). ,in (The maximum gas storage structure volume in the database) is then used to calculate the fusion eigenvalues ​​based on weights. ,in, and These represent the weight values ​​of each information source and the standardized characteristic parameter values, respectively. This value characterizes the comprehensive intensity of the "formation-gas" characteristics in the unexcavated area.

[0051] In the feature layer fusion stage, key features are extracted from multi-source data and classified and combined: the first type of features revolves around lithology, integrating information such as lithology type, rock mass porosity, and fracture density; the second type of features focuses on gas-bearing structures, covering the location, size, and integrity of karst caves; the third type of features is associated with the real-time status of the excavation face, including cutterhead vibration frequency and torque data. Subsequently, data dimensionality reduction technology is used to remove redundant information and retain the core features that play a crucial role in gas prediction. These core features can cover most of the key information in the original data, effectively reducing the complexity of subsequent calculations.

[0052] Step 33: Integration of decision-making levels; Decision-level fusion is the result of feature-level fusion. It employs a multi-dimensional weighted similarity matching algorithm to match the results with a geological-gas matching database, enabling advanced gas prediction. Specifically, this includes: (1) Multi-dimensional similarity calculation: In the geological-gas matching identification database, four key dimensions corresponding to the "current formation-gas feature vector" (including lithology F1, porosity F2, gas reservoir structure scale F4, and exploration well gas content F3) are extracted, and the similarity of each dimension is calculated: a. Lithological similarity (S1) If the current lithology is consistent with the lithology of the cases in the database (e.g., both are limestone), S1=1; if the lithology is similar (e.g., limestone and dolomite), S1=0.8; if the difference is large (e.g., limestone and mudstone), S1=0.3. b. Porosity similarity (S2) The calculation formula is S2 = 1 - |current porosity - porosity of cases in the library| / porosity of cases in the library. If the result is negative, take 0. c. Similarity of gas content in exploration wells (S3) Similarly, S3 = 1 - |current borehole content - case content in the database| / case content in the database. If the result is negative, take 0. d. Similarity in scale of gas storage structures (S4) Similarly, S4 = 1 - |current gas storage structure size - size of cases in the storage facility| / size of cases in the storage facility. If the result is negative, take 0. Based on the weights determined in step 32, calculate the overall similarity S: S = (S1 × F1 weight) + (S2 × F2 weight) + (S3 × F3 weight) + (S4 × F4 weight).

[0053] (2) Match the optimal case and output the preliminary prediction results. From the geological-gas matching database, the top 3 cases with the highest "overall similarity S" are selected, and the intersection of "gas type, content range, and risk level" of these 3 cases is taken. a. Gas type If all three cases point to "methane", the initial prediction is methane; if two points to methane and one points to ethane, then methane with the highest percentage is selected; if all three cases point to different types of gas, then the number of cases is increased from 3 to 5 until the gas type with the highest percentage is selected.

[0054] b. Gas content Take the average of the concentration ranges of the three cases (e.g., Case 1 is 800-1000 ppm, Case 2 is 750-950 ppm, Case 3 is 850-1050 ppm, and the preliminary predicted concentration is 850-980 ppm). c. Risk Level The risk levels are categorized as "high," "medium," and "low." The highest risk level among the cases is output. For example, if "high risk" appears in 3 cases, the initial risk level is "high risk"; if the risk levels of 3 cases consist of "medium risk" and "low risk," the initial risk level is "medium risk."

[0055] (3) Matching confidence verification a. If the overall similarity S ≥ 0.8, it is judged as high confidence, and the above preliminary prediction results are directly output; b. If 0.5 ≤ S < 0.8, it is judged as medium confidence. The backup algorithm in the identification library is called. The mature random forest algorithm can be used. The current feature vector is input into the model. The model is trained based on the pattern of historical data and corrects the preliminary prediction result. For example, if the preliminary prediction content is 850-980ppm, the model is corrected to 820-950ppm. c. If S < 0.5, it is considered a low confidence level. In this case, the current feature vector and similar cases in the database need to be sent to the geological engineer, who will then adjust the prediction results based on their field experience.

[0056] (4) Output the final advanced prediction results Based on the confidence level verification results, the final output is the advanced prediction result, expressed as "gas type + concentration range (with error interval, such as 820-950ppm, error ±5%) + risk level + matching confidence level (such as 0.85)".

[0057] In the geological-gas matching identification database, the corresponding gas types are screened based on the lithology of the unexcavated area obtained from the data layer fusion model. Specifically, firstly, the lithology identification results for the unexcavated area are extracted from the output of the data layer fusion model; secondly, the above lithology types are used as "keywords" or "indexes" to query the constructed "geological-gas matching identification database," and the system will retrieve a list of all potential gas types associated with the input lithology.

[0058] This disclosure inputs the core features obtained from feature layer fusion into a pre-constructed geological-gas matching identification library, and calls the correlation prediction model for the target gas within the identification library. This model is built based on a large amount of historical geological data and corresponding gas monitoring data, and can estimate the approximate content of the target gas based on the input geological feature parameters. Simultaneously, gas sensors deployed inside the tunnel monitor the actual content of the target gas near the current working face in real time, and compare the actual monitoring results with the model prediction results. If the deviation between the two is within a preset reasonable range, it indicates that the model prediction result is reliable, thereby determining the predicted content of the target gas at the target location ahead of the tunnel.

[0059] Step 4: Based on the advanced gas prediction results, use the minimum safe thickness calculation model for toxic gas spillage to determine the minimum thickness of the strata to be retained in the unexcavated area, and take protective measures against toxic gases.

[0060] To further quantify the risk of gas leakage, a calculation model for the minimum safe thickness of toxic gas spillage is established to determine the minimum thickness of the strata that must be retained in the unexcavated area. This means that when the tunnel boring machine reaches this thickness, excavation must be stopped and protective measures must be taken. The formula is as follows:

[0061] in, Minimum safe thickness for toxic gas spillage (unit: m); Predicted gas content in unexcavated areas (unit: ppm); Gas storage structure volume (unit: m) 3 ); Formation pore pressure (unit: MPa); Formation permeability coefficient (unit: m / s); Tunnel ventilation volume (unit: m / s); Safety response time (unit: seconds) is the time from the detection of a gas leak to the completion of protective measures.

[0062] This disclosure, based on actual tunnel engineering parameters, the mechanical properties of the strata and rock mass, and the gas storage pressure within the strata, uses professional mechanical calculation formulas to calculate the minimum stratum thickness required to prevent gas from breaching the stratum barrier and overflowing into the tunnel. Combining the location information of the karst cave obtained from advanced geological forecasting and the actual dip angle of the strata at that location, the remaining thickness of the strata above and around the karst cave is calculated. The calculated actual remaining stratum thickness is compared with the minimum safe thickness, and combined with previously predicted gas content: if the actual remaining stratum thickness is greater than the minimum safe thickness, and the predicted gas content is far below the safety threshold, then it is determined that there is no risk of gas overflow in the area; otherwise, the risk level is determined based on the specific degree of deviation.

[0063] The final output yields advanced prediction results (gas type, content, risk level, safe thickness, and warning time). Based on the gas prediction results and risk assessment level, targeted safety control measures are formulated: for risk-free or low-risk areas, the current ventilation system operation intensity is maintained to ensure air circulation within the tunnel, while real-time monitoring of gas concentration changes and geological data such as cutterhead vibration and torque continues to be maintained, ensuring dynamic monitoring of the construction environment. After the tunnel boring machine completes a certain distance of excavation, the newly collected multi-source geological information and corresponding real-time gas monitoring data within that section are fed back into the geological-gas matching identification library. By retraining the associated prediction model in the identification library and adjusting the model parameters, the model can adapt to new geological conditions, continuously optimizing prediction accuracy and providing more reliable support for advanced gas prediction in subsequent excavation sections, ensuring the timeliness and accuracy of the prediction method throughout the entire construction process.

[0064] Example 2 One embodiment of this disclosure provides a shield tunnel toxic gas prediction system based on multi-source information, including: The information acquisition module is used to acquire multi-source geological information of the shield tunnel and preprocess various types of information respectively; The database construction module is used to analyze the correlation between geological features and the types and contents of toxic and harmful gases, and to build a geological-gas matching and identification database. The information fusion module is used to fuse preprocessed multi-source geological information into a formation-gas feature vector by adopting a multi-source information fusion mechanism based on hierarchical fusion and weight allocation. The gas prediction module is used to match the formation-gas feature vector with the geological-gas matching and identification library to obtain the gas type and output the advanced gas prediction results. The risk quantification module is used to determine the minimum thickness of the strata to be retained in unexcavated areas based on the results of advanced gas prediction and using a minimum safe thickness calculation model for toxic gas spills, and to implement protective measures against toxic gases. Example 3 One embodiment of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method for predicting toxic gases in shield tunnels based on multi-source information.

[0065] Example 4 One embodiment of this disclosure provides a non-transitory computer-readable storage medium for storing computer instructions. When these computer instructions are executed by a processor, they implement the aforementioned method for predicting toxic gases in shield tunnels based on multi-source information.

[0066] Example 5 One embodiment of this disclosure provides an electronic device, including a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to implement the method for predicting toxic gases in shield tunnels based on multi-source information.

[0067] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0068] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0069] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. 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 this disclosure are still within the scope of protection of this disclosure.

Claims

1. A method for advanced prediction of toxic gases in shield tunneling based on multi-source information, characterized in that, include: Acquire multi-source geological information of the shield tunnel and preprocess each type of information separately; Analyze the correlation between geological features and the types and contents of toxic and harmful gases, and construct a geological-gas matching identification library; A multi-source information fusion mechanism based on hierarchical fusion and weight allocation is adopted to fuse the preprocessed multi-source geological information into a formation-gas feature vector; The formation-gas feature vector is matched with the geological-gas matching and identification library to determine the gas type, and the advanced gas prediction results are output. Based on the results of advanced gas prediction, the minimum thickness of the strata to be retained in the unexcavated area is determined by using the calculation model of the minimum safe thickness for toxic gas spillage, and protective measures against toxic gases are taken.

2. The method for predicting toxic gases in shield tunnels based on multi-source information as described in claim 1, characterized in that, The acquisition of multi-source geological information of the shield tunnel and the preprocessing of various types of information include: The multi-source geological information covers the entire process of shield tunnel construction, including pre-construction geological survey, advanced geological forecasting during construction, and real-time identification of the excavation face. It includes pre-construction geological survey information, advanced geological forecasting information during shield construction, stratum identification information at the excavation face, and gas environment monitoring data inside the tunnel. Preprocessing operations are performed on the collected multi-source geological information to eliminate noise interference and data redundancy.

3. The method for predicting toxic gases in shield tunnels based on multi-source information as described in claim 1, characterized in that, Constructing a geological-gas matching identification database, which includes building a basic relational database and a dynamic update process, specifically including: Collect engineering cases of shield tunnels covering different geological conditions and different gas types, extract geological parameters such as stratum lithology, porosity, permeability, gas storage structure type and scale from the cases, and multiple gas detection data of gas type and content range in the corresponding areas, and establish correlations between them to build a basic correlation database; The preprocessed multi-source geological information is input into the dynamic update module to update the basic correlation database in real time, resulting in a geological-gas matching identification library.

4. The method for predicting toxic gases in shield tunnels based on multi-source information as described in claim 1, characterized in that, The multi-source information fusion mechanism based on hierarchical fusion and weight allocation is adopted to fuse preprocessed multi-source geological information into a stratigraphic-gas feature vector, including three levels: data layer fusion, feature layer fusion, and decision layer fusion. Specifically: Data layer fusion is the spatial matching and fusion of information from previous geological exploration and advanced geological forecasting. The feature layer fusion process involves extracting key feature parameters from various geological information sources, assigning and fusing them with weights, and introducing stage coefficients to dynamically adjust the basic weights. Specifically, the analytic hierarchy process is used to determine the weights, score the importance of each feature parameter, construct a judgment matrix and calculate the weights, and then calculate the fused feature values ​​according to the weights. The decision-level fusion process involves fusing the feature layers to obtain fused feature values, then using a multi-dimensional weighted similarity matching algorithm to match them with a geological-gas matching identification database to achieve advanced prediction and output advanced gas prediction results.

5. The method for predicting toxic gases in shield tunnels based on multi-source information as described in claim 1, characterized in that, The determination of the minimum thickness of the strata to be retained in the unexcavated area based on the advanced gas prediction results and using the minimum safe thickness calculation model for toxic gas spillage includes: Quantify the risk of gas leakage, establish a calculation model for the minimum safe thickness of toxic gas spill, determine the minimum thickness of the strata to be retained in the unexcavated area, as well as the leakage risk level and the time of risk occurrence, and obtain all the final advanced prediction results; When the tunnel boring machine reaches the minimum thickness, it must stop tunneling and take protective measures.

6. The method for predicting toxic gases in shield tunnels based on multi-source information as described in claim 5, characterized in that, The final advanced prediction results include gas type, content, risk level, safety thickness, and warning time.

7. A shield tunnel toxic gas prediction system based on multi-source information, characterized in that, include: The information acquisition module is used to acquire multi-source geological information of the shield tunnel and preprocess various types of information respectively; The database construction module is used to analyze the correlation between geological features and the types and contents of toxic and harmful gases, and to build a geological-gas matching and identification database. The information fusion module is used to fuse preprocessed multi-source geological information into a formation-gas feature vector by adopting a multi-source information fusion mechanism based on hierarchical fusion and weight allocation. The gas prediction module is used to match the formation-gas feature vector with the geological-gas matching and identification library to obtain the gas type and output the advanced gas prediction results. The risk quantification module is used to determine the minimum thickness of the strata to be retained in the unexcavated area based on the advanced gas prediction results and the minimum safe thickness calculation model for toxic gas spillage, and to take protective measures against toxic gases.

8. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for predicting toxic gases in shield tunnels based on multi-source information as described in any one of claims 1-6.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the shield tunnel toxic gas advance prediction method based on multi-source information as described in any one of claims 1-6.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to perform the method for predicting toxic gases in shield tunnels based on multi-source information as described in any one of claims 1-6.

Images