Tunnel rock and gas outburst danger grade evaluation method based on multi-source information fusion

Abstract

The invention discloses a tunnel rock and gas outburst danger level evaluation method based on multi-source information fusion, which comprises the following steps of: 1) detecting the geological condition in front of a tunnel face by the following means, and converting a detection result into corresponding danger level data; 2) selecting evaluation indexes for evaluating the danger level of tunnel rock and gas outburst through attribute reduction and importance ranking; 3) calculating the rock and gas outburst danger level of the tunnel under various single evaluation indexes through the following formula; 4) calculating a comprehensive posteriori evaluation grade of a certain danger grade under the action of all evaluation indexes; and finally, determining the danger level with the maximum occurrence probability under the combined action of all the currently selected evaluation indexes as the final danger level of the tunnel rock and gas outburst. According to the evaluation method, various objective data are integrated, and accurate evaluation of the rock and gas outburst danger level is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of geological exploration and rock and gas outburst hazard prevention, and in particular to a tunnel rock and gas outburst hazard level evaluation method based on multi-source information fusion. Background Art

[0002] At present, gas tunnels mainly use geophysical exploration technologies such as TSP tunnel seismic detection, geological radar detection and transient electromagnetic method for advanced detection. For coal-bearing strata, advanced drilling technology can accurately grasp the occurrence of coal seams, coal seam gas parameters and other characteristics, and is the main advanced detection method. However, there are very few geophysical exploration methods used for advanced geological prediction in railway gas tunnels.

[0003] Comprehensive advance geological prediction, based on geological surveys and primarily using advance drilling, combined with geophysical and basement exploration, can provide long-term, medium-term, and short-term forecasts of the tunnel surrounding rock ahead of the tunnel face. However, the geophysical methods used are limited, and can only detect the lithology and rock structure ahead, but cannot accurately detect abnormal geological bodies such as water-rich areas and goafs. Advance drilling, which detects the geological conditions of the surrounding rock ahead, can predict gas pressure, flow, concentration, and attenuation coefficient ahead of the unexcavated tunnel face with high accuracy, making it suitable for field use in most tunnels. However, it also suffers from the disadvantage of a single detection method. Transient electromagnetic method has a relatively good effect in detecting goaf areas at the bottom of tunnels, proving that transient electromagnetic method is an effective method; using TSP prediction in coal seam gas sections has a good effect on predicting the mechanical properties of the rock ahead; combining geological logging with advance drilling is also a coal seam detection method, and combined with engineering examples, it can ensure the effectiveness of this method, combining macro and micro perspectives, but the means are also single, and geological logging has huge uncertainty and inaccuracy; TSP advance detection is supplemented by geological radar for refined detection, which is necessary The method of intensive advance drilling at the same time realizes comprehensive prediction and is a relatively comprehensive detection with greater accuracy and breadth, but there are still unexplored areas; the combination of advance drilling, TSP tunnel seismic detection and infrared detection method can predict the geological conditions and harmful gases ahead, and has made breakthroughs in the detection range, but is insufficient in accuracy; the combination of geological radar advance preliminary exploration and advance geological drilling precision exploration can make geological advance predictions for the tunnel's gas-bearing coal-bearing strata and overlapping karst fracture sections, and obtain the karst and coal seam gas characteristics ahead of the heading face.

[0004] The above methods only combine one or two geophysical exploration technologies, which cannot fully detect the geological conditions of gas tunnels and are not accurate enough to accurately evaluate the rock and gas outburst risk level of the complex geological conditions of the tunnels. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a tunnel rock and gas outburst hazard level evaluation method based on multi-source information fusion to solve the technical problem of accurately evaluating the risk level of rock and gas outburst disasters in tunnels.

[0006] The tunnel rock and gas outburst hazard level evaluation method based on multi-source information fusion of the present invention comprises the following steps:

[0007] 1) Explore the geological conditions ahead of the tunnel face using the following methods and determine the geological hazard level based on the exploration results:

[0008] ① Study the tunnel geological survey data to determine the hazard level of the hydrogeological conditions ahead of the tunnel face, the hazard level of the geological structure, the hazard level of the stratum lithology, the hazard level of unfavorable geology and special rock and soil, the hazard level of gas, and the hazard level of the surrounding rock;

[0009] ② Use advance drilling method to conduct detection and obtain the corresponding hazard level based on the detection results;

[0010] ③ Use transient electromagnetic method to conduct detection and obtain the corresponding hazard level based on the detection results;

[0011] ④ Use TSP detection method to detect and obtain the corresponding danger level according to the detection results;

[0012] ⑤ Use geological radar to conduct detection and obtain the corresponding danger level based on the detection results;

[0013] 2) Using Rosetta software to perform attribute simplification on the 10 hazard level data obtained in step 1), identify indicators with a support number equal to 100, and then rank the attribute importance based on the total number of times each indicator appears in the final attribute simplification result. The indicators ranked in the top n in attribute importance are selected as the evaluation indicators for evaluating the tunnel rock and gas outburst hazard level;

[0014] 3) The rock and gas outburst hazard level P(y) of the tunnel under a certain evaluation index i is calculated by the following formula: ij |x i ):

[0015]

[0016] Where x i Indicates the calculated value of the evaluation index, the subscript i represents each evaluation index; y ij represents the hazard level evaluation standard value set for each evaluation index, subscript j represents the rock and gas outburst hazard level, j = 1, 2, 3, 4, hazard level 1 represents no hazard, hazard level 2 represents weak hazard, hazard level 3 represents medium hazard, hazard level 4 represents strong hazard; P(y ij) represents the prior probability of rock and gas outburst hazard level under the evaluation index i; P(x i |y ij ) indicates that the rock and gas outburst hazard level threshold is y ij When the calculated value of the evaluation index is x i The probability of P(y ij |x i ) indicates that the calculated value of the evaluation index is x i When the rock and gas outburst hazard level threshold reaches y ij probability;

[0017] The calculation process is as follows:

[0018] (1) In the initial calculation stage, the prior probability P(y ij ) are the same, that is:

[0019]

[0020] (2) Then calculate P(x i |y ij ):

[0021]

[0022] Where, L ij =|x i -y ij |, L ij The smaller the value, the greater the probability that the measured index belongs to the corresponding rock and gas outburst level;

[0023] (3) Finally calculate P(y ij |x i );

[0024] 4) Calculate the comprehensive posterior evaluation level P of a certain hazard level j under the action of all evaluation indicators j :

[0025]

[0026] Where, ω i represents the weight of evaluation index i;

[0027] Then determine the P with the largest probability value by the following formula j The final hazard level for tunnel rock and gas outburst is Z:

[0028]

[0029] Furthermore, the weight ω in formula (4) i By combining weights, we get:

[0030] ω i =(1-λ)ω i1 +λω i2 (6)

[0031] Among them, ω i1 is the objective weight, ω i2 is the expert weight;

[0032] ω i1 The acquisition process is as follows:

[0033] ① For the original matrix consisting of m evaluation samples and n evaluation indicators:

[0034] Q=(q li ) m×n

[0035] Where q li is the matrix element corresponding to the i-th evaluation index of the l-th evaluation sample, l=1,2,…,m; i=1,2,…,n;

[0036] ②Matrix normalization, regardless of the positive or negative direction of the indicator, is expressed as:

[0037]

[0038] Where q max is the maximum value under the same indicator;

[0039] ③Calculate information entropy:

[0040]

[0041] ④Calculate the entropy weight of evaluation index i

[0042]

[0043] ω i2 The calculation formula is as follows:

[0044]

[0045] Where W ik is the subjective weight ratio of the kth expert to the i-th evaluation index; P k is the authority weight ratio of the kth expert in the i evaluation index; z is the number of experts.

[0046] Beneficial effects of the present invention:

[0047] The present invention is based on a tunnel rock and gas outburst hazard level prediction method based on multi-source information fusion, and combines geological survey, geophysical prospecting and advance drilling methods to achieve clarity and transparency of tunnel advance geology from the overall to the local. Based on the geological hazard level data obtained by multiple methods, the hazard level data is attribute-simplified and attribute importance-ranked to select several evaluation indicators that have the greatest impact on the risk of tunnel rock and gas outburst. The probability judgment of the rock and gas outburst hazard level is corrected by fusing prior knowledge and new evidence, thereby achieving an accurate evaluation of the rock and gas outburst hazard level. DETAILED DESCRIPTION

[0048] The present invention will be further described below with reference to the embodiments.

[0049] The tunnel rock and gas outburst hazard level assessment method based on multi-source information fusion in this embodiment includes the following steps:

[0050] 1) Explore the geological conditions ahead of the tunnel face using the following methods and determine the geological hazard level based on the exploration results:

[0051] ① Study the tunnel geological survey data to determine the hazard level of the hydrogeological conditions ahead of the tunnel face, the hazard level of the geological structure, the hazard level of the stratum lithology, the hazard level of unfavorable geology and special rock and soil, the hazard level of gas, and the hazard level of the surrounding rock.

[0052] ② Use the advanced drilling method for detection and obtain the corresponding danger level based on the detection results.

[0053] ③ Use transient electromagnetic method for detection and obtain the corresponding danger level based on the detection results.

[0054] ④ Use TSP detection method to conduct detection and obtain the corresponding danger level based on the detection results.

[0055] ⑤ Use geological radar for detection and obtain the corresponding danger level based on the detection results.

[0056] The hazard level classification table is as follows. The various detection results can be converted into corresponding hazard level data through the table below.

[0057]

[0058]

[0059] 2) Because the selection of indicators in step 1) requires full consideration of various internal and external factors affecting the hazard level, many of the selected indicators are redundant and may even generate noise that affects the accuracy of the rules. Therefore, this step uses Rosetta software to perform attribute reduction on the 10 hazard level data obtained in step 1) to identify indicators with a support value equal to 100. Through attribute reduction and attribute importance ranking, we screen out the indicators that truly affect the hazard level, improving the indicator's resistance to interference. The results of this attribute reduction step are shown in the following table:

[0060]

[0061] According to the attribute reduction results listed in the table above, 10 indicators have a support value of 100. Therefore, to improve the efficiency of actual field operations, it is necessary to rank the indicators by attribute importance and select the most important ones as the final evaluation indicators for the hazardousness of the working area. The importance of each attribute is determined by the total number of times it appears in the final results. Sorting by importance reveals the degree of influence of each conditional attribute on the decision attribute: a higher number of occurrences indicates a greater influence. The following table shows the ranking results of each conditional attribute by importance.

[0062]

[0063] Among the 10 indicators, geological structure has the greatest impact on the hazard level. Surrounding rock classification, a qualitative static indicator, has the second greatest impact. The correlations between advance drilling, unfavorable geology and special rock and soil, geological radar, TSP detection, gas level, transient electromagnetic method, hydrogeological conditions, and stratum lithology and hazard level decrease in descending order.

[0064] Then, the attribute importance is ranked according to the total number of times each indicator appears in the final result of attribute simplification, and the top five indicators in attribute importance are selected as evaluation indicators for evaluating the hazard level of tunnel rock and gas outburst.

[0065] 3) The rock and gas outburst hazard level P(y) of the tunnel under a certain evaluation index i is calculated by the following formula: ij |x i ):

[0066]

[0067] Where x i Indicates the calculated value of the evaluation index, where the subscripts i = 1, 2, 3, 4, and 5 represent the evaluation indexes respectively; ijIt represents the hazard level evaluation standard value set for each evaluation index. The subscript j represents the rock and gas outburst hazard level, j = 1, 2, 3, 4, hazard level 1 represents no hazard, hazard level 2 represents weak hazard, hazard level 3 represents medium hazard, and hazard level 4 represents strong hazard.

[0068] Conditional probability: represents the probability of event B occurring under the condition that event A has already occurred. Suppose A and B are two events, and P(A)>0, then

[0069]

[0070] It is called the conditional probability of event B occurring given event A.

[0071] In this embodiment, P(y ij ) represents the prior probability of rock and gas outburst hazard level under the evaluation index i; P(x i |y ij ) indicates that the rock and gas outburst hazard level threshold is y ij When the calculated value of the evaluation index is x i The probability of P(y ij |x i ) indicates that the calculated value of the evaluation index is x i When the rock and gas outburst hazard level threshold reaches y ij probability.

[0072] The calculation process is as follows:

[0073] (1) In the initial calculation stage, the prior probability P(y ij ) are the same, that is:

[0074]

[0075] (2) Then calculate P(x i |y ij ):

[0076]

[0077] Where, L ij =|x i -y ij |, L ij The smaller the value, the greater the probability that the measured index belongs to the corresponding rock and gas outburst level;

[0078] (3) Finally calculate P(y ij |x i );

[0079] 4) Calculate the comprehensive posterior evaluation level P of a certain hazard level j under the action of all evaluation indicators j :

[0080]

[0081] Where, ω i represents the weight of evaluation index i;

[0082] Then determine the P with the largest probability value by the following formula j The final hazard level for tunnel rock and gas outburst is Z:

[0083]

[0084] In this embodiment, the weight ω i By combining weights, we get:

[0085] ω i =(1-λ)ω i1 +λω i2 (6)

[0086] Among them, ω i1 is the objective weight, ω i2 Expert weight.

[0087] ω i1 The acquisition process is as follows:

[0088] ① For the original matrix consisting of m evaluation samples and n evaluation indicators:

[0089] Q=(q li ) m×n

[0090] Where q li is the matrix element corresponding to the i-th evaluation index of the l-th evaluation sample, l=1, 2, ..., m; i=1, 2, ..., n, and in this embodiment, n=5.

[0091] ②Matrix normalization, regardless of the positive or negative direction of the indicator, is expressed as:

[0092]

[0093] Where q max is the maximum value under the same indicator.

[0094] ③Calculate information entropy:

[0095]

[0096] ④Calculate the entropy weight of evaluation index i

[0097]

[0098] ω i2 The calculation formula is as follows:

[0099]

[0100] Where W ik is the subjective weight ratio of the kth expert to the i-th evaluation index; P k is the authority weight ratio of the kth expert in the i evaluation index; z is the number of experts.

[0101] In this implementation, the tunnel rock and gas outburst hazard level prediction method based on multi-source information fusion is used, which integrates multiple means such as geological survey, geophysical prospecting and advance drilling methods to obtain objective data to evaluate the hazard level of tunnel rock and gas outburst, and the evaluation is accurate.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A tunnel rock and gas outburst hazard level assessment method based on multi-source information fusion is characterized by: The following steps are involved: 1) Explore the geological conditions ahead of the tunnel face using the following methods and determine the geological hazard level based on the exploration results: ① Study the tunnel geological survey data to determine the hazard level of the hydrogeological conditions ahead of the tunnel face, the hazard level of the geological structure, the hazard level of the stratum lithology, the hazard level of unfavorable geology and special rock and soil, the hazard level of gas, and the hazard level of the surrounding rock; ② Use advance drilling method to conduct detection and obtain the corresponding hazard level based on the detection results; ③ Use transient electromagnetic method to conduct detection and obtain the corresponding hazard level based on the detection results; ④ Use TSP detection method to detect and obtain the corresponding danger level according to the detection results; ⑤ Use geological radar to conduct detection and obtain the corresponding danger level based on the detection results; 2) Using Rosetta software to perform attribute simplification on the 10 hazard level data obtained in step 1), identify indicators with a support number equal to 100, and then rank the attribute importance based on the total number of times each indicator appears in the final attribute simplification result. The indicators ranked in the top n in attribute importance are selected as the evaluation indicators for evaluating the tunnel rock and gas outburst hazard level; 3) The rock and gas outburst hazard level P(y) of the tunnel under a certain evaluation index i is calculated by the following formula: ij |x i ): Where x i Indicates the calculated value of the evaluation index, the subscript i represents each evaluation index; y ij represents the hazard level evaluation standard value set for each evaluation index, subscript j represents the rock and gas outburst hazard level, j = 1, 2, 3, 4, hazard level 1 represents no hazard, hazard level 2 represents weak hazard, hazard level 3 represents medium hazard, hazard level 4 represents strong hazard; P(y ij ) represents the prior probability of rock and gas outburst hazard level under the evaluation index i; P(x i |y ij ) indicates that the rock and gas outburst hazard level threshold is y ij When the calculated value of the evaluation index is x i The probability of P(y ij |x i ) indicates that the calculated value of the evaluation index is x i When the rock and gas outburst hazard level threshold reaches y ij probability; The calculation process is as follows: (1) In the initial calculation stage, the prior probability P(y ij ) are the same, that is: (2) Then calculate P(x i |y ij ): Where, L ij =|x i -y ij |, L ij The smaller the value, the greater the probability that the measured index belongs to the corresponding rock and gas outburst level; (3) Finally calculate P(y ij |x i ); 4) Calculate the comprehensive posterior evaluation level P of a certain hazard level j under the action of all evaluation indicators j : Where, ω i represents the weight of evaluation index i; Then determine the P with the largest probability value by the following formula j The final hazard level for tunnel rock and gas outburst is Z:

2. The tunnel rock and gas outburst hazard level assessment method based on multi-source information fusion according to claim 1 is characterized by: The weight ω in formula (4) i By combining weights, we get: oh i =(1-λ)ω i1 +lo i2 (6) Among them, ω i1 is the objective weight, ω i2 is the expert weight; λ is the proportional coefficient; ω i1 The acquisition process is as follows: ① For the original matrix consisting of m evaluation samples and n evaluation indicators: Q=(q li ) m×n Where q li is the matrix element corresponding to the i-th evaluation index of the l-th evaluation sample, l = 1, 2, ..., m; i=1,2,…,n; ②Matrix normalization, regardless of the positive or negative direction of the indicator, is expressed as: Where q max is the maximum value under the same indicator; ③Calculate information entropy: ④Calculate the entropy weight of evaluation index i ω i2 The calculation formula is as follows: Where W ik is the subjective weight ratio of the kth expert to the i-th evaluation index; P k is the authority weight ratio of the kth expert in the i evaluation index; z is the number of experts.

Images