Tunnel gas monitoring and early warning method

Abstract

The invention provides a tunnel gas monitoring and early warning method, which is suitable for the field of tunnel construction monitoring and comprises the steps of gas monitoring point position arrangement, monitoring data acquisition, monitoring data analysis and early warning. The tunnel gas monitoring and early warning method provided by the invention has the advantages of high reliability and good safety, and can be widely applied to the field of tunnel construction monitoring.

Description

Technical Field

[0001] The invention relates to a tunnel gas monitoring and early warning method, which is applicable to the field of tunnel construction monitoring. Background Art

[0002] During the excavation process of the tunnel, gas release may be triggered when passing through some special rock formations such as coal-bearing strata and oil shale. Gas release is divided into three types: seepage, eruption and outburst. Seepage is a slow and even release, eruption is a prominent and intense release, and outburst is a violent release of a large amount of gas in a short time.

[0003] The large-scale release of gas will cause huge hidden dangers to the safety of the tunnel, mainly concentrated in the following aspects: 1) Explosion risk. After the gas is mixed with air, it will explode within a certain concentration range, causing casualties and property losses. 2) Suffocation. The gas itself is not toxic, but excessive gas concentration will lead to a decrease in the oxygen content in the air, which will lead to suffocation of personnel. 3) Combustion. Gas is flammable. Once it burns, the flame temperature is extremely high and may cause a fire. 4) Harmful gas generation: Gas explosions will produce toxic and harmful gases, such as carbon monoxide, which pose a threat to personnel health. In view of this, it is of great significance to monitor the gas in the tunnel to ensure that it is within a safe range.

[0004] At present, scholars at home and abroad mostly use real-time monitoring methods to deal with gas in tunnels, that is, to arrange monitoring points at fixed intervals in the tunnel to obtain the concentration monitoring results of gas in the tunnel. At the same time, we need to know that the source of gas is the tunnel excavation surface, and the tunnel excavation surface is very large. If a large amount of gas is released, it will inevitably cause safety accidents. Therefore, advanced gas detection holes are also used to detect the gas situation in the front of the tunnel. However, many scholars currently only consider the changes in the advanced gas detection holes themselves, and do not consider the changes in gas concentration for the tunnel as a whole, ignoring the relevance of tunnel excavation mileage. Therefore, a gas concentration monitoring and early warning method that comprehensively considers the changes in tunnel excavation mileage is currently needed. Summary of the invention

[0005] The purpose of the present invention is to solve the current urgent need for a gas concentration monitoring and early warning method that comprehensively considers the changes in tunnel excavation mileage, and proposes a tunnel gas monitoring and early warning method.

[0006] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0007] S101, Arrangement of gas monitoring points;

[0008] The arrangement of the gas monitoring points includes arranging gas monitoring sections every x meters from the excavation entrance before tunnel excavation. Suppose a total of n gas monitoring sections are arranged, represented by A n ;

[0009] The arrangement of the gas monitoring points also includes: n During excavation, gas advance detection holes are evenly arranged according to the shape of the excavation face. Each gas monitoring section A n The number of gas advance detection holes is m, the value of m is 3-8, and the depth of the advance detection holes is x meters;

[0010] S102, acquisition of monitoring data;

[0011] The acquisition of monitoring data includes: for each excavated gas monitoring section A1, A2, ...A i , m gas advance detection holes i∈[1,n] are all subjected to gas concentration detection and gas pressure detection for 48 hours after the drilling of the gas advance detection holes. According to the gas concentration detection and gas pressure detection results, gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hours i,max And the maximum value of gas pressure detection P i,max ;

[0012] S103, analysis of monitoring data;

[0013] The analysis of the monitoring data includes obtaining gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hours i,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D1 is formed, and the gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas pressure detection P within 48 hours i,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D2 is formed;

[0014] The analysis of the monitoring data also includes constructing a prediction model for the data sets D1 and D2 using a variational autoencoder (VAE) method to obtain a next gas concentration detection prediction value y1 and a gas pressure detection maximum value prediction value y2. The steps of the method for constructing a prediction model using a variational autoencoder (VAE) method to obtain prediction values ​​are as follows:

[0015] a) Get data set D1 or D2,

[0016] b) performing normalization processing on the data set, wherein the normalization method adopts the maximum and minimum value method,

[0017] c) dividing the data set into a training set and a test set according to a certain ratio,

[0018] d) creating an input sequence and a target value, dividing the training set and the test set into a plurality of input sequences of a fixed length L, and taking the value in the next data set of each sequence as the target value, creating an input sequence and a target value,

[0019] e) A prediction model is constructed by using a variational autoencoder (VAE) method. The encoder and decoder of the prediction model have three network layers, and each layer has 64 neurons. The activation function used by the prediction model is a smoothing rectifier function Softplus, which is expressed as formula (1).

[0020] softplus(x)=log(1+e x ) (1)

[0021] f) using the prediction model constructed by the variational autoencoder VAE method described above to perform model training and model testing in combination with training set and test set data, adjusting the prediction model according to the model test results, and finally obtaining an adjusted model,

[0022] g) using the adjusted model, given an input sequence of the last L data of the data set of length L, to obtain a predicted value;

[0023] S104, the warning purpose is achieved;

[0024] The achievement of the warning purpose includes predicting the next gas section A according to the obtained gas concentration detection prediction value y1 and the gas pressure detection maximum value prediction value y2. i+1 Judge the gas situation, and combine the established gas concentration detection and early warning standards and gas pressure detection and early warning standards to determine whether to issue an early warning to achieve the purpose of guiding construction.

[0025] Furthermore, in the above S101, the gas advance detection holes are arranged along the tunnel axis direction;

[0026] Furthermore, in the above S102, the gas concentration is detected by a composite gas detector, and the gas pressure is detected by a precision gas pressure measuring instrument;

[0027] Furthermore, in the above S102, the unit of the gas concentration detection result is volume fraction / %, and the unit of the gas pressure detection result is Pascal / Pa.

[0028] Furthermore, in the above S104, the gas concentration detection warning standard is 0.5%-1%, and the gas pressure detection warning standard is 10-20Pa.

[0029] The beneficial effects of the present invention are:

[0030] The beneficial effects of the present invention are as follows: the present invention comprehensively considers the influence of tunnel mileage on gas concentration, adopts the variational autoencoder VAE method to construct a prediction model, and predicts the gas situation of the section where the advanced gas detection hole is to be excavated. An emergency plan can be prepared in advance with reference to the prediction results, thereby further improving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A flow chart of a tunnel gas monitoring and early warning method of the present invention; Figure 2 This is a graph showing the maximum value of gas concentration detection data of an embodiment of the present invention; Figure 3 This is a graph of the maximum value data results of the gas pressure detection according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings; it should be understood that the specific embodiments given here are only used to illustrate and explain the present invention and cannot be used to limit the present invention.

[0033] The following is a specific embodiment of a tunnel gas monitoring and early warning method.

[0034] like Figure 1 FIG. 1 is a flow chart of a tunnel gas monitoring and early warning method according to the present invention.

[0035] S101, Arrangement of gas monitoring points;

[0036] The arrangement of the gas monitoring points includes arranging gas monitoring sections every x meters from the excavation entrance before tunnel excavation. Suppose a total of n gas monitoring sections are arranged, represented by A n ;

[0037] The arrangement of the gas monitoring points also includes: n During excavation, gas advance detection holes are evenly arranged according to the shape of the excavation face. Each gas monitoring section A n The number of gas advance detection holes is m, the value of m is 3-8, and the depth of the advance detection holes is x meters; further, in the above S101, the gas advance detection holes are arranged along the tunnel axis;

[0038] In this embodiment, the supporting project is a tunnel project of the C2 line of the eighth division of the East Coast Rail Project of Malaysia. The total length of the tunnel is about 3,600 meters. The terrain conditions of the tunnel are located in the central part of the Malaysian Peninsula, with multiple mountain ranges parallel to the direction of the peninsula, dividing the peninsula into three geographical areas: the eastern coastal belt, the central mountain belt, and the western coastal belt. The terrain is high and low on both sides. This project is mainly located in the eastern coastal belt and the central mountain belt. The end point is the transition area from the central mountain belt to the western coastal belt. The lowest point along the line is about 3 meters above sea level, near the eastern coastal belt, and the highest point is about 1,060 meters, near the central mountain belt.

[0039] During the tunnel excavation process, a total of 25 gas monitoring sections were set up, denoted as A1~A25. The gas monitoring sections were set every 50 meters and arranged along the axis of the tunnel.

[0040] S102, acquisition of monitoring data;

[0041] The acquisition of monitoring data includes: for each excavated gas monitoring section A1, A2, ...A i , m gas advance detection holes i∈[1,n] are all subjected to gas concentration detection and gas pressure detection for 48 hours after the drilling of the gas advance detection holes. According to the gas concentration detection and gas pressure detection results, gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hours i,max And the maximum value of gas pressure detection P i,max ; Further, in the above S102, the gas concentration is detected by a composite gas detector, and the gas pressure is detected by a precision gas pressure measuring instrument; further, in the above S102, the unit of the gas concentration detection result is volume fraction / %, and the unit of the gas pressure detection result is Pascal / Pa.

[0042] In this embodiment, a total of 25 gas monitoring points are set. When the 20th gas monitoring point is excavated, the maximum gas concentration detection value Ci,max and the maximum gas pressure detection value Pi,max within 48 hours are obtained when i=10. The data results are as follows: Figure 2 and Figure 3 shown.

[0043] S103, analysis of monitoring data;

[0044] The analysis of the monitoring data includes obtaining gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hoursi,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D1 is formed, and the gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas pressure detection P within 48 hours i,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D2 is formed;

[0045] The analysis of the monitoring data also includes constructing a prediction model for the data sets D1 and D2 using a variational autoencoder (VAE) method to obtain a next gas concentration detection prediction value y1 and a gas pressure detection maximum value prediction value y2. The steps of the method for constructing a prediction model using a variational autoencoder (VAE) method to obtain prediction values ​​are as follows:

[0046] a) Get data set D1 or D2,

[0047] b) performing normalization processing on the data set, wherein the normalization method adopts the maximum and minimum value method,

[0048] c) dividing the data set into a training set and a test set according to a certain ratio,

[0049] d) creating an input sequence and a target value, dividing the training set and the test set into a plurality of input sequences of a fixed length L, and taking the value in the next data set of each sequence as the target value, creating an input sequence and a target value,

[0050] e) A prediction model is constructed by using a variational autoencoder (VAE) method. The encoder and decoder of the prediction model have three network layers, and each layer has 64 neurons. The activation function used by the prediction model is a smoothing rectifier function Softplus, which is expressed as formula (1).

[0051] softplus(x)=log(1+e x ) (1)

[0052] f) using the prediction model constructed by the variational autoencoder VAE method described above to perform model training and model testing in combination with training set and test set data, adjusting the prediction model according to the model test results, and finally obtaining an adjusted model,

[0053] g) using the adjusted model, given an input sequence of the last L data of the data set of length L, to obtain a predicted value;

[0054] In this embodiment, the prediction method is applied to obtain the predictions y1=0.19%, y2=2.0Pa.

[0055] S104, the warning purpose is achieved;

[0056] The achievement of the warning purpose includes predicting the next gas section A according to the obtained gas concentration detection prediction value y1 and the gas pressure detection maximum value prediction value y2. i+1 The gas situation is judged, and combined with the formulated gas concentration detection warning standard and gas pressure detection warning standard, it is judged whether to issue an early warning to achieve the purpose of guiding construction. Further, in the above S104, the gas concentration detection warning standard is 0.5%-1%, and the gas pressure detection warning standard is 10~20Pa.

[0057] In this embodiment, according to the obtained y1 = 0.19%, y2 = 2.0Pa, the next gas section A 21 According to the situation, combined with the early warning standard, the next gas section A is determined 21 The situation is stable and no warning is required.

[0058] In the above embodiments, the present invention discloses a tunnel gas monitoring and early warning method, including the arrangement of gas monitoring points, the acquisition of monitoring data, the analysis of monitoring data and the achievement of early warning purposes; the tunnel gas monitoring and early warning method proposed in the present invention has the advantages of high reliability and good safety, and can be widely used in the field of tunnel construction monitoring.

[0059] The above description is 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 principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A tunnel gas monitoring and early warning method, characterized in that: The steps of the tunnel gas monitoring and early warning method are as follows: 1) Arrangement of gas monitoring points; 2) Acquisition of monitoring data; 3) Analysis of monitoring data; 4) The purpose of early warning is achieved; The arrangement of the gas monitoring points includes arranging gas monitoring sections every x meters from the excavation entrance before tunnel excavation. Suppose a total of n gas monitoring sections are arranged, represented by A n ; The arrangement of the gas monitoring points also includes: n During excavation, gas advance detection holes are evenly arranged according to the shape of the excavation face. Each gas monitoring section A n The number of gas advance detection holes is m, the value of m is 3-8, and the depth of the advance detection holes is x meters; The acquisition of monitoring data includes: for each excavated gas monitoring section A1, A2, ...A i , m gas advance detection holes i∈[1,n] are all subjected to gas concentration detection and gas pressure detection for 48 hours after the drilling of the gas advance detection holes. According to the gas concentration detection and gas pressure detection results, gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hours i,max And the maximum value of gas pressure detection P i,max ; The analysis of the monitoring data includes obtaining gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas concentration detection C within 48 hours i,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D1 is formed, and the gas monitoring sections A1, A2, ...A i ,i∈[1,n] the maximum value of gas pressure detection P within 48 hours i,max Arranged from left to right according to the distance from the tunnel entrance, the gas concentration detection maximum value data set D2 is formed; The analysis of the monitoring data also includes constructing a prediction model for the data sets D1 and D2 using a variational autoencoder (VAE) method to obtain a next gas concentration detection prediction value y1 and a gas pressure detection maximum value prediction value y2. The steps of the method for constructing a prediction model using a variational autoencoder (VAE) method to obtain prediction values ​​are as follows: a) Get data set D1 or D2, b) performing normalization processing on the data set, wherein the normalization method adopts the maximum and minimum value method, c) dividing the data set into a training set and a test set according to a certain ratio, d) creating an input sequence and a target value, dividing the training set and the test set into a plurality of input sequences of a fixed length L, and taking the value in the next data set of each sequence as the target value, creating an input sequence and a target value, e) A prediction model is constructed by using a variational autoencoder (VAE) method. The encoder and decoder of the prediction model have three network layers, and each layer has 64 neurons. The activation function used by the prediction model is a smoothing rectifier function Softplus, which is expressed as formula (1). softplus(x)=log(1+e x )(1) f) using the prediction model constructed by the variational autoencoder VAE method described above to perform model training and model testing in combination with the training set and test set data, adjusting the prediction model according to the model test results, and finally obtaining the adjusted model, g) using the adjusted model, given an input sequence of the last L data of the data set of length L, to obtain a predicted value; The achievement of the warning purpose includes predicting the next gas section A according to the obtained gas concentration detection prediction value y1 and the gas pressure detection maximum value prediction value y2. i+1 Judge the gas situation, and combine the established gas concentration detection and early warning standards and gas pressure detection and early warning standards to determine whether to issue an early warning to achieve the purpose of guiding construction.

2. A tunnel gas monitoring and early warning method according to claim 1, characterized in that: In the above 1), the gas advance detection holes are arranged along the tunnel axis.

3. A tunnel gas monitoring and early warning method according to claim 1, characterized in that: In the above 2), the gas concentration is detected by a composite gas detector, and the gas pressure is detected by a precision gas pressure measuring instrument.

4. A tunnel gas monitoring and early warning method according to claim 1, characterized in that: In the above 2), the unit of the gas concentration detection result is volume fraction / %, and the unit of the gas pressure detection result is Pascal / Pa.

5. A tunnel gas monitoring and early warning method according to claim 1, characterized in that: In the above 4), the gas concentration detection and warning standard is 0.5%-1%, and the gas pressure detection and warning standard is 10-20Pa.

Images