Dynamic adjustment method for microseismic early-warning index threshold value of bed separation water disaster

Abstract

The invention provides a dynamic adjustment method for a microseismic early warning index threshold of separation water damage. The method specifically comprises the following steps: collecting related data of hydrogeology, engineering geology and coal mining in a research area; establishing a research area separation layer water disaster prediction index system; constructing a study area separation layer water disaster prediction model; microseismic data in the coal seam mining process are collected; establishing a microseismic parameter early warning index system of the separation layer water disaster in the research area; setting an initial value of a microseismic parameter early warning threshold value; dynamically adjusting a microseismic parameter early warning threshold value by adopting a machine learning calculation model; according to the method, the possibility of occurrence of the separation water disaster can be evaluated more accurately. The dynamic adjustment mode can adapt to the characteristics of bed separation water disasters in different areas and different mining stages, and the reliability and effectiveness of early warning are improved.

Description

Technical Field

[0001] The present invention relates to the field of mine water disaster monitoring and early warning, and particularly to a method for dynamically adjusting the threshold of microseismic early warning indicators for separated layer water disasters. Background Art

[0002] The dominant position of coal resources in China's energy system will not change significantly in the short term. At present, the overall situation of coal mine safety is improving, but with the transfer of coal development to the west, the problem of roof water disasters in western mining areas is becoming increasingly serious. Among them, the separated layer water disaster is characterized by large instantaneous water volume, periodicity and suddenness, with extremely great harm and extremely high prevention and control difficulty, and has become the main thorny problem faced by western mining areas.

[0003] The occurrence of mine water disaster accidents is mainly the result of the combined action of rock mass rupture and mine water seepage. Existing research shows that microseismic signals can effectively characterize the location, strength and focal mechanism of rock mass rupture. Therefore, microseismic monitoring technology is widely used in the activation and instability processes of water-conducting channel faults, collapse columns, etc. under mining disturbances, water inrush from goafs, and the gestation and development processes of mining-induced fractures in roof and floor. At present, for the monitoring and early warning of roof water disasters, the threshold of microseismic early warning indicators is selected as a fixed value. However, for roof water disasters, especially separated layer water disasters, their occurrence is closely related to coal seam mining and the geological and hydrogeological conditions of overlying strata in the working face. There are essential differences in the water inrush evolution mechanisms in different regions. Therefore, it is inappropriate to select a fixed threshold to judge the possibility of water disasters occurring during the future working face mining process. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for dynamically adjusting the threshold of microseismic early warning indicators for separated layer water disasters, which can more accurately evaluate the possibility of separated layer water disasters occurring. Among them, the dynamic adjustment method can adapt to the characteristics of separated layer water disasters in different regions and different mining stages, and improves the reliability and effectiveness of early warning.

[0005] In order to achieve the above invention purpose, the technical solution adopted by the present invention is specifically as follows: The present invention provides a method for dynamically adjusting the threshold of microseismic early warning indicators for separated layer water disasters, including the following steps:

[0006] Including the following steps: Including the following steps:

[0007] S1. Collect hydrogeological, engineering geological and coal seam mining data in the study area, and statistically analyze the spatial occurrence of overlying strata of the coal seam and hydrogeological parameters;

[0008] S2. Establish a prediction index system for separated layer water disasters in the study area, use the coal seam mining index I1, overlying strata structure index I2, stratigraphic spatial occurrence index I3 and Cretaceous aquifer index I4 as the prediction and evaluation indexes for high-level separated layer water disasters, and calculate the values of each index;

[0009] S3. Build a prediction model for separated seam water disaster in the study area, standardize the index values of boreholes near the mined areas / working faces, divide the training set and the test set, and train the model based on the PSO-SVM algorithm until the accuracy rates of the training set and the test set reach over 75%.

[0010] S4. Collect microseismic data during coal seam mining, arrange microseismic sensors in the coal seam of the mining area / working face, and monitor the development of roof fissures in real time.

[0011] S5. Establish a microseismic parameter warning index system for separated seam water disaster in the study area, select the total daily microseismic energy, the total daily number of times, and the number of energy times > 10 3 J as warning indicators, and form a ledger of the warning index system.

[0012] S6. Set the initial value of the microseismic parameter warning threshold, classify the occurrence probability of separated seam water disaster according to the prediction model of separated seam water disaster in the study area, and set the initial value in combination with the microseismic warning index situation during separated seam water inrush in different regions.

[0013] S7. Use the relative error method between the warning threshold during working face water inrush and the initial warning threshold, and combine with the machine learning calculation model to dynamically adjust the microseismic parameter warning threshold. Specifically: use the machine learning calculation model to dynamically adjust the microseismic parameter warning threshold, obtain the critical values of each microseismic warning index during each separated seam water inrush, calculate its relative error with the initial value of the warning threshold, use the relative error as the input vector and the critical values of each microseismic warning index of the next separated seam water inrush as the output vector, and adopt the GA-BP neural network method for iterative learning until the accuracy rates of the training set and the test set reach 80%, and obtain the microseismic warning thresholds before this separated seam water inrush accordingly.

[0014] Preferably, the hydrogeological, engineering geological and coal seam mining data in the study area include: coal seam thickness and burial depth, thickness of Cretaceous aquifer, permeability coefficient of Cretaceous aquifer, specific yield of Cretaceous aquifer, lithological characteristics of overlying strata of coal seam, thickness of sand / mudstone in each stratum, and distance between coal seam and Cretaceous aquifer.

[0015] Preferably, the coal seam mining index I1 is calculated according to the following formula:

[0016] I1 = ln(M c '+1)×D'

[0017] In the formula, M c is the coal seam thickness, with the unit of m; D is the coal seam burial depth, with the unit of m; M c ' is the standardized value of the coal seam thickness; D' is the standardized value of the coal seam burial depth.

[0018] Preferably, the overlying strata structure index I2 is calculated according to the following formula:

[0019]

[0020] In the formula, R s is the proportion of sandstone in the overlying strata; R n is the proportion of mudstone in the overlying strata; R′ s is the standardized value of the proportion of sandstone in the overlying strata; R' n is the standardized value of the proportion of mudstone in the overlying strata.

[0021] Preferably, the stratum space occurrence index I3 is calculated according to the following formula:

[0022]

[0023] In the formula, D ch is the distance from the Cretaceous aquifer to the coal seam, in m; M yz is the total thickness of mudstone in the Yan'an Formation and the Zhiluo Formation, in m; M a is the thickness of mudstone in the Anding Formation, in m; D′ ch is the standardized value of the distance from the Cretaceous aquifer to the coal seam; M' yz is the standardized value of the total thickness of mudstone in the Yan'an Formation and the Zhiluo Formation; M' a is the standardized value of the thickness of mudstone in the Anding Formation.

[0024] Preferably, the aquifer index I4 is calculated according to the following formula:

[0025] I4 = M′ h ·K′·q′

[0026] In the formula, M h is the thickness of the aquifer, in m; K is the permeability coefficient of the aquifer, in m / d; q is the specific yield of the aquifer, in L / (s·m); M' h is the standardized value of the thickness of the aquifer; K' is the standardized value of the permeability coefficient of the aquifer; q' is the standardized value of the specific yield of the aquifer.

[0027] Preferably, the standardization processing of each index is carried out according to the following formula:

[0028]

[0029] In the formula, x' m is the original data of each index; max(x m ) is the maximum value of each index; min(x m ) is the minimum value of each index; x' m is the normalized value of each index.

[0030] Preferably, when building the prediction model for separated seam water disaster in the study area, if the accuracy rates of the training set and the test set do not reach 75%, readjust the proportion of the training set and the test set until the accuracy rate reaches over 75%.

[0031] Preferably, when using the GA-BP neural network method for iterative learning, the samples are initially divided into a training set and a test set at a ratio of 4:1. If the accuracy rates of the training set and the test set do not reach 80%, the training set and the test set need to be re-divided at a random ratio until the accuracy rates of the training set and the test set reach 80%.

[0032] Preferably, the relative error calculation method is calculated according to the following formula:

[0033]

[0034] In the formula, RE is the relative error between the critical value of each microseismic early warning index and the initial value of the early warning threshold before each separated seam water inrush; A is the critical value of each microseismic early warning index before each separated seam water inrush; T is the initial value of the early warning threshold of each microseismic index.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] Accurately predict separated seam water disasters: By establishing a separated seam water disaster prediction index system and a prediction model, comprehensively considering various factors such as coal seam mining, overlying rock structure, stratigraphic spatial occurrence, and aquifers, the possibility of separated seam water disasters can be predicted more accurately.

[0037] Dynamically adjust the early warning threshold: Using a machine learning calculation model, according to the relative error between the critical value of the microseismic early warning index and the initial threshold during each separated seam water inrush, dynamically adjust the early warning threshold of microseismic parameters. This dynamic adjustment method can adapt to the characteristics of separated seam water disasters in different regions and different mining stages, improving the reliability and effectiveness of early warning.

[0038] Improve the prevention and control ability of separated seam water disasters: This technical solution provides a scientific and systematic method for the monitoring and early warning of separated seam water disasters, helps coal mining enterprises to detect separated seam water disaster hidden dangers in a timely manner, take effective prevention and control measures, reduce the threat of separated seam water disasters to coal mine safety production, and ensure the safe mining of coal resources. Description of the Drawings

[0039] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention.

[0040] Figure 1 It is the separated seam structure diagram of the Shaanxi Yonglong-Binchang mining area provided in the embodiments of the present invention;

[0041] Figure 2Flow chart of a dynamic adjustment method for microseismic warning index thresholds of separated seam water hazards provided by the present invention;

[0042] Figure 3 Flow chart of step S3 provided by the present invention;

[0043] Figure 4 Flow chart of step S7 provided by the present invention. Detailed implementation manners

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Of course, the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] Taking the separated seam water hazard in Yonglong, Shaanxi Binchang mining area as an example (see Figure 1 ), this embodiment provides a dynamic adjustment method for microseismic warning index thresholds of separated seam water hazards. As Figure 2 shown, it specifically includes the following steps:

[0046] S1. Collect relevant data on hydrogeology, engineering geology and coal seam mining in the study area.

[0047] In this embodiment, by collecting relevant data in the study area, the spatial occurrence of overlying strata of the coal seam and hydrogeological parameters are statistically analyzed, specifically including: coal seam thickness and burial depth, thickness of Cretaceous aquifer, permeability coefficient of Cretaceous aquifer, specific yield of Cretaceous aquifer, lithological characteristics of overlying strata of the coal seam, sand / mudstone thickness of each stratum, and distance between the coal seam and Cretaceous aquifer, etc.

[0048] S2. Establish a prediction index system for separated seam water hazards in the study area.

[0049] In this embodiment, the coal seam mining index I1, overlying strata structure index I2, stratigraphic spatial occurrence index I3 and Cretaceous aquifer index I4 are used as prediction and evaluation indexes for high-level separated seam water hazards; the index values of the coal seam mining index I1, overlying strata structure index I2, stratigraphic spatial occurrence index I3 and aquifer index I4 in the study area are calculated.

[0050] Specifically, since the overlying strata in the caving zone have a certain degree of swelling, the separated seam height must be less than the coal seam thickness. The smaller the coal seam mining thickness, the smaller the scale of separated seam in the overlying strata; the larger the coal seam thickness, the greater the deformation of the overlying strata after mining, and the larger the separated seam cavity, that is, the separated seam water accumulation space, and the greater the risk of separated seam water hazard. In addition, with the increase of the coal seam burial depth, the overlying strata stress also increases, the failure strength of the overlying strata increases, and the risk of the water-conducting fissure zone communicating with the separated seam increases. Therefore, the coal seam mining index I1 is calculated according to the following formula:

[0051] I1 = ln(M c '+ 1) × D'

[0052] Wherein, M c is the coal seam thickness, with the unit of m; D is the depth of the coal seam, with the unit of m; M c ' is the standardized value of the coal seam thickness; D' is the standardized value of the depth of the coal seam.

[0053] Specifically, the overlying rock structure affects the strata movement, the development of water-conducting fissures and separations. The strata between the coal seam and the aquifer are composed of rock masses with alternating sand and mud. The mechanical strength of sandstone is relatively high, the deformation and failure range is relatively small, and it is not easy to communicate with the overlying separation. When the overlying rock structure is mainly composed of mudstone, it is easy to break, but as a fissure channel, it is easily blocked by its own rock debris. Therefore, the overlying rock structure index I2 is calculated according to the following formula:

[0054]

[0055] Wherein, R s is the proportion of sandstone in the overlying rock; R n is the proportion of mudstone in the overlying rock; R′ s is the standardized value of the proportion of sandstone in the overlying rock; R' n is the standardized value of the proportion of mudstone in the overlying rock.

[0056] Specifically, water inrush from separation is often accompanied by a large amount of sediment flowing into the working face. The mudstone in the Jurassic Yan'an Formation and Zhiluo Formation is more likely to be cemented when encountering water compared with sandstone, and the loss of particles will further exacerbate water inrush from separation; in addition, the mudstone in the Anding Formation acts as an aquifuge and plays an inhibitory role in water inrush from separation. When the thickness of the mudstone in the Anding Formation increases, the ability to resist water inrush is stronger, and the possibility of water disaster caused by separation is smaller; in the spatial occurrence of strata, whether water inrush from separation occurs is also related to the distance between the Cretaceous aquifer and the coal seam. The smaller the distance between the Cretaceous aquifer and the coal seam, the easier it is for the water-conducting fissures to reach the separation space. Therefore, the spatial occurrence index I3 of strata is calculated according to the following formula:

[0057]

[0058] Wherein, D ch is the distance between the Cretaceous aquifer and the coal seam, with the unit of m; M yz is the total thickness of the mudstone in the Yan'an Formation and Zhiluo Formation, with the unit of m; M a is the thickness of the mudstone in the Anding Formation, with the unit of m; D′ ch is the standardized value of the distance between the Cretaceous aquifer and the coal seam; M' yz is the standardized value of the total thickness of the mudstone in the Yan'an Formation and Zhiluo Formation; M' a is the standardized value of the thickness of the mudstone in the Anding Formation.

[0059] Specifically, the water-richness of the aquifer directly affects the water inrush scale of the separated seam water. The traditional water-richness evaluation is carried out by obtaining the unit water inrush q of the borehole through a pumping test. The unit water inrush q reflects the water supply capacity of the aquifer. In addition, the water inrush in the mine is not only related to the water supply capacity of the aquifer, but also related to the water-bearing capacity of the aquifer. The thickness and permeability coefficient are the main indicators characterizing the water-bearing capacity of the aquifer. The larger the thickness of the aquifer and the larger the permeability coefficient, the greater the water-bearing capacity of the aquifer. Therefore, the aquifer index I4 is calculated according to the following formula:

[0060] I4 = M' h ·K'·q'

[0061] In the formula, M h is the thickness of the aquifer, with the unit of m; K is the permeability coefficient of the aquifer, with the unit of m / d; q is the unit water inrush of the aquifer, with the unit of L / (s·m); M' h is the standardized value of the aquifer thickness; K' is the standardized value of the aquifer permeability coefficient; q' is the standardized value of the aquifer unit water inrush.

[0062] Specifically, the standardization processing of each index is carried out according to the following formula:

[0063]

[0064] In the formula, x' m is the original data of each index; max(x m ) is the maximum value of each index; min(x m ) is the minimum value of each index; x' m is the normalized value of each index.

[0065] S3. Construct a prediction model for separated seam water disaster in the study area.

[0066] In this embodiment, first, calculate the 4 index values of the boreholes near the mined-out area / working face, and process the 4 index values using the above standardization method; then, divide the boreholes into a training set and a test set according to a quantity ratio of 3:1; finally, based on the PSO-SVM algorithm, use the matlab software to perform model training on the 4 standardized index values of the boreholes near the mined-out area / working face. When the accuracy rates of the training set and the test set reach more than 75%, the model training is completed. If the training does not meet the standard, the ratio of the training set and the test set should be readjusted until the accuracy rate reaches more than 75%.

[0067] S4. Collect microseismic data during the coal seam mining process. In this embodiment, the microseismic sensors should be arranged on the coal seam roof of the mining area / working face to achieve the purpose of real-time monitoring of the development of roof fractures.

[0068] S5. Establish a microseismic parameter warning index system for separated seam water disaster in the study area. In this embodiment, the total daily microseismic energy, the total daily number of microseisms, and the number of energy occurrences > 10 3

[0069] J are selected as the microseismic parameter warning indexes for separated seam water disaster. Through downhole microseismic sensors and ground stations, the microseismic parameters corresponding to the coal seam roof during mine production and working face excavation are monitored in real time. The collected microseismic parameters are orderly put into an excel table at 1-day time intervals to form a ledger of the microseismic parameter warning index system for separated seam water disaster. S6. Set the initial value of the microseismic parameter warning threshold.

[0070] In this embodiment, according to the above-mentioned prediction model for separated seam water disaster in the study area, the occurrence probability of separated seam water disaster in the study area is classified by the Jenks natural breakpoint method. Combining the situation of microseismic warning indexes during separated seam water inrush in different regions, the initial values of microseismic parameter warning thresholds corresponding to different prediction levels are set.

[0071] S7. Adopt a machine learning calculation model to dynamically adjust the microseismic parameter warning threshold.

[0072] In this embodiment, obtain the critical values of each microseismic warning index during each separated seam water inrush, and calculate the relative error between the critical values of each microseismic warning index before each separated seam water inrush and the initial value of the warning threshold;

[0073] Furthermore, take the relative error between the critical values of each microseismic warning index before each separated seam water inrush and the initial value of the warning threshold as the input vector, take the critical values of each microseismic warning index of the next separated seam water inrush as the output vector, divide the samples into a training set and a test set according to a ratio of 4:1, and adopt the GA-BP neural network method for iterative learning for a preset number of times until the accuracy rates of the training set and the test set reach 80%;

[0074] Furthermore, import the relative error between the critical values of each microseismic warning index before the previous separated seam water inrush and the initial value of the warning threshold into the already trained GA-BP model to obtain the warning thresholds of each microseismic index before the current separated seam water inrush.

[0075] Specifically, the relative error calculation method is calculated according to the following formula:

[0076]

[0077] In the formula, RE is the relative error between the critical value of each microseismic warning index before each separated seam water inrush and the initial value of the warning threshold; A is the critical value of each microseismic warning index before each separated seam water inrush; T is the initial value of the warning threshold of each microseismic index.

[0078] Specifically, the sample ratio can be initially divided into a training set and a test set at a ratio of 4:1. If the accuracy rates of the training set and the test set do not reach 80%, the training set and the test set need to be re-divided at a random ratio until the accuracy rates of the training set and the test set reach 80%.

[0079] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for dynamically adjusting the threshold of microseismic early warning indicators for delamination water damage, characterized in that: The method comprises the following steps: S1. Collect hydrogeological, engineering geological and coal mining data in the study area, and count the spatial distribution of coal seam overburden and hydrogeological parameters; S2. Establish a prediction index system for abscission water hazards in the study area, taking coal seam mining index I1, overburden structure index I2, stratum spatial occurrence index I3 and Cretaceous aquifer index I4 as prediction and evaluation indexes for high-level abscission water hazards, and calculate the values ​​of each index; S3. Construct a prediction model for delamination water damage in the study area, standardize the index values ​​of the boreholes near the mined area / working face, divide the training set and the test set, and train the model based on the PSO-SVM algorithm until the accuracy of the training set and the test set reaches more than 75%; S4. Collect microseismic data during coal mining, place microseismic sensors in the coal seams in the mining area / working face, and monitor the development of roof cracks in real time; S5. Establish an early warning index system for microseismic parameters of delamination water hazards in the study area, and select the daily total energy, daily total number and >10 3 The energy times of J are used as early warning indicators to form an early warning indicator system account; S6. Set the initial value of the microseismic parameter warning threshold, classify the probability of occurrence of delamination water disaster according to the delamination water disaster prediction model in the study area, and set the initial value based on the microseismic warning indicators of delamination water inrush in different regions; S7. Use a machine learning calculation model to dynamically adjust the microseismic parameter warning threshold, obtain the critical value of each microseismic warning indicator at each delamination water inrush, calculate the relative error between the critical value and the initial value of the warning threshold, use the relative error as the input vector and the critical value of each microseismic warning indicator of the next delamination water inrush as the output vector, use the GA-BP neural network method for iterative learning until the accuracy of the training set and the test set reaches 80%, and obtain the microseismic warning thresholds before the delamination water inrush.

2. The method according to claim 1, characterized in that The hydrogeological, engineering geological and coal mining data of the study area include: coal seam thickness and burial depth, Cretaceous aquifer thickness, Cretaceous aquifer permeability, Cretaceous aquifer unit yield, coal seam overburden lithology, sand / mudstone thickness of each stratum and the distance between coal seam and Cretaceous aquifer.

3. The method according to claim 1, characterized in that The coal seam mining index I1 is calculated according to the following formula: I1=ln(M c '+1)×D' Where M c is the thickness of the coal seam, in m; D is the buried depth of the coal seam, in m; M c ' is the standardized value of coal seam thickness; D' is the standardized value of coal seam burial depth.

4. The method according to claim 1, characterized in that: The overburden structure index I2 is calculated according to the following formula: In the formula, R s is the proportion of overlying sandstone; R n is the proportion of overlying mudstone; R s ' is the standardized value of the overburden sandstone ratio; R' n is the normalized value of the overlying mudstone ratio.

5. The method according to claim 1, characterized in that The stratum spatial occurrence index I3 is calculated according to the following formula: Where D ch is the distance from the Cretaceous aquifer to the coal seam, in meters; M yz is the total thickness of the mudstone in the Yan'an Formation and Zhiluo Formation, in m; M a is the thickness of the Anding Formation mudstone, in meters; D′ ch is the normalized value of the distance from the Cretaceous aquifer to the coal seam; M' yz is the standardized value of the total thickness of the Yan'an Formation and Zhiluo Formation mudstone; M' a is the standardized value of the Anding Formation mudstone thickness.

6. The method according to claim 1, characterized in that The aquifer index I4 is calculated according to the following formula: I4=M′ h ·K′·q′ Where M h is the thickness of the aquifer, in m; K is the permeability coefficient of the aquifer, in m / d; q is the unit yield of the aquifer, in L / (s·m); M' h is the standardized value of aquifer thickness; K' is the standardized value of aquifer permeability; q' is the standardized value of aquifer unit yield.

7. The method according to any one of claims 2 to 5, characterized in that: The standardized processing of each indicator is carried out according to the following formula: In the formula, x m is the original data of each indicator; max(x m ) is the maximum value of each indicator; min(x m ) is the minimum value of each indicator; x' m is the normalized value of each indicator.

8. The method according to claim 1, characterized in that When constructing the abscission water disaster prediction model for the study area, if the accuracy of the training set and the test set does not reach 75%, the ratio of the training set and the test set will be readjusted until the accuracy reaches more than 75%.

9. The method according to claim 1, characterized in that: When using the GA-BP neural network method for iterative learning, the samples are initially divided into training set and test set in a ratio of 4:

1. If the accuracy of the training set and the test set does not reach 80%, the training set and the test set need to be redivided according to a random ratio until the accuracy of the training set and the test set reaches 80%.

10. The method according to claim 1, characterized in that The relative error calculation method is calculated according to the following formula: Where RE is the relative error between the critical value of each microseismic warning indicator and the initial value of the warning threshold before each delamination water inrush; A is the critical value of each microseismic warning indicator before each delamination water inrush; T is the initial value of the warning threshold of each microseismic indicator.

Images