Coal mine disaster risk prediction method based on time sequence

Abstract

The invention relates to the technical field of coal mine disaster risk prediction, in particular to a coal mine disaster risk prediction method based on a time sequence, and the method comprises the steps: obtaining coal mine data; the average gas concentration of the coal seam is obtained by combining gas release characteristics; the distribution difference characteristic of the average gas concentration of the coal seam is combined with the characteristic of gas flow to analyze and construct a flow trend value between the coal seams; constructing the gas inrush concentration of the coal seam according to the mining depth of the coal seam; optimizing parameters of the exponential moving average algorithm; obtaining optimized average concentration data of the coal seam in combination with an index moving average algorithm; and the LSTM model is adopted to predict coal mine disasters. The method aims at optimizing actual measurement data in combination with the gas flow characteristics, more accurately capturing the sudden change moment of the gas concentration, and early warning potential safety risks in advance.

Description

Technical Field

[0001] The present application relates to the technical field of coal mine disaster risk prediction, and specifically to a coal mine disaster risk prediction method based on time series. Background Art

[0002] Coal mines refer to underground or open-pit mines where coal resources are extracted. The coal mining industry is a core component of the global energy supply chain and the world's largest natural energy source after oil. Its widespread use in power generation, metallurgy, and chemical industries has given the coal mining industry a significant global presence. However, coal mine disasters pose a significant challenge to the coal mining industry. These disasters not only cause property loss and environmental damage, but also lead to serious consequences such as casualties. Time series analysis methods play a key role in coal mine disaster prediction. By processing and predicting coal mine temporal data, the likelihood or risk of coal mine disasters can be determined. Different processing solutions are then applied to the predicted results to reduce the likelihood of disasters and the resulting loss of life and property.

[0003] Regarding the time series prediction of coal mine disaster risks, since coal mining is carried out in real time, a large amount of gas will be generated at the mining surface. Directly using the values ​​measured by the gas sensor will cause a large deviation in the prediction results, and the measured gas concentration data needs to be optimized. The exponential moving average algorithm EMA is an algorithm for optimizing data, but since the smoothing coefficient in the traditional exponential moving average algorithm is usually a fixed value selected by experience, the optimized results are inaccurate for more complex scenarios, so adaptive optimization weights are required. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a coal mine disaster risk prediction method based on time series to solve the existing problems.

[0005] The coal mine disaster risk prediction method based on time series of the present invention adopts the following technical solutions: One embodiment of the present invention provides a method for predicting coal mine disaster risks based on time series, the method comprising the following steps: Obtain coal mine data; The average gas concentration of each coal seam is obtained based on the coal mine data of each coal seam and the characteristics of gas release; the sequence composed of the average gas concentrations of all coal seams is used as the gas concentration sequence; the flow difference sequence of each coal seam is obtained based on the element distribution difference characteristics of the gas concentration sequence; the flow solid value sequence of each coal seam is obtained based on the positivity of the elements in the flow difference sequence of each coal seam; the flow trend value between each coal seam and other coal seams is obtained based on the distribution characteristics of the flow difference sequence and the flow solid value sequence between each coal seam and other coal seams and the difference characteristics of the average gas concentration; the gas influx concentration of each coal seam is obtained based on the comparison of the mining depth and average gas concentration between each coal seam and other coal seams and the flow trend value; the parameters of the exponential moving average algorithm are optimized according to the gas influx concentration of each coal seam; the average concentration data of each coal seam based on the time series is optimized according to the optimized weight of the measurement value of each coal seam and the exponential moving average algorithm; The LSTM model is used to predict coal mine disasters based on the optimized average gas concentration data of each coal seam.

[0006] Preferably, the coal mine data include but are not limited to: gas concentration of each coal seam, gas release coefficient, mining face gas emission, ground area, mining height, mining depth, coal drop rate and recovery rate.

[0007] Preferably, the average gas concentration of each coal seam is obtained based on the coal mine data of each coal seam and the characteristics of gas release, specifically: For the i-th coal seam, get the ground area of ​​the i-th coal seam , mining height , mining depth ; Unit emission of gas in the i-th coal seam The expression is:

[0008] Where, represents the gas emission volume of the i-th coal seam mining face; Indicates the coal falling rate; represents the recovery rate; represents the gas release degree coefficient of the i-th coal seam; The sum of the unit gas emission of the i-th coal seam and the gas concentration of the i-th coal seam is taken as the average gas concentration of the i-th coal seam.

[0009] Preferably, the flow difference sequence of each coal seam is obtained according to the element distribution difference characteristics of the gas concentration sequence, specifically: For the i-th coal seam, the i-th coal seam corresponds to the i-th element of the gas concentration sequence; For each element to the left of the i-th element in the gas concentration sequence, the difference between the value of each element to the left and the value of the element immediately following the value of each element to the left is used as the value of the element to the left of the i-th element in the flow difference sequence of the i-th coal seam; Set the value of the i-th element of the flow difference sequence of the i-th coal seam to 0; For the elements to the right of the i-th element in the gas concentration sequence, the difference between the values ​​of the elements to the right and the values ​​of the adjacent previous elements of the elements to the right is used as the value of the elements corresponding to the positions to the right of the i-th element in the flow difference sequence of the i-th coal seam.

[0010] Preferably, the flow solid value sequence of each coal seam is obtained according to the positivity of the elements of the flow difference value sequence of each coal seam, specifically: The element values ​​in the flow difference sequence of each coal seam that are greater than 0 are marked as 1, the element values ​​that are equal to 0 are marked as 0, and the element values ​​that are less than 0 are marked as -1. The sequence composed of the values ​​of all element marks in the flow difference sequence of each coal seam is saved as the flow solid value sequence of each coal seam.

[0011] Preferably, the flow trend value between each coal seam and other coal seams is obtained based on the distribution characteristics of the flow difference sequence and the flow solid value sequence between each coal seam and other coal seams in combination with the difference characteristics of the average gas concentration. The expression is:

[0012] Where, Indicates the flow trend value between the i-th and j-th coal seams; represents a symbolic function; 、 represent the average gas concentration of the i-th and j-th coal seams respectively; represents the value of the kth element in the flow solid value sequence of the i-th coal seam; It represents the kth element value of the flow difference sequence of the i-th coal seam; i and j represent the serial number values ​​of the coal seam.

[0013] Preferably, the method of obtaining the gas influx concentration of each coal seam based on the comparison of the mining depth and average gas concentration between each coal seam and other coal seams in combination with the flow trend value includes: For the i-th coal seam, calculate the difference in average gas concentration between other coal seams and the i-th coal seam; calculate the ratio of the mining depth of other coal seams to that of the i-th coal seam; calculate the product of the difference, ratio and flow trend value between other coal seams and the i-th coal seam; and take the sum of the products of all other coal seams and the i-th coal seam as the gas inrush concentration of the i-th coal seam.

[0014] Preferably, the optimization of the parameters of the exponential moving average algorithm according to the gas influx concentration of each coal seam includes: Obtain the weight value of the exponential moving average algorithm; use the opposite of the absolute value of the gas inrush concentration of each coal seam as the exponent of an exponential function with a natural constant as the base; calculate the difference between 1 and the calculation result of the exponential function; and use the product of the difference and the weight value as the optimized weight of the measurement value of each coal seam.

[0015] Preferably, the average concentration data of each coal seam based on the time series is optimized by combining the optimized weights of the measured values ​​of each coal seam with the exponential moving average algorithm, specifically as follows: The average gas concentration data of each coal seam based on time series is used as the input of the exponential moving average algorithm. Combined with the optimized weight of the measurement value of each coal seam, the output result of the exponential moving average algorithm is the real-time optimized average gas concentration data of each coal seam.

[0016] Preferably, the LSTM model is used to predict coal mine disasters based on the optimized average gas concentration data of each coal seam, specifically: The optimized average gas concentration data of each coal seam is used as the input of the LSTM model, and the output of the LSTM model is the prediction result of coal mine disasters, which is specifically normal, dangerous, and severely dangerous.

[0017] The present invention has at least the following beneficial effects: The present invention mainly obtains the gas outflow concentration per unit volume of the coal seam through the mining space and mining depth of the coal seam, the coal falling rate, and the coal mine gas release coefficient. First, the environmental characteristics are combined to make the measurement results more accurate; the gas flow trend value of the coal seam to other coal seams is obtained through the gas concentration of the coal seam, reflecting the flow direction of gas between the current coal seam and other coal seams; then the gas concentration of the coal seam and the gas flow trend value sequence of the coal seam are used to obtain the gas influx concentration of the coal seam, reflecting the severity of the change in the gas concentration of the coal seam; the optimization weight of the measurement value is obtained through the gas influx concentration. A higher severity requires a higher optimization weight, making the data more dependent on the predicted data. The exponential moving average algorithm is improved by the optimization weight of the measurement value. The improved exponential moving average algorithm not only takes into account the historical information of the gas concentration in the past period of time, but also combines the importance and real-time change trend of the current measurement value. This dynamic weight adjustment method can more accurately capture the moment of sudden change in gas concentration and warn of potential safety risks in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A flow chart of the coal mine disaster risk prediction method based on time series provided by the present invention; Figure 2 It is the time variation diagram of gas concentration; Figure 3 Flowchart for obtaining risk prediction. DETAILED DESCRIPTION

[0020] To further illustrate the technical means and effects employed by the present invention to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effects of the time series-based coal mine disaster risk prediction method proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable form.

[0021] Unless defined otherwise, 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 invention belongs.

[0022] The specific scheme of the coal mine disaster risk prediction method based on time series provided by the present invention is described in detail below with reference to the accompanying drawings.

[0023] An embodiment of the present invention provides a coal mine disaster risk prediction method based on time series.

[0024] Specifically, the following coal mine disaster risk prediction method based on time series is provided, please refer to Figure 1 , the method comprises the following steps: Step S001: Acquire coal mine data and pre-process the data.

[0025] This embodiment mainly considers the analysis and prediction of underground coal mining data. First, the ground area, mining height, and mining depth of each coal seam in the coal mine are obtained; an intrinsically safe deviation sensor GEJ35 is installed in the coal mine, and a sensor is installed every 15 meters to measure the gas concentration in the coal mine. Data is collected every 1 minute, and the total collection time is 2 hours. The gas concentration data of all sensors in each coal seam at the same time are averaged and saved as the gas concentration of each coal seam, such as Figure 2 As shown. Due to equipment or environmental factors, data may be missing. Third-order polynomial interpolation is used to supplement missing data. This third-order polynomial interpolation method is well known in the art and will not be described in detail in this embodiment. Professional gas detection equipment is also used to obtain the gas release coefficient of each coal seam during the coal fall process and the gas emission rate of each mining face.

[0026] In addition, the coal falling rate of the shearer per unit time (per minute) is measured, and the recovery rate data is recorded within the same period.

[0027] At this point, the coal mine data is obtained.

[0028] Step S002: Obtain the gas outburst concentration per unit volume of the coal seam through the mining space and mining depth of the coal seam, the coal falling rate, and the coal mine gas release coefficient; obtain the gas flow trend value of the coal seam to other coal seams through the gas concentration of the coal seam; then obtain the gas influx concentration of the coal seam using the gas concentration of the coal seam and the gas flow trend value of the coal seam; obtain the optimized weight of the measurement value through the gas influx concentration, and optimize the gas concentration of the coal seam in combination with the exponential moving average algorithm.

[0029] First, we analyze the gas emission rates of different coal seams. As the gas content increases with the depth of each layer, the gas emission rate also increases with the depth of the coal seam. Therefore, the gas release rate of each coal seam is different. If the current coal seam mining depth increases and the mining intensity increases, a higher weight is assigned to the current coal seam to ensure that the real-time monitoring data can promptly and effectively reflect abnormal gas changes. The unit gas emission rate of each coal seam is calculated using the data of each coal seam. The expression is:

[0030] Where, represents the unit emission of gas in the i-th coal seam; represents the gas emission volume of the i-th coal seam mining face; Indicates the coal falling rate; represents the recovery rate; represents the ground area of ​​the ith coal seam; represents the gas release degree coefficient of the i-th coal seam; represents the mining height of the i-th coal seam; Represents the mining depth of the i-th coal seam.

[0031] When the mining depth of the i-th coal seam The deeper the coal mine, the greater the amount of gas contained in a unit volume, and the amount of gas released during mining. The larger the value, the greater the unit outburst concentration component between the mining surface and the entire coal seam. The value increases; when the coal dropping rate is higher and the recovery rate is lower, the coal dropping amount increases, which increases the concentration released per unit volume of the coal seam. The deeper the coal seam, the higher the gas release coefficient of coal dropping. The larger the value, the greater the unit gas release in the goaf. ,Right now The value of increases.

[0032] The unit gas emission of each coal seam is added to the gas concentration of the coal seam to obtain the average gas concentration of each coal seam in a specific period of time. , save the sequence of average gas concentration of all coal seams as gas concentration sequence .

[0033] Gas fluid dynamics indicates that gas molecules in high-concentration areas will spontaneously diffuse toward lower-concentration areas until equilibrium is reached. By comparing the gas concentration distribution between adjacent coal seams, the direction of gas diffusion from high-concentration coal seams to low-concentration coal seams can be inferred. This diffusion direction can then be used to calculate the flow trend within the current coal seam.

[0034] Specifically, taking the i-th coal seam as an example, the i-th coal seam corresponds to the i-th element of the gas concentration sequence. For the elements to the left of the i-th element of the gas concentration sequence, the difference between the value of each element to the left and the value of the next element is used as the value of each element to the left of the i-th element of the flow difference sequence of the i-th coal seam; 0 is used as the value of the i-th element of the flow difference sequence of the i-th coal seam; for the elements to the right of the i-th element of the gas concentration sequence, the difference between the value of each element to the right and the value of the previous element is used as the value of each element to the right of the i-th element of the flow difference sequence of the i-th coal seam. For example, the gas concentration sequence is , for the element , and its flow difference sequence is ; For elements , and its flow difference sequence is . Based on the flow difference sequence of the i-th coal seam, its flow solid value sequence is obtained: the element value in the flow difference sequence of the i-th coal seam greater than 0 is marked as 1, the element value equal to 0 is marked as 0, and the element value less than 0 is marked as -1, and the sequence composed of all marked values ​​of the flow difference sequence of the i-th coal seam is saved as the flow solid value sequence of the i-th coal seam.

[0035] Furthermore, the flow trend value between the i-th coal seam and other coal seams is calculated, and the expression is:

[0036] Where, Indicates the flow trend value between the i-th and j-th coal seams; represents a symbolic function; 、 represent the average gas concentration of the i-th and j-th coal seams respectively; represents the value of the kth element in the flow solid value sequence of the i-th coal seam; It represents the kth element value of the flow difference sequence of the i-th coal seam; i and j represent the serial number values ​​of the coal seam.

[0037] When there is a difference in gas concentration between the i-th and j-th coal seams, the gas in the coal seam with higher gas concentration will eventually flow to the coal seam with lower gas concentration. The sign function determines the flow direction of the gas by calculating the flow trend value between the i-th and j-th coal seams, and judges the tendency of the gas flow direction to the i-th coal seam; for all the intermediate coal seams between the i-th and j-th coal seams, the more the number of intermediate coal seams is and the higher the difference in gas concentration between the intermediate coal seams is, the larger the value of the flow trend value between the i-th and j-th layers will be.

[0038] Calculate the flow trend value of the i-th coal seam and other coal seams. Calculate the gas influx concentration of the i-th coal seam by the difference in gas concentration between the i-th coal seam and other coal seams and the gas flow trend value. The expression is:

[0039] Where, represents the gas influx concentration of the i-th coal seam; Indicates the number of coal seams in a coal mine; 、 represent the average gas concentration of the i-th and j-th coal seams respectively; 、 Respectively represent the mining depth of the i-th and j-th coal seams; Represents the flow trend value between the i-th and j-th coal seams.

[0040] If the gas concentration of the i-th coal seam is low, the gas concentration of other coal seams is high, making the gas concentration difference between other coal seams and the i-th coal seam larger, the gas from other coal seams will flow to the i-th coal seam, and the flow trend sequence value between the i-th coal seam and other coal seams will be larger, making the gas influx concentration of the i-th coal seam increase. The higher the other coal seams are, the less likely it is that the gas from other coal seams will flow into the i-th coal seam, that is, the possibility The smaller the value is, that is, the lower the current coal seam is, the smaller the gas influx concentration in the coal seam is.

[0041] In coal mine gas monitoring, to improve prediction accuracy and real-time responsiveness, the traditional exponential moving average algorithm can be improved by introducing a dynamic weight optimization mechanism. This allows it to adapt to changing conditions and effectively reflect the weight of the current coal seam's true gas conditions. The improved exponential moving average algorithm not only considers historical information on gas concentrations over a period of time, but also incorporates the importance of the current measurement value and its real-time changing trends. The optimized weight of the measurement data is calculated using the expression:

[0042] Where, represents the optimization weight of the measurement value of the i-th coal seam; represents an exponential function with the natural constant e as the base; represents the gas influx concentration of the i-th coal seam; It should be noted that the weight value of the original algorithm is 0.9 in this embodiment.

[0043] When the gas influx concentration in the coal seam is greater, it means that the gas concentration in the coal seam changes more dramatically. When optimizing the gas concentration data, we should rely more on the predicted value, and the optimization weight value will be larger. On the contrary, when the difference in gas concentration between layers is small, it means that the gas does not have much fluidity, the gas sensor detection data is more accurate, the optimized data value should tend to the measured value, and the optimization weight value will be smaller.

[0044] By improving the exponential moving average algorithm using optimized weights for measured values, the time-series average gas concentration data for each coal seam is used as input to the improved exponential moving average algorithm. The output of the improved exponential moving average algorithm is the real-time optimized average gas concentration data for the current coal seam. The exponential moving average algorithm is well known in the art and will not be described in detail here.

[0045] Step S003: Obtain a large amount of optimized gas data, use the LSTM network for training, use the real-time gas concentration as the input of the trained model, and complete the risk prediction of coal mine disasters.

[0046] Obtain a large amount of optimized gas concentration data, and manually label the gas concentration data as three labels: normal, dangerous, and extremely dangerous. Then use the labeled gas concentration data as the training data of the LSTM model, use cross entropy loss as the loss function of the model, and Adam as the optimizer of the model to obtain the model training model through training. Then obtain the optimized coal seam gas data in real time as input and input it into the trained LSTM model. If the output result is normal, it indicates that there is no disaster risk in the coal mine. If the output is dangerous, it indicates that there may be a coal mine disaster risk, and the miners in the coal mine need to leave the mine immediately; if it is extremely dangerous, the coal mine is very likely to have a coal mine disaster risk, and the mine needs to immediately activate the disaster emergency plan to face possible coal mine disasters. Among them, the LSTM model is an existing well-known technology and will not be repeated in this embodiment. The process of obtaining risk prediction is as follows: Figure 3 shown.

[0047] At this point, the coal mine disaster risk forecast is completed.

[0048] In summary, the embodiment of the present invention mainly obtains the gas outburst concentration per unit volume of the coal seam through the mining space and mining depth of the coal seam, the coal falling rate, and the coal mine gas release coefficient. First, the environmental characteristics are combined to make the measurement results more accurate; the gas flow trend value of the coal seam to other coal seams is obtained through the gas concentration of the coal seam, reflecting the flow direction of the gas between the current coal seam and other coal seams; then the gas concentration of the coal seam and the gas flow trend value sequence of the coal seam are used to obtain the gas influx concentration of the coal seam, reflecting the severity of the change in the gas concentration of the coal seam; the optimization weight of the measurement value is obtained through the gas influx concentration. A higher severity requires a higher optimization weight, making the data more dependent on the predicted data. The exponential moving average algorithm is improved by the optimization weight of the measurement value. The improved exponential moving average algorithm not only takes into account the historical information of the gas concentration in the past period of time, but also combines the importance and real-time change trend of the current measurement value. This dynamic weight adjustment method can more accurately capture the moment of sudden change in gas concentration and warn of potential safety risks in advance.

[0049] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0050] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0051] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them. Modifications to the technical solutions described in the aforementioned embodiments, or equivalent replacements of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A coal mine disaster risk prediction method based on time series, characterized by: The method comprises the following steps: Obtain coal mine data; The average gas concentration of each coal seam is obtained based on the coal mine data of each coal seam and the characteristics of gas release; the sequence composed of the average gas concentrations of all coal seams is used as the gas concentration sequence; the flow difference sequence of each coal seam is obtained based on the element distribution difference characteristics of the gas concentration sequence; the flow solid value sequence of each coal seam is obtained based on the positivity of the elements in the flow difference sequence of each coal seam; the flow trend value between each coal seam and other coal seams is obtained based on the distribution characteristics of the flow difference sequence and the flow solid value sequence between each coal seam and other coal seams and the difference characteristics of the average gas concentration; the gas influx concentration of each coal seam is obtained based on the comparison of the mining depth and average gas concentration between each coal seam and other coal seams and the flow trend value; the parameters of the exponential moving average algorithm are optimized according to the gas influx concentration of each coal seam; the average concentration data of each coal seam based on the time series is optimized according to the optimized weight of the measurement value of each coal seam and the exponential moving average algorithm; The LSTM model is used to predict coal mine disasters based on the optimized average gas concentration data of each coal seam.

2. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The coal mine data include but are not limited to: gas concentration of each coal seam, gas release coefficient, mining face gas emission, ground area, mining height, mining depth, coal drop rate and recovery rate.

3. The method for predicting coal mine disaster risk based on time series according to claim 2, characterized in that: The average gas concentration of each coal seam is obtained based on the coal mine data of each coal seam and the characteristics of gas release, specifically: For the i-th coal seam, get the ground area of ​​the i-th coal seam Mining height Mining depth ; Unit emission of gas in the i-th coal seam The expression is: Where, represents the gas emission volume of the i-th coal seam mining face; Indicates the coal falling rate; represents the recovery rate; represents the gas release degree coefficient of the i-th coal seam; The sum of the unit gas emission of the i-th coal seam and the gas concentration of the i-th coal seam is taken as the average gas concentration of the i-th coal seam.

4. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The flow difference sequence of each coal seam is obtained based on the element distribution difference characteristics of the gas concentration sequence, specifically: For the i-th coal seam, the i-th coal seam corresponds to the i-th element of the gas concentration sequence; For each element to the left of the i-th element in the gas concentration sequence, the difference between the value of each element to the left and the value of the element immediately following the value of each element to the left is used as the value of the element to the left of the i-th element in the flow difference sequence of the i-th coal seam; Set the value of the i-th element of the flow difference sequence of the i-th coal seam to 0; For the elements to the right of the i-th element in the gas concentration sequence, the difference between the values ​​of the elements to the right and the values ​​of the adjacent previous elements of the elements to the right is used as the value of the elements corresponding to the positions to the right of the i-th element in the flow difference sequence of the i-th coal seam.

5. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The flow solid value sequence of each coal seam is obtained according to the positive and negative elements of the flow difference sequence of each coal seam, specifically: The element values ​​in the flow difference sequence of each coal seam that are greater than 0 are marked as 1, the element values ​​that are equal to 0 are marked as 0, and the element values ​​that are less than 0 are marked as -1. The sequence composed of the values ​​of all element marks in the flow difference sequence of each coal seam is saved as the flow solid value sequence of each coal seam.

6. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The flow trend value between each coal seam and other coal seams is obtained based on the distribution characteristics of the flow difference sequence and the flow solid value sequence between each coal seam and other coal seams, combined with the difference characteristics of the average gas concentration. The expression is: Where, Indicates the flow trend value between the i-th and j-th coal seams; represents a symbolic function; 、 represent the average gas concentration of the i-th and j-th coal seams respectively; represents the value of the kth element in the flow solid value sequence of the i-th coal seam; It represents the kth element value of the flow difference sequence of the i-th coal seam; i and j represent the serial number values ​​of the coal seam.

7. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The method of obtaining the gas influx concentration of each coal seam based on the comparison of the mining depth and average gas concentration of each coal seam with other coal seams in combination with the flow trend value includes: For the i-th coal seam, calculate the difference in average gas concentration between other coal seams and the i-th coal seam; calculate the ratio of the mining depth of other coal seams to that of the i-th coal seam; calculate the product of the difference, ratio and flow trend value between other coal seams and the i-th coal seam; and take the sum of the products of all other coal seams and the i-th coal seam as the gas inrush concentration of the i-th coal seam.

8. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The optimization of the parameters of the exponential moving average algorithm according to the gas influx concentration of each coal seam includes: Obtain the weight value of the exponential moving average algorithm; use the opposite of the absolute value of the gas inrush concentration of each coal seam as the exponent of an exponential function with a natural constant as the base; calculate the difference between 1 and the calculation result of the exponential function; and use the product of the difference and the weight value as the optimized weight of the measurement value of each coal seam.

9. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The optimization weights of the measured values ​​of each coal seam are combined with the exponential moving average algorithm to optimize the average concentration data of each coal seam based on the time series, specifically: The average gas concentration data of each coal seam based on time series is used as the input of the exponential moving average algorithm. Combined with the optimized weight of the measurement value of each coal seam, the output result of the exponential moving average algorithm is the real-time optimized average gas concentration data of each coal seam.

10. The method for predicting coal mine disaster risk based on time series according to claim 1, characterized in that: The LSTM model is used to predict coal mine disasters based on the optimized average gas concentration data of each coal seam, specifically: The optimized average gas concentration data of each coal seam is used as the input of the LSTM model, and the output of the LSTM model is the prediction result of coal mine disasters, which is specifically normal, dangerous, and severely dangerous.

Images