Coal mine roof pressure prediction method and system

Abstract

The invention provides a coal mine roof pressure prediction method and system, and belongs to the technical field of coal mine roof pressure prediction. According to the technical scheme, the coal mine roof pressure prediction method and system are provided, and the method comprises the steps that coal mine pressure real-time data are stored in a database, filtering processing is conducted on the stored data through a denoising model, and roadway historical mine pressure data used for analysis are obtained; based on the denoised historical mine pressure data, establishing a nonlinear regression model with a denoising function; generating a prediction report based on the established nonlinear regression model; the generated prediction report is compared with a preset alarm threshold value, and if the prediction report exceeds the corresponding threshold value, alarm reminding is carried out; the method is applied to coal mine roof pressure prediction.

Description

Technical Field

[0001] The invention provides a method and system for predicting coal mine roof pressure, belonging to the technical field of coal mine roof pressure prediction. Background Art

[0002] With the continuous improvement of coal mining equipment manufacturing and mining technology, in coal mine safety production, improving the automation and intelligence level of mining equipment and gradually realizing intelligent mining has become the development trend of coal comprehensive mining technology; roof falls accidents occurring during the mining process are one of the main causes of casualties, which not only threatens the lives of miners, but also causes repeated shutdowns of the comprehensive mining face, causing huge economic losses.

[0003] Whether the pressure of the coal mine roof is stable will directly affect the occurrence of roof fall accidents. However, it is currently impossible to predict the roof pressure in advance and to warn of roof fall accidents in advance, which seriously restricts the intelligent mining of the fully mechanized mining face. Although some fully mechanized mining faces are currently connected to the mine pressure monitoring system, and the massive monitoring data collected by setting up pressure sensors provides data support for the early warning of roof fall accidents on the fully mechanized mining face, it is still impossible to accurately predict the roof pressure. In addition, the goaf area generated during the advancement of the coal mining face and the increase in the exposed span of the old roof lead to periodic changes in roof pressure. If the support strength of the working face is insufficient, roof fall accidents may occur. At present, coal mines monitor the roof pressure through the mine pressure monitoring system, but they cannot make timely and accurate predictions of the roof pressure.

[0004] Therefore, researching and developing a solution that can accurately predict coal mine roof pressure is of great significance for reducing roof fall accidents, protecting the lives of miners, and improving coal mine production efficiency. Summary of the invention

[0005] In order to overcome the deficiencies in the prior art, the present invention aims to solve the technical problem of providing a method and system for predicting coal mine roof pressure.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for predicting coal mine roof pressure, comprising the following prediction steps:

[0007] Step 1: Store the real-time data of coal mine pressure into the database, filter the stored data through the denoising model, and obtain the historical mine pressure data of the tunnel for analysis;

[0008] Step 2: Based on the denoised historical mine pressure data, a nonlinear regression model with denoising function is established;

[0009] Step 3: Generate a forecast report based on the established nonlinear regression model;

[0010] Step 4: Based on the generated forecast report, compare it with the preset alarm threshold. If it exceeds the corresponding threshold, an alarm reminder will be issued.

[0011] The specific method for performing filtering processing in step 1 is:

[0012] Define a set of measured data as: t1, t2, t3, t4, …, t i ,…,t n and u1,u2,u3,u4,…,u i ,…,u n ;

[0013] Among them, t i For each time point, u i is the detection value corresponding to each time point, n is the number of monitoring values, and this set of data is used for denoising filtering;

[0014] Use the time point t of the measured data i and the corresponding detection value u i , use the function u = f(t) to represent the relationship of the measured data, use the least squares method to determine a fitting straight line g = at + b, and use this fitting straight line to calculate the i-th point t = t i g i =at i +b, as g after linear sliding i The specific calculation steps are as follows:

[0015] Using the three-point linear sliding model, take the detection data of the i-th point and the two points before and after it: (t i-1 ,u i-1 )、(t i ,u i )、(t i+1 ,u i+1 ) to perform sliding analysis, changing the independent variable t, and let Get i-1 , t i , t i+1 The corresponding z i-1 =1, z i =0, z i+1 =1, then Δz = z i -z i-1 =1, so (z i-1 ,u i-1 )、(z i ,u i )、(z i+1 ,u i+1 ) is fitted with a straight line g = a1z + b1;

[0016] Establish the error equation V i=u i -g i (i=1,2,…,n), according to the least squares principle, Q=V T V=min, which satisfies:

[0017]

[0018] Since a1 and b1 are parameters, and u and z are known monitoring values, the extreme values ​​are derived by respectively calculating the partial values ​​of a1 and b1 based on the above formula. The calculation formula is:

[0019]

[0020] Let the partial derivative in the above formula be equal to 0, and change z i-1 =1, z i =0, z i+1 =1 into the above formula, calculate the values ​​of a1 and b1, and substitute the values ​​of a1 and b1 into g, the expression is:

[0021]

[0022] When i=1 or n, the two adjacent data are used to calculate and determine the value of g. The calculation formula is:

[0023]

[0024] Use the above formula to remove noise from the direct monitoring value and get the data g after removing noise. i ,i∈(1,n), and apply g i Do relevant forecasting work.

[0025] The specific method for establishing the nonlinear regression model in step 2 is:

[0026] The exponential function is used to establish the model, and the expression is:

[0027]

[0028] In the formula, u is pressure, a and b are regression coefficients, and t is measurement time;

[0029] When using the exponential function model, the above formula needs to be transformed. First, remove the logarithm on both sides of the above formula, and the expression is:

[0030]

[0031] Then let v = ln u, w = 1 / t, c = ln a, then the two sets of data are transformed into: w1, w2, w3, w4, ..., w i ,…,w nand v1,v2,v3,v4,…,v i ,…,v n , so that the above formula is transformed into:

[0032] v = c + b·w;

[0033] In this way, the exponential function model is converted into a linear regression model. The two sets of transformed data can be calculated by calculating the linear function. The calculation formula is:

[0034]

[0035] The values ​​obtained by the above formula are calculated respectively. Substitute the values ​​of b and c into the formula v = c + b·w to obtain a linear equation;

[0036] Considering that the sublinear equation is obtained through transformation, after inverse transformation, the calculation formula is:

[0037] a=e c ;

[0038] Substitute the values ​​of a and b into the formula Thus, the corresponding nonlinear regression model equation is obtained.

[0039] The specific method for generating the forecast report in step 3 is:

[0040] According to the nonlinear regression model equation obtained in step 2 Let t in the equation be equal to the difference between the date to be predicted and the current date, calculate the predicted value u for each date to be predicted, and draw a prediction curve.

[0041] The specific method of performing the alarm reminder in step 4 is:

[0042] Based on the data summarized in actual production, the coal mine specifies the roof pressure alarm threshold as θ. If the predicted value u>θ in the prediction report, the main console unit will remind relevant personnel to strengthen observation and monitoring of the alarm location through voice broadcast.

[0043] The prediction system used to realize the coal mine roof pressure prediction method includes a data acquisition unit, an interface gateway unit, and a main control console unit;

[0044] The data acquisition unit calculates the mine pressure data of each sensor in real time by accessing multiple mine pressure sensors, and simultaneously transmits the acquired mine pressure data to the main control console unit through the interface gateway unit;

[0045] The interface gateway unit connects to the coal mine ring network to realize the connection between the system and the ring network, and merges and summarizes the data collected by the data acquisition unit and sends it to the main control console unit;

[0046] The main control console unit filters, stores, and analyzes the data based on the data sent back by the acquisition unit, and generates real-time monitoring reports and prediction reports for mine pressure. The real-time monitoring reports can be used to monitor the mine pressure conditions in the corresponding tunnels in real time, and the prediction reports can be used to predict abnormal mine pressure conditions in the corresponding tunnels in advance, so that corresponding disposal measures can be arranged early.

[0047] A mine pressure sensor is also installed in the monitored tunnel, and the mine pressure sensor is connected to the data acquisition unit through a wire.

[0048] The beneficial effects of the present invention compared with the prior art are as follows: the present invention provides a solution for predicting the roof pressure of a coal mine, mainly by denoising and filtering massive mine pressure data to improve the quality of collected data, by selecting a reasonable prediction model to improve the prediction accuracy, and based on the fitting to obtain the mine pressure prediction curve, the prediction result can be displayed intuitively; the prediction method adopted by the present invention can integrate the data of the entire mine, combined with the specified alarm threshold, and automatically monitor. If the prediction exceeds the threshold, the management personnel will be notified in time for early processing, so as to solve the roof safety problem in the bud stage and ensure the safe production of the coal mine. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention will be further described below in conjunction with the accompanying drawings:

[0050] Figure 1 It is a structural schematic diagram of the coal mine roof pressure prediction system of the present invention;

[0051] Figure 2 It is a flow chart of the steps of the method for predicting coal mine roof pressure of the present invention;

[0052] Figure 3 This is a schematic diagram of a report for real-time monitoring of mine pressure in an embodiment of the present invention;

[0053] Figure 4 It is a schematic diagram of a report for predicting mine pressure in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] The present invention provides a coal mine roof pressure prediction scheme, which solves the problems of untimely early warning and low prediction accuracy of the existing mine pressure monitoring system; the prediction system adopted includes a data acquisition unit, an interface gateway unit, and a main control console unit. The received sensor data is converted and transmitted to the main control console unit through the data acquisition unit and the interface gateway unit. The main control console unit establishes a prediction model by denoising and filtering the data, predicts future roof pressure data through the model, compares it with the alarm threshold, and timely warns of possible dangers, thereby guiding the roof management of the coal mine;

[0055] like Figure 1As shown, the present invention provides a coal mine roof pressure prediction system, including a data acquisition unit, an interface gateway unit, and a main control console unit;

[0056] The data acquisition unit calculates the mine pressure data of each sensor in real time by accessing multiple mine pressure sensors, and simultaneously transmits the acquired mine pressure data to the main control console unit through the interface gateway unit;

[0057] The interface gateway unit connects to the coal mine ring network to realize the connection between the system and the ring network, and merges and summarizes the data collected by the data acquisition unit and sends it to the main control console unit;

[0058] The main control console unit filters, stores and analyzes the data according to the data sent back by the acquisition unit, and generates a real-time monitoring report and a prediction report of the mine pressure. The real-time monitoring report can monitor the mine pressure of the corresponding tunnel in real time, and the prediction report can predict the abnormal mine pressure of the corresponding tunnel in advance, so as to arrange corresponding disposal measures in advance.

[0059] The monitored tunnel is equipped with a mine pressure sensor, which is connected to a data acquisition unit.

[0060] like Figure 2 As shown, the prediction method adopted by the present invention is based on massive monitoring data and uses an effective prediction algorithm to monitor and accurately predict the roof pressure in real time, solve the technical problems of excessive fluctuations and low quality of real-time monitoring data, low accuracy of roof pressure prediction, and inability to effectively specify roof management, so as to achieve management of coal mine roof stability. The specific prediction steps are:

[0061] Step 1: Store the real-time data of mine pressure into the database, filter the stored data through the denoising model, and obtain the historical mine pressure data of the tunnel for analysis;

[0062] Step 2: Based on the denoised historical mine pressure data, a nonlinear regression model with denoising function is established;

[0063] Step 3: Generate a forecast report based on the established nonlinear regression model;

[0064] Step 4: Based on the generated forecast report, compare it with the preset alarm threshold. If it exceeds the corresponding threshold, an alarm reminder will be issued.

[0065] The steps of performing denoising and filtering on the unprocessed historical mine pressure data in step 1 are as follows:

[0066] Assume there is a set of measured data: t1, t2, t3, t4, …, t i ,…,t n and u1,u2,u3,u4,…,u i ,…,un , t i For each time point, u i is the detection value corresponding to each time point, n is the number of monitoring values, and this set of data is used for denoising filtering;

[0067] Using the measured data t mentioned above i and u i , use the function u = f(t) to represent the relationship of the measured data, determine a fitting straight line g = at + b using the principle of least squares, and use this fitting straight line to calculate the i-th point t = t i g i =at i +b as g after linear sliding i The specific calculation steps are as follows:

[0068] Using the three-point linear sliding model, take the detection data of the i-th point and the two points before and after it (t i-1 ,u i-1 )、(t i ,u i )、(t i+1 ,u i+1 ) to do sliding analysis, here we change the independent variable t, let Get i-1 , t i , t i+1 The corresponding z i-1 =1, z i =0, z i+1 =1, then Δz = z i -z i-1 =1, so (z i-1 ,u i-1 )、(z i ,u i )、(z i+1 ,u i+1 ) is fitted with a straight line g=a1z+b1.

[0069] Establish the equation V i =u i -g i (i=1,2,…,n), according to the least squares principle, Q=V T V=min, which satisfies:

[0070]

[0071] Since a1 and b1 are parameters, and u and z are known monitoring values, we can use formula (1) to derive the extreme values ​​of a1 and b1, namely:

[0072]

[0073] Let the partial derivative in equation (2) be equal to 0, and z i-1 =1, z i =0, z i+1 =1 is substituted into formula (2) to calculate the values ​​of a1 and b1, and then the values ​​of a1 and b1 are substituted into g, as shown in formula (3):

[0074]

[0075] When i=1 or n, the value of g can be determined by using two adjacent data, as shown in the following formula (4):

[0076]

[0077] In this way, using formula (4), the direct monitoring value can be denoised to obtain the data g after denoising. i ,i∈(1,n), and apply g i Do relevant forecasting work.

[0078] The steps of establishing the nonlinear regression model in step 2 are as follows:

[0079] In actual tunnel engineering, the distribution relationship between the measured pressure and time is generally complex, and there is a certain nonlinear correlation between them. The analysis of such problems generally uses a nonlinear regression model to process and analyze them. The exponential function in engineering analysis is used to establish the model. The exponential function formula is shown in the following formula (5):

[0080]

[0081] Where u is pressure (KPa), a and b are regression coefficients, and t is measurement time (days).

[0082] When using the exponential function model, it is necessary to transform equation (5). First, take the logarithm of both sides of equation (5), and we get:

[0083]

[0084] Then let v = ln u, w = 1 / t, c = ln a, then the two sets of data are transformed into: w1, w2, w3, w4, ..., w i ,…,w n and v1,v2,v3,v4,…,v i ,…,v n , transform formula (6) into:

[0085] v = c + b·w (7);

[0086] In this way, the exponential function model is converted into a linear regression model. The two sets of transformed data can be calculated by calculating the linear function. The values ​​that need to be calculated are:

[0087]

[0088] The value obtained by formula (8) is used to calculate Substituting the obtained values ​​of b and c into equation (7), we get a linear equation.

[0089] Considering that the sublinear equation is obtained through transformation, after inverse transformation, calculate: a = e c , substituting the values ​​of a and b into equation (5), we can get the corresponding nonlinear regression model equation.

[0090] like Figure 3 and Figure 4 As shown, the steps for generating a forecast report in step three are as follows:

[0091] According to the nonlinear regression model equation (5) obtained in step 2, let t in the equation be equal to the difference between the date to be predicted and the current date, calculate the predicted value u of each date to be predicted, and draw a prediction curve.

[0092] The steps for triggering the alarm in step 4 are as follows:

[0093] The coal mine specifies the roof pressure alarm threshold θ based on the data summarized in actual production. If the predicted value u>θ in the prediction report, the main control console unit will remind relevant personnel to strengthen observation and monitoring of the alarm location through voice broadcast.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements 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 invention.

Claims

1. A method for predicting coal mine roof pressure, characterized in that: The prediction steps include the following: Step 1: Store the real-time data of coal mine pressure into the database, filter the stored data through the denoising model, and obtain the historical mine pressure data of the tunnel for analysis; Step 2: Based on the denoised historical mine pressure data, a nonlinear regression model with denoising function is established; Step 3: Generate a forecast report based on the established nonlinear regression model; Step 4: Based on the generated forecast report, compare it with the preset alarm threshold. If it exceeds the corresponding threshold, an alarm reminder will be issued.

2. A method for predicting coal mine roof pressure according to claim 1, characterized in that: The specific method for performing filtering processing in step 1 is: Define a set of measured data as: t1, t2, t3, t4, …, t i ,…,t n and u1,u2,u3,u4,…,u i ,…,u n ; Among them, t i For each time point, u i is the detection value corresponding to each time point, n is the number of monitoring values, and this set of data is used for denoising filtering; Use the measured data at time t i and the corresponding detection value u i , use the function u = f(t) to represent the relationship of the measured data, use the least squares method to determine a fitting straight line g = at + b, and use this fitting straight line to calculate the i-th point t = t i g i =at i +b, as g after linear sliding i The specific calculation steps are as follows: Using the three-point linear sliding model, take the detection data of the i-th point and the two points before and after it: (t i-1 ,u i-1 )、(t i ,u i )、(t i+1 ,u i+1 ) to perform sliding analysis, changing the independent variable t, and let Get i-1 , t i , t i+1 The corresponding z i-1 =1, z i =0, z i+1 =1, then Δz = z i -z i-1 =1, so (z i-1 ,u i-1 )、(z i ,u i )、(z i+1 ,u i+1 ) is fitted with a straight line g = a1z + b1; Establish the error equation V i =u i -g i (i=1,2,…,n), according to the principle of least squares, Q=V T V=min, which satisfies: Since a1 and b1 are parameters, and u and z are known monitoring values, the extreme values ​​are derived by respectively calculating the partial values ​​of a1 and b1 based on the above formula. The calculation formula is: Let the partial derivative in the above formula be equal to 0, and z i-1 =1, z i =0, z i+1 =1 into the above formula, calculate the values ​​of a1 and b1, and substitute the values ​​of a1 and b1 into g, the expression is: When i=1 or n, the two adjacent data are used to calculate and determine the value of g. The calculation formula is: Use the above formula to remove noise from the direct monitoring value and get the data g after removing noise. i ,i∈(1,n), and apply g i Do relevant forecasting work.

3. A method for predicting coal mine roof pressure according to claim 2, characterized in that: The specific method for establishing the nonlinear regression model in step 2 is: The exponential function is used to establish the model, and the expression is: In the formula, u is pressure, a and b are regression coefficients, and t is measurement time; When using the exponential function model, the above formula needs to be transformed. First, remove the logarithm on both sides of the above formula, and the expression is: Then let v = ln u, w = 1 / t, c = ln a, then the two sets of data are transformed into: w1, w2, w3, w4, ..., w i ,…,w n and v1,v2,v3,v4,…,v i ,…,v n , so that the above formula is transformed into: v = c + b·w; In this way, the exponential function model is converted into a linear regression model. The two sets of transformed data can be calculated by calculating the linear function. The calculation formula is: The values ​​obtained by the above formula are calculated respectively. Substitute the values ​​of b and c into the formula v = c + b·w to obtain a linear equation; Considering that the sublinear equation is obtained through transformation, after inverse transformation, the calculation formula is: a=e c ; Substitute the values ​​of a and b into the formula Thus, the corresponding nonlinear regression model equation is obtained.

4. A method for predicting coal mine roof pressure according to claim 3, characterized in that: The specific method for generating the forecast report in step 3 is: According to the nonlinear regression model equation obtained in step 2 Let t in the equation be equal to the difference between the date to be predicted and the current date, calculate the predicted value u for each date to be predicted, and draw a prediction curve.

5. A method for predicting coal mine roof pressure according to claim 4, characterized in that: The specific method of performing the alarm reminder in step 4 is: Based on the data summarized in actual production, the coal mine specifies the roof pressure alarm threshold as θ. If the predicted value u>θ in the prediction report, the main console unit will remind relevant personnel to strengthen observation and monitoring of the alarm location through voice broadcast.

6. A prediction system for realizing the method for predicting coal mine roof pressure according to claim 1, characterized in that: It includes data acquisition unit, interface gateway unit and main control console unit; The data acquisition unit calculates the mine pressure data of each sensor in real time by accessing multiple mine pressure sensors, and simultaneously transmits the acquired mine pressure data to the main control console unit through the interface gateway unit; The interface gateway unit connects to the coal mine ring network to realize the connection between the system and the ring network, and merges and summarizes the data collected by the data acquisition unit and sends it to the main control console unit; The main control console unit filters, stores, and analyzes the data based on the data sent back by the acquisition unit, and generates real-time monitoring reports and prediction reports for mine pressure. The real-time monitoring reports can be used to monitor the mine pressure conditions in the corresponding tunnels in real time, and the prediction reports can be used to predict abnormal mine pressure conditions in the corresponding tunnels in advance, so that corresponding disposal measures can be arranged early.

7. A coal mine roof pressure prediction system according to claim 6, characterized in that: A mine pressure sensor is also installed in the monitored tunnel, and the mine pressure sensor is connected to the data acquisition unit through a wire.

Images