CN106845447A - A kind of face gas concentration prediction method for early warning - Google Patents

Abstract

The invention discloses a method for predicting and early warning of gas concentration in a working face, which is characterized in that, in order to predict the concentration of gas monitoring and early warning in a coal mine working face, the gas monitoring data is firstly subjected to preprocessing such as replacing abnormal data, supplementing missing data, and denoising. Realized the perfection of the original input data of gas concentration; then used the Lyapunov exponent to identify the chaos and reconstruct the phase space of the original gas emission time series, so as to establish the chaotic prediction mathematical model of gas emission; according to the gas concentration prediction The results and prediction intervals are used to calculate the warning threshold, and the corresponding warning levels are divided to complete the warning information; it can be used for gas concentration prediction and early warning work in the working face.

Description

A Method for Prediction and Early Warning of Gas Concentration in Working Face

technical field

The invention relates to mining safety, in particular to gas concentration prediction and early warning work on working faces.

Background technique

With the development of technology, the alarm system equipped with real-time monitoring has been realized in most coal mines in my country. However, the current monitoring and monitoring system is mainly limited to the local monitoring and management of the underground, and lacks effective processing and specific analysis of the monitoring data for the purpose of realizing the prediction and early warning function. According to the actual situation and needs of the mine, by using the law of mine gas monitoring data as an analysis method, and specifically analyzing the abnormal conditions in the process of gas gushing, the technical research on the prediction and early warning analysis of mine gas concentration is carried out, and a true and effective prediction is given. As a result, realizing the real-time prediction of gas early warning, determining the level of early warning and releasing the early warning information, so that the gas early warning can be transferred from the back to the front.

1 Mine gas monitoring data processing

Due to the special and complex production environment of coal mines and the limitations of the monitoring system itself, the monitoring data may have data anomalies, data missing and noise. Therefore, the gas monitoring data collected by the monitoring and monitoring system will show complex and nonlinear characteristics. Through data processing processes such as abnormal data replacement, missing data completion, and noise elimination, the prediction accuracy can be improved as much as possible.

1.1 Gas monitoring abnormal data processing

Abnormal values may appear in the gas concentration monitoring data at a certain time period or time point. For the maximum value in the abnormal data, it may be caused by catastrophe and the probability of its occurrence is very small, so it cannot be simply eliminated, and it should be processed according to its frequency of occurrence; for the data of zero in the abnormal data, it may be due to the detection signal If it is disturbed, it cannot be simply eliminated.

Let the gas concentration time series composed of original gas monitoring data be {x _t , t=1,2,...,N}, where N is the sequence length. When there is an abnormal value at a certain moment t=t ₀ , and the abnormal value is zero value or low frequency high gas concentration measurement value, the first N ₀ numbers in the sequence where the point is located can be used to pass the cubic spline The interpolation calculation is replaced by the corresponding value in the interpolation sequence value; and the replacement value can also be calculated by performing cubic spline interpolation optimization again on the basis of the first N ₀ data at the time t=t ₀ . The actual monitoring data is selected as the basis for interpolation calculation to avoid transmission calculation errors.

1.2 Handling of missing gas monitoring data

In the real-time monitoring data of the coal mine monitoring and monitoring system, there may be missing data at a certain period or time point. According to the characteristics of the exponential smoothing algorithm, it is suitable for the short-term forecast calculation under the long-term quadratic parabolic growth trend of the series. It has the function of resisting and weakening abnormal data. Therefore, the three-time exponential smoothing method of time series, that is, the Brown quadratic polynomial exponential smoothing method, is used to complete the data.

1.3 Gas monitoring data denoising processing

Due to the joint action of human, environmental and other uncontrollable factors in the process of data acquisition, transmission, storage and processing, the signal contains noise effects. The existence of noise data not only affects its own accuracy and reliability, but also makes its sequence show dispersion. Therefore, it is necessary to perform wavelet denoising processing on the collected data.

(1) After the gas concentration monitoring data is decomposed by wavelet, and the outlier processing and missing data filling are performed on the data, there are still large fluctuations in the monitoring data, and the distribution of the maximum value is not concentrated. Through the wavelet denoising process, it is decomposed by wavelet, and the scale coefficients c _j,k and wavelet coefficients d _j,k are obtained respectively. Therefore, the authenticity and validity of the data must still be maintained, so the maximum number of decomposition layers required for denoising cannot exceed 5.

(2) Threshold processing of wavelet decomposition coefficients. For the gas concentration monitoring data, on the basis of ensuring the authenticity of the data, the unbiased risk estimation criterion method with adaptive characteristics is used to determine the threshold, that is, for a given threshold λ, find The corresponding risk value, and choose the minimum value. Arrange the square values of the wavelet coefficients obtained in the previous step from small to large to form a vector W={w(k),k=1,2,...,N _k }, and define its risk value as:

(3) Wavelet inverse transformation reconstruction, the wavelet decomposition coefficient obtained through the above calculation and scale coefficients c _{j, k} to perform wavelet inverse transformation, and the reconstructed sequence is the gas concentration time series obtained after wavelet denoising.

2 Time series correlation analysis of gas concentration in working face

2.1 Identification of chaotic time series

Using Lyapunov exponent to judge the chaos of time series, if the largest Lyapunov exponent is greater than 0, then the dynamical system is chaotic. For the calculation of the Lyapunov index, the Wolf method is used to calculate: set the embedding dimension to m, and the time delay to τ, then repeat the time series x(1), x(2),...,x(t),... The phase space can be expressed as:

X(t)={x(t),x(t+t),...,x[t+(m-1)t]},t=1,2,...,M (2)

Denote the initial time as t ₀ , the current time as t _i , the end time as t _M , M=N-(m+1)τ, and the end point of the sequence as N. Denote the initial point as X(t ₀ ), the distance from the nearest point X ₀ (t ₀ ) is L ₀ , after the time t _i , the distance exceeds the predetermined threshold ε (ε>0), and L ₀ =| X(t _l )-X(t ₀ )|>ε. Find another point X _l (t _l ) in the field of X(t _l ), so that L _l ＝|X(t _l )-X _l (t _l )|<ε, trace its evolution process until x(t) to time The sequence end point N, the total number of iterations is t _M -t ₀ , then the maximum Lyapunov exponent λ ₁ is:

Among them, L' _i ＝|X(t _i )-X(t _i-1 )|, L _i ＝|X(t _i )-X _i (t _i )|, Xi (t _{i )} is the time t _i , in the state X(t _i ) is a point in the field of radius ε.

2.2 Phase space reconstruction of chaotic time series

Takens proposed the delayed coordinate embedding theorem in 1981, which proved that the phase space of a system can be reconstructed from a single time series. Takens theorem: M is a d-dimensional manifold, φ: M→M, y: M→R, y has a second-order continuous derivative, φ(φ,y):M→R ^{2d +1} , f(φ,y)＝ (y(x),y(φ(x)),y(φ ² (x)),...,y(φ ^2d (x))), then φ(φ,y) is M to R ^2d+ An embedding of ¹ .

From the above theorem, we can know that there exists an embedding dimensional space that can recover regular trajectories. Thus, a phase space is reconstructed, and the state of this space can be used to estimate the original system φ. The time delay τ of the embedding dimension and the embedding dimension m are obtained through the calculation and analysis of the C-C method and the G-P method.

2.3 Chaotic time series forecasting

The chaotic time series prediction principle is to fit the adjacent state model by restoring the attractor after reconstructing the phase space. The first-order weighted local method is mainly used to construct the chaotic prediction mathematical model of gas emission. The method of first-order weighted local prediction is to find out the vector set as a reference set by calculating the Euclidean distance between each neighborhood point and X(t) in the phase space, which is: X(t _i ) ,i=1,2,...,N, and use d _i to represent the distance between point X(t _i ) and X(t), and let d _min be the minimum value in d _i , so as to calculate the definition Point X(t _i ) weight, its formula is:

Among them, a is a constant coefficient, here take a=1, then its linear fitting form is:

X(t _i +1)=aR+bX(t _i )(5)

where R=[1,1,...,1] ^T . Then each phase point can be decomposed into:

in, Represents the jth dimension component of X(t _i +1), m is the embedding dimension.

According to the weighted least square method:

Consider formula (7) as a one-variable function about a and b, and calculate partial derivatives for a and b at the same time, then:

Solve the equations to obtain a and b, analyze the validity of the fit, obtain the prediction of the evolution phase point, and extract the last one-dimensional component as the predicted value.

Contents of the invention

1. A gas concentration prediction and early warning method in a working face, characterized in that, in order to carry out gas monitoring concentration prediction and early warning in a coal mine working face, the gas monitoring data is firstly subjected to preprocessing such as abnormal data replacement, missing data completion, and denoising, which realizes The integrity of the original input data of gas concentration; then use the Lyapunov index to identify the chaos and reconstruct the phase space of the time series of the original gas emission, so as to establish a chaotic prediction mathematical model of gas emission; according to the prediction results of gas concentration and The warning threshold is calculated for the prediction interval, and the corresponding warning levels are divided to complete the warning information; it can be used for gas concentration prediction and early warning work on the working face.

2. A method for predicting and early warning of gas concentration in a working face according to claim 1, characterized in that the determination of the basic indicators of early warning of gas emission at the monitoring point is carried out by the mean value μ _x and variance σ _x of the gas concentration monitoring data Calculate and determine the time t _h during which the predicted value of gas gushing continues to be too large; according to different situations, the corresponding confidence intervals at different confidence levels can be divided into the following three types: the confidence interval at the 95% confidence level [μ- 1.96σ, μ+1.96σ], the confidence interval [μ-1.44σ, μ+1.44σ] at the 85% confidence level, and the confidence interval [μ-σ, μ+σ] at the 68.3% confidence level.

3. A kind of working face gas concentration prediction and early warning method according to claim 1, characterized in that, early warning level I: if the gas concentration pre-prediction value and the upper bound of the prediction interval x _* + δ _* ≤ μ _x + σ _x , Judging that the predicted value is at the 68.3% confidence level, it can be regarded as a normal situation, and no warning will be issued; under the premise of x _* <w _s , when x _* +δ _* ∈[μ _x +σ _x ,μ _x +1.44σ _x ], it is judged that the predicted value is under the 85% confidence level, which can be regarded as a normal situation, and no warning is given; otherwise, it is considered that the gas concentration is too high, and if the predicted value continues to be too high within 1 consecutive hour And the time is t _h , when t _h >30min, it is an abnormal situation, and the level is determined as the warning level I, and the warning time is set as t _h ≤ t _w ≤ 1h, and t _w represents the warning time.

4. A method for predicting and early warning of gas concentration in a working face according to claim 1, characterized in that, early warning level II: when x _* +δ ₂ ∈ [μ _x +1.44σ _x , μ _x +1.96σ _x ], _Judging that the predicted value _is between 85% and 95% confidence level, it can be considered as a normal situation, and there _is no need to make an early warning _; The predicted value continues to be too large within 1 hour and the time is t _h . When t _h > 30min, it is judged that the predicted value is between 85% and 95% confidence level, and the predicted value has a trend of continuing to be too large, which is an abnormal situation , and determine the level as early warning level II, and set the warning time interval as: t _h ≤ t _w ≤ 1h.

5. A method for predicting and early warning of gas concentration in a working face according to claim 1, characterized in that, early warning level III: when x _* + δ ₂ > μ _x + 1.96σ _x does not reach the alarm limit value, in the continuous If _{t h >} 60min, it means that the predicted value has not exceeded the alarm concentration and is not under the 95% confidence level, and the predicted value has a trend of continuous large, it is an abnormal situation, and The level is determined as early warning level III, and the warning time interval is set as: t _h ≤ t _w ≤ 2h; if 30min<t _h <60min, the level can be determined as early warning level II.

Description of drawings

Figure 1 Part of the original data time series data segment of the 1# return air alleyway of the 2101 working face

Figure 2 Enlarged view of time series after outlier processing

Figure 3 Time series after wavelet denoising processing

Fig. 4 Gas gushing warning information display at the 1# return air alleyway of 2101 working face

detailed description

Taking Changcun Coal Mine 2101 working face 1# return air alleyway as an example, its gas monitoring point is from June 1, 2015 to June 15, 2015, a total of 15 days, 4320 data, the time length is 21600min, and the average time interval is In 5 minutes, the maximum value of the gas concentration is 0.311%, and the minimum value is 0.216%. As the original data, the data preprocessing method is used to process these data. Part of its time series data is shown in Figure 1.

The gas concentration time series is processed by the cubic spline interpolation method for outliers, and the enlarged image of the processed time series is shown in Figure 2.

The original data of the gas concentration monitoring is complete and does not need to be supplemented, but the sequence needs to be de-noised by wavelet. The obtained gas concentration time series is shown in Figure 3. On the basis of maintaining the average trend of the original time series, the data curve becomes smooth, and it is easy to extract the instantaneous characteristics of the time series.

The time delay τ is calculated using the C-C method. With the help of Matlab software, it is obtained that the time delay of the 1# return air lane of the 2101 working face is 10. The embedding dimension m is obtained by the G-P method, and the result is obtained by running the self-compiled program. The embedding dimension m=20 of the 1# return air alleyway of the working face 2101. Use the time delay τ obtained from the above calculation and analysis and the embedding dimension m to reconstruct the phase space. According to the small amount of data algorithm steps, the Lyapunov exponent λ=0.0406 of the 1# return air alley pilot site is calculated. Then the maximum predictable time is within 25 moments.

Based on the predicted value calculated by using the prediction method for the gas concentration monitoring data at the 1# return airway entrance of the 2101 working face, the average value of the predicted value of the gas emission at the 2101 working face at the 95% confidence level is μ _x =0.399344, the variance is σ _x =0.009989, then the prediction interval: [x _* -δ _* ,x _* +δ _* ] is: [0.389355,0.409334]. The gas concentration warning information at the 1# return air lane entrance of the 2101 working face obtained above is displayed, as shown in Figure 4.

It can be seen from Figure 4 that the comparison between the predicted value and the measured value of the gas emission at the 1# return air lane of the 2101 working face in the time zone of 5 hours shows that the predicted values of the predicted points are all within the range of the early warning interval, so the predicted value is consistent with the actual value. The monitoring situation is consistent, and the prediction method is feasible.

Claims (5)

1. A working face gas concentration prediction early warning method is characterized in that in order to carry out gas monitoring concentration prediction early warning on a coal mine working face, preprocessing such as abnormal data substitution, missing data completion and noise elimination is carried out on gas monitoring data, and therefore the integrity of original input data of the gas concentration is achieved; then, performing chaos recognition and phase space reconstruction on the time sequence of the original gas emission quantity by using a Lyapunov index so as to establish a gas emission quantity chaos prediction mathematical model; calculating an early warning threshold according to the gas concentration prediction result and the prediction interval thereof, and dividing early warning grades corresponding to the early warning threshold to finish early warning information; the method can be used for working face gas concentration prediction and early warning work.

2. The working face gas concentration prediction and early warning method according to claim 1, characterized in that the basic index determination of the early warning of the gas emission quantity at the monitoring point is carried out through the mean value mu of the monitoring data of the gas concentration_xSum variance σ_xCalculating and determining the time t when the predicted value of the gas emission is continuously larger_h(ii) a According to different situations, corresponding confidence intervals under different confidence levels^[10]The method can be divided into the following three types: confidence intervals at 95% confidence level [ mu-1.96 sigma,. mu. +1.96 sigma ]]Confidence interval at 85% confidence level [ mu-1.44 sigma,. mu. +1.44 sigma ]]And confidence intervals at 68.3% confidence level [ mu-sigma, [ mu ] + sigma ]]。

3. The working face gas concentration prediction and early warning method as claimed in claim 1, wherein the early warning I stage comprises: if the gas concentration is predicted and the upper bound x of the prediction interval_*+_*≤μ_x+σ_xJudging that the predicted value is at a confidence level of 68.3%, and judging that the predicted value is normal, and not carrying out early warning; at x_*<w_sOn the premise of (1), when x_*+_*∈[μ_x+σ_x,μ_x+1.44σ_x]If so, judging that the predicted value is under the 85% confidence level, and considering the predicted value as a normal condition without warning; otherwise, the gas concentration is considered to be larger, if the predicted value is larger continuously within 1 hour and the time is t_hWhen t is_h>When the time is 30min, the abnormal condition is determined, the grade is determined to be early warning I grade, and warning time t is set_h≤t_w≤1h，t_wIndicating an alert time.

4. The working face gas concentration prediction and early warning method as claimed in claim 1, wherein the early warning level II comprises: when x is_*+₂∈[μ_x+1.44σ_x,μ_x+1.96σ_x]Judging the confidence level between the predicted value of 85 percent and 95 percent, considering the normal condition, and not carrying out early warning; when x is_*+₂>μ_x+1.96σ_xIf the predicted value is continuously larger and the time is t within 1 continuous hour when the alarm threshold value is not reached_hWhen t is_h>And when the time is 30min, judging the confidence level between 85% and 95% of the predicted value, determining that the predicted value has a continuous greater trend, determining that the predicted value is abnormal, determining the grade as early warning II grade, and setting a warning time interval as follows: t is t_h≤t_w≤1h。

5. The working face gas concentration prediction and early warning method as claimed in claim 1, wherein early warning level III: when x is_*+₂>μ_x+1.96σ_xWhen the alarm threshold value is not reached, the alarm continues to be larger within 2 continuous hours for t_hIf t is_h>And (5) 60min, indicating that the predicted value does not exceed the alarm concentration and is not at the 95% confidence level, and the predicted value has a continuous greater trend, determining the abnormal condition, determining the grade as early warning grade III, and setting an alarm time interval as follows: t is t_h≤t_wLess than or equal to 2 h; if it is 30min<t_h<And (5) determining the grade as early warning II grade after 60 min.