Mine earthquake positioning method based on improved pigeon inspired optimization algorithm

Abstract

A mine earthquake positioning method based on an improved pigeon inspired optimization algorithm comprises the following steps: 1) preprocessing original mine earthquake waveform data by using ACF to obtain an ACF sequence; 2) performing first arrival time pickup on the processed data by using a long and short time window algorithm STA / LTA; (3) storing the station coordinates and the first arrival time picked in the step (2) into a file according to the format of station coordinates X, station coordinates Y, station coordinates Z and first arrival time; 4) reading the file generated in the step 3), and setting an upper boundary and a lower boundary of a search space; 5) selecting a combination of a time-of-arrival method and a variance function as a target function; and 6) using the improved pigeon inspired optimization algorithm FC-PIO to carry out seismic source positioning operation: minimizing the target function in the step 5) through the improved pigeon inspired optimization algorithm FC-PIO, finding an optimal solution, and obtaining the occurrence position of a mine earthquake event. Through the method, the mine earthquake positioning precision and efficiency can be remarkably improved in practical application, and the reliability of mine safety monitoring is ensured.

Description

Technical Field

[0001] The present invention relates to the field of microseismic positioning in the process of coal mine exploitation, and is a method for mine seismic positioning that improves swarm intelligence algorithms. Background Art

[0002] Mine seismic positioning technology is an important part of mine safety monitoring and is widely used in fields such as mine accident prediction, geological exploration, and mine disaster early warning. Accurate mine seismic positioning can not only provide the precise location of the microseismic source, help evaluate risks in a timely manner, but also provide key data for the early warning system of mine disasters. Traditional mine seismic positioning methods include the Geiger method, grid search method, genetic algorithm, etc. Although these methods are relatively mature, in the mine environment, due to factors such as the complexity of the geological structure, medium inhomogeneity, and multipath effects, the propagation of mine seismic signals is significantly affected, and phenomena such as signal attenuation, reflection, and refraction make the precise positioning of the source location more difficult.

[0003] In recent years, mine seismic positioning methods based on swarm intelligence optimization algorithms have gradually received extensive attention. These methods can effectively solve the problems of accuracy and efficiency existing in traditional positioning methods by simulating the collective behavior in nature. However, these optimization algorithms still face some challenges in practical applications, such as slow convergence speed and easy to fall into local optimal solutions. Especially in complex mine environments, the accuracy and stability of positioning results are greatly affected.

[0004] To address this problem, the present invention improves the traditional pigeon flock optimization algorithm to achieve mine seismic positioning. On the basis of the traditional pigeon flock optimization algorithm, an improved Fuch chaotic mapping is introduced to ensure that the population covers the entire search space as much as possible, improving the global search ability of the algorithm; a cross and vertical intersection strategy is introduced to avoid falling into local minima during optimization, improving the positioning accuracy of the algorithm; a target function combining the time difference of arrival method and the variance function is adopted to improve the stability of the algorithm. This algorithm can significantly improve the positioning accuracy, convergence speed, and stability in mine seismic positioning, providing more reliable technical support for mine safety monitoring. Summary of the Invention

[0005] The present invention proposes a method for mine seismic positioning that improves pigeon flock optimization, solves problems such as low positioning accuracy, slow convergence speed, and easy to fall into local optimal solutions, and provides a more accurate, stable, and efficient solution for mine seismic positioning.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A method for mine seismic positioning that improves the pigeon flock optimization algorithm, the steps are as follows:

[0008] 1) Preprocess the original mine seismic waveform data.

[0009] Due to the complex mine environment, mine tremor signals often have the characteristics of low signal-to-noise ratio and insignificant changes in amplitude and frequency. Using ACF for preprocessing mine tremor signals can better identify the subtle differences between signals and noise, improve the resolution between signals and noise, and better amplify the weak changes in signal amplitude and frequency, thereby improving the accuracy of picking. The relevant formula of ACF is as follows:

[0010]

[0011] X(i) = |x(i) / X max | (2)

[0012]

[0013] Where: x(i) is a time series, R(i) is a weight constant that changes with signal characteristics, X max is the maximum absolute amplitude of the time series, X(i) is the relative time series, and α is the relative energy coefficient, which is used to distinguish noise and weak amplitude regions.

[0014] 2) Use the short-term and long-term window algorithm STA / LTA to pick the first arrival time of the processed data.

[0015] 2.1) Set the p-wave arrival time picking threshold E0;

[0016] 2.2) Calculate the STA / LTA ratio of the ACF sequence obtained in step 1). The relevant formula is as follows:

[0017]

[0018] Where: E(i) is the STA / LTA ratio of each point, T1 is the starting point of the long-term window, T2 is the end point of the short-term window, T0 is the division point of the long-term window and the short-term window, m is the number of data in the long-term window, n is the number of data in the short-term window, and k is the sensitivity coefficient. In order to make E(i) more sensitive to the weak amplitude changes of the signal and reduce the picking error.

[0019] 2.3) Compare the values in the E(i) sequence with the picking threshold E0 one by one, and obtain the first value greater than E0 in the sequence. The corresponding time is the time when the mine tremor event first arrives at the station.

[0020] 3) Save the station coordinates and the first arrival time picked in step 2) to the relevant file.

[0021] Save the station coordinates and the first arrival time picked in step 2) to the file in the format of "station coordinate X, station coordinate Y, station coordinate Z, first arrival time".

[0022] 4) Read the file generated in step 3) and set the boundary of the search space.

[0023] Based on the file generated in step 3), set the boundaries of the search space. The lower boundary lb (lbx, lby, lbz) of the search space is set as follows:

[0024] lbx = min[x1, x2, x3,..., x n-1 , x n (6)

[0025] lby = min[y1, y2, y3,..., y n-1 , y n (7)

[0026] lbz = min[z1, z2, z3,..., z n-1 , z n (8)

[0027] The upper boundary ub (ubx, uby, ubz) of the search space is set as follows:

[0028] ubx = max[x1, x2, x3,..., x n-1 , x n (9)

[0029] uby = max[y1, y2, y3,..., y n-1 , y n (10)

[0030] ubz = max[z1, z2, z3,..., z n-1 , z n (11)

[0031] Where: n is the number of stations, and (x i , y i , z i ) is the position coordinate of the i-th station.

[0032] 5) Select the combination of time difference of arrival method and variance function as the objective function.

[0033] FC-PIO is a seismic source location algorithm based on the TDOA method. Select the combination of time difference of arrival method and variance function as the objective function. The relevant formula of the objective function is as follows:

[0034]

[0035] Where: (x0, y0, z0) is the position coordinate of the simulated seismic source, (x i , y i , z i ) is the position coordinate of station i, D iis the distance between the simulated seismic source point and station i; t0 is the event occurrence time, v is the wave velocity, then t i ' is the calculated arrival time at station i; t i is the actual arrival time at station i. Since the smaller the difference between the actual arrival time and the calculated arrival time, the smaller the positioning error, and because the calculated arrival time of the station is affected by the event occurrence time t0, the time difference between the arrival times of two stations is used as the objective function, which can not only eliminate the influence of t0 on the positioning result, but also effectively evaluate the positioning error of the simulated seismic source point. In addition, to enhance the stability of the positioning result, the time difference arrival method and the variance function are combined as the objective function. The first term in the objective function is the L2 norm, which is used to evaluate the accuracy of the acoustic emission positioning result, and the second term is used to evaluate the stability of the overall result.

[0036] 6) Use the improved pigeon flock optimization algorithm FC-PIO for seismic source positioning operation: Minimize the objective function through the improved pigeon flock optimization algorithm FC-PIO to find the optimal solution and obtain the location where the mine seismic event occurs.

[0037] 6.1) Use the improved Fuch chaotic mapping to improve the population initialization stage of the pigeon flock optimization algorithm PIO. The relevant formula is as follows:

[0038] X = lb + tanh(cos(1 / r 2 ))(ub - lb)(16)

[0039] where: r is a random number in the interval [0,1].

[0040] 6.2) Use the horizontal crossover of the cross search optimization strategy CSO to improve the map and compass operator stages of the PIO algorithm, improve the global search ability of the algorithm, and avoid falling into local optima. The relevant formula is as follows:

[0041] V i (t) = V i (t - 1)·e +Rt + rand·(X g - X i (t - 1))(17)

[0042] X i (t) = X i (t - 1)+ V i (t)(18)

[0043]

[0044] where: t is the current iteration number, R is the map and compass factor, rand is a random number between (0,1), a and b are random numbers uniformly distributed between (0,1), c and d are random numbers uniformly distributed between (-1,1), Xg The current global optimal position obtained by comparing the positions of all pigeons, V i (t) describes the magnitude and direction of the velocity of an individual after t iterations, X i (t) describes the position of individual i after t iterations, X i1j and X i2j respectively represent the j-th dimension of individual X i1 and X i2 . and respectively represent the j-th dimension of the offspring generated by the horizontal crossover of X i1 and X i2 on the j-th dimension. The generated offspring are compared with their parents respectively, and the individuals with better performance are retained.

[0045] 6.3) Update the number of pigeon flocks capable of identifying directions, calculate the central position of the pigeon flocks capable of identifying directions, and use the landmark operator of the PIO algorithm combined with the vertical crossover of the CSO strategy to update the individual positions, improve the local search ability of the algorithm, and accelerate the convergence speed of the algorithm. The relevant formulas are as follows:

[0046] half = N / 2 (21)

[0047]

[0048] X i (t) = X i (t - 1) + rand·(X c (t) - X i (t - 1)) (23)

[0049]

[0050] where: N is the population size, X i (t) represents the position of individual i after t iterations, fitness represents the objective function value of this individual, X c (t) is the central position of the pigeon flocks capable of identifying directions at the t-th iteration, r is a random number uniformly distributed between (0, 1), represents operating only on the j1 dimension of individual X i , and the other dimensions are the same as those of the parent. Similarly, the generated offspring are compared with their parents, and the individuals with better performance are retained.

[0051] 6.4) When the algorithm reaches the iteration termination condition, output the final positioning result.

[0052] The beneficial effects of the present invention are as follows: The present invention proposes an improved pigeon flock optimization method for mine tremor location. Firstly, in the population initialization stage of applying the pigeon flock optimization algorithm for mine tremor location, an improved Fuch chaotic mapping is proposed to ensure that the population covers the entire search space as much as possible, improving the global search ability of the algorithm. Then, a cross strategy is introduced during the optimization process to avoid falling into local minima during optimization and improve the location accuracy of the algorithm. Finally, an objective function combining the time difference of arrival method and the variance function is adopted to improve the stability of the algorithm. The method of the present invention verifies the effectiveness of the improved pigeon flock optimization algorithm through slab knocking location experiments. Compared with other swarm intelligence algorithms, it has higher accuracy, convergence speed and stability. Brief Description of the Drawings

[0053] Figure 1 Structural diagram of the improved pigeon flock optimization algorithm;

[0054] Figure 2 Overall flowchart of mine tremor location;

[0055] Figure 3 Distribution map of the positions of stations and simulated seismic source points;

[0056] Figure 4 Convergence process diagram of the improved pigeon flock optimization algorithm for locating simulated seismic source point #1;

[0057] Figure 5 Location result map. Detailed Embodiment

[0058] I. Theoretical Basis of the Solution of the Present Invention

[0059] 1. Short-Time Average / Long-Time Average Method

[0060] The Short-Time Average / Long-Time Average (STA / LTA) method is a method for analyzing the signal change characteristics through windows of different time scales. The core idea of this method is to use the ratio of the average energy within a short-time window and a long-time window to identify the arrival of vibration events.

[0061] The average energy within the long-time window is shown in formula (23), where T L is the number of samples in the long-time window. The long-time window is mainly used to extract the global features of the signal, including the overall trend, low-frequency components or long-term change rules; the average energy within the short-time window is shown in formula (24), where, T S is the number of samples in the short-time window. The short-time window is used to capture local changes, including mutation points, transient features or high-frequency components.

[0062]

[0063] Wherein, CF is the absolute amplitude of the signal. When a vibration event occurs, the STA / LTA ratio increases significantly. Therefore, this method sets a threshold. When the STA / LTA ratio exceeds the set threshold, it can be determined that vibration has occurred. The long-short time window method can effectively balance the global trend and local details, detect abnormal points or key change points of the signal, and is applicable to fields such as mutation signal detection, state recognition, and non-stationary signal analysis.

[0064] 2. Pigeon-Inspired Optimization Algorithm

[0065] The Pigeon-Inspired Optimization (PIO) algorithm is a swarm intelligence optimization algorithm based on the navigation behavior of pigeons proposed by Duan et al. in 2014. Pigeons have a special homing ability. Research shows that pigeons combine the sun, the earth's magnetic field, and landmarks to determine the direction when homing, and use different navigation tools in different behavioral stages. Based on this phenomenon, Duan proposed the pigeon-inspired optimization algorithm. This algorithm mainly designs two behavioral stages. The first stage is the map and compass operator stage, and the second stage is the landmark operator stage.

[0066] In the map and compass operator stage, this algorithm simulates the behavior of pigeons adjusting their flight directions under the influence of the earth's magnetic field induction and the direction of the sun. The algorithm uses formula (25) for speed update and formula (26) for position update.

[0067] V i (t) = V i (t - 1)·e -Rt + rand·(X g - X i (t - 1)) (25)

[0068] X i (t) = X i (t - 1)+ V i (t)(26)

[0069] Wherein, t is the current iteration number, R is the map and compass factor, rand is a random number between (0, 1), X g is the current global optimal position obtained by comparing the positions of all pigeons, V i (t) describes the magnitude and direction of the velocity of the individual after t iterations, and X i (t) describes the position of the individual i after t iterations.

[0070] In the landmark operator stage, the algorithm simulates the flight behavior of pigeons approaching the destination based on landmarks such as buildings, lakes, and mountains. However, some pigeons are far from the destination and can only fly by following pigeons closer to the destination. Therefore, the flight of these pigeons will not affect the subsequent adjustment of the flight direction, and the number of pigeons will decrease with each iteration, as shown in formula (27).

[0071]

[0072] Let X c (t) be the central position of the pigeon flock capable of distinguishing directions at the t-th iteration. The calculation process is shown in formula (28).

[0073]

[0074] In the formula, N p (t) represents the current number of pigeons, X i (t) represents the position of individual i after t iterations, and fitness represents the objective function value of this individual. Then, the pigeons update their individual positions according to the central position of the pigeon flock using formula (29).

[0075] X i (t) = X i (t - 1) + rand·(X c (t) - X i (t - 1)) (29)

[0076] II. The method of the present invention has the following steps:

[0077] 1) Preprocess the original mine seismic waveform data.

[0078] Due to the complex mine environment, mine seismic signals often have the characteristics of low signal-to-noise ratio and insignificant amplitude and frequency changes. Using ACF to preprocess the mine seismic signals can better identify the subtle differences between signals and noise, improve the resolution between signals and noise, and better amplify the weak changes in amplitude and frequency of the signals, thereby improving the accuracy of picking. The relevant formulas of ACF are as follows:

[0079]

[0080] X(i) = |x(i) / X max | (2)

[0081]

[0082] ACF(i) = xα(i) + R·(x(i) - x(i - 1))α (4)

[0083] where: x(i) is a time series, R(i) is a weight constant that varies with the signal characteristics, X max is the maximum absolute amplitude of the time series, X(i) is the relative time series, and α is the relative energy coefficient, which is used to distinguish noise and weak amplitude regions.

[0084] 2) Use the short-term and long-term window algorithm STA / LTA to pick the first arrival time of the processed data.

[0085] 2.1) Set the p-wave arrival time picking threshold E0;

[0086] 2.2) Calculate the STA / LTA ratio of the ACF sequence obtained in step 1), and the relevant formula is as follows:

[0087]

[0088] where: E(i) is the STA / LTA ratio of each point, T1 is the starting point of the long-term window, T2 is the end point of the short-term window, T0 is the division point of the long-term window and the short-term window, m is the number of data in the long-term window, n is the number of data in the short-term window, and k is the sensitivity coefficient. In order to make E(i) more sensitive to the weak amplitude change of the signal and reduce the picking error.

[0089] 2.3) Compare the values in the E(i) sequence with the picking threshold E0 one by one, and get the first value greater than E0 in the sequence. The corresponding time is the time when the mine tremor event first arrives at the station.

[0090] 3) Save the station coordinates and the first arrival time picked in step 2) to the relevant file.

[0091] Save the station coordinates and the first arrival time picked in step 2) to the file in the format of "station coordinate X, station coordinate Y, station coordinate Z, first arrival time".

[0092] 4) Read the file generated in step 3) and set the boundary of the search space.

[0093] According to the file generated in step 3), set the boundary of the search space. The lower boundary lb(lbx, lby, lbz) of the search space is set as:

[0094] lbx = min[x1, x2, x3,..., x n-1 , x n (6)

[0095] lby = min[y1, y2, y3,..., y n-1 , y n (7)

[0096] lbz = min[z1, z2, z3,..., zn-1 , z n (8)

[0097] The upper boundary ub(ubx, uby, ubz) of the search space is set as follows:

[0098] ubx = max[x1, x2, x3,..., x n-1 , x n (9)

[0099] uby = max[y1, y2, y3,..., y n-1 , y n (10)

[0100] ubz = max[z1, z2, z3,..., z n-1 , z n (11)

[0101] Where: n is the number of stations, and (x i , y i , z i ) is the position coordinate of the i-th station.

[0102] 5) Select the combination of the time difference of arrival method and the variance function as the objective function.

[0103] FC-PIO is a seismic source location algorithm based on the TDOA method. The combination of the time difference of arrival method and the variance function is selected as the objective function. The relevant formulas of the objective function are as follows:

[0104]

[0105] Where: (x0, y0, z0) is the position coordinate of the simulated seismic source, (x i , y i , z i ) is the position coordinate of station i, and D i is the distance between the simulated seismic source point and station i; t0 is the event occurrence time, v is the wave velocity, then t i ' is the calculated arrival time of station i; t i is the actual arrival time of station i. Since the smaller the difference between the actual arrival time and the calculated arrival time, the smaller the positioning error, and because the calculated arrival time of the station is affected by the event occurrence time t0, the time difference between the arrival times of two stations is used as the objective function, which can not only eliminate the influence of t0 on the positioning result, but also effectively evaluate the positioning error of the simulated seismic source point. In addition, to enhance the stability of the positioning result, the combination of the time difference of arrival method and the variance function is used as the objective function. The first term in the objective function is the L2 norm, which is used to evaluate the accuracy of the acoustic emission positioning result, and the second term is used to evaluate the stability of the overall result.

[0106] 6) Use the improved pigeon flock optimization algorithm FC-PIO for seismic source location operation: Minimize the objective function through the improved pigeon flock optimization algorithm FC-PIO to find the optimal solution and obtain the location where the mine tremor event occurs.

[0107] 6.1) Improve the population initialization stage of the pigeon flock optimization algorithm PIO using the improved Fuch chaotic mapping. The relevant formula is as follows:

[0108] X = lb + tanh(cos(1 / r 2 ))(ub - lb)(16)

[0109] Where: r is a random number in the interval [0, 1].

[0110] 6.2) Use the horizontal crossover of the cross search optimization (CSO) strategy to improve the map and compass operator stages of the PIO algorithm, enhance the global search ability of the algorithm, and avoid falling into local optima. The relevant formula is as follows:

[0111]

[0112] Where: t is the current iteration number, R is the map and compass factor, rand is a random number between (0, 1), a and b are random numbers uniformly distributed between (0, 1), c and d are random numbers uniformly distributed between (-1, 1), X g is the current global optimal position obtained by comparing the positions of all pigeons, V i (t) describes the magnitude and direction of the velocity of the individual after t iterations, X i (t) describes the position of individual i after t iterations, X i1j and X i2j respectively represent the j-th dimension of individual X i1 and X i2 . and respectively represent the j-th dimension of the offspring generated by the horizontal crossover of X i1 and X i2 in the j-th dimension. Compare the generated offspring with their parents respectively and retain the individuals with better performance.

[0113] 6.3) Update the number of pigeon flocks capable of distinguishing directions, calculate the center position of the pigeon flocks capable of distinguishing directions, and use the landmark operator of the PIO algorithm combined with the vertical crossover of the CSO strategy to update the individual positions, improve the local search ability of the algorithm, and accelerate the convergence speed of the algorithm. The relevant formula is as follows:

[0114] half = N / 2(19)

[0115]

[0116] X i X(t) = i X(t - 1)+rand·(X c (t)-X i (t - 1))(21)

[0117]

[0118] where: N is the population size, X i (t) represents the position of individual i after t iterations, fitness represents the objective function value of this individual, X c (t) is the central position of the pigeon flock capable of identifying directions at the t-th iteration, r is a random number uniformly distributed between (0, 1), denotes that only the j1 dimension of individual X i is operated on, and other dimensions are the same as those of the parent generation. Similarly, the generated offspring are compared with their parents, and the individuals with better performance are retained.

[0119] 6.4) When the algorithm reaches the iteration termination condition, output the final positioning result.

[0120] III. Embodiment:

[0121] In this paper, a simulated mine tremor positioning experiment is designed to verify the effectiveness of the FC-PIO acoustic emission positioning method. The positioning space range set in this simulation experiment is 1000m×1000m×1000m, 8 stations are set up and 5 simulated seismic source points are randomly generated, and its spatial structure diagram is as Figure 3 shown.

[0122] Step 1: Record the coordinates of 8 stations and 5 simulated seismic source points, as shown in Table 1 and Table 2;

[0123] Step 2: Calculate the time difference of arrival between 5 simulated seismic source points and each station, as shown in Table 3;

[0124] Step 3: According to the station position coordinates and time difference of arrival information, use the improved pigeon flock optimization algorithm FC-PIO to locate 5 simulated seismic source points respectively, and the positioning results are shown in Table 4 and Figure 5 shown.

[0125] Table 1 Station coordinate table

[0126] Station number X / mm Y / mm Z / mm CH01 0 0 0 CH02 0 1000 0 CH03 1000 1000 0 CH04 1000 0 0 CH05 0 0 1000 CH06 0 1000 1000 CH07 1000 1000 1000 CH08 1000 0 1000

[0127] Table 2 Simulated seismic source point coordinate table

[0128] Source number X / m Y / m Z / m #1 260 870 330 #2 120 190 100 #3 980 950 790 #4 600 490 550 #5 240 570 960

[0129] Table 3 Time difference of arrival table between each station

[0130]

[0131] Table 4 Positioning Results

[0132] Source number X / m Y / m Z / m Location error / m #1 264.57 865.39 334.62 7.96 #2 124.55 187.78 94.46 7.50 #3 988.7 951.11 788.75 8.85 #4 604.26 486.99 547.89 5.62 #5 241.91 567.77 953.73 6.92

Claims

1. A method for locating mine tremors by improving the pigeon flock optimization algorithm, characterized in that, The steps are as follows: 1) Use ACF to preprocess the original mine seismic waveform data to obtain the ACF sequence; 2) Use the short-term average / long-term average (STA / LTA) algorithm to pick the first arrival time of the processed data; 3) Save the station coordinates and the first arrival time picked in step 2) to a file in the format of "station coordinate X, station coordinate Y, station coordinate Z, first arrival time"; 4) Read the file generated in step 3) and set the upper and lower boundaries of the search space; 5) Select the combination of arrival time difference method and variance function as the objective function; 6) Use the improved pigeon-inspired optimization (FC-PIO) algorithm to perform the seismic source location operation: minimize the objective function in 5) through the improved pigeon-inspired optimization (FC-PIO) algorithm to find the optimal solution and obtain the location where the mine seismic event occurs.

2. An improved pigeon flock optimization algorithm-based mine tremor location method according to claim 1, characterized in that: In the above-mentioned 1), the specific method is as follows: Use ACF to preprocess the mine seismic signal, identify the difference between the signal and the noise, and amplify the weak changes in the amplitude and frequency of the signal. The relevant formula of ACF is as follows: X(i) = |x(i) / X max | (2) ACF(i) = x a (i) + R·(x(i) - x(i - 1)) α (4) where: x(i) is a time series, R(i) is a weight constant that varies with signal characteristics, X max is the maximum absolute amplitude of the time series, X(i) is the relative time series, and α is the relative energy coefficient used to distinguish noise and weak amplitude regions.

3. An underground mine tremor location method for improving pigeon flock optimization algorithm according to claim 1, characterized in that: In the above-mentioned 2), the specific method is as follows: 2.1) Set the p-wave arrival time picking threshold E0; 2.2) Calculate the STA / LTA ratio of the ACF sequence obtained in step 1). The relevant formula is as follows: Where: E(i) is the STA / LTA ratio of each point, T1 is the starting point of the long-time window, T2 is the end point of the short-time window, T0 is the segmentation point of the long-time window and the short-time window, m is the number of data in the long-time window, n is the number of data in the short-time window, and k is the sensitivity coefficient. In order to make E(i) more sensitive to the weak amplitude change of the signal and reduce the picking error; 2.3) Compare the values in the E(i) sequence with the picking threshold E0 one by one, and obtain the first value greater than E0 in the sequence. The time corresponding to this value is the time when the mine seismic event first arrives at the station, that is, the first arrival time.

4. An underground mine tremor location method for improving pigeon flock optimization algorithm according to claim 1, characterized in that: In the above-mentioned 4), the specific method is as follows: According to the file generated in step 3), set the boundaries of the search space. The lower boundary lb(lbx, lby, lbz) of the search space is set as: lbx = min[x1, x2, x3,..., x n-1 , x n (6) lby = min[y1, y2, y3,..., y n-1 , y n (7) lbz = min[z1, z2, z3, ..., z n-1 , z n (8) The upper boundary ub(ubx, uby, ubz) of the search space is set as: ubx = max[x1, x2, x3,..., x n-1 , x n (9) uby = max[y1, y2, y3,..., y n-1 , y n (10) ubz = max[z1, z2, z3, ..., z n-1 , z n (11) where: n is the number of stations, (x i , y i , z i ) are the position coordinates of the i-th station.

5. An improved pigeon flock optimization algorithm-based mine tremor location method according to claim 1, characterized in that: In the above-mentioned 5), the specific method is as follows: Select the combination of arrival time difference method and variance function as the objective function. The relevant formula of the objective function is as follows: Where: (x0, y0, z0) is the coordinate of the simulated seismic source position, (x i , y i , z i ) is the coordinate of station i, and D i is the distance between the simulated seismic source point and station i; t0 is the event occurrence time, v is the wave velocity, then t i ' is the calculated arrival time of station i; t i is the actual arrival time of station i; since the smaller the difference between the actual arrival time and the calculated arrival time, the smaller the positioning error, and at the same time the calculated arrival time of the station is affected by the event occurrence time t0, the time difference between the arrival times of two stations is used as the objective function to eliminate the influence of t0 on the positioning result, and effectively evaluate the positioning error of the simulated seismic source point; to enhance the stability of the positioning result, the arrival time difference method and the variance function are combined as the objective function. The first term in the objective function is the L2 norm, which is used to evaluate the accuracy of the acoustic emission positioning result, and the second term is used to evaluate the stability of the overall result.

6. According to the mine seismic location method with an improved pigeon-inspired optimization algorithm described in claim 1, it is characterized in that: in the above-mentioned 6), the specific method is as follows: 6.1) Use the improved Fuch chaotic mapping to improve the population initialization stage of the pigeon-inspired optimization (PIO) algorithm. The formula is as follows: X = lb + tanh(cos(1 / r 2 ))(ub - lb)(16) Where: r is a random number in the interval [0, 1]; 6.2) Use the horizontal crossover of the cross search optimization (CSO) strategy to improve the map and compass operator stages of the PIO algorithm. The relevant formula is as follows: V i V(t) = i V(t - 1)·e -Rt + rand·(X g - X i (t - 1))(17) X i X(t) = i X(t - 1)+V i X(t)(18) Where: t is the current iteration number, R is the map and compass factor, rand is a random number between (0, 1), a and b are random numbers uniformly distributed between (0, 1), c and d are random numbers uniformly distributed between (-1, 1), X g is the current global optimal position obtained by comparing the positions of all pigeons, V i (t) describes the magnitude and direction of the velocity of the individual after t iterations, X i (t) describes the position of individual i after t iterations, X i1j and X i2j respectively represent the j-th dimension of individual X i1 and X i2 ; and respectively represent the j-th dimension of the offspring generated by horizontal crossover of X i1 and X i2 in the j-th dimension; Compare the generated offspring with their parents respectively, and retain the individuals with better performance; 6.3) Update the number of pigeon flocks capable of distinguishing directions, calculate the central position of the pigeon flocks capable of distinguishing directions, and use the landmark operator of the PIO algorithm combined with the vertical crossover of the CSO strategy to update the individual positions. The relevant formula is as follows: half=N / 2(21) X i X(t) = i X(t - 1)+rand·(X c (t)-X i (t - 1))(23) Where: N is the population size, X i (t) represents the position of individual i after t iterations, fitness represents the objective function value of this individual, X c (t) is the central position of the pigeon flock capable of identifying directions at the t-th iteration, r is a random number uniformly distributed between (0, 1), represents operating only on the j1 dimension of individual X i while keeping other dimensions the same as the parent; similarly, the generated offspring is compared with its parent, and the individual with better performance is retained. 6.4) When the algorithm reaches the iteration termination condition, output the final location result.

Images