A rock burst early warning effect evaluation method and system after eliminating data processing errors

Abstract

The application provides a rock burst early warning effect evaluation method and system after eliminating data processing errors, and the method comprises the following steps: obtaining daily cumulative microseismic event numbers and daily cumulative microseismic release energy data in a warning period, outputting daily rock burst early warning grades, calculating early warning occurrence probabilities of each rock burst early warning grade in the warning period and accuracy rates of data processing stages, and calculating early warning occurrence probabilities of each rock burst early warning grade after eliminating data processing errors; and evaluating rock burst early warning effects based on the early warning occurrence probabilities of each rock burst early warning grade after eliminating data processing errors and actual occurrence probabilities. The application proposes a scientific error elimination method for errors that may be introduced in the data processing stage of a rock burst early warning system, the method can effectively eliminate interference caused by input data errors, and the evaluation of early warning effects is more focused on the performance of the system itself rather than covering up or exaggerating the actual ability of the system due to improper data processing.

Description

Technical Field

[0001] This invention relates to the field of rockburst early warning and prevention, and more specifically, to a method and system for evaluating the effectiveness of rockburst early warning after eliminating data processing errors. Background Technology

[0002] In the field of rockburst early warning and prevention, rockburst early warning systems are widely used to assess potential rockburst risks so that timely countermeasures can be taken to ensure the safety of personnel and property at construction sites. Currently, commonly used rockburst early warning systems in China mainly rely on sensors to collect microseismic signals from construction sites. By analyzing the characteristics of these microseismic signals within the warning time domain (such as the number of microseismic events, microseismic energy release, and the location of microseismic events), potential rockburst risks are predicted. Microseismic signal processing is a crucial part of the early warning system. Currently, the preliminary processing of microseismic signals mainly relies on manual processing by experienced personnel or the use of automated processing technologies to process the input information for the early warning system. Specifically, this includes filtering microseismic signals collected from construction sites by microseismic sensors to identify precursory rockburst information with specific waveforms, and then calibrating the first arrival times of P-waves and S-waves based on the specific waveforms to locate microseismic events and calculate their energy release. The raw data collected by microseismic sensors at the construction site includes a wealth of information, such as micro-vibrations caused by on-site construction, electrical signal noise, and rock fracture signals. The fracture signals released by rock micro-fractures, which provide precursory information for rockburst risk prediction, are crucial. The input process requires filtering these rock micro-fracture signals from the raw signal set. Subsequently, the first arrival times of P-waves and S-waves are selected based on the specific waveforms to locate the microseismic event and calculate its released energy. Therefore, the input information processing stage plays a decisive role in the selection of effective microseismic signals and the calculation of energy.

[0003] Currently, common signal processing methods include manual processing and automated processing. Errors generated during the manual processing stage directly affect the identification of valid signals and the statistical analysis of energy, thus directly impacting the accuracy of the early warning system. With the development of computer technology, intelligent methods such as machine learning are increasingly being used for signal screening and processing. However, errors generated by automated processing techniques in this process also directly affect the early warning results.

[0004] Current evaluations of rockburst early warning effectiveness typically combine rockburst warnings with actual rockburst conditions at the construction site, focusing solely on the final accuracy of the warning and neglecting the potential impact of errors generated during data processing. For example, in a rockburst early warning system based on microseismic information, manually or automatically processed microseismic data serves as input. Therefore, errors generated during manual or automated data processing also enter the system, affecting the final warning result. Consequently, scientifically eliminating the interference of input data errors on the warning effectiveness becomes a crucial issue in evaluating the performance of the early warning system itself. Summary of the Invention

[0005] This invention addresses the technical problems existing in the prior art by providing a method and system for evaluating the effectiveness of rockburst early warning after eliminating data processing errors.

[0006] According to a first aspect of the present invention, a method for evaluating the effectiveness of rockburst early warning after eliminating data processing errors is provided, comprising:

[0007] The daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are obtained during the early warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals.

[0008] Based on the threshold range of different rockburst warning levels, the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are used to determine the rockburst warning level.

[0009] Based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level, the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage are calculated.

[0010] Based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage, the probability of occurrence of each rockburst warning level after removing data processing errors is calculated.

[0011] The effectiveness of rockburst warnings is evaluated based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level.

[0012] According to a second aspect of the present invention, a rockburst early warning effect evaluation system after eliminating data processing errors is provided, comprising:

[0013] The acquisition module is used to acquire the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals.

[0014] The early warning module is used to classify rockburst early warning levels by outputting the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data according to the threshold range of different rockburst early warning levels.

[0015] The first calculation module is used to calculate the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level.

[0016] The second calculation module is used to calculate the probability of occurrence of each rockburst warning level after removing data processing errors, based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage.

[0017] The evaluation module is used to evaluate the effectiveness of rockburst warnings based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level.

[0018] This invention provides a method and system for evaluating the effectiveness of rockburst early warning systems after eliminating data processing errors. The method involves acquiring the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period, classifying these data into rockburst early warning levels, and calculating the probability of occurrence and the accuracy of the data processing stage for each level based on the number of microseismic events and the microseismic energy release data for each level. It also calculates the probability of occurrence for each level after eliminating data processing errors. Based on the probability of occurrence for each level after eliminating data processing errors and the actual probability of occurrence for each level, the effectiveness of the rockburst early warning system is evaluated. This invention proposes a scientific error elimination method to address potential errors introduced during the data processing stage of the rockburst early warning system. This method effectively eliminates interference caused by input data errors, allowing the evaluation of the early warning effect to focus more on the system's performance itself, rather than masking or exaggerating the system's actual capabilities due to improper data processing. Attached Figure Description

[0019] Figure 1 A flowchart of a method for evaluating the effectiveness of rockburst early warning after eliminating data processing errors, provided by the present invention;

[0020] Figure 2 This is a schematic diagram of the structure of a rockburst early warning effect evaluation system after eliminating data processing errors, provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. In addition, the technical features of the various embodiments or individual embodiments provided by the present invention can be arbitrarily combined with each other to form feasible technical solutions. Such combinations are not constrained by the order of steps and / or structural composition patterns, but must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0022] Currently, the commonly used rockburst early warning system in tunnel and water conservancy projects is the microseismic rockburst early warning system based on micro-signals of rock mass fracture. The early warning process of this system can be summarized as follows: 1. Signal acquisition stage: Microseismic sensors collect all collectable microseismic signals at the construction site and upload the data to the system; 2. Rockburst early warning system signal processing stage: Rock mass fracture signals are filtered from the raw microseismic signals, and the location and energy release of these signals are calculated. The rock mass fracture signals after filtering, location, and calculation are called valid signals or valid events; 3. Early warning stage: The input information (data after filtering, location, calculation, and statistics, i.e., data after the data processing stage) is input into the rockburst early warning system, and the system outputs the early warning result. Currently, the sum of the number of valid events and the sum of the microseismic energy release are mainly used as the primary early warning indicators for rockburst early warning.

[0023] During the signal acquisition phase, all micro-vibration sources on-site may generate micro-vibration signals, including vibrations from construction machinery, micro-vibrations from construction procedures, noise signals such as electrical signals, and micro-signals generated by rock fractures. Among these, the precursory information of rockburst incubation is primarily the micro-vibration signals released by rock micro-fractures; a single micro-vibration signal is also called a micro-vibration event. Currently, rockburst early warning systems generally process input signals in two ways: manual processing and automated processing. Manual processing is typically handled by experienced personnel at the construction site, while automated processing typically employs methods such as machine learning. The system filters all micro-vibration signals acquired on-site and calibrates the P-wave and S-wave first arrival times of each rock fracture signal. Based on the calibration results, the system automatically calculates the location of the rock fracture signal and the micro-vibration release energy. Valid signals have unique waveforms, distinguishing them from other noise signals. However, it cannot be ruled out that noise sources may also generate waveforms similar to valid signals under certain circumstances. In such cases, various processing methods cannot easily distinguish between valid and noise signals based solely on waveform characteristics, and there is a possibility of mistaking noise signals for valid signals during the screening of valid events. Meanwhile, due to issues such as signal attenuation and distortion during the propagation and transmission of microseismic signals in the medium, and limitations imposed by the performance of on-site sensors, the accuracy of microseismic signals is also limited to a certain range. Under the influence of various factors, it is difficult to accurately match the location of rock fracture signals and the calculation of released energy to real-world conditions.

[0024] In summary, the errors in the results of rockburst early warning systems produced by using data with inherent errors do not necessarily originate entirely from the early warning system itself. It is necessary to consider the propagation of errors generated from the input information during the system's operation. Therefore, to scientifically and effectively evaluate the early warning performance of a rockburst early warning system, it is necessary to quantify the impact of errors generated during the data processing stage on the early warning results, thereby eliminating the influence of these errors and thus evaluating the overall effectiveness of the rockburst early warning system.

[0025] Based on this, the present invention provides a method for evaluating the effectiveness of rockburst early warning after eliminating data processing errors, see [link to relevant documentation]. Figure 1 The method includes the following steps:

[0026] Step 1: Obtain the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals.

[0027] Understandably, in the data processing stage of a rockburst early warning system, the most direct result of various processing methods is the generation of the cumulative number of microseismic events and the cumulative microseismic energy release, both of which are important criteria for judging the rockburst early warning results. Therefore, this study selects the probability of the cumulative number of microseismic events and the cumulative energy release as the research object for quantifying the error of the input information. However, the errors that are unavoidable in the signal acquisition and screening process exhibit strong uncertainty and randomness. Therefore, a method combining historical statistical data and probabilistic statistical models is adopted to quantify the distribution patterns of the cumulative effective number of events and the cumulative energy release.

[0028] During the warning period, the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are generated by processing the original microseismic signals.

[0029] Step 2: Based on the threshold range of different rockburst warning levels, the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are used to determine the rockburst warning level.

[0030] Understandably, rockburst warning levels are determined by the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data. In one possible embodiment of the present invention, based on the threshold ranges of different rockburst warning levels, the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are used to determine the rockburst warning level, including:

[0031] Based on the daily cumulative number of microseismic events collected during the early warning period, the daily cumulative number of valid events is calculated. The valid event refers to the total number of times the original microseismic signal was a rock fracture signal.

[0032] Based on the threshold range into which the daily cumulative valid event count falls, the rockburst warning level corresponding to the daily cumulative valid event count is obtained, wherein the threshold range of the cumulative valid event count corresponding to each rockburst warning level is obtained through the rockburst warning system.

[0033] Specifically, when using a rockburst early warning system for rockburst level prediction, the system can predict the rockburst warning level by inputting the processed cumulative microseismic event count and cumulative microseismic energy release data. Then, by combining historical cumulative microseismic event counts and cumulative microseismic energy release data with the predicted rockburst warning level from the system, the threshold ranges for the cumulative effective event count and cumulative microseismic energy release data corresponding to each rockburst warning level can be summarized.

[0034] The daily cumulative number of valid events is calculated from the daily cumulative number of microseismic events. Based on the threshold range of the daily cumulative number of valid events corresponding to each rockburst warning level, the rockburst warning level corresponding to the daily cumulative number of valid events is obtained.

[0035] Similarly, based on the threshold range into which the daily cumulative microseismic energy release data falls during the collected warning period, the rockburst warning level corresponding to the daily cumulative microseismic energy release data is obtained.

[0036] Step 3: Based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level, calculate the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage.

[0037] Understandably, based on the rockburst warning level corresponding to the daily cumulative number of valid events and the rockburst warning level corresponding to the daily cumulative microseismic release energy data in step 2, the daily cumulative number of microseismic events and the daily cumulative microseismic release energy data for each rockburst warning level within the warning period are statistically calculated. Based on this, the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage are calculated.

[0038] Specifically, based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level, the probability of occurrence for each rockburst warning level is calculated as follows:

[0039] Based on the number of valid events per day for each rockburst warning level, the number of valid events for each rockburst warning level during the warning period is counted, and the ratio of the number of rockburst warnings for each level to the total number of warnings during the warning period is calculated as the probability of the first warning occurring for each rockburst warning level.

[0040] Based on the daily microseismic energy release data for each rockburst warning level, the microseismic energy release data for each rockburst warning level during the warning period is statistically analyzed. The ratio of the number of warnings for each rockburst warning level to the total number of warnings for microseismic energy release data during the warning period is calculated as the probability of the second warning occurring for each rockburst warning level.

[0041] Understandably, based on the number of effective events per day for each rockburst warning level, the number of effective events for each rockburst warning level over multiple days within the warning period is counted, and the ratio of the number of warnings for each rockburst warning level based on the number of effective events to the total number of warnings within the warning period is calculated as the probability of the first warning occurring for each rockburst warning level.

[0042] Similarly, based on the daily microseismic energy release data for each rockburst warning level, the microseismic energy release data for each rockburst warning level during the warning period is statistically analyzed. The ratio of the number of warnings for each rockburst warning level based on the microseismic energy release data to the total number of warnings during the warning period is calculated as the probability of the second warning occurring for each rockburst warning level.

[0043] Specifically, the probability of an early warning for each rockburst level is calculated based on the daily cumulative number of events and the daily cumulative microseismic energy release, according to the warning threshold. Calculations reveal that, generally, the probabilities of the daily cumulative events and the daily cumulative microseismic energy release for each rockburst level deviate from the actual early warning probabilities for each rockburst level. Furthermore, the first early warning probability calculated based on the daily microseismic energy release data for each rockburst warning level may differ from the second early warning probability calculated based on the daily microseismic energy release data for each rockburst warning level. Therefore, weighting coefficients are introduced for the first and second early warning probabilities. Based on these weighting coefficients, the first and second early warning probabilities are weighted and summed to obtain the early warning probability for each rockburst warning level. Here, the early warning probability for each rockburst warning level is the output probability of the rockburst early warning system before data processing errors are removed.

[0044] The probability of the first warning based on the number of microseismic events and the probability of the second warning based on the microseismic energy release data were calculated. Then, the accuracy of the data processing of the number of microseismic events and the accuracy of the data processing of the microseismic energy release data were calculated.

[0045] In one possible embodiment of the present invention, the step of calculating the accuracy of the data processing stage based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level includes: calculating the accuracy of processing the number of microseismic events and the accuracy of processing the microseismic energy release data in the data processing stage, respectively.

[0046] Specifically, the accuracy of processing microseismic event numbers during the data processing stage includes:

[0047] The first two-dimensional confusion matrix records the number of microseismic events whose event status after data processing matches the actual event status and the number of microseismic events whose event status does not match the actual event status in the daily cumulative microseismic events; wherein, the event status includes valid event status and invalid event status;

[0048] The first accuracy rate of the microseismic event count is calculated based on the number of microseismic events whose event state after the data processing stage is consistent with the actual event state and the number of microseismic events whose event state is inconsistent with the actual event state.

[0049] Specifically, the accuracy of processing the cumulative microseismic event count can be evaluated using a two-dimensional confusion matrix. First, the daily data processing error is calculated, including the number of valid events marked as invalid and the number of invalid events marked as valid. In the two-dimensional confusion matrix, P represents valid events and N represents invalid events, and the F1 score is used as the accuracy rate of the processed data.

[0050] The two-dimensional confusion matrix for processing the cumulative microseismic event count is shown in Table 1 below:

[0051] Table 1

[0052]

[0053] Calculate the accuracy based on Table 1:

[0054] Recall rate:

[0055] F1 score:

[0056] F1 represents the accuracy of processing the number of microseismic events.

[0057] Specifically, the accuracy of processing microseismic release energy data during the data processing stage includes:

[0058] The second two-dimensional confusion matrix records the number of daily accumulated microseismic energy release data that are consistent with the actual data state and the number of inconsistent data states after the data processing stage; wherein, the data state includes valid data state and invalid data state;

[0059] The second accuracy rate of the microseismic release energy data processing stage is calculated based on the number of microseismic release energy data that are consistent with the actual data state after the data processing stage and the number of inconsistent microseismic release energy data.

[0060] The accuracy of processing cumulative microseismic release energy can be evaluated using a two-dimensional confusion matrix. Calculating the cumulative microseismic release energy for each rock fracture event is complex and difficult to define accurately. However, the calculation of microseismic release energy for rock fracture events is related to the event's location; therefore, events located within a valid range, and whose calculated microseismic release energy is within a reasonable range, are considered valid events. First, the daily error in processing microseismic release energy is statistically analyzed, including the number of valid events marked as invalid and the number of invalid events marked as valid. In the two-dimensional confusion matrix, P represents valid events, N represents invalid events, and the F1 score is used as the accuracy rate of the processed data, referred to as the probability of the first warning occurring. The two-dimensional confusion matrix for processing cumulative microseismic release energy data is shown in Table 2 below:

[0061] Table 2

[0062]

[0063] Calculate the accuracy based on Table 2:

[0064] Recall rate:

[0065] F1 score:

[0066] F1 represents the accuracy of processing the microseismic energy release data, and is referred to as the probability of the second early warning occurring.

[0067] Step 4: Based on the predicted occurrence probability of each rockburst warning level and the accuracy of the data processing stage, calculate the predicted occurrence probability of each rockburst warning level after removing data processing errors.

[0068] Understandably, step 3 calculates the probability and accuracy of the first warning, as well as the probability and accuracy of the second warning, and then calculates the predicted probability of each rockburst warning level after removing data processing errors.

[0069] Based on the probability of occurrence of the first warning and the first accuracy rate, after eliminating the error in the number of microseismic events during the data processing stage, the probability of occurrence of the first warning P(B) for each rockburst warning level is calculated. 1i ), and based on the probability of the second warning occurring and the second accuracy rate, after removing the errors in the microseismic release energy data during the data processing stage, the probability of the second warning occurring for each rockburst warning level, P(B), is calculated. 2i Then, based on the probability of the first warning occurring P(B) 1i ) and the probability of the second warning occurring P(B) 2i ), calculate the probability of occurrence P(B) for each rockburst warning level after removing data processing errors. i ).

[0070] Specifically, based on Bayesian theory, after eliminating errors in the data processing stage regarding the number of microseismic events, the probability of the first warning occurring for each rockburst warning level is P(B). 1i ):

[0071]

[0072] In the formula, P(B) 1i |A1) represents the probability of the first warning occurring at rockburst warning level i after the data processing stage; P(A1|B 1i The first accuracy rate in the data processing stage for rockburst early warning level i is ).

[0073] P(B 1i P(A1) represents the probability of the first warning of rockburst warning level i after removing the error in the number of microseismic events during the data processing stage; P(A1) represents the overall accuracy of the number of microseismic events during the data processing stage.

[0074] After removing errors in the microseismic energy release data during the data processing stage, the probability of a second warning occurring for each rockburst warning level is calculated as P(B). 2i );

[0075]

[0076] In the formula, P(B) 2i |A2) represents the probability of the second warning occurring at rockburst warning level i after the data processing stage; P(A2|B 2i The second accuracy rate in the data processing stage for rockburst early warning level i is [data missing].

[0077] P(B 2i P(A2) represents the probability of the second warning occurring at rockburst warning level i after removing errors in the microseismic release energy during the data processing stage; P(A2) represents the overall accuracy of the microseismic release energy data during the data processing stage.

[0078] Based on the probability of the first warning occurring P(B) 1i ) and the probability of the second warning occurring P(B) 2i ), calculate the probability of occurrence P(B) for each rockburst warning level after removing data processing errors. i ):

[0079] P(B i )=α·P(B 1i )+β·P(B 2i )

[0080] Where α and β are the probabilities of the first warning occurring, P(B). 1i The weighting coefficient of the second warning and the probability of its occurrence P(B) 2i The weighting coefficients, α and β, are obtained by least squares fitting.

[0081] Step 5: Evaluate the effectiveness of rockburst warnings based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level.

[0082] Specifically, after obtaining the probability of occurrence of each rockburst warning level after removing data processing errors, the probability of occurrence of each rockburst warning level is compared with the actual probability of occurrence of each rockburst warning level at the construction site to establish an evaluation index for the warning effect of the rockburst warning system.

[0083]

[0084] In the formula, S1 is the warning deviation value, α i P(B) represents the scoring weight for rockburst warning level i. i ) RP(B) represents the actual probability of rockburst warning level i occurring, i.e., the number of times rockburst warning level i occurs on-site / the total number of rockbursts. i ) represents the probability of rockburst warning level i after removing data processing errors.

[0085] Where -1 ≤ S1 ≤ 1, the closer S1 is to 0, the more accurate and effective the rockburst early warning system is. S1 < 0 indicates insufficient safety in the early warning system. S1 > 0 indicates sufficient safety in the early warning system, but the closer S1 is to 1, the greater the economic losses at the field.

[0086] Evaluation factors for the impact of data processing errors on rockburst early warning results:

[0087] S2=P(B i )-P(B i ) *

[0088] Among them, P(B) i P(B) represents the probability of rockburst warning level i occurring after removing data processing errors; i ) * This represents the probability of a rockburst warning occurring at level i before data processing errors are removed.

[0089] The following example illustrates the rockburst early warning effect evaluation method provided by this invention.

[0090] 1. Statistically calculate the cumulative number of valid events and the probability distribution of cumulative released energy.

[0091] In the early warning process of a rockburst warning system, the most direct result of data processing is the generation of the number of valid events and the cumulative released energy of valid events. The daily number of valid events and the cumulative released energy are also important criteria for judging the rockburst warning results. Therefore, this study selects the probability of the cumulative number of valid events and the cumulative released energy as the research object for quantifying data processing errors. However, the errors that inevitably occur during signal acquisition and screening exhibit strong uncertainty and randomness. Therefore, a method combining historical statistical data and probabilistic statistical models is adopted to quantify the distribution patterns of the cumulative number of valid events and the cumulative released energy. Rockburst warning data collected over one month is shown in Table 3 below:

[0092] Table 3

[0093]

[0094] Based on Table 3, the frequency of each rockburst warning level is statistically analyzed and the probability of its occurrence is calculated:

[0095]

[0096] The statistical and calculation results are shown in Table 4 below:

[0097] Table 4

[0098]

[0099]

[0100] 1.1 Statistical patterns of cumulative valid events.

[0101] To calculate the probability of the daily cumulative number of valid events in different warning intervals since the rockburst early warning system began operation, it is first necessary to collect data on the daily cumulative number of events since the system began operation, define the threshold for the number of valid events for each rockburst level warning, and divide the warning intervals accordingly. Then, the frequency of data points within each interval is calculated and normalized using the total number of days to obtain the probability for each interval. The ratio of the frequency counted in each warning interval to the total number of events is used as the probability of a warning for each rockburst level. The warning threshold standards are shown in Table 5 below.

[0102] Table 5

[0103] Rockburst level Number of microseismic events Cumulative radiant energy (J) Maximum radiant energy (J) No rockburst ＜10 ＜1E+04 ＜5E+03 Minor rock burst [10，25) [1E+04, 1E+05] [5E+03, 1E+04) Moderate rockburst [25，50) [1E+05, 1E+06) [1E+04, 1E+05] Intense rock burst ＞50 ＞1E+06 ＞1E+05 This monitoring section 14 2.12E+05 4.60E+04

[0104]

[0105] The statistical and calculation results are shown in Table 6 below:

[0106] Table 6

[0107]

[0108] 1.2 Statistical laws governing the release of energy from accumulated microseismic events.

[0109] To statistically analyze the probability distribution of daily cumulative energy release since the rockburst early warning system began operation, it is first necessary to collect data on daily cumulative energy release. The collected data is processed, and considering the characteristic of a large number of days with low energy release and a small number of days with high energy release, it is preliminarily determined that the cumulative microseismic energy release follows an exponential distribution, which can be verified using actual data. The threshold ranges for cumulative energy release for each rockburst level are then defined, and energy intervals are divided accordingly. The probability distribution of cumulative microseismic energy release within each warning threshold range can then be calculated, and normalization is used to obtain the probability for each energy interval. The warning threshold standards are shown in Table 7 below.

[0110] Table 7

[0111] Rockburst level Number of microseismic events Cumulative radiant energy (J) Maximum radiant energy (J) No rockburst ＜10 ＜1E+04 ＜5E+03 Minor rock burst [10，25) [1E+04, 1E+05] [5E+03, 1E+04) Moderate rockburst [25，50) [1E+05, 1E+06) [1E+04, 1E+05] Intense rock burst ＞50 ＞1E+06 ＞1E+05 This monitoring section 14 2.12E+05 4.60E+04

[0112]

[0113] The statistical and calculation results are shown in Table 8 below:

[0114] Table 8

[0115]

[0116] 2. Establish an early warning model based on historical early warning data.

[0117] The probability calculations for the daily cumulative number of events and daily cumulative microseismic energy release based on the warning threshold for each rockburst level were completed. Calculations revealed that, generally, the probabilities of the daily cumulative number of events and daily cumulative microseismic energy release based on the warning threshold for each rockburst level deviated from the actual warning probabilities for each rockburst level. Therefore, weighting coefficients were introduced to adjust the model results to approximate the warning results. Since there are four warning levels but only two location weighting coefficients, this means four equations need to solve for two unknown coefficients, constituting an overdetermined system of equations. Therefore, the least squares method was used to optimize the unknown coefficients in the model, obtaining the optimal solution by minimizing the sum of squared prediction errors. After calculating the warning weighting coefficients α and β for the daily cumulative number of events and daily cumulative microseismic energy release, the data-processed rockburst warning probability model was completed. The probability of occurrence for each rockburst warning level is shown in Table 9 below.

[0118] Table 9

[0119]

[0120] Based on the above statistical results, a set of overdetermined equations can be listed:

[0121]

[0122] We obtain the coefficient matrix A and the constant vector b of the overdetermined system of equations:

[0123]

[0124] Solving overdetermined systems of equations using the least squares method:

[0125] x=(A T A) -1 A T b

[0126] Solve for α and β:

[0127] α = 0.504, β = 0.496,

[0128] Using the formula:

[0129] P i =α·P 1i +β·P 2i

[0130] In the formula: P 1iP represents the probability of rockburst warning level i occurring, derived from the cumulative number of microseismic events. 2i The probability of an i-level rockburst occurring is determined by the cumulative microseismic release energy.

[0131] The early warning probabilities after processing the quantified rockburst levels are shown in Table 10 below:

[0132] Table 10

[0133]

[0134] Where i = 1, 2, 3, 4, corresponding to no rockburst, slight rockburst, moderate rockburst, and severe rockburst, respectively.

[0135] 3. Establish a rockburst early warning probability model based on historical early warning data and after removing errors.

[0136] The input to the early warning model based on historical early warning data is processed microseismic data. Therefore, the result of this early warning model is a coupling of the error caused by data processing and the early warning effect of the model. To accurately evaluate the early warning effect of the model, it is necessary to remove the error caused by data processing. Therefore, based on the early warning model of historical early warning data, a two-dimensional confusion matrix is ​​first used to evaluate the accuracy of the processed data, and then Bayesian theory is used to remove the error caused by the processed data.

[0137] 3.1 Evaluation of the accuracy of the cumulative number of microseismic events in the data processing.

[0138] The accuracy of processing the cumulative microseismic event count can be evaluated using a two-dimensional confusion matrix. First, the daily processing error is calculated, including the number of valid events marked as invalid and the number of invalid events marked as valid. In the two-dimensional confusion matrix, P represents valid events and N represents invalid events, and the F1 score is used as the accuracy rate of the processed data. The two-dimensional confusion matrix for processing the cumulative microseismic event count is shown in Table 11 below.

[0139] Table 11

[0140]

[0141] Based on Table 11, calculate the precision, recall, and F1 score for processing the cumulative number of microseismic events. The F1 score represents the precision for processing the cumulative number of microseismic events.

[0142] Accuracy:

[0143] Recall rate:

[0144] F1 score:

[0145] The F1 scores for evaluating the accuracy of processed data for each early warning rockburst level i are shown in Table 12 below.

[0146] Table 11

[0147]

[0148] 3.2. Evaluate the accuracy of processing the accumulated microseismic release energy.

[0149] The accuracy of processing cumulative microseismic release energy can be evaluated using a two-dimensional confusion matrix. Calculating the cumulative microseismic release energy for each rock fracture event is complex and difficult to define accurately. However, the calculation of microseismic release energy for rock fracture events is related to the location of the event; therefore, events located within a valid range, and whose calculated microseismic release energy is within a reasonable range, are considered valid events. First, the daily error in processing microseismic release energy is statistically analyzed, including the number of valid events marked as invalid and the number of invalid events marked as valid. In the two-dimensional confusion matrix, P represents valid events, N represents invalid events, and the F1 score is used as the accuracy rate of the processed data. The two-dimensional confusion matrix for processing cumulative microseismic release energy data is shown in Table 12 below.

[0150] Table 12

[0151]

[0152] Based on Table 12, the precision, recall, and F1 beam splitting of the cumulative microseismic energy release data were calculated. The F1 score represents the precision of the cumulative microseismic energy release data.

[0153] Accuracy:

[0154] Recall rate:

[0155] F1 score:

[0156] The F1 score for evaluating the accuracy of cumulative microseismic energy release after calculation for each early warning rockburst level i is shown in Table 13 below.

[0157] Table 13

[0158]

[0159] 3.3 Eliminate errors in processing the cumulative number of microseismic events.

[0160] Based on the above, the probability P of each level of rockburst can be calculated. 1iThe calculation result is a result of the coupling between errors caused by data processing and the early warning operation of the rockburst early warning system. To evaluate the early warning effectiveness of the rockburst early warning system itself, the influence of these errors needs to be eliminated. Let P... 1i Considering the probability of each rockburst level occurring under the premise of data processing, i.e., the conditional probability P(B) 1i |A1). Using Bayesian theory, the probability P(B) of each level of rockburst occurrence after removing errors can be calculated. 1i ).

[0161]

[0162] In the formula, P(B) 1i |A1) represents the probability of the first warning occurring at rockburst warning level i after the data processing stage; P(A1|B 1i The first accuracy rate in the data processing stage for rockburst early warning level i is ).

[0163] P(B 1i P(A1) represents the probability of the first warning of rockburst warning level i after removing the error in the number of microseismic events during the data processing stage; P(A1) represents the overall accuracy of the number of microseismic events during the data processing stage, as shown in Table 14 below.

[0164] Table 14

[0165]

[0166] By calculating and normalizing, the probability of each level of rockburst warning occurring after removing errors can be obtained, as shown in Table 15 below.

[0167] Table 15

[0168]

[0169] 3.4 Eliminate errors in processing accumulated microseismic energy release.

[0170] Based on the above, the probability P of each level of rockburst can be calculated. 2i The calculation result is a result of the coupling between errors caused by data processing and the early warning operation of the rockburst early warning system. To evaluate the early warning effectiveness of the rockburst early warning system itself, the influence of these errors needs to be eliminated. Let P... 2i Considering the probability of each rockburst level occurring under the premise of data processing, i.e., the conditional probability P(B) 2i |A2). Using Bayesian theory, the probability P(B) of each level of rockburst occurrence after removing errors can be calculated. 2i ).

[0171]

[0172] In the formula, P(B) 2i|A1) represents the probability of the second warning occurring at rockburst warning level i after the data processing stage; P(A2|B 2i The second accuracy rate in the data processing stage for rockburst early warning level i is [data missing].

[0173] P(B 2i P(A2) represents the probability of the second warning of rockburst warning level i after removing the error of microseismic release energy in the data processing stage; P(A2) represents the overall accuracy of the microseismic release energy data in the data processing stage, as shown in Table 16 below.

[0174] Table 16

[0175]

[0176] By calculating and normalizing, the probability of each level of rockburst warning occurring after removing errors can be obtained, as shown in Table 17 below.

[0177] Table 17

[0178]

[0179] 3.5 Eliminate errors in the processing stage.

[0180] By using the aforementioned least squares method to quantify the rockburst early warning results, the cumulative number of microseismic events and the cumulative microseismic release energy, along with their weights α and β in the early warning process, can be used to obtain the probability of rockburst early warning for each level after eliminating errors in the data processing stage.

[0181] Using the formula:

[0182] P(B i )=α·P(B 1i )+β·P(B 2i )

[0183] In the formula: P(B) 1i P(B) represents the probability of a rockburst warning level i, calculated based on the cumulative number of microseismic events after removing errors; 2i The rockburst warning level i, after removing errors, is the probability of occurrence of the warning derived from the accumulated microseismic release energy, P(B). i The table shows the probability of occurrence for each rockburst warning level after removing data processing errors. The resulting warning probabilities after removing data processing errors are shown in Table 18.

[0184] Table 18

[0185]

[0186] 4. Establish evaluation indicators for the early warning effect of the rockburst early warning system.

[0187] After obtaining the rockburst warning probabilities for each level after removing data processing errors, these warning probabilities are compared with the actual probabilities of rockbursts of each level occurring at the construction site, thus establishing an evaluation index for the warning effect of the rockburst warning system.

[0188]

[0189] In the formula, S1 is the deviation value of rockburst early warning effect, and α i P(B) represents the scoring weight for rockburst warning level i. i ) R P(B) represents the actual probability of rockburst warning level i occurring, i.e., the number of times rockburst warning level i occurs on-site / the total number of rockbursts. i ) represents the probability of rockburst warning level i after removing data processing errors.

[0190] Based on on-site feedback, the actual probability of each level of rock burst occurring is shown in Table 19 below.

[0191] Table 19

[0192]

[0193] Before eliminating processing errors, the probability of rockburst warnings at each level is shown in Table 20.

[0194] Table 20

[0195]

[0196] After eliminating data processing errors, the probability of rockburst warnings at each level is shown in Table 21.

[0197] Table 21

[0198]

[0199]

[0200] According to the formula:

[0201]

[0202] The final evaluation index for rockburst early warning effect can be calculated.

[0203] Where -1 ≤ S1 ≤ 1, the closer S1 is to 0, the more accurate and effective the rockburst early warning system is. S1 < 0 indicates insufficient safety in the early warning system. S1 > 0 indicates sufficient safety in the early warning system, but the closer S1 is to 1, the greater the economic losses at the field.

[0204] See Figure 2 The present invention provides a rockburst early warning effect evaluation system, the system comprising:

[0205] The acquisition module 201 is used to acquire the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals.

[0206] The early warning module 202 is used to classify rockburst early warning levels by outputting the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data according to the threshold range of different rockburst early warning levels.

[0207] The first calculation module 203 is used to calculate the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level.

[0208] The second calculation module 204 is used to calculate the probability of occurrence of each rockburst warning level after removing data processing errors, based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage.

[0209] Evaluation module 205 is used to evaluate the effectiveness of rockburst warnings based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level.

[0210] It is understood that the rockburst early warning effect evaluation system provided by the present invention corresponds to the rockburst early warning effect evaluation method provided in the foregoing embodiments. The relevant technical features of the rockburst early warning effect evaluation system can be referred to the relevant technical features of the rockburst early warning effect evaluation method, and will not be repeated here.

[0211] This invention provides a method and system for evaluating the effectiveness of rockburst early warning systems. The method acquires the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period, classifies these data into rockburst early warning levels, calculates the probability of occurrence for each level and the accuracy of the data processing stage, and calculates the probability of occurrence for each level after removing data processing errors. Based on the probability of occurrence for each level after removing data processing errors and the actual probability of occurrence for each level, the effectiveness of the rockburst early warning system is evaluated. This invention proposes a scientific error removal method to address potential errors introduced during the data processing stage of the rockburst early warning system. This method effectively eliminates interference caused by input data errors, allowing the evaluation of the early warning effect to focus more on the system's performance itself, rather than masking or exaggerating the system's actual capabilities due to improper data processing.

[0212] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0213] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0214] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for evaluating the effectiveness of rockburst early warning after eliminating data processing errors, characterized in that, include: The daily cumulative number of microseismic events and the daily cumulative microseismic energy release data are obtained during the warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals. Based on the threshold ranges of different rockburst warning levels, output the rockburst warning level corresponding to the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data; Based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level, the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage are calculated. Based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage, the probability of occurrence of each rockburst warning level after removing data processing errors is calculated. The effectiveness of rockburst warnings is evaluated based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level. The calculation of the probability of occurrence and the accuracy of data processing for each rockburst warning level, based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data, includes: Based on the number of valid events per day for each rockburst warning level, the number of valid events for each rockburst warning level during the warning period is counted, and the ratio of the number of rockburst warnings for each level to the total number of warnings during the warning period is calculated as the probability of the first warning occurring for each rockburst warning level. Based on the daily microseismic energy release data for each rockburst warning level, the microseismic energy release data for each rockburst warning level during the warning period is statistically analyzed. The ratio of the number of warnings for the microseismic energy release data for each rockburst warning level to the total number of warnings for the microseismic energy release data during the warning period is calculated as the probability of the second warning occurring for each rockburst warning level. The calculation of the probability of occurrence and the accuracy of data processing for each rockburst warning level, based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data, includes: The first two-dimensional confusion matrix records the number of microseismic events whose event status after data processing matches the actual event status and the number of microseismic events whose event status does not match the actual event status in the daily cumulative microseismic events; wherein, the event status includes valid event status and invalid event status; The first accuracy rate of the microseismic event count is calculated based on the number of microseismic events whose event status is consistent with the actual event status and the number of microseismic events whose event status is inconsistent after the data processing stage. The second two-dimensional confusion matrix records the number of daily cumulative microseismic energy release data that are consistent with the actual data state and the number of inconsistent data states after the data processing stage; wherein, the data state includes valid data state and invalid data state; The second accuracy rate of the microseismic release energy data processing stage is calculated based on the number of microseismic release energy data that are consistent with the actual data state after the data processing stage and the number of microseismic release energy data that are inconsistent. Based on the first warning occurrence probability and the first accuracy rate, after eliminating the error in the number of microseismic events during the data processing stage, the first warning occurrence probability for each rockburst warning level is calculated. : In the formula, The probability of the first warning occurring at rockburst warning level i after the data processing stage; The first accuracy rate in the data processing stage for rockburst early warning level i; To eliminate the error in the number of microseismic events during the data processing stage, the probability of the first warning of rockburst warning level i is calculated. This represents the overall accuracy of the number of microseismic events during the data processing phase. Based on the second warning occurrence probability and the second accuracy rate, after removing errors in the microseismic energy release data during the data processing stage, the second warning occurrence probability for each rockburst warning level is calculated. ; In the formula, The probability of the second early warning occurring at rockburst early warning level i after the data processing stage; The second accuracy rate in the data processing stage for rockburst early warning level i; To eliminate errors in the microseismic energy release during the data processing stage, the probability of the second warning occurring at rockburst warning level i is calculated. This represents the overall accuracy of the microseismic release energy data during the data processing stage. Based on the probability of the first warning occurring and the probability of the second warning occurring Calculate the probability of occurrence of each rockburst warning level after removing data processing errors. : in, and The probability of the first warning occurring Weighting coefficient and probability of second warning occurrence The weighting coefficients, and The result was obtained by fitting using the least squares method.

2. The method for evaluating the effectiveness of rockburst early warning according to claim 1, characterized in that, The method of outputting the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data corresponding to the rockburst warning level based on the threshold range of different rockburst warning levels includes: Based on the daily cumulative number of microseismic events collected during the early warning period, the daily cumulative number of valid events is calculated. The number of valid events refers to the total number of events where the original microseismic signal is a rock fracture signal. Based on the threshold range into which the daily cumulative valid event count falls, the rockburst warning level corresponding to the daily cumulative valid event count is output, wherein the threshold range of the cumulative valid event count corresponding to each rockburst warning level is obtained through the rockburst warning system.

3. The method for evaluating the effectiveness of rockburst early warning according to claim 1, characterized in that, The method of outputting the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data corresponding to the rockburst warning level based on the threshold range of different rockburst warning levels includes: Based on the threshold range into which the daily cumulative microseismic energy release data falls during the collected warning period, the rockburst warning level corresponding to the daily cumulative microseismic energy release data is output. The threshold range of the cumulative microseismic energy release data corresponding to each rockburst warning level is obtained through the rockburst warning system.

4. The method for evaluating the effectiveness of rockburst early warning according to claim 1, characterized in that, The calculation of the probability of occurrence of each rockburst warning level after removing data processing errors, based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage, includes: Based on the first warning occurrence probability and the first accuracy rate, after eliminating the error in the number of microseismic events during the data processing stage, the first warning occurrence probability for each rockburst warning level is calculated. ; Based on the second warning occurrence probability and the second accuracy rate, after removing errors in the microseismic energy release data during the data processing stage, the second warning occurrence probability for each rockburst warning level is calculated. ; Based on the probability of the first warning occurring and the probability of the second warning occurring Calculate the probability of occurrence of each rockburst warning level after removing data processing errors. .

5. The method for evaluating the effectiveness of rockburst early warning according to claim 1, characterized in that, The evaluation of rockburst warning effectiveness is based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level, including: In the formula, S1 is the rockburst early warning deviation value. The scoring weight for rockburst warning level i, This represents the actual probability of rockburst warning level i occurring, which is the number of times rockburst warning level i occurred on-site divided by the total number of rockburst occurrences. The probability of rockburst warning level i after removing data processing errors.

6. The method for evaluating the effectiveness of rockburst early warning according to claim 5, characterized in that, The evaluation of rockburst warning effectiveness, based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level, also includes calculating the impact evaluation factor of data processing errors on rockburst warning results. in, The probability of rockburst warning level i after removing data processing errors; This represents the probability of a rockburst warning occurring at level i before data processing errors are removed.

7. A rockburst early warning effect evaluation system, applied to the rockburst early warning effect evaluation method described in claim 1 after removing data processing errors, characterized in that, include: The acquisition module is used to acquire the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data during the warning period. The cumulative number of microseismic events and the cumulative microseismic energy release data are obtained by data processing of the original microseismic signals. The early warning module is used to output the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data to determine the rockburst early warning level based on the threshold range of different rockburst early warning levels. The first calculation module is used to calculate the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage based on the daily cumulative number of microseismic events and the daily cumulative microseismic energy release data for each rockburst warning level. The second calculation module is used to calculate the probability of occurrence of each rockburst warning level after removing data processing errors, based on the probability of occurrence of each rockburst warning level and the accuracy of the data processing stage. The evaluation module is used to evaluate the effectiveness of rockburst warnings based on the probability of occurrence of each rockburst warning level after removing data processing errors and the actual probability of occurrence of each rockburst warning level.

Images