Mine water disaster intelligent early warning system and method

Abstract

The application discloses a mine water disaster intelligent early warning system and method, and relates to the technical field of mine water disaster monitoring. The early warning method comprises water inrush risk early warning, mine flooding risk early warning and water environment risk early warning. The water inrush risk early warning processes water inrush amount time series data through a sliding window, and triggers a three-level water inrush risk early warning for three times of abnormal excitation, and can upgrade the early warning and predict the water inrush amount. The mine flooding risk early warning sets three threshold values based on the double ratios of water inrush amount, drainage capacity and water sump volume. The water environment risk early warning monitors water quality according to the national standard, and performs early warning according to a pollution risk index in three levels. The early warning system comprises monitoring, data management, water disaster early warning and visualization units. The application can significantly improve the accuracy, timeliness and intelligent level of water disaster early warning through multi-parameter collaborative monitoring, intelligent analysis and expert judgment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine water disaster monitoring, and particularly relates to a mine water disaster intelligent early warning system and method. BACKGROUND

[0002] Mine water disaster is one of the major threats to the safety production of coal mines and metal mines, and water inrush accidents often cause serious casualties and economic losses. The traditional mine water disaster monitoring technology mainly relies on manual periodic data collection and experience judgment, and has problems such as response lag and insufficient early warning accuracy. In recent years, with the development of artificial intelligence technology, some research attempts to use machine learning, deep learning and other algorithms to build water inrush prediction models, and uses the advantages of data processing, model analysis and intelligent decision-making to improve the accuracy, real-time performance and reliability of prediction. For example, the Chinese patent application with the publication number CN117688873A discloses a mine water inrush point and simulation model parameter identification method based on artificial intelligence, which generates training sample data set and test sample data set of the substitute model by using numerical model and parameter prior information, and constructs and trains the neural network of the substitute model. The Chinese patent application with the publication number CN117313987A discloses a coal mine water inrush prediction method based on a gated recurrent neural network, which can learn the change law of a series of sample data and the influence of such change law on water inrush, thereby improving the correctness and reliability of prediction. However, in actual application, the above methods have the following defects: 1. High data dependency and insufficient data, artificial intelligence models need a large amount of historical data for training, but mine water inrush events are small probability events, and effective samples are scarce, resulting in poor model generalization ability; 2. The black box characteristics of deep machine learning models result in unclear mechanism, the decision-making process of artificial intelligence models lacks explainability, and cannot be associated with hydrogeological mechanism, so that the early warning result is difficult to be trusted by engineering personnel and guide water disaster rescue; 3. Early warning function failure, existing systems are mostly limited to data monitoring level (replacing early warning with monitoring), or use static threshold alarm, without considering the dynamic change characteristics of water inrush, and often have false alarms and missed alarms, such as invalid alarms triggered by short-term flow fluctuations, or failure to identify real gradual water inrush risks. SUMMARY

[0003] The present application aims to provide a mine water disaster intelligent early warning system and method, which is based on expert experience, water control technology standards and underground water scientific theories and methods, uses mine water inflow as the main monitoring data, uses groundwater level, precipitation, ground subsidence and water chemistry as response data, constructs a multi-source data fusion multi-level water disaster evaluation model, defines mine water disaster from three aspects of mine water inrush risk, mine flooding risk and mine water environment risk, and can provide practical guidance for mine water disaster prevention, has strong operability and high early warning reliability.

[0004] Based on the above invention purpose, the present application adopts the following technical scheme: in the first aspect, the present application proposes a mine water disaster intelligent early warning method, the water disaster intelligent early warning method includes water inrush risk early warning, the water inrush risk early warning includes the following steps: according to the set fixed time length, the water inflow in the monitoring area of the mine is collected, and the water inflow time series data is formed;Set a time window with a fixed width, calculate the average value and standard deviation of the water inflow time series data in the time window;The fixed width is N times of the fixed time length;Adopt sliding window mechanism, when the next new water inflow data is sequentially added, the oldest water inflow data in the window is removed;With the time window sliding forward, the time series data and the average value and standard deviation in the time window are updated sequentially;When the value of the next new water inflow data exceeds the average water inflow in the current time window plus 2 times or 3 times standard deviation, it is determined that the new water inflow data is the first abnormal water inflow;The subsequent sliding data is also compared with the average water inflow in the current time window plus 2 times or 3 times standard deviation to determine whether it is abnormal water inflow;If three or more abnormal water inflows occur continuously, it is determined that the continuous abnormality is water inrush risk abnormality, and the previous water inflow of the first abnormal water inflow is taken as the water inrush early warning starting point to trigger the risk early warning based on water inrush amount;The water inrush amount is the difference between the single abnormal water inflow and the water inflow at the water inrush early warning starting point.

[0005] As a further improvement of the present application, the specific rules of the risk early warning based on water inrush amount are as follows: the water inrush risk early warning adopts a three-level progressive threshold design, threshold I < threshold II < threshold III;When the water inrush amount satisfies: threshold I ≤ water inrush amount < threshold II, output mild water inrush risk early warning;When the water inrush amount satisfies: threshold II ≤ water inrush amount < threshold III, output moderate water inrush risk early warning;When the water inrush amount is greater than or equal to threshold III, output severe water inrush risk early warning;If three continuous abnormalities are all below threshold I, only monitoring is needed, and no water inrush risk early warning is triggered;If the abnormal water inflow is less than three times continuously, it is determined that the abnormal water inflow is only accidental abnormality or isolated abnormality, and no water inrush risk early warning is triggered;If the abnormal water inflow is less than three times continuously, it is determined that the abnormal water inflow is only accidental abnormality or isolated abnormality, and no water inrush risk early warning is triggered.

[0006] As a further improvement of the present application, after the water inrush risk early warning is started, the water inrush amount at the subsequent three time points is predicted, and the water inrush risk early warning is carried out according to the predicted water inrush amount.

[0007] As a further improvement of the present application, the water inrush risk early warning further comprises a synchronous monitoring and early warning upgrade mechanism for response data, specifically as follows: monitoring and synchronously acquiring response data related to water inrush risk, the response data types including groundwater level, precipitation, ground subsidence and water chemical index value; the response data are processed by using the same sliding window mechanism as the water gushing quantity monitoring, specifically as follows: setting a time window with the same fixed width as the water gushing quantity monitoring, the time window being composed of continuously monitored response data; when the time window slides into new response data, the oldest response data in the current window is removed to maintain the constant width of the time window; if the newly slid-in response data is greater than the average value of the response data in the current time window plus 2 times or 3 times standard deviation, it is determined that the newly slid-in response data constitutes abnormal data of the response data in the current time window; if the subsequent slid-in data constitutes continuous three times or more abnormal data, it is determined that the abnormal data belongs to abnormal response synchronous with the water gushing quantity, and the water inrush risk early warning level is raised in combination with the response amplitude of the abnormal data, i.e. the original mild water inrush risk early warning is upgraded to moderate water inrush risk early warning, the original moderate water inrush risk early warning is upgraded to severe water inrush risk early warning, and the original severe water inrush risk early warning remains unchanged; at the same time, after judging that the response data has synchronous abnormal response, the water inrush quantity at the subsequent three time points is predicted based on the response relationship between the water inrush quantity and the abnormal response data, and the water inrush risk early warning is performed according to the predicted water inrush quantity.

[0008] As a further improvement of the present application, the water inrush risk early warning further comprises a synchronous monitoring and early warning upgrade mechanism for response data, specifically as follows: monitoring and synchronously acquiring response data related to water inrush risk, the response data types including groundwater level, precipitation, ground subsidence and water chemical index value; the response data are processed by using the same sliding window mechanism as the water gushing quantity monitoring, specifically as follows: setting a time window with the same fixed width as the water gushing quantity monitoring, the time window being composed of continuously monitored response data; when the time window slides into new response data, the oldest response data in the current window is removed to maintain the constant width of the time window; if the newly slid-in response data is greater than the average value of the response data in the current time window plus 2 times or 3 times standard deviation, it is determined that the newly slid-in response data constitutes abnormal data of the response data in the current time window; if the subsequent slid-in data constitutes continuous three times or more abnormal data, it is determined that the abnormal data belongs to abnormal response synchronous with the water gushing quantity, and the water inrush risk early warning level is raised in combination with the response amplitude of the abnormal data, i.e. the original mild water inrush risk early warning is upgraded to moderate water inrush risk early warning, the original moderate water inrush risk early warning is upgraded to severe water inrush risk early warning, and the original severe water inrush risk early warning remains unchanged; at the same time, after judging that the response data has synchronous abnormal response, the water inrush quantity at the subsequent three time points is predicted based on the response relationship between the water inrush quantity and the abnormal response data, and the water inrush risk early warning is performed according to the predicted water inrush quantity. . , respectively, to set three-level early warning, output the water inrush risk, and form a double water inrush risk early warning mechanism.

[0009] As a further improvement of the present application, the water disaster intelligent early warning method further comprises a water environment risk early warning, which comprises the following steps: determining a water pollutant control project for monitoring mine discharge water and a maximum allowable discharge concentration standard, monitoring the mine discharge water quality index in real time online according to a specified position, time period and frequency, and sampling and testing the mine discharge water quality index according to a specified sampling frequency, the water quality index including conventional physicochemical indexes and water pollution discharge control indexes; analyzing the change trend and abnormal value of the discharge water quality index in real time, calculating the daily average discharge concentration and the annual total discharge amount; defining a pollution risk index γ as the ratio of the daily average discharge concentration of the discharge water pollutant to the maximum allowable discharge concentration, and when the ratio > 1, it indicates that the discharge standard has been exceeded and there is a pollution risk; based on the calculated value of the pollution risk index, the water pollution risk early warning level is divided, when the pollution risk index satisfies: 0.8≤γ<0.9, a potential pollution risk early warning is output, when the pollution risk index satisfies: 0.9≤γ<1.0, a critical pollution risk early warning is output, and when the pollution risk index γ≥1.0, an over-standard pollution risk early warning is output.

[0010] As a further improvement of the present application, the mine monitoring area is the entire mine, panel or working face.

[0011] In a second aspect, the present application further provides a mine water disaster intelligent early warning system, which comprises: a monitoring unit, the monitoring unit is used for tracking and monitoring the data including water inflow, underground water level and water quality, mine water chemistry, mine water pressure relief, atmospheric precipitation, ground subsidence, drainage system and water gate data; a data management unit, the data management unit is used for storing, cleaning, analyzing, calculating and outputting the tracking and monitoring data in the monitoring unit; a water disaster early warning unit, the water disaster early warning unit comprises a water inrush risk early warning module, a flooded mine risk early warning module and a water environment risk early warning module, and executes a mine water disaster intelligent early warning method.

[0012] As a further improvement of the present application, it further comprises a visualization unit, which sets a standardized functional drawing in the field of water prevention and control, and centrally displays the mine water filling state, water filling characteristics and water disaster early warning results; and sets a mine water filling three-dimensional digital twin model, which is used for real-time display of the hardware environment, running state and early warning results of the monitoring unit.

[0013] As a further improvement of the present application, the monitoring data of the monitoring unit is divided into real-time online input and manual input, the manual input includes system terminal direct input, batch input, and underground personnel uploading water inrush point image data and manually measuring data through mobile phone software, and the manual input data and real-time monitoring data are fused and analyzed.

[0014] As a further improvement of the present application, the early warning information display and push mode of the water disaster early warning unit includes system interface display, directional mobile phone short message push and directional electronic mail sending.

[0015] Compared with the prior art, the present application has the following technical effects: (1) The water inrush risk early warning method of the present application is driven by real-time monitoring data, realizes dynamic monitoring and early warning of water disasters, realizes dynamic trend and anomaly analysis by real-time monitoring of water inflow data, intelligent dynamic data processing, calculation of average water inflow by sliding window mechanism, improves the timeliness of early warning; the sliding window mechanism eliminates old data, keeps the time span constant, avoids false alarms caused by single-point data fluctuations, and improves the accuracy of early warning; real-time updating of average water inflow and standard deviation eliminates seasonal fluctuation interference and can accurately capture real anomalies; a double filtering mechanism is adopted: first, determine whether the water inflow is within the abnormal range, and then further identify whether it constitutes continuous abnormal water inflow, so as to identify real hydrogeological process anomalies and prevent water engineering significance anomalies, and improve the sensitivity and integrity of early warning.

[0016] (2) After the water inrush risk early warning is started, the water inrush amount of the subsequent three time points can be predicted, compared with relying only on real-time monitoring data, short-term prediction can reveal the subsequent water inrush trend in advance, and discover that the water inrush amount will break through the drainage capacity threshold within a few hours, which can provide sufficient response time for preventing secondary disasters such as flooded wells, and can guide the flexible deployment of drainage equipment and adjustment of personnel on duty key areas on site, avoiding excessive investment or insufficient allocation of emergency resources.

[0017] (3) The present application carries out intelligent multi-source data fusion analysis, the water inrush flow data is combined with underground water level data, precipitation, ground subsidence and water chemical index value data, and synchronous analysis is carried out through a unified sliding window mechanism, which can improve the comprehensiveness and accuracy of early warning, avoid the limitations of single data source, reduce false positives and false negatives; the water disaster mechanism and early warning mechanism of multi-source data anomaly analysis and data fusion model are clear and simple, the model and calculation are simple, the complexity of big data is avoided, the early warning result is good in interpretation, and the reliability and operability of early warning are improved.

[0018] (4) The present application can dynamically evaluate the drainage capacity, update the mine drainage capacity in real time, prevent flooded well accidents caused by insufficient drainage; by comparing the water inflow and the predicted water inrush amount with the drainage capacity, setting the flooded well risk threshold, discovering potential flooded well risks in advance, and optimizing drainage scheduling; at the same time, by comparing the water sump volume calculated by the standard empirical formula with the actual water sump volume, setting the flooded well risk threshold caused by water sump, a water disaster flooding early warning system is constructed from two dimensions of dynamic drainage capacity and static storage capacity, automatic hierarchical decision-making is implemented, and the system has obvious scientificity and practicality.

[0019] (5) The present application can monitor water quality compliance, set water quality early warning thresholds based on national standards, ensure that the effluent water meets environmental protection requirements, avoid pollution accidents, and through water quality trend analysis, potential pollution risks can be found in advance to assist decision-making.

[0020] (6) The present application adopts a hierarchical early warning management mode, and differentiates monitoring for different regions such as mines, panels and working faces, effectively improving the pertinence of early warning; the water inflow collection frequency (fixed time length) can be flexibly adjusted according to the size of the monitoring area, enhancing the operability of early warning; at the same time, the water inflow collection frequency is correspondingly reduced after the water inrush risk early warning is started, so as to improve the timeliness of data collection and increase the early warning response speed.

[0021] (7) The early warning system of the present application is visualized and interactive, the hydrogeological map can intuitively show the water filling state, such as water inflow and groundwater level curve, the three-dimensional digital twin model simulates the mine environment in real time, and the hardware environment, running state and early warning results are monitored in real time, which assists personnel to quickly locate the risk point.

[0022] (8) The early warning system of the present application can realize intelligent man-machine collaborative input, support software manual uploading of data such as water inrush point image data, make up for the blind area of automatic monitoring, and improve the data integrity.

[0023] (9) The early warning system of the present application forms a closed-loop management from data collection, analysis to visualization and early warning, and improves the intelligent prevention and control ability of mine water disaster. BRIEF DESCRIPTION OF DRAWINGS

[0024] The present application will be further described in detail below in combination with the drawings and specific embodiments.

[0025] Figure 1 is a mine water disaster intelligent early warning method flow chart of the present application.

[0026] Figure 2 is a water inrush risk early warning dynamic monitoring mechanism principle diagram in the mine water disaster intelligent early warning method of the present application.

[0027] Figure 3 is a water inrush risk early warning hierarchical early warning schematic diagram in the mine water disaster intelligent early warning method of the present application.

[0028] Figure 4 is a flooded mine risk early warning schematic diagram in the mine water disaster intelligent early warning method of the present application.

[0029] Figure 5 is a water environment risk early warning schematic diagram in the mine water disaster intelligent early warning method of the present application.

[0030] Figure 6 is a mine water disaster intelligent early warning system structure diagram of the present application. DETAILED DESCRIPTION

[0031] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art belong to the scope protected by the present application.

[0032] Mine water inrush refers to the phenomenon that groundwater or surface water continuously and slowly seeps into the mine space during the mining process; mine water inrush refers to the phenomenon that groundwater or surface water suddenly and rapidly inrushes into the mine through water-conducting structures, which has instantaneous and catastrophic nature, and its direct reflection is the abnormal inrush of water in a short time.

[0033] Please refer to Figure 1 The present application proposes a mine water disaster intelligent early warning method. The water disaster of the present application is divided into three types, i.e. water inrush, mine flooding and water environment, and the corresponding early warnings are water inrush risk early warning, mine flooding risk early warning and water environment risk early warning. The mechanisms of the three types of water disaster risk early warning will be introduced in detail below.

[0034] For the first type of water disaster risk early warning, the water inrush risk early warning includes the following steps: real-time monitoring and obtaining the inrush flow time series data in the mine monitoring area, collecting the inrush flow in the mine monitoring area according to the set fixed time length. The mine monitoring area can be the entire mine, a panel or a working face according to the monitoring range, and the fixed time length can be flexibly set, such as 1 day (such as the entire mine), 4 hours (such as a panel) or 1 hour (such as a working face) for data collection, to form the time series data of inrush flow, so as to increase the flexibility and timeliness of inrush flow data collection.

[0035] A time window with a fixed width is set, and the average value (i.e. average inrush flow) and 2 or 3 times standard deviation of the time series data of inrush flow in the time window are calculated. The fixed width is N times of the fixed time length, and N is an integer between 7 and 10. A sliding window mechanism is adopted, and when a new inrush flow data slides in, the oldest inrush flow data in the window is removed. With the sliding of the time window, the time series data and its average value and standard deviation are updated continuously.

[0036] When a newly-included water inflow data exceeds the average water inflow in the current time window plus 2 or 3 times the standard deviation, it is determined that the data is an abnormal water inflow data, and the subsequent sliding data is compared with the average water inflow in the current time window plus 2 or 3 times the standard deviation to determine whether it is an abnormal water inflow. If three or more abnormal water inflows occur continuously, the water inflow before the first abnormal water inflow is taken as the water inrush early warning starting point, and the water inrush risk early warning based on the water inrush amount of the water inrush early warning starting point is triggered. The water inrush amount is the difference between the single abnormal water inflow and the water inflow of the water inrush early warning starting point. The water inrush risk early warning has three progressive water inrush risk early warning thresholds: threshold I < threshold II < threshold III, with the unit being m³ / h. When the water inrush amount satisfies: threshold I ≤ water inrush amount < threshold II, a mild water inrush risk early warning is output. When the water inrush amount satisfies: threshold II ≤ water inrush amount < threshold III, a moderate water inrush risk early warning is output. When the water inrush amount ≥ threshold III, a severe water inrush risk early warning is output. It should be noted that when the water inrush risk early warning is started, the fixed time should be shortened accordingly to increase the frequency of water inflow data collection, such as from the original four hours of collection once to two hours or one hour of collection once, to improve the data effectiveness and strengthen the risk response speed.

[0037] Please refer to Figure 2 and Figure 3 The above method is described below by way of example: a fixed-width time window is set, such as a time length of one day, N = 7, and the width of the time window is 7. It is assumed that from the first day to the seventh day of the beginning of the month as the sliding window starting point, there are 7 days of 7 water inflow data in the time window, and the average water inflow and the standard deviation are calculated based on the 7 data. Using the sliding window mechanism, when the latest water inflow data is added every day, the oldest water inflow data in the previous day window is removed, and the window time span is maintained constant, i.e. when the time reaches the eighth day, the water inflow data of the first day is removed, and then the average water inflow and the standard deviation of the 7 days from the second day to the eighth day are calculated. Figure 2In the middle, time window 1, time window 2, time window 3 and time window 4 represent the process of the same time window sliding forward. At the same time, whether the average water inflow of the previous day exceeds the average water inflow of the previous day plus 3 times the standard deviation is used as the judgment standard to identify whether the latest water inflow constitutes the abnormal water inflow of the previous day. The standard deviation is an index for measuring the degree of deviation of the overall data from the mean value. The average water inflow plus 3 times the standard deviation is used as the upper limit value of the average water inflow of each day, and the sliding time window updates the daily water inflow, the average water inflow and the upper limit value of the average water inflow in real time. Assuming that the water inflow of the 8th day exceeds the average water inflow of the previous 7 days plus 3 times the standard deviation, the water inflow is marked as an abnormal water inflow, and when the water inflow of the 9th day also exceeds the average water inflow of the previous 7 days plus 3 times the standard deviation, and the water inflow of the 10th day also exceeds the average water inflow of the previous 7 days plus 3 times the standard deviation, it is determined that there are three consecutive abnormal water inflows. At this time, the water inflow of the 7th day is used as the starting point of the water inrush warning, the water inrush is calculated and marked, and the water inrush risk warning is triggered. The water inrush is the difference between the abnormal water inflow and the water inflow of the starting point of the water inrush warning, i.e. the water inflow of the 8th day minus the water inflow of the 7th day, forming a water inrush data, the water inflow of the 9th day minus the water inflow of the 7th day, forming a water inrush data, and the water inflow of the 10th day minus the water inflow of the 7th day, forming a water inrush data. The water inrush risk warning is set to three thresholds, i.e. threshold I, threshold II and threshold III, the three thresholds increase in turn, for example, threshold I is 60 m 3 / h, threshold II is 120 m 3 / h, and threshold III is 200 m 3 / h. The water inrush of the abnormal water inflow produced by the 8th, 9th and 10th days is judged by threshold, when the water inrush is greater than or equal to 60 m 3 / h and less than 120 m 3 / h, it is set to a mild water inrush warning, and a yellow warning mark is used for warning identification; when the water inrush is greater than or equal to 120 m 3 / h and less than 200 m 3 / h, it is set to a moderate water inrush warning, and an orange warning mark is used for warning identification; when the water inrush is greater than or equal to 200 m 3 / h, it is set to a severe water inrush warning, and a red warning mark is used for warning identification, i.e. the 8th, 9th and 10th days will produce three different or same risk warning marks. The determination of the three approaching thresholds should be based on the mine hydrogeological conditions, water control technical standards, expert experience and the degree of water inrush disaster.

[0038] If there are three consecutive abnormalities, but all are below threshold I, only monitoring is needed, and no risk warning is triggered. When the abnormal water inflow is less than three times in a row, it is determined to be an accidental event, and no risk warning is triggered, but it also needs to be recorded or marked.

[0039] After the flood inrush risk warning is activated, the flood inrush volume at three subsequent time points is predicted. The early stage of the flood inrush is the core stage of the formation and expansion of the flood inrush channel and the surge in flood volume, which is the most dangerous. Predicting the flood inrush volume at the subsequent three time points allows us to focus on this critical stage, understand the growth rate and magnitude of the flood volume in advance, and avoid the risk prediction delay caused by an excessively long prediction period. The specific prediction method includes the following steps: Step 1: Determine the starting point of the flood inrush warning: Mark the time t corresponding to the previous flood volume before the initial abnormal flood volume. 起 With traffic Q 起 This serves as the starting point for flood inrush warning. The second step involves real-time calculation of core parameters: During the abnormal flood inrush phase, the flood inrush time t is calculated in real-time based on the monitoring time t and the inrush volume Q. 突 With the amount of water inrush Q 突 , where t 突 =t- t 起 Q 突 = Q - Q 起 And draw and update with t 突 Q is the x-coordinate 突 The scatter plot shows the water inrush process on the ordinate. The third step is to construct a linear model: using a linear regression tool, specifically Python's statsmodels, to perform a regression analysis on (Q... 突 , t 突 Perform regression analysis on the data, outputting regression coefficients a and b. Simultaneously calculate and output model validation indices: the t-test p-value of the regression coefficients, the F-test p-value of the regression model, and the goodness-of-fit R². If the p-value < 0.05 and R² > 0.6, the linear model is determined to be Q. 突 =at 突 + b. Step 4: Prediction and Visualization: Based on the established primary model, predict the inrush volume (Q) at the subsequent three time points. 突 , t 突 ), and in (Q 突 , t 突 The scatter plot displays the predicted water inrush volume data synchronously using different legends. Step 5: Model Error Judgment and Update: When new monitoring data enters the time window, calculate the average relative error (MRE) between the actual water inrush volume and the previously predicted value. If MRE < 20%, continue using the current model prediction, and synchronously output the current and predicted water inrush risk warning levels; if MRE ≥ 20%, recalculate (Q... 突 , t 突 Perform linear regression, repeat the model testing process in step 3, and establish an updated primary model; if the updated model MRE < 20%, then use the same model to continue prediction and output the warning level simultaneously.

[0040] Mine water inrush emergency needs rapid response, only to predict the next three time points, can be short cycle prediction results as the basis, for the site whether to start drainage system upgrade, personnel evacuation, water gate closed and other emergency measures to provide direct data support, to avoid the uncertainty of over extrapolation to affect the decision efficiency. Short-term prediction can quickly pass through the actual water inrush of the new time window, calculate the error of prediction value such as MRE, verify whether the current regression model is suitable for the current water inrush process, provide the basis for judging whether the model needs to be updated, and ensure the reliability of the prediction method. The risk warning of the predicted water inflow refers to the above rules.

[0041] Mine water inflow is related to groundwater, surface water and other factors, so the change of water inflow is also related to the change of groundwater level, precipitation and ground subsidence. Abnormal groundwater level is often an important signal of increased water inrush risk. When there is more rainfall or the aquifer recharge increases, the groundwater level will rise. If the water level rises close to or exceeds the safety warning line, it may cause water inrush accident. A lot of precipitation may make surface water seep into the ground through fissures, faults and other channels, increase the water inflow of mine, and thus increase the water inrush risk. Especially in a short period of heavy rainfall, the impact on the mine is more significant. During monitoring, the above data are analyzed synchronously with the water inflow data to determine whether they are synchronous abnormal response, and the water inrush risk warning level is adjusted according to the response factor type and response amplitude. The specific method is as follows: obtain the response data related to water inrush risk, including groundwater level, precipitation, ground subsidence and water chemical index value. The response data are processed by the same sliding window mechanism as the water inflow monitoring, which is as follows: set the same fixed width time window as the water inflow monitoring, which is composed of continuous daily records of response data; adopt sliding window mechanism, when the new response data is slid in every day, the oldest response data in the previous day window is removed, and the window width of the time window is kept constant. According to the average value and standard deviation of the response data in the previous day time window, it is identified whether the latest response data constitutes abnormal data of the response data in the previous day time window; if the response data appears continuous abnormality for three times or more in the time window, it is determined that the response data appears continuous abnormality.

[0042] When the water inrush risk warning is triggered, the response data of groundwater level, precipitation, ground subsidence and water chemistry is analyzed synchronously to determine whether continuous anomalies also occur. The correlation between the anomalies of the response data and the water inrush volume is further determined (using Pearson correlation coefficient, the core function of Pearson correlation coefficient is to quantify the strength and direction of the linear correlation between two continuous variables) to determine whether it is a synchronous and correlated abnormal response. If it is a synchronous and correlated abnormal response, the water inrush risk warning level is raised according to the response data type and response amplitude, i.e. the original mild water inrush warning is upgraded to moderate water inrush warning, the original moderate water inrush warning is upgraded to severe water inrush warning, and the original severe water inrush warning remains unchanged. At the same time, after determining that the response data has a synchronous abnormal response, a multivariate correlation prediction model of water inrush volume and related synchronous anomalies is established in real time based on the relationship between water inrush volume and response data anomalies, such as using stepwise regression method, to predict the water inrush volume at the next three time points and to conduct water inrush risk warning according to the predicted water inrush volume. If there is no synchronous anomaly, it is determined whether to maintain or adjust the warning level according to the specific response characteristics.

[0043] After the water inrush risk warning occurs, the well flooding risk triggered by water inrush is further determined.

[0044] For the second water hazard risk warning, the well flooding risk warning includes the following steps: real-time acquisition of water inrush volume data in the monitored area of the mine; when the water inrush risk warning occurs, trigger the water inrush well flooding risk determination, and simultaneously acquire the water inrush volume prediction value and the water inrush volume prediction value in real time; acquire the operation of the drainage system equipment and the water storage capacity, and update the drainage capacity of the monitored area of the mine in real time. For example, the real-time operation data of the drainage system is synchronously collected, including the number of water pumps in use, the rated flow of a single pump (m³ / h), the actual effective drainage flow (considering pipeline loss, calculated at 90% of the rated flow), the standby pump state (whether it can be started immediately), and equipment failure information (such as shutdown, overload). The total volume and current water storage data of the existing water storage are collected in real time to obtain the effective water storage capacity.

[0045] Please refer to Figure 4 , according to the ratio α of the water inrush volume (including water inrush volume prediction) in the monitored area of the mine obtained by real-time monitoring and the real-time drainage capacity, three-level early warning is set according to 0.8 times, 0.9 times and 1.0 of α, and well flooding risk is output. For example, the maximum drainage capacity of a certain mine is 600 m 3 / h, it is set that when the ratio α of the water inrush volume and the drainage capacity is in the interval of 0.8-0.9, i.e. 480 m 3 / h-540 m 3 / h, a first warning (yellow) is issued, and the monitored area has a potential risk of flooding; when α is in the interval of 0.9-1.0, i.e. 540 m 3 / h-600 m 3When the ratio of the dynamic drainage capacity to the static water storage capacity of the water sump is greater than 0.8, a first warning (green) is issued, indicating that the water sump is in a normal state; when the ratio is between 0.8 and 1.0, a second warning (orange) is issued, indicating that the water sump is in a critical state of being flooded; when the ratio is greater than 1.0, a third warning (red) is issued, indicating that the drainage capacity is exceeded, and there is a risk of flooding the well, requiring the highest level of warning, and the staff is required to take emergency measures, such as shutting down the power supply in some areas, organizing personnel to evacuate, etc. 3 When the ratio of the dynamic drainage capacity to the static water storage capacity of the water sump is greater than 1.0, a third warning (red) is issued, indicating that the drainage capacity is exceeded, and there is a risk of flooding the well, requiring the highest level of warning, and the staff is required to take emergency measures, such as shutting down the power supply in some areas, organizing personnel to evacuate, etc.

[0046] At the same time, the ratio of the water sump volume calculated by the standard empirical formula to the existing water sump volume is , and the three-stage warning is set to 0.8 times, 0.9 times, and 1.0, respectively, to output the water sump flooding risk. Specifically, the standard empirical formula is as follows: when the normal water inflow is less than 1000m 3 / h, the standard empirical formula is V=8×Q, where V is the effective volume of the water sump, and Q is the water inflow; when the normal water inflow is greater than 1000m 3 / h, the standard empirical formula is V=2×(Q+3000). In actual operation, each mine can adaptively select the standard empirical formula according to different technical specifications.

[0047] The present application can greatly improve the accuracy and timeliness of the flooding risk warning by monitoring the dynamic drainage capacity and the static water storage capacity of the water sump.

[0048] For the third water hazard risk warning, the water environment risk warning specifically includes the following steps: real-time analysis of the change trend and abnormality of the monitored discharge water quality index, determination of the monitoring mine discharge water water pollutant control project and the highest allowable discharge concentration standard according to the "Integrated Wastewater Discharge Standard" (GB 8978) or local environmental protection requirements, real-time online monitoring and sampling testing of mine discharge water quality index at the specified location, time period and frequency, water quality index including conventional physicochemical index and water pollution discharge control index, real-time analysis of the change trend and abnormal value of the discharge water monitoring and testing index, calculation of the daily average discharge concentration and the annual total discharge.

[0049] Please refer to Figure 5 , the pollution risk index γ is defined as the ratio of the daily average discharge concentration of the discharge water pollutant to the highest allowable discharge concentration. When the ratio is greater than 1, it indicates that the discharge standard has been exceeded, and there is a pollution risk. Based on the pollution risk index calculation value, the water pollution risk warning level is divided. When the pollution risk index γ satisfies: 0.8≤γ<0.9, a potential pollution risk warning is output; when the pollution risk index γ satisfies: 0.9≤γ<1.0, a critical pollution risk warning is output; when the pollution risk index γ≥1.0, an over-standard pollution risk warning is output.

[0050] The mine monitoring area in the embodiment is the entire mine, and can also be a mining working face or a panel. According to the actual situation of the mine area, special area early warning is performed, so that the flexibility of water disaster early warning arrangement is improved.

[0051] Please refer to Figure 6 The application further provides a mine water disaster intelligent early warning system, which comprises a monitoring unit, a data management unit, a water disaster early warning unit and a visualization unit. The monitoring unit is used for tracking and monitoring water inflow, underground water, water chemistry, mine drainage pressure reduction, precipitation, ground subsidence and drainage and water gate. The data management unit is used for storing, cleaning, analyzing, calculating and outputting the tracking and monitoring data in the monitoring unit. The water disaster early warning unit comprises a water inrush risk early warning module, a water inrush hazard early warning module and a water environment risk early warning module, and the water disaster early warning unit executes the mine water disaster intelligent early warning method. The visualization unit comprises a hydrogeological map and a three-dimensional digital twin model used for centrally displaying the water filling state and water filling characteristics. The hydrogeological map comprises the relationship curves of mine water inflow and related underground water level, mining area, roadway length, precipitation and ground subsidence, is a core visualization map of the water disaster early warning system, and is used for displaying the water disaster monitoring state in real time and analyzing the mine water disaster risk. The three-dimensional digital twin model is used for displaying the hardware environment and operating state of the monitoring unit in real time. The visualization unit has a man-machine interaction function and a manual intervention window, and is used for eliminating accidental abnormalities, manually marking abnormal reasons or water inrush reasons, manually intervening in termination of a water inrush process, manually adding water inrush response measures and the like. Meanwhile, a management personnel can view data details in a specific time period, enlarge or reduce a graph by means of mouse clicking and dragging, so as to more carefully analyze data changes. In addition, the visualization unit sets a special early warning information display area on a system interface, displays various early warning information issued by the water disaster early warning unit in real time, highlights the early warning information by means of a conspicuous color and a flashing effect, and ensures that the management personnel can timely find the early warning situation. Meanwhile, the visualization unit also supports exporting the visualization map in a picture or PDF format, so as to facilitate reporting and archiving.

[0052] In order to more comprehensively obtain mine water disaster related information, the system supports underground personnel to input data by means of a special mobile phone software. In a daily inspection process, if the underground personnel find a water inrush point or other abnormal situation, the underground personnel can use the mobile phone software to shoot water inrush point image data, and manually measure and record water inrush time, water inrush position, water inrush preliminary estimated flow and the like. The data are uploaded to the data management unit in real time through a wireless network in the mine, and are fused and analyzed with real-time monitoring data. The data management unit verifies effectiveness of the manually input data and converts a format of the manually input data, so as to ensure that the manually input data can be seamlessly connected with the automatic monitoring data, and to provide a more abundant and accurate information source for water disaster early warning.

[0053] The warning information display and push mode of the water disaster early warning unit includes system interface display, directional mobile phone short message push and directional email sending.

[0054] In summary, the present application can significantly improve the timeliness, accuracy, comprehensiveness and operability of mine water disaster early warning through dynamic data processing, multi-source data fusion analysis, hierarchical intelligent decision-making, scenario adaptation and man-machine collaborative interaction, and provides efficient and practical technical support for mine water disaster prevention.

[0055] The above has made a detailed description of the embodiments of the present application in combination with the drawings, but the present application is not limited thereto, and various changes can be made within the knowledge range of those skilled in the art without departing from the purpose of the present application, which are all within the protection scope of the claims of the present application.

Claims

1. A mine water disaster intelligent early warning method, characterized in that, The water disaster intelligent early warning method comprises water inrush risk early warning, and the water inrush risk early warning comprises the following steps. Water inflow in the mine monitoring area is collected according to a set fixed time length, and time series data of the water inflow is formed; A time window with a fixed width is set, and the average value and the standard deviation of the time series data of the water inflow in the time window are calculated; the fixed width is N times of the fixed time length; A sliding window mechanism is adopted, when a next new water inflow data is sequentially added, the oldest water inflow data in the window is removed; with the time window sliding forward, the time series data and the average value and the standard deviation in the time window are sequentially updated; When the value of the next new water inflow data exceeds the average water inflow in the current time window plus 2 times or 3 times the standard deviation, it is determined that the new water inflow data is a first abnormal water inflow; subsequent sliding data is also compared with the average water inflow in the current time window plus 2 times or 3 times the standard deviation to determine whether it is an abnormal water inflow; If three or more abnormal water inflows occur continuously, it is determined that the continuous abnormality is a water inrush risk abnormality, the previous water inflow of the first abnormal water inflow is taken as a water inrush early warning starting point, and a risk early warning based on a water inrush amount is triggered; the water inrush amount is the difference between the single abnormal water inflow and the water inflow at the water inrush early warning starting point; The water inrush risk early warning also comprises a synchronous monitoring and early warning upgrading mechanism of response data, and specifically as follows: response data related to the water inrush risk is monitored and synchronously acquired, and the response data types include underground water level, precipitation, ground subsidence and water chemical index value; The response data is processed by adopting the same sliding window mechanism as the water inflow monitoring, and specifically as follows: a time window with the same fixed width as the water inflow monitoring is set, and the time window is composed of the continuously monitored response data; when the time window slides into new response data, the oldest response data in the current window is removed, and the window width of the time window is kept constant; If the newly slid-in response data is greater than the average value of the response data in the current time window plus 2 times or 3 times the standard deviation, it is determined that the newly slid-in response data constitutes abnormal data of the response data in the current time window; if the subsequent slid-in data constitutes three or more abnormal data continuously, it is determined that the abnormal data belongs to synchronous abnormal response of the water inflow, and the water inrush risk early warning level is improved according to the response amplitude of the abnormal data, that is, the original mild water inrush risk early warning is upgraded to moderate water inrush risk early warning, the original moderate water inrush risk early warning is upgraded to severe water inrush risk early warning, and the original severe water inrush risk early warning remains unchanged; meanwhile, after judging that the response data has a synchronous abnormal response, the water inrush amount at the next three time points is predicted based on the response relationship between the water inrush amount and the abnormal response of the response data, and the water inrush risk early warning is performed according to the predicted water inrush amount; The water disaster intelligent early warning method also comprises a flooded mine risk early warning, and the flooded mine risk early warning comprises the following steps. Real-time monitoring and acquisition of water inflow and predicted water inrush amount in the mine monitoring area are performed; The operation of the drainage system equipment and the water storage volume are acquired, and the drainage capacity of the mine monitoring area is updated in real time; According to the ratio α of the mine monitoring area water inflow and the predicted water inrush volume to the real-time drainage capacity obtained by real-time monitoring, three-stage early warning is set with α being 0.8 times, 0.9 times and 1.0 respectively, and the mine flooding risk is output. At the same time, the ratio of the water storage volume required by the monitoring water inflow and the predicted water inrush volume to the existing water storage volume is calculated in real time according to the specification empirical formula , The three-stage early warning is set with α being 0.8 times, 0.9 times and 1.0 respectively, and the mine flooding risk is output, forming a double mine flooding risk early warning mechanism. The water disaster intelligent early warning method also comprises a water environment risk early warning, and the water environment risk early warning comprises the following steps. Determine the water pollutant control project and the maximum allowable emission concentration standard for monitoring the mine discharge water, and monitor the mine discharge water quality index in real time online according to the specified position, time period and frequency, and sample and test the mine discharge water quality index according to the specified sampling frequency, the water quality index including the conventional physicochemical index and the water pollution emission control index; Real-time analysis of the change trend and abnormal value of the discharge water quality index, calculation of the daily average emission concentration and the annual total emission; Define the pollution risk index γ as the ratio of the daily average emission concentration of the discharge water pollutant to the maximum allowable emission concentration, and the ratio > 1 indicates that the emission standard has been exceeded and there is a pollution risk; based on the pollution risk index calculation value, divide the water pollution risk early warning level, when the pollution risk index satisfies: 0.8 ≤ γ < 0.9, output the potential pollution risk early warning; when the pollution risk index satisfies: 0.9 ≤ γ < 1.0, output the critical pollution risk early warning; when the pollution risk index γ ≥ 1.0, output the over-standard pollution risk early warning.

2. The intelligent mine water disaster early warning method according to claim 1, characterized in that, The specific rules of the water inrush risk early warning based on the water inrush amount are as follows: The water inrush risk early warning adopts a three-level progressive threshold design, threshold I < threshold II < threshold III; when the water inrush amount satisfies: threshold I ≤ water inrush amount < threshold II, output the mild water inrush risk early warning; When the water inrush amount satisfies: threshold II ≤ water inrush amount < threshold III, output the moderate water inrush risk early warning; When the water inrush amount ≥ threshold III, output the severe water inrush risk early warning; If three consecutive abnormalities are all below the lower limit of threshold I, only monitoring is required, and the water inrush risk early warning is not triggered; If the abnormal water inrush amount is less than three times in a row, it is determined that the abnormal water inrush amount is only an accidental anomaly or an isolated anomaly, and the water inrush risk early warning is not triggered.

3. The intelligent mine water disaster early warning method according to claim 2, characterized in that, After the water inrush risk early warning is started, the water inrush amount at the subsequent three time points is predicted, and the water inrush risk early warning is performed according to the predicted water inrush amount.

4. A mine water disaster intelligent early warning system, which adopts the mine water disaster intelligent early warning method according to any one of claims 1-3, characterized in that, The system comprises: A monitoring unit for tracking and monitoring the data including the water inrush amount, the underground water level and water quality, the mine water chemistry, the mine water drainage pressure reduction, the atmospheric precipitation, the ground subsidence, the drainage system and the water gate data; A data management unit for storing, cleaning, analyzing, calculating and outputting the tracking and monitoring data in the monitoring unit; A water disaster early warning unit comprising a water inrush risk early warning module, a flooded mine risk early warning module and a water environment risk early warning module, and executing a mine water disaster intelligent early warning method.

5. The intelligent mine water disaster early warning system according to claim 4, characterized in that, It also comprises a visualization unit, which sets a standardized functional drawing in the water control field, and centrally displays the mine water filling state, water filling characteristics and water disaster early warning results; and sets a three-dimensional digital twin model of mine water filling, which is used to display the hardware environment, running state and early warning results of the monitoring unit in real time.

6. The intelligent mine water disaster early warning system according to claim 4, characterized in that, The monitoring data of the monitoring unit is divided into real-time online input and manual input, the manual input includes system terminal direct input, batch input, and underground personnel uploading water inrush point image data and manually measured data through mobile phone software, and the manual input data and real-time monitoring data are fused and analyzed.

7. The intelligent mine water disaster early warning system according to claim 4, characterized in that, The early warning information display and push mode of the water disaster early warning unit includes system interface display, targeted mobile phone short message push and targeted electronic mail sending.

Images