CN-121982854-A - Multi-parameter mine hydrologic automatic monitoring alarm system

CN121982854ACN 121982854 ACN121982854 ACN 121982854ACN-121982854-A

Abstract

The invention relates to the technical field of mine monitoring, in particular to a multi-parameter mine hydrologic automatic monitoring alarm system which comprises the following steps of generating a multi-parameter prediction sequence of each monitoring point in a future preset period through a space-time prediction model based on historical hydrologic time sequence data and real-time hydrologic parameters of each monitoring point of a mine, inputting a dynamic digital twin model for calculation, constructing a spatial amplification response of prediction uncertainty in a roadway, fault and water-containing structure based on the consistency of variation trend and time sequence discrete characteristics in the multi-parameter prediction sequence, outputting a mine integral water burst risk space-time distribution map in the future preset period, and automatically generating and executing alarm instructions corresponding to the spatial position and risk level of a high risk area. According to the method, the spatial cascade enhancement identification of the uncertainty risk is realized, and the risk amplification area is accurately defined through grid coupling with a preset geological structure sensitivity distribution map.

Inventors

LU XIN
WANG ZHANJUN
SHI ZHIJIANG
GUO HAILONG
ZHANG LE
LIU JUNXU
LIU PENG
ZHANG LEI

Assignees

国能蒙西煤化工股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260130

Claims (10)

1. A multi-parameter mine hydrologic automatic monitoring alarm system, characterized in that the alarm system comprises the following: Generating a multi-parameter prediction sequence of each monitoring point in a future preset period through a space-time prediction model based on historical hydrological time sequence data and real-time hydrological parameters of each monitoring point of the mine, wherein the multi-parameter prediction sequence at least comprises a predicted value sequence of water level, water pressure and water quality conductivity; Inputting the multi-parameter prediction sequence into a pre-constructed dynamic digital twin model reflecting the hydrogeology of the mine to carry out calculation, wherein the calculation process comprises the steps of simulating the space-time conduction and superposition effects of the prediction hydrographic parameters in a roadway network under the well, constructing the space amplification response of the prediction uncertainty in the roadway, faults and water-containing structures based on the consistency of the variation trend and the time sequence discrete characteristics in the multi-parameter prediction sequence, and identifying a risk amplification area sensitive to the prediction disturbance by coupling mapping the prediction uncertainty and the structural sensitivity of the hydrogeology of the mine so as to output the integral water burst risk space-time distribution map of the mine in a future preset period; and when a continuous high-risk area exceeding a preset risk threshold appears in the water gushing risk space-time distribution diagram, automatically generating and executing an alarm instruction corresponding to the space position and the risk level of the high-risk area.
2. The multi-parameter mine hydrologic automatic monitoring alarm system according to claim 1 is characterized in that historical hydrologic time sequence data and real-time hydrologic parameters of all monitoring points of a mine are obtained and at least comprise time sequences of water level, water pressure and water quality conductivity, and preprocessing is carried out on the historical hydrologic time sequence data and the real-time hydrologic parameters to form a unified model input data set.
3. The multi-parameter mine hydrologic automatic monitoring alarm system according to claim 2 is characterized in that the space-time prediction model is based on a space-time diagram neural network of an attention mechanism, a mine roadway network topological diagram is built according to the space position relation of monitoring points and is used as diagram structure input of the diagram neural network, the model input data set is used as node characteristic input, and the model learning hydrologic parameter evolution rule in the time dimension is related to conduction in the space roadway network through training.
4. The multi-parameter mine hydrologic automatic monitoring alarm system according to claim 3 is characterized in that the latest real-time hydrologic parameter sequence and historical sequence segments are input into a trained space-time prediction model, the space-time prediction model outputs predicted values of water level, water pressure and water quality conductivity of each time step of each monitoring point in a preset time period in the future to form the multi-parameter prediction sequence, and simultaneously, a prediction confidence interval corresponding to each predicted value in the multi-parameter prediction sequence is output.
5. The multi-parameter mine hydrologic automatic monitoring and alarming system according to claim 1, wherein the input of the dynamic digital twin model comprises taking a multi-parameter prediction sequence as an initial boundary condition and a source and sink item, and based on a three-dimensional geological structure, a roadway network topology and hydrogeological parameters in the dynamic digital twin model, a numerical simulation method is adopted to calculate the conduction and diffusion of the prediction hydrologic parameters in a downhole roadway network along with time and superposition effects from different source points, so as to generate basic hydrologic state space-time distribution data of each node of mine space in a future preset period.
6. The multi-parameter mine hydrologic automatic monitoring alarm system according to claim 5 is characterized in that based on the multi-parameter prediction sequence, a change trend consistency coefficient between different parameter prediction value sequences in the multi-parameter prediction sequence and the dispersion of a single parameter prediction value sequence in different future time steps are calculated, the consistency coefficient and the dispersion are fused, an original prediction uncertainty index at each monitoring point position is obtained quantitatively, the original prediction uncertainty index of each monitoring point is taken as input, the propagation process of the original prediction uncertainty index on a physical conduction path is simulated according to roadway connectivity, fault water guide characteristics and water-containing construction space spread defined in the dynamic digital twin model, and when the propagation path passes through a fault, a fracture intensive zone or a water-containing construction boundary, nonlinear enhancement is carried out according to a preset amplification coefficient, so that an amplification response distribution map of the original prediction uncertainty in a mine three-dimensional space is generated.
7. The multi-parameter mine hydrographic automatic monitoring and alarming system of claim 6, further comprising calling a pre-stored structural sensitivity distribution map of mine hydrogeology, wherein the structural sensitivity distribution map characterizes response intensity of hydrographic states of different spatial positions to key geological parameter disturbance, performing grid layer superposition and fusion calculation on the amplified response distribution map and the structural sensitivity distribution map, identifying spatial units with high uncertainty amplification response and high geological structure sensitivity, and defining the spatial units as risk amplification areas sensitive to predicted disturbance.
8. The multi-parameter mine hydrologic automatic monitoring and alarming system according to claim 7 is characterized in that basic water gushing risk values of each space unit are calculated according to the basic hydrologic state space-time distribution data, the basic water gushing risk values are adjusted upwards and corrected according to the uncertainty and sensitivity coupling strength of the space units in the risk amplifying area, and a mine overall water gushing risk space-time distribution map with risk confidence information in a future preset period is generated by integrating the risk values after correction of all the space units.
9. The multi-parameter mine hydrologic automatic monitoring and alarming system according to claim 1, wherein the identification of the continuous high risk area comprises spatial clustering analysis of a mine overall water burst risk space-time distribution map, identification of an area which is in spatial communication and in which a risk value exceeds a preset risk threshold value, and definition of the area as the continuous high risk area, and extraction of characteristic parameters of the continuous high risk area, wherein the characteristic parameters comprise geometric center coordinates, spatial boundary outlines, the highest risk level, time expected to reach a risk peak value and the type of the area with risk amplification.
10. The multi-parameter mine hydrologic automatic monitoring alarm system of claim 9, wherein a preset alarm strategy knowledge base is called according to characteristic parameters of the continuous high risk area, the alarm strategy knowledge base is constructed based on historical catastrophe cases and emergency treatment plans, corresponding alarm content templates are matched for combinations of different spatial positions, risk levels and risk amplification types, the alarm content templates are automatically filled, and a structured alarm instruction comprising at least one of the following is generated: Three-dimensional coordinates and influence range of the high-risk area, risk level and emergency degree, and suggested on-site treatment measures; and triggering a hierarchical alarm action according to the risk level and the emergency degree in the structural alarm instruction.

Description

Multi-parameter mine hydrologic automatic monitoring alarm system Technical Field The invention relates to the technical field of mine monitoring, in particular to a multi-parameter mine hydrologic automatic monitoring alarm system. Background Mine water-gushing disasters are one of the most important geological disasters in coal mines and other underground mines, and are complicated in inducement, hidden in evolution process and sudden, and once the water-gushing disasters occur, the life safety of operators is seriously threatened, and great economic loss is caused. Therefore, the mine hydrologic monitoring and risk early warning system becomes a core guarantee means for mine safety production. The existing mine hydrologic monitoring technology mainly relies on arranging a water level gauge, a water pressure gauge or a water quality sensor at a typical position, and carrying out state judgment by collecting data periodically or at intervals. The monitoring parameter dimension is single, the hydrologic evolution process cannot be comprehensively reflected, most of the existing systems focus on the single index of water level or water pressure, and important parameters reflecting water quality abnormality such as conductivity are ignored, so that the identification capability on sudden water-containing structures, fault water diversion or pollution water gushing events is weak. The traditional system often uses the current actual measurement value as a judgment basis, cannot predict the propagation and superposition trend of hydrologic parameters in time and space, and especially has insufficient identification capability for potential risk areas of deep roadways, fault zones and blind areas. Disclosure of Invention The invention provides an automatic monitoring and alarming system for multi-parameter mine hydrology, which integrates multi-source hydrology parameter prediction, digital twin risk calculation and closed-loop alarm response, so as to realize accurate prediction of future risks, early identification of sensitive areas and rapid execution of multi-level emergency strategies, thereby comprehensively improving intelligent prevention and control capability of mine hydrologic damage. A multi-parameter mine hydrologic automatic monitoring alarm system, the alarm system comprising performing the following: Generating a multi-parameter prediction sequence of each monitoring point in a future preset period through a space-time prediction model based on historical hydrological time sequence data and real-time hydrological parameters of each monitoring point of the mine, wherein the multi-parameter prediction sequence at least comprises a predicted value sequence of water level, water pressure and water quality conductivity; Inputting the multi-parameter prediction sequence into a pre-constructed dynamic digital twin model reflecting the hydrogeology of the mine to carry out calculation, wherein the calculation process comprises the steps of simulating the space-time conduction and superposition effects of the prediction hydrographic parameters in a roadway network under the well, constructing the space amplification response of the prediction uncertainty in the roadway, faults and water-containing structures based on the consistency of the variation trend and the time sequence discrete characteristics in the multi-parameter prediction sequence, and identifying a risk amplification area sensitive to the prediction disturbance by coupling mapping the prediction uncertainty and the structural sensitivity of the hydrogeology of the mine so as to output the integral water burst risk space-time distribution map of the mine in a future preset period; and when a continuous high-risk area exceeding a preset risk threshold appears in the water gushing risk space-time distribution diagram, automatically generating and executing an alarm instruction corresponding to the space position and the risk level of the high-risk area. Optionally, acquiring historical hydrologic time sequence data and real-time hydrologic parameters of each monitoring point of the mine, wherein the historical hydrologic time sequence data and the real-time hydrologic parameters at least comprise time sequences of water level, water pressure and water quality conductivity, and preprocessing the historical hydrologic time sequence data and the real-time hydrologic parameters to form a unified model input data set. Optionally, the space-time prediction model is based on a space-time diagram neural network of an attention mechanism, a mine roadway network topological diagram is built according to a monitoring point space position relationship and is used as diagram structure input of the diagram neural network, the model input data set is used as node characteristic input, and the model is trained to learn the relation between the evolution rule of hydrologic parameters in a time dimension and the conduction in the space roadway network. Optional