CN-121352252-B - Mine catastrophe period ventilation emergency decision-making system based on artificial intelligence

CN121352252BCN 121352252 BCN121352252 BCN 121352252BCN-121352252-B

Abstract

The invention discloses an artificial intelligence-based mine catastrophe period ventilation emergency decision system which comprises a data acquisition module, a catastrophe sensing module, a decision execution module and a decision making module, wherein the data acquisition module acquires multi-source real-time data of a mine catastrophe period, the data acquisition module acquires standardized mine space-time data through data cleaning and alignment preprocessing operation, the catastrophe sensing module extracts space and data association information based on a standardized space-time data set to construct a mine topological graph structure, the mine topological graph structure is based on a mine topological graph structure, a dynamic three-dimensional airflow field is constructed through a graph neural network fused with physical constraints, the decision optimization module regards the dynamic three-dimensional airflow field as a dynamically changing physical background field, defines catastrophe events as disturbance sources in the physical background field, and carries out deduction catastrophe evolution based on core disturbance propagation rules to acquire catastrophe situation sensing data, and the decision execution module carries out catastrophe early warning based on the catastrophe situation sensing data to realize mine catastrophe period optimal ventilation emergency decision.

Inventors

LI HONGYE
CHEN RUI
YAN CHAOLIANG
ZHAO XIAOXI
Xi Mengbo
XU HAOYU

Assignees

诺文科风机(北京)有限公司

Dates

Publication Date: 20260512
Application Date: 20251217

Claims (5)

1. An artificial intelligence-based mine catastrophe period ventilation emergency decision system, which is characterized by comprising: The data acquisition module acquires multi-source real-time data in a mine catastrophe period, and obtains standardized mine space-time data through data cleaning and alignment pretreatment operation; the process for obtaining the standardized mine space-time data comprises the following steps: A sensor network is deployed to collect wind speed data and gas concentration data, time alignment is carried out, and time stamp t is used as a unified reference to carry out time dimension matching on the wind speed data and the gas concentration data; comparing the acquisition time of each data point with the system standard time, if the deviation exceeds Adjusting the corresponding time stamp to the latest standard time stamp to form time alignment data; Based on the space coordinates (x, y, z) of the monitoring points, associating the time alignment data with the space positions, binding the corresponding wind speed data and gas concentration data of each monitoring point with the space coordinates (x, y, z) to form a three-dimensional data structure of the space coordinates, time and parameters, and obtaining standardized mine space-time data; the catastrophe perception module is used for constructing a mine topological graph structure based on a standardized space-time data set and extracting space and data association information, and constructing a dynamic three-dimensional airflow field through a graph neural network fused with physical constraints based on the mine topological graph structure; The graph neural network fused with the physical constraints is realized by introducing a first physical constraint and a second physical constraint, and embedding the first physical constraint and the second physical constraint into a message transmission process of the original graph neural network; the first physical constraint describes the air volume conservation of any space area in unit time, and aims at a mine topological graph Any node in (a) In the corresponding space region, the total air quantity flowing in is equal to the total air quantity flowing out, and the formula is as follows: ; Wherein, the Is a node Is defined by a set of contiguous nodes, Is a node And (3) with Wind speed in between; The second physical constraint describes the relationship between the wind pressure difference and the wind speed in the roadway and the resistance loss, and is used for mine topological diagrams Any edge of (a) The corresponding roadway, the wind pressure difference satisfies the formula: ; Wherein, the Is a node Is used for the standard wind pressure of the air conditioner, Is a node Is used for the standard wind pressure of the air conditioner, Is a node Is set in the air velocity of (1), Is a node Is set in the air velocity of (1), For the air density in the mine, Is the roughness coefficient of the roadway and is used for controlling the roughness coefficient of the roadway, Is the equivalent diameter of the roadway and is equal to the equivalent diameter of the roadway, As the average wind speed of the roadway, The roadway length is the roadway length; The decision optimization module regards the dynamic three-dimensional wind flow field as a dynamic physical background field, defines a catastrophe event as a disturbance source in the physical background field, and carries out deduction catastrophe evolution based on a core disturbance propagation rule to obtain catastrophe situation perception data; and the decision executing module is used for carrying out disaster early warning on the mine based on the disaster situation awareness data so as to realize the optimal ventilation emergency decision in the disaster period of the mine.
2. The artificial intelligence based mine catastrophe period ventilation emergency decision system of claim 1, wherein the process of constructing the mine topology is: Extracting space coordinates of all monitoring points and ventilation equipment based on a standardized space-time data set, associating the space coordinates of the ventilation equipment with corresponding equipment numbers, taking each monitoring point and ventilation equipment as independent nodes, and constructing a graph node set Constructing a graph edge set E based on the actual roadway connection relation and the spatial distance of the mine, and passing through a graph node set With edge set Construction of mine topology 。
3. The artificial intelligence-based mine catastrophe period ventilation emergency decision system according to claim 1, wherein the process of constructing a dynamic three-dimensional airflow field is: Message transmission process of embedding first physical constraint and second physical constraint into graph neural network, and setting total layer number of graph neural network as ; The node characteristics of the initial layer are the initial characteristics of the nodes in the mine topological graph G; Then through iterative updating of the graph neural network until the first The node characteristics of the layers are all converged, corresponding wind speed and wind pressure data are extracted, and a field wind speed value and a field wind pressure value are obtained; And obtaining the field wind speed values and the field wind pressure values of all the node space positions, and complementing the field wind speed values and the field wind pressure values of all the space positions in the mine by adopting a space interpolation method to form a dynamic three-dimensional wind flow field.
4. The artificial intelligence based mine catastrophe period ventilation emergency decision system according to claim 3, wherein the construction process of the core disturbance propagation rule is as follows: Disassembling disturbance propagation into two core physical processes, including a natural diffusion process and a wind flow carrying process, wherein a core disturbance propagation rule is expressed as: ; Wherein, the Represents the core catastrophe parameter, D is the diffusion coefficient, Indicating the course of carrying the wind current, Indicating the course of the natural diffusion, Representing the source of the catastrophic disturbance, Is the source intensity of the disturbance.
5. The artificial intelligence based ventilation emergency decision system for mine catastrophe according to claim 4, wherein the process of obtaining catastrophe situation awareness data is: Dynamic three-dimensional airflow field Regarding as a physical background field of a mine environment, defining a catastrophe event as a catastrophe disturbance source in the physical background field, carrying out evolution deduction according to a core disturbance propagation rule, and recording core catastrophe parameters of the reaction mine catastrophe on each grid point (x, y, z, t) under each time stamp t; The core catastrophe parameters comprise the concentration of the catastrophe gas, the deviation of the wind speed of a field and the deviation of the wind pressure of the field; setting deviation threshold values to label risk levels, and integrating core catastrophe parameters of each region with corresponding risk levels to form catastrophe situation awareness data 。

Description

Mine catastrophe period ventilation emergency decision-making system based on artificial intelligence Technical Field The invention relates to the technical field of new generation information, in particular to an artificial intelligence-based mine catastrophe period ventilation emergency decision system. Background The evolution of mine ventilation emergency decision-making technology is always in need of mine safety, and is limited by the lack of monitoring means in the early stage, ventilation emergency decision-making mainly depends on manual experience and static plans, and technicians formulate a fixed emergency plan according to mine basic design parameters such as roadway trend and fan power, and when sudden disasters such as gas leakage and fire disaster occur, ventilation equipment can be adjusted only according to a preset flow, and the dynamic evolution characteristics of the disasters cannot be adapted. In the existing ventilation emergency decision system, a neural network model driven by pure data is mostly adopted, characteristic fitting is carried out only by depending on monitoring data, and the relation between the core physical rules of mine ventilation such as air conservation and air pressure difference and resistance loss is not integrated, so that the modeling mode is easy to cause physical distortion problems, for example, the deduced air flow direction is contrary to the actual roadway ventilation rule, or the air pressure distribution does not accord with the hydrodynamic principle, so that the disaster evolution deduction result has larger actual deviation; In addition, the conventional deduction method mostly regards catastrophe parameters as independent diffusion substances, neglects the forced carrying effect of a mine dynamic air flow field on catastrophe diffusion, and in practice, gas can be quickly brought to an air return roadway by the air flow, but the conventional model always predicts the uniform diffusion of the gas, and cannot accurately describe the double evolution rule of natural diffusion and air flow driving, so that the pre-judging precision of the catastrophe influence range and the diffusion rate is insufficient, and the accurate pre-warning is difficult to support; Accordingly, an artificial intelligence based mine disaster period ventilation emergency decision system is presented herein. Disclosure of Invention In order to overcome the defects in the prior art, the invention provides the following technical scheme for achieving the purposes: an artificial intelligence based mine catastrophe period ventilation emergency decision system, comprising: The data acquisition module acquires multi-source real-time data in a mine catastrophe period, and obtains standardized mine space-time data through data cleaning and alignment pretreatment operation; the catastrophe perception module is used for constructing a mine topological graph structure based on a standardized space-time data set and extracting space and data association information, and constructing a dynamic three-dimensional airflow field through a graph neural network fused with physical constraints based on the mine topological graph structure; The graph neural network fused with the physical constraints is realized by introducing a first physical constraint and a second physical constraint, and embedding the first physical constraint and the second physical constraint into a message transmission process of the original graph neural network; The decision optimization module regards the dynamic three-dimensional wind flow field as a dynamic physical background field, defines a catastrophe event as a disturbance source in the physical background field, and carries out deduction catastrophe evolution based on a core disturbance propagation rule to obtain catastrophe situation perception data; and the decision executing module is used for carrying out disaster early warning on the mine based on the disaster situation awareness data so as to realize the optimal ventilation emergency decision in the disaster period of the mine. The process for obtaining the standardized mine space-time data comprises the following steps: A sensor network is deployed to collect wind speed data and gas concentration data, time alignment is carried out, and time stamp t is used as a unified reference to carry out time dimension matching on the wind speed data and the gas concentration data; comparing the acquisition time of each data point with the system standard time, if the deviation exceeds Adjusting the corresponding time stamp to the latest standard time stamp to form time alignment data; And (3) based on the space coordinates (x, y, z) of the monitoring points, correlating the time alignment data with the space positions, binding the corresponding wind speed data and gas concentration data of each monitoring point with the space coordinates (x, y, z) to form a three-dimensional data structure of the space coordinates, t