Disaster early warning system for coal mine area

Abstract

The invention provides a disaster early warning system for a coal mine area, relates to the technical field of disaster early warning, aims to realize efficient and more accurate disaster early warning, and comprises a historical data module used for storing historical disaster data; the real-time data module is used for periodically collecting and storing real-time monitoring environment data; the credibility module is used for acquiring the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster; the historical data prediction module is used for performing first disaster prediction according to the historical disaster data; the real-time data prediction module is used for performing second disaster prediction according to the real-time monitoring environment data; and the prediction fusion module is used for fusing the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence. The method has the advantage of improving the accuracy of disaster prediction in the coal mine area.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster early warning, and in particular to a disaster early warning system for a coal mining area. Background Art

[0002] In coal mine production, safety issues have always been the focus of attention in the industry.

[0003] With the increase in mining depth and the complexity of the mining environment, coal mining areas are facing a variety of potential disasters, such as gas explosions and earthquakes. The occurrence of these disasters not only threatens the lives and safety of miners, but also has a serious impact on production efficiency and the mining environment, so it is necessary to establish disaster warnings. Generally speaking, disasters are caused by natural disasters and non-natural disasters. In conventional technical solutions, disaster warnings are generally carried out through the monitoring of environmental data. However, for some disasters, especially natural disasters, it is difficult to monitor and warn in time through environmental data, which will reduce the overall reliability of warnings.

[0004] Therefore, it is particularly important to establish an efficient and accurate disaster early warning system. Summary of the invention

[0005] The purpose of the present invention is to provide a disaster warning system for coal mining areas, which can achieve efficient and more accurate disaster warning.

[0006] The present invention is achieved through the following technical solutions:

[0007] The present invention provides a disaster early warning system for coal mining areas, comprising:

[0008] Historical data module, used to store historical disaster data;

[0009] Real-time data module, used to periodically collect and store real-time monitoring environment data;

[0010] A credibility module, used to respectively obtain the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster, wherein the sum of the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster is 1;

[0011] A historical data prediction module, used for performing a first disaster prediction based on the historical disaster data;

[0012] A real-time data prediction module, used for performing a second disaster prediction based on the real-time monitoring environment data;

[0013] A prediction fusion module is used to fuse the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence.

[0014] Preferably, the historical disaster data includes various types of disasters and corresponding occurrence times.

[0015] Preferably, the method for obtaining the prediction credibility of each disaster in the historical disaster data is:

[0016] Based on the historical disaster data, the occurrence frequency sequence of each disaster in different periods is obtained respectively:

[0017]

[0018]

[0019] in, is the occurrence frequency sequence of the i-th disaster in the j-th period, T j represents the length of the jth cycle, t is the total time length of the historical disaster data, and ts i,j,n Representative according to the T j The number of occurrences of the i-th disaster in the n-th cycle when obtaining, n = 1, 2, ..., m, m is an intermediate parameter;

[0020] Get the autocorrelation parameters of each occurrence sequence separately:

[0021]

[0022] in, represents the autocorrelation parameter of the occurrence frequency sequence of the i-th disaster in the j-th period, cov(.) is the function for finding the covariance, var(.) is the function for finding the variance, and Respectively represent the From 1 to The number of parameters and the to The array formed by the parameters, and are functions for downward and upward integration respectively;

[0023] Obtain the credibility of the historical disaster data based on the autocorrelation parameter:

[0024]

[0025] Among them, γ i is the prediction credibility of the historical disaster data for the i-th disaster, α is the basic credibility, and J is the total number of categories in the period.

[0026] Preferably, the period includes one day, one month, one quarter, one year and ten years.

[0027] Preferably, the method for obtaining the credibility of the real-time monitoring environment data is:

[0028] δ i =1-γ i ;

[0029] Among them, δ i Represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster.

[0030] Preferably, the method for performing the first disaster prediction based on the historical disaster data is:

[0031] The historical disaster data is used as input to output the probability of occurrence of the corresponding disaster on the target day through a recurrent neural network.

[0032] Preferably, the method for performing the second disaster prediction based on the real-time monitoring environment data is:

[0033] Based on the real-time monitoring environmental data, the probability of occurrence of the corresponding disaster on the target day is output through a long short-term memory network.

[0034] Preferably, the method for fusing the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence is:

[0035] ω i =γ i * i +δ i *σ i ;

[0036] Among them, ω i is the prediction result of the probability of the i-th disaster occurring on the target day after fusion, γ i is the prediction credibility of the historical disaster data for the i-th disaster, ρ i is the prediction result of the first disaster prediction for the probability of the i-th disaster occurring on the target day, δ i represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster, σ i The second disaster prediction is the prediction result of the probability of the i-th disaster occurring on the target day.

[0037] Preferably, if the prediction credibility of the historical disaster data for the disaster is lower than a preset threshold, the disaster is predicted only by the real-time data prediction module.

[0038] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0039] The present invention uses a weighted combination of historical data modules and real-time data modules to enable the system to obtain disaster information more comprehensively. Historical data provides long-term trends and rules, while real-time data reflects the current environmental status, making predictions of natural disasters that are difficult to predict more referenceable and accurate.

[0040] The present invention can calculate the prediction credibility for historical data and real-time data respectively through the credibility module, so that the system can weight according to the reliability of different data sources, thereby ensuring that the prediction results are reasonably balanced among multiple sources of information, thereby improving the accuracy of early warning;

[0041] The present invention can realize multi-level disaster prediction through historical data prediction module and real-time data prediction module, and can make predictions from two levels: long time span and short time fluctuation, so as to further improve the response sensitivity of the system;

[0042] The accurate and real-time disaster prediction and early warning system of the present invention can help mining areas take preventive measures in a timely manner, reduce human and property losses, and thus greatly improve the overall safety of mining areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a schematic diagram of the principles of a disaster warning system for a coal mining area provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0045] Example 1

[0046] The embodiment of the present invention is a disaster early warning system for coal mining areas, referring to Figure 1 ,include:

[0047] Historical data module, used to store historical disaster data;

[0048] Real-time data module, used to periodically collect and store real-time monitoring environment data;

[0049] A credibility module, used to respectively obtain the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster, wherein the sum of the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster is 1;

[0050] A historical data prediction module, used for performing a first disaster prediction based on the historical disaster data;

[0051] A real-time data prediction module, used for performing a second disaster prediction based on the real-time monitoring environment data;

[0052] A prediction fusion module is used to fuse the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence.

[0053] This embodiment combines historical data and real-time data to improve the accuracy and timeliness of disaster warnings through multi-dimensional prediction mechanisms and data fusion. The historical data module mainly identifies long-term disaster trends and patterns. For example, some natural disasters may be related to specific seasonal changes or the laws of nature's movement. By analyzing these data, the system can predict the possibility of some natural disasters occurring in the future. The real-time data module can periodically collect and store real-time environmental monitoring data, such as gas concentration in the air, rock pressure, temperature and humidity, surface vibration, etc., and transmit the data to the system in real time. Combining the above two data sources can predict disasters such as earthquakes, fires, floods, and explosions.

[0054] Based on the two types of data prediction, the credibility module assigns different credibility to the disaster predictions of historical data and real-time data according to their characteristics. For example, the reliability of historical data will increase in disasters that are greatly affected by seasonal or natural changes, while the reliability of historical data will decrease in unconventional sudden disasters (such as gas leaks). Through the credibility module, the system can flexibly weight different data sources to generate more reasonable disaster prediction results.

[0055] In summary, the disaster warning system of this embodiment makes full use of the stability of historical data and the agility of real-time data, and solves the conflict or inconsistency between the two data sources through the credibility weighting mechanism. Through the multi-level and multi-source disaster prediction method, the mining area can maintain the early warning capability of potential disasters in the short and long term. This design not only improves the accuracy of the prediction, but also greatly reduces the losses caused by sudden disasters.

[0056] In this embodiment, the historical disaster data includes various types of disasters and corresponding occurrence times.

[0057] As a preferred solution, the method for obtaining the prediction credibility of each disaster in the historical disaster data is as follows:

[0058] Based on the historical disaster data, the occurrence frequency sequence of each disaster in different periods is obtained respectively:

[0059]

[0060]

[0061] in, is the occurrence frequency sequence of the i-th disaster in the j-th period, T j represents the length of the jth cycle, t is the total time length of the historical disaster data, and ts i,j,n Representative according to the T j The number of occurrences of the i-th disaster in the n-th cycle when obtaining, n = 1, 2, ..., m, m is an intermediate parameter;

[0062] Get the autocorrelation parameters of each occurrence sequence separately:

[0063]

[0064] in, represents the autocorrelation parameter of the occurrence frequency sequence of the i-th disaster in the j-th period, cov(.) is the function for finding the covariance, var(.) is the function for finding the variance, and Respectively represent the From 1 to The number of parameters and the to The array formed by the parameters, and are functions for downward and upward integration respectively;

[0065] Obtain the credibility of the historical disaster data based on the autocorrelation parameter:

[0066]

[0067] Among them, γ i is the prediction credibility of the historical disaster data for the i-th disaster, α is the basic credibility, and J is the total number of categories in the period.

[0068] Furthermore, the cycle includes one day, one month, one quarter, one year and ten years.

[0069] In other words, the credibility of historical disaster data predictions for each disaster is mainly determined by judging the degree of periodicity of disaster occurrence based on a large amount of historical data collected. The higher the degree of periodicity, the higher the weight given.

[0070] On this basis, the method for obtaining the credibility of the real-time monitoring environment data is:

[0071] δ i =1-γ i ;

[0072] Among them, δ i Represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster.

[0073] Next, the method for performing the first disaster prediction based on the historical disaster data is preferably:

[0074] The historical disaster data is used as input to output the probability of occurrence of the corresponding disaster on the target day through a recurrent neural network. This prediction is mainly based on the time period.

[0075] On the other hand, the method for performing a second disaster prediction based on the real-time monitoring environmental data is:

[0076] Based on the real-time monitoring environmental data, the probability of occurrence of the corresponding disaster on the target day is output through the long short-term memory network. For example: gas explosion prediction through air temperature, wind speed, air pressure and temperature; landslide prediction through amplitude, support stress, surface subsidence, etc.; explosion prediction based on dust concentration, humidity, wind speed, oxygen concentration and temperature; water disaster prediction based on water level, rainfall, soil moisture content, geological fissure conditions and mine water volume.

[0077] As a preferred solution of this embodiment, a method for fusing the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence is:

[0078] ω i =γ i * i +δ i *σ i ;

[0079] Among them, ω i is the prediction result of the probability of the i-th disaster occurring on the target day after fusion, γ i is the prediction credibility of the historical disaster data for the i-th disaster, ρ i is the prediction result of the first disaster prediction for the probability of the i-th disaster occurring on the target day, δ i represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster, σ i The second disaster prediction is the prediction result of the probability of the i-th disaster occurring on the target day.

[0080] As a further preferred solution, if the prediction credibility of the historical disaster data for the disaster is lower than a preset threshold, the disaster is predicted only by the real-time data prediction module. This approach can avoid unnecessary waste of computing resources.

[0081] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A disaster early warning system for coal mining areas, characterized in that: include: Historical data module, used to store historical disaster data; Real-time data module, used to periodically collect and store real-time monitoring environment data; A credibility module, used to respectively obtain the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster, wherein the sum of the prediction credibility of the historical disaster data and the real-time monitoring environment data for each disaster is 1; A historical data prediction module, used for performing a first disaster prediction based on the historical disaster data; A real-time data prediction module, used for performing a second disaster prediction based on the real-time monitoring environment data; A prediction fusion module is used to fuse the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence.

2. A disaster early warning system for coal mining areas according to claim 1, characterized in that: The historical disaster data includes various types of disasters and corresponding occurrence times.

3. A disaster early warning system for coal mining areas according to claim 2, characterized in that: The method for obtaining the prediction credibility of each disaster in the historical disaster data is as follows: Based on the historical disaster data, the occurrence frequency sequence of each disaster in different periods is obtained respectively: in, is the occurrence frequency sequence of the i-th disaster in the j-th period, T j represents the length of the jth cycle, t is the total time length of the historical disaster data, and ts i,j,n Representative according to the T j The number of occurrences of the i-th disaster in the n-th cycle when obtaining, n = 1, 2, ..., m, m is an intermediate parameter; Get the autocorrelation parameters of each occurrence sequence separately: in, represents the autocorrelation parameter of the occurrence frequency sequence of the i-th disaster in the j-th period, cov(.) is the function for finding the covariance, var(.) is the function for finding the variance, and Respectively represent the From 1 to The number of parameters and the to The array formed by the parameters, and are functions for downward and upward integration respectively; Obtain the credibility of the historical disaster data based on the autocorrelation parameter: Among them, γ i is the prediction credibility of the historical disaster data for the i-th disaster, α is the basic credibility, and j is the total number of categories in the period.

4. A disaster early warning system for coal mining areas according to claim 3, characterized in that: The cycles include a day, a month, a quarter, a year, and a decade.

5. A disaster early warning system for coal mining areas according to claim 4, characterized in that: The method for obtaining the credibility of the real-time monitoring environment data is: d i =1-c i ; Among them, δ i Represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster.

6. A disaster early warning system for coal mining areas according to claim 1, characterized in that: The method for performing the first disaster prediction based on the historical disaster data is: The historical disaster data is used as input to output the probability of occurrence of the corresponding disaster on the target day through a recurrent neural network.

7. A disaster early warning system for coal mining areas according to claim 6, characterized in that: The method for performing the second disaster prediction based on the real-time monitoring environmental data is: Based on the real-time monitoring environmental data, the probability of occurrence of the corresponding disaster on the target day is output through a long short-term memory network.

8. A disaster early warning system for coal mining areas according to claim 7, characterized in that: The method for fusing the prediction results of the first disaster prediction and the second disaster prediction according to the prediction confidence is: oh i =c i *r i +d i *s i ; Among them, ω i is the prediction result of the probability of the i-th disaster occurring on the target day after fusion, γ i is the prediction credibility of the historical disaster data for the i-th disaster, ρ i is the prediction result of the first disaster prediction for the probability of the i-th disaster occurring on the target day, δ i represents the credibility of the real-time monitoring environment data of the historical disaster data for the i-th disaster, σ i The second disaster prediction is the prediction result of the probability of the i-th disaster occurring on the target day.

9. A disaster early warning system for coal mining areas according to claim 1, characterized in that: If the prediction credibility of the historical disaster data for the disaster is lower than a preset threshold, the disaster is predicted only by the real-time data prediction module.

Images