Multi-sensor-based Comprehensive Analysis and Early Warning System for Coal Mine Safety Data

Abstract

The present invention relates to the technical field of data prediction, and specifically relates to a comprehensive analysis and early warning system for coal mine safety data based on multi-sensors; obtaining the degree of fluctuation according to the gas concentration within a preset recent short window; obtaining the concentration difference according to the concentration at the current moment and any historical moment in the gas concentration sequence; obtaining a first concentration prediction value according to the concentration difference and the data within the preset recent short window; obtaining a second concentration prediction value according to the concentration difference and the gas concentration sequence; obtaining a final prediction value according to the first concentration prediction value, the second concentration prediction value and the degree of fluctuation. The present invention obtains an early warning value according to the final prediction value and the concentration change characteristics within the preset recent short window and makes a first early warning judgment; obtaining a ventilation anomaly index according to the wind speed sequence and the air pressure sequence; making a second early warning judgment on the gas concentration state according to the ventilation anomaly index, thereby improving the accuracy of gas concentration early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of data prediction, and in particular to a comprehensive analysis and early warning system for coal mine safety data based on multi-sensors. Background Art

[0002] During the coal mining process, gas is generated. When the gas concentration is too high, it is easy to cause mine explosions and accidents. Therefore, in order to detect gas anomalies in a timely manner and ensure mine safety, it is necessary to monitor and warn the gas concentration. Since the change of gas concentration shows a trend, the gas concentration at a future moment can be predicted based on the historical change of gas concentration, and early warning analysis can be carried out according to the prediction results, so as to increase the response time and reduce the occurrence of accidents. For data prediction, the existing moving average method is usually used, but the prediction accuracy of this algorithm for data sequences with volatility is poor, which further results in low early warning accuracy of gas concentration. In addition, the change of gas concentration is related to the ventilation device in the mine. Conducting early warning analysis only based on a single factor of the gas concentration prediction result will also affect the early warning accuracy. Summary of the Invention

[0003] In order to solve the technical problems that predicting the gas concentration by the existing moving average method will lead to low early warning accuracy, and at the same time, only a single factor of the gas concentration prediction result will also affect the early warning accuracy, the purpose of the present invention is to provide a comprehensive analysis and early warning system for coal mine safety data based on multi-sensors. The specific technical solutions adopted are as follows:

[0004] A data acquisition module, configured to acquire a gas concentration sequence, a wind speed sequence, and a wind pressure sequence within a currently preset time period;

[0005] A concentration analysis module, configured to obtain a fluctuation degree according to the gas concentration fluctuation characteristics within a preset recent short window in the gas concentration sequence; obtain a concentration difference according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence; obtain a first concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics within the preset recent short window; obtain a second concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics of the gas concentration sequence; obtain a final prediction value according to the first concentration prediction value, the second concentration prediction value, and the fluctuation degree;

[0006] A first early warning module, configured to obtain an early warning value according to the final prediction value and the concentration change characteristics within the preset recent short window; perform a first early warning judgment on the gas concentration state according to the early warning value and the final prediction value;

[0007] The second warning module is used to obtain a ventilation anomaly index according to the change correlation characteristics of the wind speed sequence and the wind pressure sequence, and perform a second warning judgment on the gas concentration state according to the ventilation anomaly index.

[0008] Further, the step of obtaining the fluctuation degree according to the gas concentration fluctuation characteristics within a preset recent short window in the gas concentration sequence includes:

[0009] Within the preset recent short window, calculate the absolute value of the difference between the gas concentration value at each moment and the average gas concentration value to obtain a concentration difference characteristic value; calculate the sum of the standard deviation of the gas concentration within the preset recent short window and a preset extremely small positive number to obtain a discrete reference; calculate the ratio of the concentration difference characteristic value within the preset recent short window to the discrete reference and normalize it to obtain a deviation characteristic value; normalize the concentration difference characteristic value and calculate the product with a preset first weight to obtain a first fluctuation factor; calculate the product of the deviation characteristic value and a preset second weight to obtain a second fluctuation factor; calculate the average value of the sum of the first fluctuation factor and the second fluctuation factor corresponding to each moment within the preset recent short window to obtain the fluctuation degree.

[0010] Further, the step of obtaining the concentration difference according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence includes:

[0011] Calculate the difference between the gas concentration at the current moment and any historical moment in the gas concentration sequence to obtain the concentration difference between the current moment and any historical moment.

[0012] Further, the step of obtaining the first concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics within the preset recent short window includes:

[0013]

[0014] In the formula, R represents the first concentration prediction value, H represents the average gas concentration within the preset recent short window, N represents the number of historical moments within the preset recent short window, e represents the natural constant, represents the time interval coefficient of the nth historical moment from the current moment, represents the time interval weight of the nth historical moment from the current moment, represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference of the nth historical moment; represents the comprehensive weighted concentration difference.

[0015] Further, the step of obtaining the second concentration prediction value at a future moment according to the data distribution characteristics of the concentration difference and the gas concentration sequence includes:

[0016]

[0017] In the formula, W represents the second concentration prediction value, E represents the average gas concentration of the gas concentration sequence, G represents the number of historical moments in the gas concentration sequence, e represents the natural constant, represents the time interval coefficient of the nth historical moment from the current moment, represents the time interval weight of the nth historical moment from the current moment, represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference of the nth historical moment, represents the comprehensive weighted concentration difference.

[0018] Further, the step of obtaining the final prediction value according to the first concentration prediction value, the second concentration prediction value and the degree of fluctuation includes:

[0019] Calculate the difference between the constant 1 and the degree of fluctuation to obtain the proportionality coefficient; calculate the product of the degree of fluctuation and the first concentration prediction value to obtain the first prediction factor; calculate the product of the proportionality coefficient and the second concentration prediction value to obtain the second prediction factor; calculate the sum of the first prediction factor and the second prediction factor to obtain the final prediction value at the future moment.

[0020] Further, the step of obtaining the warning value according to the final prediction value and the concentration change characteristics within the preset recent short window includes:

[0021] Calculate the product of the standard deviation of the gas concentration within the preset recent short window and the preset coefficient to obtain the adjustment value; calculate the sum of the final prediction value and the adjustment value to obtain the warning value at the future moment.

[0022] Further, the step of making the first warning judgment on the gas concentration state according to the warning value and the final prediction value includes:

[0023] When the final prediction value or the actual gas concentration value at a future moment exceeds the preset safe concentration value, the gas concentration at the future moment is abnormal and a warning is issued, and the second warning judgment is not made; when the actual gas concentration value at a future moment exceeds the warning value, the gas concentration at the future moment is suspected to be abnormal and the second warning judgment is required.

[0024] Further, the step of obtaining the ventilation abnormality index according to the change correlation characteristics of the wind speed sequence and the wind pressure sequence includes:

[0025] Calculate the joint entropy of the wind speed sequence and the wind pressure sequence and normalize it to obtain the first ventilation anomaly factor; calculate the absolute value of the Pearson correlation coefficient of the wind speed sequence and the wind pressure sequence and perform a negative correlation mapping to obtain the second ventilation anomaly factor; calculate the sum value of the first ventilation anomaly factor and the second ventilation anomaly factor to obtain the ventilation anomaly index.

[0026] Further, the step of making a second warning judgment on the gas concentration state according to the ventilation anomaly index includes:

[0027] When the ventilation anomaly index does not exceed the preset anomaly threshold, the gas concentration at a future moment is abnormal and a warning is issued.

[0028] The present invention has the following beneficial effects:

[0029] In the present invention, obtaining the degree of fluctuation can characterize the fluctuation characteristics of the recent gas concentration, and thus be used to determine the weights of prediction results based on different data lengths; obtaining the concentration difference can determine the concentration differences between different historical moments and the current moment, so as to consider the influence of gas concentration fluctuations on the prediction results in the process of calculating the first concentration prediction value and the second concentration prediction value, improving the prediction accuracy and the accuracy of concentration warning. The first concentration prediction value represents the prediction result obtained according to recent data, and the second concentration prediction value represents the prediction result obtained according to long-term data. Obtaining the final prediction value can reasonably allocate the weights of the first concentration prediction value and the second concentration prediction value according to the degree of fluctuation, further improving the accuracy of the prediction result and the accuracy of concentration warning. Obtaining the warning value can be used to determine whether there is a trend of rapid increase in gas concentration; making a first warning judgment can preliminarily judge whether the gas concentration is abnormal; obtaining the ventilation anomaly index can characterize whether the ventilation device is faulty, and further determine whether the ventilation device is the cause of the rapid increase in concentration. Finally, the accuracy of concentration warning is improved according to the first warning judgment and the second warning judgment. Description of the Drawings

[0030] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0031] Figure 1 It is a block diagram of a coal mine safety data comprehensive analysis and warning system based on multi-sensors provided by an embodiment of the present invention. Detailed Embodiments

[0032] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following specifically describes, with reference to the accompanying drawings and preferred embodiments, a multi-sensor-based comprehensive analysis and early warning system for coal mine safety data, including its specific implementation manner, structure, features, and effects. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0034] The following specifically describes, with reference to the accompanying drawings, the specific solution of a multi-sensor-based comprehensive analysis and early warning system for coal mine safety data provided by the present invention.

[0035] Please refer to Figure 1 , which shows a block diagram of a multi-sensor-based comprehensive analysis and early warning system for coal mine safety data provided by an embodiment of the present invention. The system includes the following modules:

[0036] A data acquisition module S1, configured to acquire a gas concentration sequence, a wind speed sequence, and a wind pressure sequence for a currently preset time period.

[0037] In an embodiment of the present invention, the implementation scenario is to give an early warning of the gas concentration in the mine to improve the accuracy of the early warning. First, acquire a gas concentration sequence, a wind speed sequence, and a wind pressure sequence for a currently preset time period; collect the gas concentration in the mine through a gas concentration sensor, the wind speed at the ventilation equipment through a wind speed sensor, and the wind pressure at the ventilation equipment through a differential pressure sensor; in an embodiment of the present invention, the currently preset time period is a time range of the historical three minutes including the current moment. After each update of the current moment, the data within the corresponding preset time period is updated once, and the data is collected and updated every 3 seconds. The implementer can determine it according to the implementation scenario.

[0038] A concentration analysis module S2, configured to obtain the degree of fluctuation according to the gas concentration fluctuation characteristics within a preset recent short window in the gas concentration sequence; obtain the concentration difference according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence; obtain the first concentration prediction value for a future moment according to the concentration difference and the data distribution characteristics within the preset recent short window; obtain the second concentration prediction value for a future moment according to the concentration difference and the data distribution characteristics of the gas concentration sequence; obtain the final prediction value according to the first concentration prediction value, the second concentration prediction value, and the degree of fluctuation.

[0039] When predicting the gas concentration at a future moment based on the gas concentration sequence, the degree of data fluctuation has a great impact on the prediction accuracy. If the recent concentration fluctuates greatly, in order to improve the accuracy and reflect the recent fluctuation characteristics, more reliance on recent data is needed during prediction. If the recent concentration fluctuates little, more data is required during prediction to consider the overall concentration characteristics. Therefore, it is necessary to analyze the fluctuation of the recent gas concentration before prediction to adaptively adjust the prediction strategy. Thus, the degree of fluctuation is obtained according to the fluctuation characteristics of the gas concentration within a preset recent short window in the gas concentration sequence.

[0040] Preferably, in the embodiment of the present invention, the steps of obtaining the degree of fluctuation include: calculating the absolute value of the difference between the gas concentration value at each moment and the average gas concentration within the preset recent short window to obtain the concentration difference characteristic value. In the embodiment of the present invention, the preset recent short window is the gas concentration data range within the most recent 1 minute including the current moment. When the concentration difference characteristic value is larger, it means that the difference between the gas concentration value at this moment and the average gas concentration within this window is larger, and the fluctuation is more obvious. Calculating the sum value of the standard deviation of the gas concentration within the preset recent short window and a preset extremely small positive number to obtain the discrete benchmark. The standard deviation of the gas concentration reflects the degree of data fluctuation within the preset recent short window. In the embodiment of the present invention, the preset extremely small positive number is 0.1, aiming to avoid the situation of a constant 0 when the standard deviation is used as the denominator. Calculating the ratio of the concentration difference characteristic value within the preset recent short window to the discrete benchmark and normalizing it to obtain the deviation characteristic value. When the deviation characteristic value is larger, it means that the deviation degree of the gas concentration at this moment is higher, and the fluctuation is more obvious. Normalizing the concentration difference characteristic value and calculating the product with a preset first weight to obtain the first fluctuation factor. In the embodiment of the present invention, the preset first weight is 0.6, and the implementer can determine it according to the implementation scenario. Calculating the product of the deviation characteristic value and a preset second weight to obtain the second fluctuation factor. In the embodiment of the present invention, the preset second weight is 0.4, and the implementer can determine it according to the implementation scenario. Calculating the average value of the sum values of the first fluctuation factor and the second fluctuation factor corresponding to each moment within the preset recent short window to obtain the degree of fluctuation. When the sum value of the first fluctuation factor and the second fluctuation factor at each moment is larger, it means that the gas concentration value at this moment is more discrete, and further the degree of fluctuation is larger, indicating that the data fluctuation within the preset recent short window is stronger.

[0041] Furthermore, since the existing moving average method has poor prediction accuracy when the data volatility is relatively obvious, this algorithm only smooths the data fluctuations and is likely to obscure the short-term trend features in the data. If the data fluctuation characteristics are considered in the prediction, the rise and fall of the gas concentration can be captured more sensitively. Therefore, it is necessary to improve the moving average method; first, obtain the concentration difference according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence; preferably, in the embodiment of the present invention, the step of obtaining the concentration difference includes: calculating the difference between the gas concentration at the current moment and the gas concentration at any historical moment in the gas concentration sequence to obtain the concentration difference between the current moment and any historical moment; the current moment is the previous moment of the future moment, and the gas concentration value at the future moment is relatively close to that at the current moment. Therefore, analyzing the concentration difference between the current moment and the historical moment can be used to reflect the influence of recent concentration fluctuations on the gas concentration at the future moment during the prediction process, thereby improving the prediction accuracy. After obtaining the concentration difference, the first concentration prediction value at the future moment can be obtained according to the concentration difference and the data distribution characteristics within the preset nearest short window; preferably, in the embodiment of the present invention, the step and formula for obtaining the first concentration prediction value include:

[0042]

[0043] In the formula, R represents the first concentration prediction value, H represents the average gas concentration within the preset nearest short window, and this average gas concentration represents the average level of the concentration within the preset nearest short window. The gas concentration value at the future moment fluctuates based on this average value. N represents the number of historical moments within the preset nearest short window, e represents the natural constant, represents the time interval coefficient of the nth historical moment from the current moment. For example, if n is 1 for the previous moment of the current moment, the time interval coefficient for this previous moment is 1, represents the time interval weight of the nth historical moment from the current moment. The closer to the current moment, the greater the time interval weight, and the higher the contribution degree of the corresponding concentration data; the implementer can determine the values of the time interval coefficient and the time interval weight according to the implementation scenario. represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference of the nth historical moment; the weighted concentration difference characterizes the concentration difference between the nth historical moment and the current moment in combination with the time interval weight. represents the comprehensive weighted concentration difference, and this value characterizes the overall difference degree between the gas concentrations at different moments within the preset nearest short window and the current moment; furthermore, calculating the sum value of the average gas concentration within the preset nearest short window and the comprehensive weighted concentration difference can accurately predict the gas concentration value at the future moment while considering the concentration fluctuation characteristics during this period.

[0044] Further, the first concentration prediction value is calculated based on the concentration data within a shorter preset recent short window. If the volatility of the gas concentration is small, the prediction accuracy based on longer concentration data is higher. Therefore, the second concentration prediction value at a future moment can be obtained according to the concentration difference and the data distribution characteristics of the gas concentration sequence. The specific formula is as follows:

[0045]

[0046] In the formula, W represents the second concentration prediction value, E represents the average gas concentration of the gas concentration sequence, G represents the number of historical moments in the gas concentration sequence, and the data calculation range of the second concentration prediction value is longer than that of the first concentration prediction value. e represents the natural constant, represents the time interval coefficient of the nth historical moment from the current moment, represents the time interval weight of the nth historical moment from the current moment, represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference of the nth historical moment, represents the comprehensive weighted concentration difference. It should be noted that the calculation logic of the second concentration prediction value is the same as that of the first concentration prediction value, and the specific steps will not be elaborated.

[0047] If the fluctuation of the gas concentration is obvious, the prediction accuracy based on the gas concentration data in the relatively recent time is higher; if the fluctuation is weak, the prediction based on the longer gas concentration data can better reflect the comprehensive concentration trend. Therefore, the final prediction value can be obtained according to the first concentration prediction value, the second concentration prediction value, and the degree of fluctuation. Preferably, in the embodiment of the present invention, the steps of obtaining the final prediction value include: calculating the difference between the constant 1 and the degree of fluctuation to obtain the proportionality coefficient; the greater the degree of fluctuation, the smaller the proportionality coefficient. Calculating the product of the degree of fluctuation and the first concentration prediction value to obtain the first prediction factor; the greater the degree of fluctuation, the greater the importance of the concentration data within the preset recent short window, and the greater the weight of the first concentration prediction value. Calculating the product of the proportionality coefficient and the second concentration prediction value to obtain the second prediction factor; the greater the degree of fluctuation, the lower the reliability of the data with a longer interval, and the smaller the weight of the second concentration prediction value. Therefore, the greater the degree of fluctuation, the greater the weight of the first concentration prediction value; the smaller the degree of fluctuation, the greater the weight of the second concentration prediction value. Calculating the sum of the first prediction factor and the second prediction factor to obtain the final prediction value at a future moment; this final prediction value has higher accuracy compared to the existing moving average method, thereby improving the accuracy of gas concentration early warning.

[0048] The first early warning module S3 is used to obtain the early warning value according to the final prediction value and the concentration change characteristics within the preset recent short window; and perform the first early warning judgment on the gas concentration state according to the early warning value and the final prediction value.

[0049] Since the gas in the mine may suddenly leak due to construction, resulting in a rapid increase in gas concentration and safety accidents, it is necessary to quickly give an early warning for such a sudden increase in concentration. Therefore, the early warning value can be obtained according to the final predicted value and the concentration change characteristics within the preset recent short window; preferably, in the embodiment of the present invention, the steps of obtaining the early warning value include: calculating the product of the standard deviation of the gas concentration within the preset recent short window and the preset coefficient to obtain an adjustment value; in the embodiment of the present invention, the preset coefficient is 0.2, and the implementer can determine it by himself according to the implementation scenario. Calculating the sum value of the final predicted value and the adjustment value to obtain the early warning value for the future moment; since the gas concentration will normally fluctuate, if the upward floating degree is too high, there may be a situation where the gas concentration rapidly increases. Therefore, an adjustment value is set as the upward floating interval based on the final predicted value. If the actual concentration value at the future moment exceeds the early warning value, it means that the upward floating degree is too high. Even if it does not exceed the set upper limit value of the gas concentration, it is necessary to give an early warning in time to avoid an explosion.

[0050] Furthermore, the first early warning judgment on the gas concentration state can be made according to the early warning value and the final predicted value; preferably, in the embodiment of the present invention, the steps of making the first early warning judgment include: when the final predicted value or the actual gas concentration value at the future moment exceeds the preset safety concentration value, the gas concentration at the future moment is abnormal and an early warning is given, and the second early warning judgment is not made; at this time, it means that the gas concentration is already too high and exceeds the safety range, and an early warning needs to be given in time to increase the response time of the on-site personnel; the implementer can determine the preset safety concentration value by himself according to the implementation scenario. When the actual gas concentration value at the future moment exceeds the early warning value, the gas concentration at the future moment is suspected to be abnormal. At this time, it means that the gas concentration increases too fast. Although it does not reach the preset safety concentration value, there is a development trend of abnormal concentration; and the rapid increase in concentration may be due to a sudden large leakage of gas or a malfunction of the ventilation device in the mine, affecting the discharge of gas. Therefore, the second early warning judgment needs to be made to determine the specific reason and improve the early warning accuracy.

[0051] The second early warning module S4 is used to obtain the ventilation abnormality index according to the change correlation characteristics of the wind speed sequence and the wind pressure sequence; and make the second early warning judgment on the gas concentration state according to the ventilation abnormality index.

[0052] When the ventilation equipment fails, the rapid increase in gas concentration may be due to blocked airflows or gas accumulation. If a warning is issued directly, it may lead to a large-scale evacuation of on-site personnel and even trigger other safety accidents. In this case, it is only necessary to report the failure of the ventilation device and carry out repairs. When the ventilation device is operating normally, the data characteristics of the wind speed sequence and the wind pressure sequence are relatively stable, and it is difficult for complex and variable data to occur. Moreover, the correlation between the changes in the wind speed sequence and the wind pressure sequence is relatively obvious. Therefore, the ventilation anomaly index can be obtained based on the correlation characteristics of the changes in the wind speed sequence and the wind pressure sequence.

[0053] Preferably, in the embodiment of the present invention, the steps of obtaining the ventilation anomaly index include: calculating and normalizing the joint entropy of the wind speed sequence and the wind pressure sequence to obtain the first ventilation anomaly factor. It should be noted that the joint entropy belongs to the prior art, and the specific calculation steps will not be elaborated. When the data combinations at the corresponding moments of the two sequences are more random, the joint entropy is larger, and the first ventilation anomaly factor is larger, which means that the changes in the wind speed sequence and the wind pressure sequence are less regular, and the possibility of ventilation device failure is greater. Calculating the absolute value of the Pearson correlation coefficient of the wind speed sequence and the wind pressure sequence and performing a negative correlation mapping to obtain the second ventilation anomaly factor. It should be noted that the calculation of the Pearson correlation coefficient belongs to the prior art, and the specific calculation steps will not be elaborated. When the correlation between the wind speed and the wind pressure is worse, the absolute value of the Pearson correlation coefficient is closer to 0, and the second ventilation anomaly factor is larger, indicating that the ventilation device is more likely to fail. Calculating the sum value of the first ventilation anomaly factor and the second ventilation anomaly factor to obtain the ventilation anomaly index. When the ventilation anomaly index is larger, it means that the possibility of ventilation device failure is higher, the effect of gas emission is worse, and gas accumulation is more likely to occur.

[0054] Furthermore, the second warning judgment of the gas concentration state can be made according to the ventilation anomaly index, which specifically includes: when the ventilation anomaly index does not exceed the preset anomaly threshold, it means that the ventilation device is operating normally, and the rapid increase in gas concentration is due to a large amount of gas leakage. The gas concentration at future moments will be abnormal and a warning will be issued, and on-site personnel need to evacuate quickly. In the embodiment of the present invention, the preset anomaly threshold is 0.6, and the implementer can determine it according to the implementation scenario. If the ventilation anomaly index exceeds the preset anomaly threshold, the failure of the ventilation device needs to be reported quickly. Thus, by analyzing the volatility of the gas concentration to determine the weights of the prediction results under different data lengths, and considering the volatility of the gas concentration in the prediction results, the accuracy of concentration prediction is improved, and the accuracy of gas concentration warning is initially improved; at the same time, whether there is a failure in the ventilation device is analyzed, excluding the influence of non-gas leakage factors, and further improving the accuracy of concentration warning.

[0055] In summary, the embodiment of the present invention provides a comprehensive analysis and early warning system for coal mine safety data based on multi-sensors; obtaining the degree of fluctuation according to the gas concentration within a preset recent short window; obtaining the concentration difference according to the concentration at the current moment and any historical moment in the gas concentration sequence; obtaining a first concentration prediction value according to the concentration difference and the data within the preset recent short window; obtaining a second concentration prediction value according to the concentration difference and the gas concentration sequence; obtaining a final prediction value according to the first concentration prediction value, the second concentration prediction value and the degree of fluctuation. The present invention obtains an early warning value according to the final prediction value and the concentration change characteristics within the preset recent short window and makes a first early warning judgment; obtains a ventilation anomaly index according to the wind speed sequence and the air pressure sequence; makes a second early warning judgment on the gas concentration state according to the ventilation anomaly index, improving the accuracy of gas concentration early warning.

[0056] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0057] Each embodiment in this specification is described in a progressive manner, and the same or similar parts among the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A coal mine safety data comprehensive analysis and early warning system based on multiple sensors, characterized in that: The system includes the following modules: A data acquisition module is used to obtain the gas concentration sequence, wind speed sequence and wind pressure sequence of the current preset period; A concentration analysis module, used to obtain a fluctuation degree according to the gas concentration fluctuation characteristics within a preset recent short window in the gas concentration sequence; obtain a concentration difference according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence; obtain a first concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics within the preset recent short window; obtain a second concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics of the gas concentration sequence; obtain a final prediction value according to the first concentration prediction value, the second concentration prediction value and the fluctuation degree; A first warning module, used to obtain a warning value according to the final predicted value and the concentration change characteristics within the preset most recent short window; Performing a first warning judgment on the gas concentration state according to the warning value and the final predicted value; The second warning module is used to obtain a ventilation abnormality index according to the change correlation characteristics of the wind speed sequence and the wind pressure sequence; and perform a second warning judgment on the gas concentration state according to the ventilation abnormality index; The step of obtaining the degree of fluctuation according to the gas concentration fluctuation characteristics within a preset recent short window in the gas concentration sequence comprises: Calculate the absolute value of the difference between the gas concentration value at each moment and the average value of the gas concentration within the preset most recent short window to obtain a concentration difference characteristic value; calculate the sum of the gas concentration standard deviation within the preset most recent short window and a preset minimum positive number to obtain a discrete benchmark; calculate the ratio of the concentration difference characteristic value within the preset most recent short window to the discrete benchmark and normalize it to obtain a deviation characteristic value; normalize the concentration difference characteristic value and calculate the product with the preset first weight to obtain a first fluctuation factor; calculate the product of the deviation characteristic value and the preset second weight to obtain a second fluctuation factor; calculate the average value of the sum of the first fluctuation factor and the second fluctuation factor corresponding to each moment within the preset most recent short window to obtain the fluctuation degree; The step of obtaining a first concentration prediction value at a future moment according to the concentration difference value and the data distribution characteristics in the preset most recent short window comprises: , where R represents the first concentration prediction value, H represents the average gas concentration in the preset recent short window, N represents the number of historical moments in the preset recent short window, and e represents a natural constant. Indicates the time interval coefficient from the nth historical moment to the current moment, Represents the time interval weight from the nth historical moment to the current moment, represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference at the nth historical moment; represents the comprehensive weighted concentration difference; The step of obtaining a second concentration prediction value at a future moment according to the concentration difference and the data distribution characteristics of the gas concentration sequence comprises: , where W represents the second concentration prediction value, E represents the average gas concentration of the gas concentration sequence, G represents the number of historical moments in the gas concentration sequence, and e represents a natural constant. Indicates the time interval coefficient from the nth historical moment to the current moment, Represents the time interval weight from the nth historical moment to the current moment, represents the concentration difference between the current moment and the nth historical moment, represents the weighted concentration difference at the nth historical moment, Represents the comprehensive weighted concentration difference.

2. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of obtaining the concentration difference value according to the concentration difference characteristics between the current moment and any historical moment in the gas concentration sequence comprises: The difference between the gas concentration at the current moment and any historical moment in the gas concentration sequence is calculated to obtain the concentration difference between the current moment and any historical moment.

3. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of obtaining a final predicted value according to the first concentration predicted value, the second concentration predicted value and the degree of fluctuation comprises: Calculate the difference between constant 1 and the degree of fluctuation to obtain the proportionality coefficient; calculate the product of the degree of fluctuation and the first concentration prediction value to obtain the first prediction factor; calculate the product of the proportionality coefficient and the second concentration prediction value to obtain the second prediction factor; calculate the sum of the first prediction factor and the second prediction factor to obtain the final prediction value at the future time.

4. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of obtaining the warning value according to the final predicted value and the concentration change characteristics within the preset recent short window comprises: The product of the standard deviation of the gas concentration in the preset most recent short window and the preset coefficient is calculated to obtain the adjustment value; the sum of the final prediction value and the adjustment value is calculated to obtain the warning value at the future moment.

5. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of performing a first warning judgment on the gas concentration state according to the warning value and the final prediction value comprises: When the final predicted value or actual gas concentration value at a future moment exceeds the preset safety concentration value, the gas concentration at the future moment is abnormal and an early warning is issued, and no second early warning judgment is made; when the actual gas concentration value at the future moment exceeds the early warning value, the gas concentration at the future moment is suspected to be abnormal, and a second early warning judgment is required.

6. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of obtaining the ventilation abnormality index according to the change correlation characteristics of the wind speed sequence and the wind pressure sequence comprises: The joint entropy of the wind speed sequence and the wind pressure sequence is calculated and normalized to obtain a first ventilation abnormality factor; the absolute value of the Pearson correlation coefficient of the wind speed sequence and the wind pressure sequence is calculated and negatively correlated to obtain a second ventilation abnormality factor; the sum of the first ventilation abnormality factor and the second ventilation abnormality factor is calculated to obtain the ventilation abnormality index.

7. The multi-sensor based coal mine safety data comprehensive analysis and early warning system according to claim 1 is characterized in that: The step of performing a second early warning judgment on the gas concentration state according to the ventilation abnormality index comprises: When the ventilation abnormality index does not exceed the preset abnormality threshold, the gas concentration at a future time will be abnormal and an early warning will be issued.

Images