Sensor-based mine dust concentration on-line monitoring and dust falling system

Abstract

The invention relates to the technical field of data processing, in particular to a sensor-based mining dust concentration on-line monitoring dust fall system, which comprises a data acquisition module for acquiring dust concentration data to obtain a dust concentration data sequence; the data analysis module is used for acquiring an important coefficient of each dust concentration data according to each dust concentration data difference, and acquiring lagging data in a process of predicting a dust concentration data sequence by using an autoregression model; the weight acquisition module is used for obtaining a local fitting straight line of each piece of lagged data according to the importance coefficient of each piece of dust concentration data, obtaining the importance degree of each piece of lagged data according to the local fitting straight line, and obtaining the adaptive weight of each piece of lagged data according to the importance degree; and the data prediction module is used for obtaining dust concentration prediction data at the next moment according to the self-adaptive weight of each piece of lagging data, and taking a dust falling measure on a mine hole, so that the accuracy of predicting the dust concentration is improved.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a sensor-based online monitoring and dust reduction system for mining dust concentration. Background Art

[0002] Mine dust refers to the tiny particles of solid materials such as rocks and mine dust generated by blasting, mechanical chiseling, cutting, friction and other operations during the mining process. These dust particles can be suspended in the air or settle in the tunnel, causing serious harm to the health of miners and the safety of mines. In order to prevent and reduce the hazards of mine dust, it is necessary to monitor the dust concentration and predict the dust concentration for early warning.

[0003] Traditionally, an autoregressive model is used to predict the dust concentration at the mine entrance. The autoregressive model performs a straight-line fit on historical data, assigns weights to each historical data based on the fitted data, and then predicts the change in dust concentration in the future. However, since the speed of dust particles after blasting changes with time, the rate of increase of dust concentration at the mine entrance will also change with the speed of dust particles. Therefore, the change in the rate of increase will cause some turning points to appear in the historical data. These turning points have a great impact on the data after the turning point. However, since these turning points have a small impact on the straight-line fit, the traditional autoregressive model will assign lower weights to these turning points, resulting in inaccurate predictions of dust concentration in the future.

[0004] Therefore, how to obtain the adaptive weight of each data to improve the accuracy of dust concentration prediction using the autoregressive model has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a sensor-based online monitoring and dust reduction system for mining dust concentration to solve the problem of how to obtain an adaptive weight for each data to improve the accuracy of predicting dust concentration using an autoregressive model.

[0006] An embodiment of the present invention provides a sensor-based online monitoring and dust reduction system for mining dust concentration, the system comprising the following steps: The data collection module is used to collect dust concentration data in the historical period after blasting at the mine entrance to obtain a dust concentration data sequence; A data analysis module, for obtaining important coefficients of each dust concentration data according to the difference of dust concentration data in the dust concentration data sequence, performing autocorrelation analysis on the dust concentration data sequence in the process of predicting the dust concentration data sequence using an autoregressive model to obtain the order of the autoregressive model, and obtaining the lagged data in the dust concentration data sequence according to the order of the autoregressive model; A weight acquisition module, used for performing straight line fitting on all dust concentration data within a local range of each lag data according to the importance coefficient of each dust concentration data, and correspondingly obtaining a local fitting straight line for each lag data, obtaining the importance of each lag data according to the local fitting straight line of each lag data, and obtaining an adaptive weight of each lag data according to the importance; The data prediction module is used to obtain dust concentration prediction data at the next moment according to the adaptive weight of each of the lagging data, and take dust reduction measures at the mine entrance according to the dust concentration prediction data.

[0007] Furthermore, the data analysis module obtains important coefficients of each dust concentration data according to the difference of dust concentration data in the dust concentration data sequence, including: Constructing a curve graph according to the dust concentration data, wherein the horizontal axis of the curve graph represents time and the vertical axis represents the dust concentration data; For any dust concentration data in the curve graph, the mutation degree of the dust concentration data is obtained according to the variation trend of the dust concentration data in its local range; According to the similarity between the previous and next data of the dust concentration data, the local similarity of the dust concentration data is obtained; The mean of the mutation degree and the local similarity is calculated, and the important coefficient of the dust concentration data is obtained by subtracting the mean from a constant 1.

[0008] Furthermore, the data analysis module obtains the degree of mutation of the dust concentration data according to the variation trend of the dust concentration data in its local range, including: In the curve graph, if the dust concentration data is not the first dust concentration data or the last dust concentration data in the curve graph, respectively obtain the left dust concentration data adjacent to the left of the dust concentration data, and the right dust concentration data adjacent to the right of the dust concentration data, and calculate the left slope between the dust concentration data and the left dust concentration data, and the right slope between the dust concentration data and the right dust concentration data; Calculate the absolute value of the difference between the left slope and the right slope, add the absolute value of the difference with a preset multiple to a preset constant, use the reciprocal of the addition result as a first variable, subtract the first variable from a constant 1, and obtain the mutation degree of the dust concentration data; If the dust concentration data is the first dust concentration data in the curve graph, obtain the right slope corresponding to the dust concentration data, add the right slope to the preset constant, use the reciprocal of the addition result as the first variable, subtract the first variable from the constant 1, and obtain the mutation degree of the dust concentration data; If the dust concentration data is the last dust concentration data in the curve graph, obtain the left slope corresponding to the dust concentration data, add the left slope to the preset constant, take the inverse of the addition result as the first variable, subtract the first variable from the constant 1, and obtain the mutation degree of the dust concentration data.

[0009] Furthermore, the data analysis module obtains the local similarity of the dust concentration data according to the similarity between the previous and next data of the dust concentration data, including: In the dust concentration data sequence, if the number of dust concentration data before the dust concentration data is greater than or equal to n, and the number of dust concentration data after the dust concentration data is greater than or equal to m, a straight line fitting is performed on the n dust concentration data before the dust concentration data to obtain a first fitting straight line, and a straight line fitting is performed on the m dust concentration data after the dust concentration data to obtain a second fitting straight line, wherein n and m are both greater than a first preset number, and the Spearman correlation coefficient between the first fitting straight line and the second fitting straight line is calculated to obtain the local similarity of the dust concentration data; If the number of dust concentration data before the dust concentration data is less than n, obtain a second fitting straight line corresponding to the dust concentration data, obtain a target fitting value at the sampling time of the dust concentration data according to the second fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data; If the number of dust concentration data after the dust concentration data is less than m, obtain the first fitting straight line corresponding to the dust concentration data, obtain the target fitting value at the sampling time of the dust concentration data according to the first fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data.

[0010] Furthermore, the weight acquisition module performs straight line fitting on all dust concentration data within a local range of each lag data according to the important coefficient of each dust concentration data, and obtains a local fitting straight line for each lag data, including: For any lag data, in the dust concentration data sequence, the important coefficients of u dust concentration data before the lag data are used as the weights of the corresponding dust concentration data to obtain u weighted data before the lag data, and a straight line is fitted to the u weighted data before the lag data to obtain a front local fitting straight line of the lag data, wherein u is less than or equal to the order; If the number of dust concentration data after the lag data is greater than or equal to v, the important coefficients of the v dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain a posterior local fitting straight line of the lag data, wherein v is greater than or equal to a second preset number; If the number of dust concentration data after the lag data is less than v, and the lag data is not the last dust concentration data in the dust concentration data sequence, then the important coefficients of the lag data and all the dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain the lag data and v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain the posterior local fitting straight line of the lag data.

[0011] Furthermore, the weight acquisition module obtains the importance of each of the lagged data according to the local fitting straight line of each of the lagged data, including: For any lag data, according to the previous local fitting straight line of the lag data, a first fitting value at the sampling time of the lag data is obtained, and the absolute value of the difference between the lag data and the first fitting value is calculated to obtain a first difference index; If there is no local fitting straight line for the lagged data, subtract the reciprocal of the first difference index from the constant 1, calculate the average of the importance coefficient of the lagged data and the reciprocal, and obtain the importance of the lagged data; If there is a rear local fitting straight line for the lag data, the importance of the lag data is obtained according to the difference between the front local fitting straight line and the rear local fitting straight line of the lag data.

[0012] Furthermore, the weight acquisition module obtains the importance of the lag data according to the difference between the front local fitting straight line and the rear local fitting straight line of the lag data, including: According to the posterior local fitting straight line of the lagged data, a second fitting value at the sampling time of the lagged data is obtained, and the absolute value of the difference between the lagged data and the second fitting value is calculated to obtain a second difference index; Calculating the reciprocal of the mean of the first difference index and the second difference index to obtain a second variable, and subtracting the second variable from a constant 1 to obtain a local difference degree of the lagged data; Calculate the Spearman correlation coefficient between the front local fitting straight line and the rear local fitting straight line of the lag data, and subtract the Spearman correlation coefficient from a constant 1 to obtain a discrimination coefficient of the lag data; The importance coefficient, local difference degree and average value of the distinction coefficient of the lag data are calculated to obtain the importance degree of the lag data.

[0013] Furthermore, the weight acquisition module acquires the adaptive weight of each of the lagging data according to the importance, including: Get the average importance of all lagged data to get the average importance; For any lagging data, the ratio between the importance of the lagging data and the average importance is calculated to obtain the adaptive weight of the lagging data.

[0014] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The present invention provides a sensor-based online dust concentration monitoring and dust reduction system for mining, comprising a data acquisition module for collecting dust concentration data in a historical period after blasting at a mine entrance to obtain a dust concentration data sequence; a data analysis module for respectively obtaining important coefficients of each dust concentration data according to the difference of dust concentration data in the dust concentration data sequence, and in the process of predicting the dust concentration data sequence using an autoregressive model, performing autocorrelation analysis on the dust concentration data sequence to obtain the order of the autoregressive model, and obtaining the dust concentration data sequence according to the order of the autoregressive model. Lag data; a weight acquisition module, used to perform straight-line fitting on all dust concentration data within the local range of each lag data according to the importance coefficient of each dust concentration data, and obtain the corresponding local fitting straight line of each lag data; according to the local fitting straight line of each lag data, obtain the importance of each lag data; according to the importance, obtain the adaptive weight of each lag data; a data prediction module, used to obtain the dust concentration prediction data at the next moment according to the adaptive weight of each lag data, and take dust reduction measures for the mine entrance according to the dust concentration prediction data. Among them, according to the difference of dust concentration data in the dust concentration data sequence, the important coefficient of each dust concentration data is obtained; in the process of using the autoregressive model to predict the dust concentration data at the next moment, the dust concentration data sequence is subjected to autocorrelation analysis to obtain the lagged data, and a straight line fitting is performed according to the dust concentration data in the local range of the lagged data, so that in the process of straight line fitting, the weight of the suspected noise data in the dust concentration data sequence is smaller, reducing the interference to the prediction result; according to the difference between each lagged data and its corresponding fitting result (local fitting straight line), the importance of each lagged data is obtained; according to the importance, the adaptive weight of each lagged data is obtained, and then the dust concentration data at the next moment is predicted according to the adaptive weight of each lagged data, thereby improving the accuracy of dust concentration prediction using the autoregressive model. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 It is a structural block diagram of a sensor-based online monitoring and dust reduction system for mining dust concentration provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0017] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0018] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0019] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0020] See also Figure 1 , is a structural block diagram of a sensor-based online monitoring and dust reduction system for mining dust concentration provided in the first embodiment of the present invention, such as Figure 1 As shown, the system may include: The data collection module 11 is used to collect dust concentration data in a historical period after blasting at the mine entrance to obtain a dust concentration data sequence.

[0021] The dust concentration data within 10 minutes after the blasting at the mine entrance is collected through sensors. The collection frequency is once per second, and a dust concentration data sequence is obtained. There is no restriction here, and the implementer can set it according to the specific implementation scenario.

[0022] The data analysis module 12 is used to obtain the important coefficients of each dust concentration data according to the dust concentration data differences in the dust concentration data sequence, and in the process of predicting the dust concentration data sequence using the autoregressive model, perform autocorrelation analysis on the dust concentration data sequence to obtain the order of the autoregressive model, and obtain the lagged data in the dust concentration data sequence according to the order of the autoregressive model.

[0023] When a mine is blasted, a large amount of dust particles will be generated. Under the action of high temperature, high pressure and airflow, these dust particles will quickly spread around, and the dust concentration at the mine entrance will increase rapidly. As time changes, the temperature gradually decreases, the diffusion rate of dust particles gradually decreases, and the increase rate of dust concentration at the mine entrance slows down. Therefore, the dust concentration at the mine entrance will first maintain a low value for a period of time, then increase rapidly, and then increase slowly. Since the dust concentration will be reflected as a turning point when the diffusion rate of dust particles changes, the turning point can reflect the turning point of dust concentration, so the data of the turning point is more important and should be given a higher weight.

[0024] However, due to the complex environment of the mine, the collected dust concentration data may contain noise data, so it is necessary to first reduce the importance of the noise data to avoid unnecessary interference with the prediction results.

[0025] The characteristics of noise data are similar to those of turning points. However, since noise data is data that causes the original data to deviate from the normal state due to interference, the noise data does not change the trend of the original data. The data before and after the noise data should have similar changing trends, and there is a large correlation between the data before and after the noise data. The changing trends of the data before and after the turning point will be quite different, and the correlation is small. Therefore, based on these characteristics, the important coefficient of each dust concentration data in the dust concentration data sequence is calculated to reduce the importance of the noise data and avoid unnecessary interference to the prediction results.

[0026] In an example of the present invention, a curve graph is constructed based on a dust concentration data sequence, wherein the horizontal axis of the curve graph represents time and the vertical axis represents dust concentration data. For the i-th dust concentration data in the curve graph, the slopes between the i-th dust concentration data and the two dust concentration data adjacent to its left and right are respectively obtained, and the mutation degree of the i-th dust concentration data is obtained based on the difference between the two slopes, which is used to characterize the difference in the changing trend of the data before and after the i-th dust concentration data; the local similarity of the i-th dust concentration data is obtained through the Spearman correlation coefficient between the data before and after the i-th dust concentration data, which is used to characterize the correlation between the data before and after the i-th dust concentration data, wherein the Spearman correlation coefficient is a prior art and will not be repeated here. Furthermore, based on the mutation degree and local similarity of the i-th dust concentration data, the important coefficient of the i-th dust concentration data is obtained, which is used to reflect the importance of the i-th dust concentration data, so as to reduce the importance of the noise data, and the important coefficient of the i-th dust concentration data is obtained as follows: (1) According to the changing trend of the i-th dust concentration data within its local range, the mutation degree of the i-th dust concentration data is obtained.

[0027] Specifically, in the curve graph, if the dust concentration data is not the first dust concentration data or the last dust concentration data in the curve graph, the left dust concentration data adjacent to the left of the dust concentration data and the right dust concentration data adjacent to the right of the dust concentration data are respectively obtained, and the left slope between the dust concentration data and the left dust concentration data and the right slope between the dust concentration data and the right dust concentration data are calculated; The absolute value of the difference between the left slope and the right slope is calculated, the absolute value of the difference with a preset multiple is added to a preset constant, the reciprocal of the addition result is used as the first variable, and the constant 1 is subtracted from the first variable to obtain the degree of mutation of the dust concentration data.

[0028] In one embodiment, the calculation formula for the mutation degree of the i-th dust concentration data is:

[0029] in, Indicates the mutation degree of the i-th dust concentration data, represents the left slope between the ith dust concentration data and the left dust concentration data, represents the right slope between the ith dust concentration data and the right dust concentration data, 1 represents a constant, c represents an arbitrary constant, Represents the absolute value symbol.

[0030] It should be noted that setting , there is no restriction here, and the implementer can set it according to the implementation scenario. c is used to prevent the denominator from being 0. is the first variable, The larger the value of is, the greater the difference in the changing trend between the data before and after the i-th dust concentration data is, and the smaller the value of the first variable is. The larger the value of , the greater the probability that the i-th dust concentration data is a noise point.

[0031] In particular, if the dust concentration data is the first dust concentration data in the curve graph, the right slope corresponding to the dust concentration data is obtained, the right slope is added to the preset constant, the reciprocal of the addition result is used as the first variable, and the constant 1 is subtracted from the first variable to obtain the mutation degree of the dust concentration data, that is, the calculation formula for the mutation degree of the i-th dust concentration data is: .

[0032] If the dust concentration data is the last dust concentration data in the curve graph, the left slope corresponding to the dust concentration data is obtained, the left slope is added to the preset constant, the reciprocal of the addition result is used as the first variable, and the constant 1 is subtracted from the first variable to obtain the mutation degree of the dust concentration data, that is, the calculation formula for the mutation degree of the i-th dust concentration data is: .

[0033] (2) Based on the similarity between the previous and next data of the i-th dust concentration data, the local similarity of the i-th dust concentration data is obtained.

[0034] Specifically, in the dust concentration data sequence, if the number of dust concentration data before the dust concentration data is greater than or equal to n, and the number of dust concentration data after the dust concentration data is greater than or equal to m, a straight line fitting is performed on the n dust concentration data before the dust concentration data to obtain a first fitting straight line, and a straight line fitting is performed on the m dust concentration data after the dust concentration data to obtain a second fitting straight line, wherein n and m are both greater than a first preset number, and the Spearman correlation coefficient between the first fitting straight line and the second fitting straight line is calculated to obtain the local similarity of the dust concentration data; If the number of dust concentration data before the dust concentration data is less than n, obtain a second fitting straight line corresponding to the dust concentration data, obtain a target fitting value at the sampling time of the dust concentration data according to the second fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data; If the number of dust concentration data after the dust concentration data is less than m, obtain the first fitting straight line corresponding to the dust concentration data, obtain the target fitting value at the sampling time of the dust concentration data according to the first fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data.

[0035] It should be noted that setting , The first preset number is 2, which is not limited here. The implementer can set it according to the implementation scenario. The values ​​of n and m may not be equal. The linear fitting adopts the least squares method. The least squares method is used for linear fitting in the prior art and will not be described here.

[0036] (3) According to the mutation degree and local similarity of the i-th dust concentration data, the important coefficient of the i-th dust concentration data is obtained.

[0037] Specifically, the mean of the mutation degree and the local similarity is calculated, and the mean is subtracted from a constant 1 to obtain the important coefficient of the dust concentration data.

[0038] In one embodiment, the calculation formula of the important coefficient of the i-th dust concentration data is:

[0039] in, Represents the important coefficient of the i-th dust concentration data, Indicates the mutation degree of the i-th dust concentration data, represents the local similarity of the i-th dust concentration data, and 1 represents a constant.

[0040] It should be noted that The larger the value of is, the greater the probability that the i-th dust concentration data is a noise point. The smaller the value of , the less important the i-th dust concentration data is; The larger the value of , the more likely the data before and after the ith dust concentration data are to be noise points. The smaller the value, the less important it is in the dust concentration data series.

[0041] Similarly, the important coefficient of each dust concentration data in the dust concentration data sequence is calculated. In the subsequent process of using the autoregressive model to predict the dust concentration data at future moments, the important coefficient of each dust concentration data is used as the weight corresponding to each dust concentration data. A straight line fitting is performed based on the weighted data to reduce the interference of noise data on the prediction results and improve the accuracy of the prediction.

[0042] Considering that the autoregressive model is simple to calculate, can provide real-time data feedback, and can intelligently predict the dust concentration changes in the future based on historical data and current environmental change trends, the autoregressive model is used to predict the dust concentration at future times. Since the autoregressive model needs to be predicted based on the order, where the order refers to the number of historical data required to predict future data, the dust concentration data sequence is analyzed based on the information criterion, or the autocorrelation function and partial correlation function to obtain the order of the autoregressive model, and the order is recorded as , the last dust concentration data as the lag data, that is, assuming that there are N dust concentration data in the dust concentration data sequence, the lag data is the dust concentration data from the first to the Nth dust concentration data, wherein the order of the autoregressive model is obtained according to the information criterion, or the autocorrelation function and the partial correlation function is the prior art and will not be described in detail here.

[0043] The weight acquisition module 13 is used to perform straight-line fitting on all dust concentration data within the local range of each lag data according to the importance coefficient of each dust concentration data, and obtain the corresponding local fitting straight line of each lag data; according to the local fitting straight line of each lag data, obtain the importance of each lag data; according to the importance, obtain the adaptive weight of each lag data.

[0044] When using the autoregressive model to predict dust concentration at future moments, the prediction is based on the lagged data, and the weight of the lagged data is obtained according to the difference between the lagged data and the fitting line in its local range. Since the change in the speed of dust particles will cause the increase rate of dust concentration to change, it is reflected in the lagged data as a turning point in dust concentration, that is, there may be turning points in the lagged data. These turning points will have lower weights in the fitting line, but in fact these turning points have a greater impact on the data after the turning point, and are more important, so they should be given higher weights.

[0045] Since the turning point is the change in the dust concentration caused by the change in the speed of dust particles, there is a big difference in the data change trend before and after the turning point, that is, the turning point has a big difference with a fitting straight line of the data before and after, and the similarity between the fitting straight lines of the data before and after the turning point is low. Therefore, the importance of each lagged data can be obtained according to the difference and similarity, and then the adaptive weight is given to each lagged data according to its importance, so that the greater the importance of the lagged data, the greater its corresponding adaptive weight, which improves the accuracy of dust concentration prediction using the autoregressive model.

[0046] In the example of the present invention, for the j-th lag data, first obtain the fitting straight line of the j-th lag data in its local range. In order to reduce the interference of noise data, the important coefficients of the dust concentration data are used as weights to obtain the corresponding weighted data, and the weighted data are used to perform straight line fitting to obtain the fitting straight line (local fitting straight line) of the j-th lag data in its local range. Specifically, in the dust concentration data sequence, the important coefficients of the u dust concentration data before the lag data are used as the weights of the corresponding dust concentration data to obtain the u weighted data before the lag data, and the u weighted data before the lag data are linearly fitted to obtain the front local fitting straight line of the lag data, wherein u is less than or equal to the order; If the number of dust concentration data after the lag data is greater than or equal to v, the important coefficients of the v dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain a posterior local fitting straight line of the lag data, wherein v is greater than or equal to a second preset number; If the number of dust concentration data after the lag data is less than v, and the lag data is not the last dust concentration data in the dust concentration data sequence, then the important coefficients of the lag data and all the dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain the lag data and v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain the posterior local fitting straight line of the lag data.

[0047] It should be noted that setting , The second preset number is 2, which is not limited here. The implementer can set it according to the implementation scenario. The values ​​of u and v may not be equal. The linear fitting adopts the least squares method. The least squares method is used for linear fitting in the prior art and will not be described here.

[0048] Further, according to the difference between the j-th lag data and the fitting straight line in its local range, the importance of the j-th lag data is obtained. Specifically, for any lag data, according to the first local fitting straight line of the lag data, a first fitting value at the sampling time of the lag data is obtained, and the absolute value of the difference between the lag data and the first fitting value is calculated to obtain a first difference index; If there is a posterior local fitting straight line for the lag data, a second fitting value at the sampling time of the lag data is obtained according to the posterior local fitting straight line of the lag data, and the absolute value of the difference between the lag data and the second fitting value is calculated to obtain a second difference index; Calculating the reciprocal of the mean of the first difference index and the second difference index to obtain a second variable, and subtracting the second variable from a constant 1 to obtain a local difference degree of the lagged data; Calculate the Spearman correlation coefficient between the front local fitting straight line and the rear local fitting straight line of the lag data, and subtract the Spearman correlation coefficient from a constant 1 to obtain a discrimination coefficient of the lag data; The importance coefficient, local difference degree and average value of the distinction coefficient of the lag data are calculated to obtain the importance degree of the lag data.

[0049] In one embodiment, the calculation formula for the importance of the j-th lagging data is:

[0050] in, Indicates the importance of the j-th lagged data, represents the first difference index, represents the second difference index, represents the Spearman correlation coefficient between the front local fitting line and the back local fitting line of the j-th lag data, represents the important coefficient of the j-th lag data, 1 represents a constant, Represents the absolute value symbol.

[0051] It should be noted that is the second variable, That is, the local difference degree of the j-th lag data. The greater the difference between the j-th lag data and its previous and subsequent fitted values, the smaller the value of the second variable, indicating that the j-th lag data is more likely to be a turning point, and thus the local difference degree of the j-th lag data is greater. The larger the value of is, the greater the importance of the j-th lagged data; is the discrimination coefficient of the jth lag data, The smaller the value, the lower the similarity between the front and back fitting lines of the j-th lag data, which means that the j-th lag data is more likely to be a turning point, and thus the greater the difference coefficient of the j-th lag data. The larger the value of , the greater the importance of the j-th lagged data.

[0052] In particular, if there is no local fitting straight line for the lagged data, the reciprocal of the first difference index is subtracted from the constant 1, and the importance coefficient of the lagged data and the average of the reciprocal are calculated to obtain the importance of the lagged data, that is, the calculation formula for the importance of the j-th lagged data is: .

[0053] Similarly, the importance of each lagged data is obtained, and then a corresponding adaptive weight is assigned to each lagged data. Specifically, the average value of the importance of all lagged data is obtained to obtain the average importance; for any lagged data, the ratio between the importance of the lagged data and the average importance is calculated to obtain the adaptive weight of the lagged data.

[0054] In one implementation, taking the jth lagged data as an example, the calculation formula for the adaptive weight of the jth lagged data is:

[0055] in, represents the adaptive weight of the j-th lagged data, Indicates the importance of the j-th lagged data, Indicates the importance of the kth lagged data, Represents the number of all lagged data, that is, the order.

[0056] It should be noted that That is, the average importance, The larger the value of is, the greater the possibility that the jth lagged data is a turning point. The larger the value of , the greater the adaptive weight of the j-th lag data.

[0057] Similarly, the adaptive weight of each lagged data in the dust concentration data series is obtained.

[0058] The data prediction module 14 is used to obtain dust concentration prediction data at the next moment according to the adaptive weight of each of the lagging data, and take dust reduction measures at the mine entrance according to the dust concentration prediction data.

[0059] The weight acquisition module 13 obtains the adaptive weight of each lagged data in the dust concentration data sequence. According to the adaptive weight of each lagged data, the autoregressive model is used to obtain the dust concentration prediction data at the next moment, so as to improve the accuracy of the dust concentration prediction using the autoregressive model. The alarm concentration is set according to the specific implementation scenario. When the dust concentration prediction data reaches the alarm concentration, an alarm is triggered to remind the staff to take dust reduction measures in time, where the dust reduction measures include but are not limited to: transporting water to the rotating nozzles installed in each area through the high-pressure main engine, turning on the switch of the alarm area, spraying a large area of ​​water mist particles, and combining with the dust particles to increase the dust weight, so as to achieve rapid dust reduction.

[0060] It is worth noting that the focus of the present invention is how to obtain the adaptive weight of each lagged data according to the importance of each dust concentration data in the dust concentration data sequence, and then use the autoregressive model to predict the dust concentration data at the next moment, so as to improve the accuracy of the dust concentration prediction using the autoregressive model. It is a prior art to set the alarm concentration according to the specific implementation scenario and take dust reduction measures in time, which will not be repeated here.

[0061] In summary, the present invention provides a sensor-based online monitoring and dust reduction system for mining dust concentration, including a data acquisition module for collecting dust concentration data in a historical period after blasting at a mine entrance to obtain a dust concentration data sequence; a data analysis module for obtaining important coefficients of each dust concentration data according to the difference in dust concentration data in the dust concentration data sequence, and in the process of predicting the dust concentration data sequence using an autoregressive model, performing autocorrelation analysis on the dust concentration data sequence to obtain the order of the autoregressive model, and obtaining the dust concentration data sequence according to the order of the autoregressive model. The lagging data in the column; a weight acquisition module, used to perform straight-line fitting on all dust concentration data within the local range of each of the lag data according to the importance coefficient of each of the dust concentration data, and obtain the local fitting straight line of each of the lag data; the importance of each of the lag data is obtained according to the local fitting straight line of each of the lag data; and the adaptive weight of each of the lag data is obtained according to the importance; a data prediction module, used to obtain the dust concentration prediction data of the next moment according to the adaptive weight of each of the lag data, and take dust reduction measures for the mine entrance according to the dust concentration prediction data. Among them, according to the difference of dust concentration data in the dust concentration data sequence, the important coefficient of each dust concentration data is obtained; in the process of using the autoregressive model to predict the dust concentration data at the next moment, the dust concentration data sequence is subjected to autocorrelation analysis to obtain the lagged data, and a straight line fitting is performed according to the dust concentration data in the local range of the lagged data, so that in the process of straight line fitting, the weight of the suspected noise data in the dust concentration data sequence is smaller, reducing the interference to the prediction result; according to the difference between each lagged data and its corresponding fitting result (local fitting straight line), the importance of each lagged data is obtained; according to the importance, the adaptive weight of each lagged data is obtained, and then the dust concentration data at the next moment is predicted according to the adaptive weight of each lagged data, thereby improving the accuracy of dust concentration prediction using the autoregressive model.

[0062] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A sensor-based online monitoring and dust reduction system for mining dust concentration, characterized in that: The sensor-based mining dust concentration online monitoring and dust reduction system includes: The data collection module is used to collect dust concentration data in the historical period after blasting at the mine entrance to obtain a dust concentration data sequence; A data analysis module, for obtaining important coefficients of each dust concentration data according to the difference of dust concentration data in the dust concentration data sequence, performing autocorrelation analysis on the dust concentration data sequence in the process of predicting the dust concentration data sequence using an autoregressive model to obtain the order of the autoregressive model, and obtaining the lagged data in the dust concentration data sequence according to the order of the autoregressive model; A weight acquisition module, used for performing straight line fitting on all dust concentration data within a local range of each lag data according to the importance coefficient of each dust concentration data, and correspondingly obtaining a local fitting straight line for each lag data, obtaining the importance of each lag data according to the local fitting straight line of each lag data, and obtaining an adaptive weight of each lag data according to the importance; The data prediction module is used to obtain dust concentration prediction data at the next moment according to the adaptive weight of each of the lagging data, and take dust reduction measures at the mine entrance according to the dust concentration prediction data.

2. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 1 is characterized in that: The data analysis module obtains important coefficients of each dust concentration data according to the difference of dust concentration data in the dust concentration data sequence, including: Constructing a curve graph according to the dust concentration data, wherein the horizontal axis of the curve graph represents time and the vertical axis represents the dust concentration data; For any dust concentration data in the curve graph, the mutation degree of the dust concentration data is obtained according to the variation trend of the dust concentration data in its local range; According to the similarity between the previous and next data of the dust concentration data, the local similarity of the dust concentration data is obtained; The mean of the mutation degree and the local similarity is calculated, and the important coefficient of the dust concentration data is obtained by subtracting the mean from a constant 1.

3. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 2 is characterized in that: The data analysis module obtains the mutation degree of the dust concentration data according to the change trend of the dust concentration data in its local range, including: In the curve graph, if the dust concentration data is not the first dust concentration data or the last dust concentration data in the curve graph, respectively obtain the left dust concentration data adjacent to the left of the dust concentration data, and the right dust concentration data adjacent to the right of the dust concentration data, and calculate the left slope between the dust concentration data and the left dust concentration data, and the right slope between the dust concentration data and the right dust concentration data; Calculate the absolute value of the difference between the left slope and the right slope, add the absolute value of the difference with a preset multiple to a preset constant, use the reciprocal of the addition result as a first variable, subtract the first variable from a constant 1, and obtain the mutation degree of the dust concentration data; If the dust concentration data is the first dust concentration data in the curve graph, obtain the right slope corresponding to the dust concentration data, add the right slope to the preset constant, use the reciprocal of the addition result as the first variable, subtract the first variable from the constant 1, and obtain the mutation degree of the dust concentration data; If the dust concentration data is the last dust concentration data in the curve graph, obtain the left slope corresponding to the dust concentration data, add the left slope to the preset constant, take the inverse of the addition result as the first variable, subtract the first variable from the constant 1, and obtain the mutation degree of the dust concentration data.

4. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 2 is characterized in that: The data analysis module obtains the local similarity of the dust concentration data according to the similarity between the previous and next data of the dust concentration data, including: In the dust concentration data sequence, if the number of dust concentration data before the dust concentration data is greater than or equal to n, and the number of dust concentration data after the dust concentration data is greater than or equal to m, a straight line fitting is performed on the n dust concentration data before the dust concentration data to obtain a first fitting straight line, and a straight line fitting is performed on the m dust concentration data after the dust concentration data to obtain a second fitting straight line, wherein n and m are both greater than a first preset number, and the Spearman correlation coefficient between the first fitting straight line and the second fitting straight line is calculated to obtain the local similarity of the dust concentration data; If the number of dust concentration data before the dust concentration data is less than n, obtain a second fitting straight line corresponding to the dust concentration data, obtain a target fitting value at the sampling time of the dust concentration data according to the second fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data; If the number of dust concentration data after the dust concentration data is less than m, obtain the first fitting straight line corresponding to the dust concentration data, obtain the target fitting value at the sampling time of the dust concentration data according to the first fitting straight line, calculate the absolute value of the difference between the dust concentration data and the target fitting value, and obtain the local similarity of the dust concentration data.

5. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 1 is characterized in that: The weight acquisition module performs straight line fitting on all dust concentration data within a local range of each lag data according to the important coefficient of each dust concentration data, and obtains a local fitting straight line for each lag data, including: For any lag data, in the dust concentration data sequence, the important coefficients of u dust concentration data before the lag data are used as the weights of the corresponding dust concentration data to obtain u weighted data before the lag data, and a straight line is fitted to the u weighted data before the lag data to obtain a front local fitting straight line of the lag data, wherein u is less than or equal to the order; If the number of dust concentration data after the lag data is greater than or equal to v, the important coefficients of the v dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain a posterior local fitting straight line of the lag data, wherein v is greater than or equal to a second preset number; If the number of dust concentration data after the lag data is less than v, and the lag data is not the last dust concentration data in the dust concentration data sequence, then the important coefficients of the lag data and all the dust concentration data after the lag data are used as the weights of the corresponding dust concentration data to obtain the lag data and v weighted data after the lag data, and a straight line is fitted to the v weighted data after the lag data to obtain the posterior local fitting straight line of the lag data.

6. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 5 is characterized in that: The weight acquisition module obtains the importance of each of the lagged data according to the local fitting straight line of each of the lagged data, including: For any lag data, according to the previous local fitting straight line of the lag data, a first fitting value at the sampling time of the lag data is obtained, and the absolute value of the difference between the lag data and the first fitting value is calculated to obtain a first difference index; If there is no local fitting straight line for the lagged data, subtract the reciprocal of the first difference index from the constant 1, calculate the average of the importance coefficient of the lagged data and the reciprocal, and obtain the importance of the lagged data; If there is a rear local fitting straight line for the lag data, the importance of the lag data is obtained according to the difference between the front local fitting straight line and the rear local fitting straight line of the lag data.

7. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 6 is characterized in that: The weight acquisition module obtains the importance of the lag data according to the difference between the front local fitting straight line and the rear local fitting straight line of the lag data, including: According to the posterior local fitting straight line of the lagged data, a second fitting value at the sampling time of the lagged data is obtained, and the absolute value of the difference between the lagged data and the second fitting value is calculated to obtain a second difference index; Calculating the reciprocal of the mean of the first difference index and the second difference index to obtain a second variable, and subtracting the second variable from a constant 1 to obtain a local difference degree of the lagged data; Calculate the Spearman correlation coefficient between the front local fitting straight line and the rear local fitting straight line of the lag data, and subtract the Spearman correlation coefficient from a constant 1 to obtain a discrimination coefficient of the lag data; The importance coefficient, local difference degree and average value of the distinction coefficient of the lag data are calculated to obtain the importance degree of the lag data.

8. The sensor-based online monitoring and dust reduction system for mining dust concentration according to claim 1 is characterized in that: The weight acquisition module acquires the adaptive weight of each of the lagging data according to the importance, including: Get the average importance of all lagged data to get the average importance; For any lagging data, the ratio between the importance of the lagging data and the average importance is calculated to obtain the adaptive weight of the lagging data.

Images