Intelligent assessment method for fire risk of high-altitude tunnel

Abstract

The invention relates to the technical field of data processing, in particular to a high-altitude tunnel fire risk intelligent assessment method, which comprises the following steps: acquiring a monitoring data sequence of each fire monitoring index at each monitoring point of reference tunnels at different altitudes; acquiring a multi-dimensional risk assessment characteristic value of each reference tunnel according to the data change trend in each monitoring data sequence and the monitoring data difference of each fire monitoring index at different monitoring points at the same monitoring moment; according to the altitude and the fire accident score of each reference tunnel and the multi-dimensional risk assessment feature value of each reference tunnel, obtaining the importance degree of each dimensional risk assessment feature value of the target tunnel; and according to the corresponding multi-dimensional risk assessment feature value of the target tunnel in the current time period, the importance degree of each dimensional risk assessment feature value and the fire accident score of each reference tunnel, the fire risk of the current tunnel is assessed, and the accuracy of the fire risk assessment result of the high-altitude tunnel is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent method for assessing the risk of fires in high-altitude tunnels. Background Technology

[0002] Tunnels are important hubs in modern transportation. However, due to their unique structural characteristics such as enclosed space and poor ventilation, dense smoke and high temperatures can spread rapidly in the event of a fire, making evacuation difficult. Therefore, assessing the fire risk in tunnels in advance and eliminating fire safety hazards is a crucial step in ensuring traffic safety within tunnels.

[0003] Existing methods for assessing tunnel fire risk primarily involve training a model using historical fire risk monitoring and assessment data to obtain a tunnel fire risk assessment model. The assessment results are then obtained by substituting the fire risk monitoring data into this model. However, in high-altitude environments, due to environmental characteristics such as low air pressure, low oxygen, and complex weather, the combustion conditions, flame propagation patterns, and evacuation difficulties during a fire differ from those in ordinary tunnels. Therefore, it is impossible to obtain accurate assessment results using a unified tunnel fire risk assessment model.

[0004] Therefore, how to effectively improve the accuracy of fire risk assessment results for high-altitude tunnels has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide an intelligent assessment method for fire risk in high-altitude tunnels to address the problem of effectively improving the accuracy of fire risk assessment results for high-altitude tunnels.

[0006] This invention provides an intelligent assessment method for fire risk in high-altitude tunnels, which includes the following steps: A preset number of reference tunnels at different altitudes are obtained. For any reference tunnel, during the historical period when a fire accident occurs in any reference tunnel, the monitoring data of each fire monitoring index at each monitoring point of any reference tunnel at each monitoring time are obtained, and the monitoring data sequence of each fire monitoring index at each monitoring point is obtained. Based on the data change trends in each monitoring data sequence and the differences in monitoring data of each fire monitoring indicator at different monitoring points at the same monitoring time, the multidimensional risk assessment characteristic value of any reference tunnel is obtained. Based on the altitude and fire accident score of each reference tunnel, as well as the multidimensional risk assessment feature value of each reference tunnel, the relationship curve between altitude and the importance of each dimension of risk assessment feature value is obtained. Based on the altitude of the target tunnel and the relationship curve between altitude and the importance of each dimension of risk assessment feature value, the importance of each dimension of risk assessment feature value of the target tunnel is obtained. Based on the multi-dimensional risk assessment characteristic value of the target tunnel in the current time period, the importance of each risk assessment characteristic value of the target tunnel, and the fire accident score of each reference tunnel, the fire risk score of the current tunnel in the current time period is obtained.

[0007] Preferably, the step of obtaining the multidimensional risk assessment feature value of any reference tunnel based on the data change trend in each monitoring data sequence and the monitoring data differences of each fire monitoring indicator at different monitoring points at the same monitoring time includes: For any fire monitoring indicator, the monitoring data sequence of the fire monitoring indicator at each monitoring point of the reference tunnel is recorded as the target monitoring data sequence. Linear fitting is performed on the data in each target monitoring data sequence to obtain the fitting line of each target monitoring data sequence. The slope of each fitting line is obtained. Among all slopes, the maximum value is taken as the trend value of the fire monitoring indicator. The trend value is linearly normalized to obtain the change trend characteristic value of the fire monitoring indicator. Based on the differences in monitoring data of any fire monitoring index at different monitoring points of any reference tunnel at the same monitoring time, the abnormal propagation speed of any fire monitoring index is obtained. Obtain the abnormal propagation speed of each fire monitoring index under any reference tunnel, calculate the average value of the abnormal propagation speed of all fire monitoring indices as the fire spread speed of any reference tunnel, and linearly normalize the fire spread speed to obtain the fire spread characteristic value of any reference tunnel. Obtain the trend characteristic value of each fire monitoring indicator, and combine the trend characteristic values ​​of all fire monitoring indicators with the fire spread characteristic value to form the multidimensional risk assessment characteristic value of any reference tunnel.

[0008] Preferably, the step of obtaining the abnormal propagation speed of any fire monitoring indicator based on the difference in monitoring data of any fire monitoring indicator at different monitoring points of any reference tunnel at the same monitoring time includes: For any given monitoring time, calculate the standard deviation of the monitoring data of any fire monitoring index at all monitoring points at that given monitoring time, and record it as the data difference degree of any fire monitoring index at that given monitoring time; The data difference degree of each fire monitoring indicator at each monitoring time is obtained to obtain a data difference degree sequence. The absolute value of the difference between each two adjacent data in the data difference degree sequence is calculated. All the absolute values ​​of the difference are accumulated to obtain the abnormal propagation speed of each fire monitoring indicator.

[0009] Preferably, the step of obtaining the relationship curve between altitude and the importance of each dimension of risk assessment feature value based on the altitude and fire accident score of each reference tunnel, and the multidimensional risk assessment feature value of each reference tunnel, includes: Based on the elevation similarity between each reference tunnel, all reference tunnels are divided into at least three elevation similarity groups; For any one-dimensional risk assessment feature value, based on the correlation between the risk assessment feature value of the reference tunnel in each altitude similarity group and the fire accident score, the importance of the risk assessment feature value of the reference tunnel in each altitude similarity group to the fire accident score is obtained respectively. The average elevation of all reference tunnels in each similar elevation group is obtained and recorded as the average elevation of each similar elevation group. Based on the average elevation of each similar elevation group and the importance of any one-dimensional risk assessment feature value in the fire accident score under each similar elevation group, a scatter plot is constructed. The horizontal axis of the scatter plot represents the average elevation, and the vertical axis represents the importance of any one-dimensional risk assessment feature value in the fire accident score. Obtain the weight of each data point in the scatter plot. Based on the weight of each data point in the scatter plot, use the weighted least squares method to fit the data points in the scatter plot to obtain the relationship curve between altitude and the importance of any one-dimensional risk assessment feature value.

[0010] Preferably, the step of dividing all reference tunnels into at least three altitude similarity groups based on the altitude similarity between each reference tunnel includes: The elevation of each reference tunnel is mapped onto a number axis. For any data point on the number axis, the reciprocal of the distance between the data point and other data points on the number axis is calculated. All reciprocals are summed to obtain the density of the position of the data point on the number axis. Obtain the density corresponding to each data point on the number axis, and record the data point corresponding to the maximum value among all densities as the target data point. On the number axis, form an altitude similarity group by the number of target data points closest to the target data point and the reference tunnel corresponding to the target data point. Remove the data points corresponding to the altitude similarity group from the number axis to obtain a new number axis. Use the new number axis as the number axis and repeat the steps of obtaining the density of each data point on the number axis to obtain the altitude similarity group corresponding to the number axis until the number of data points on the number axis is less than the target number, and obtain at least three altitude similarity groups.

[0011] Preferably, the step of obtaining the importance of any dimension of risk assessment feature value to the fire accident score for each altitude similarity group based on the correlation between the reference tunnel's risk assessment feature value and the fire accident score includes: For any altitude similarity group, the risk assessment feature values ​​of any dimension of all reference tunnels in the altitude similarity group are combined into a feature value sequence, and the fire accident scores of all reference tunnels in the altitude similarity group are combined into an accident score sequence. In the feature value sequence and the accident score sequence, the reference tunnels corresponding to two data points at the same location are the same. Calculate the absolute value of the Pearson correlation coefficient between the feature value sequence and the accident score sequence to obtain the importance of any dimension of risk assessment feature value to the fire accident score under any altitude similarity group.

[0012] Preferably, obtaining the weight of each data point in the scatter plot includes: For any data point in the scatter plot, the altitude similarity group corresponding to the data point is recorded as the target altitude similarity group. The density of each reference tunnel in the target altitude similarity group is obtained in the number axis. The average density of all reference tunnels in the target altitude similarity group is linearly normalized to obtain the weight of any data point in the scatter plot.

[0013] Preferably, the step of obtaining the fire risk score of the current tunnel in the current time period based on the multi-dimensional risk assessment feature value corresponding to the target tunnel in the current time period, the importance of each dimension of the risk assessment feature value under the target tunnel, and the fire accident score of each reference tunnel includes: The multidimensional risk assessment feature value of the target tunnel in the current time period and the multidimensional risk assessment feature value of each reference tunnel are mapped to a multidimensional coordinate system, where each coordinate axis of the multidimensional coordinate system represents the risk assessment feature value of each dimension. In the process of clustering based on the distance between every two data points in the multidimensional coordinate system, the importance of each risk assessment feature value under the target tunnel is used as a weight to obtain the weighted distance between every two data points in the multidimensional coordinate system. Based on the weighted distance between every two data points in the multidimensional coordinate system, the data points in the multidimensional coordinate system are clustered to obtain the target cluster to which the target tunnel belongs. Based on the weighted distance between each reference tunnel in the target cluster and the corresponding data point of the target tunnel, and the fire accident score of all reference tunnels in the target cluster, the fire risk score of the target tunnel in the current time period is obtained.

[0014] Preferably, the step of obtaining the fire risk score of the target tunnel in the current time period based on the weighted distance between the corresponding data points of each reference tunnel in the target cluster and the target tunnel, and the fire accident scores of all reference tunnels in the target cluster, includes: The reciprocal of the weighted distance between each reference tunnel in the target cluster and the corresponding data point of the target tunnel is linearly normalized to obtain the reference degree of each reference tunnel in the target cluster to the target tunnel. Using the degree of reference of each reference tunnel in the target cluster to the target tunnel as a weight, the weighted average of the fire accident scores of all reference tunnels in the target cluster is calculated as the fire risk score of the target tunnel in the current time period.

[0015] Preferably, the multidimensional risk assessment feature values ​​also include wind speed feature values, then obtaining the wind speed feature value corresponding to any reference tunnel includes: Obtain the wind speed monitoring data sequence at each monitoring point of any reference tunnel, calculate the average value of all data in all monitoring data sequences corresponding to the wind speed, and linearly normalize the average value to obtain the wind speed characteristic value of any reference tunnel.

[0016] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: This invention acquires a predetermined number of reference tunnels at different altitudes. For any reference tunnel, during the historical period of a fire accident in that reference tunnel, it acquires monitoring data for each fire monitoring indicator at each monitoring point of that reference tunnel at each monitoring time, resulting in a monitoring data sequence for each fire monitoring indicator at each monitoring point. Based on the data change trend in each monitoring data sequence and the differences in monitoring data for each fire monitoring indicator at different monitoring points at the same monitoring time, it acquires a multidimensional risk assessment feature value for that reference tunnel. Based on the altitude and fire accident score of each reference tunnel, and the multidimensional risk assessment feature value of each reference tunnel, it acquires a relationship curve between altitude and the importance of each dimension of risk assessment feature value. Based on the altitude of the target tunnel and the relationship curve between altitude and the importance of each dimension of risk assessment feature value, it acquires the importance of each dimension of risk assessment feature value of the target tunnel. Based on the multidimensional risk assessment feature value of the target tunnel in the current time period, the importance of each dimension of risk assessment feature value under the target tunnel, and the fire accident score of each reference tunnel, it acquires the fire risk score of the current tunnel in the current time period. Specifically, based on the altitude, fire accident score, and multi-dimensional risk assessment feature value of each reference tunnel, a relationship curve between altitude and the importance of each risk assessment feature value is obtained. This curve reflects the influence of altitude changes on the importance of different risk assessment feature values. Then, based on the altitude of the target tunnel, the importance of each risk assessment feature value of the target tunnel is obtained. This is used to adjust the weight of each risk assessment feature value when conducting risk assessment on the target tunnel, and obtain the fire risk score of the target tunnel in the current time period. This effectively improves the accuracy of fire risk assessment results for high-altitude tunnels. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for intelligent assessment of fire risk in high-altitude tunnels provided in Embodiment 1 of the present invention. Detailed Implementation

[0019] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0020] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0021] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0022] See Figure 1 This is a flowchart of a method for intelligent assessment of fire risk in high-altitude tunnels provided in Embodiment 1 of the present invention. Figure 1 As shown, the method may include: Step S101: Obtain a preset number of reference tunnels at different altitudes. For any reference tunnel, during the historical period in which a fire accident occurred in the reference tunnel, obtain the monitoring data of each fire monitoring index at each monitoring point of the reference tunnel at each monitoring time, and obtain the monitoring data sequence of each fire monitoring index at each monitoring point.

[0023] Existing methods for assessing tunnel fire risk primarily involve training a model using historical fire risk monitoring and assessment data to obtain a tunnel fire risk assessment model. The assessment results are then obtained by substituting the fire risk monitoring data into this model. However, in high-altitude environments, due to environmental characteristics such as low air pressure, low oxygen, and complex weather, the combustion conditions, flame propagation patterns, and evacuation difficulties during a fire differ from those in ordinary tunnels. Therefore, it is impossible to obtain accurate assessment results using a unified tunnel fire risk assessment model.

[0024] Therefore, in this embodiment of the invention, tunnels that have experienced fire accidents and are located at different altitudes are first identified and designated as reference tunnels. Then, based on the historical fire risk monitoring data of each reference tunnel before the historical fire accident, multidimensional risk assessment feature values ​​are obtained for each reference tunnel to reflect the environmental characteristics of each reference tunnel before the fire accident. By comparing the differences in the multidimensional risk assessment feature values ​​of reference tunnels at different altitudes, the influence of altitude changes on the importance of each dimension of risk assessment feature value is analyzed. Subsequently, based on the altitude of the tunnel to be assessed, the influence weight of each dimension of risk assessment feature value of the tunnel to be assessed is adjusted to conduct a fire risk assessment of the tunnel to be assessed, thereby improving the accuracy of fire risk assessment results for high-altitude tunnels.

[0025] All tunnels that have experienced fires within the last 5 years and are located at different altitudes are designated as reference tunnels. To improve the stability and accuracy of the analysis results, the number of reference tunnels must not be less than 30. If the number of reference tunnels that have experienced fires within the last 5 years is less than 30, the time range needs to be expanded. There is no limit to this; implementers can set the number of reference tunnels according to the proposed scenario.

[0026] Then, historical fire risk monitoring data for each reference tunnel is acquired to obtain multidimensional risk assessment characteristic values ​​for each reference tunnel. This is based on the understanding that fire risk assessment within tunnels primarily relies on temperature and gases (oxygen). ,carbon dioxide carbon monoxide (etc.) concentration, smoke concentration, etc., therefore, in the embodiments of the present invention, temperature, concentration, concentration, Concentration and smoke concentration are used as fire monitoring indicators, and are detected by temperature sensors, smoke detectors, sensor, sensor, The sensor is set to a monitoring frequency of 0.2Hz to acquire monitoring data of each fire monitoring indicator at each monitoring moment within 5 minutes before the fire occurs in each reference tunnel. That is, the 5 minutes before the fire occurs in each reference tunnel is taken as the historical period of the fire in each reference tunnel. There are no restrictions here. Implementers can set the types, number, monitoring frequency, and duration of historical periods according to the scenario.

[0027] Considering that the tunnel is relatively long and the data at different locations inside may differ, in this embodiment of the invention, taking the w-th reference tunnel as an example, the historical fire risk monitoring data of each fire monitoring index of each monitoring point of the w-th reference tunnel at each monitoring time within the historical period are obtained to obtain the monitoring data sequence of each monitoring point under each fire monitoring index, which is used to obtain the multidimensional risk assessment feature value of each reference tunnel. The distance between any two adjacent monitoring points is 100m, which is not limited here and can be set by the implementer according to the specific scenario.

[0028] Step S102: Based on the data change trend in each monitoring data sequence and the differences in monitoring data of each fire monitoring indicator at different monitoring points at the same monitoring time, obtain the multidimensional risk assessment feature value of any reference tunnel.

[0029] When a fire occurs, it usually starts from one point and then spreads to other places. Therefore, the environmental data at the point of ignition is more representative. In this embodiment of the invention, taking the c-th fire monitoring index as an example, the least squares method is used to linearly fit the data in the monitoring data sequence of the c-th fire monitoring index at each monitoring point of the w-th reference tunnel, so as to obtain the fitted straight line of the c-th fire monitoring index at each monitoring point of the w-th reference tunnel. The slope of each fitted straight line is obtained. The more obvious the change in the data of the c-th fire monitoring index at the monitoring point corresponding to the fitted straight line with the largest slope, that is, the closer the monitoring point corresponding to the fitted straight line with the largest slope is to the point of ignition. Therefore, among all the slopes, the maximum value is taken as the trend value of the c-th fire monitoring index, which is used to characterize the data change characteristics of the c-th fire monitoring index before the fire occurs. For the convenience of subsequent analysis, the trend value of the c-th fire monitoring index is linearly normalized and recorded as the change trend characteristic value of the c-th fire monitoring index. Similarly, the characteristic values ​​of the changing trends of each fire monitoring indicator under the w-th reference tunnel are obtained, and these characteristic values ​​are used to form the multidimensional risk assessment characteristic values ​​of the w-th reference tunnel, including: the characteristic value of the changing trend of temperature, denoted as... The trend value of smoke concentration change is denoted as ; The trend value of concentration change is denoted as ; The trend value of concentration change is denoted as ; The trend value of concentration change is denoted as Linear fitting using the least squares method is an existing technique and will not be elaborated upon here.

[0030] When a fire occurs, the changes in monitoring data at different monitoring points reflect the speed of fire spread and the degree of smoke or toxic gas accumulation. Therefore, it is necessary to obtain the fire spread characteristic value of the w-th reference tunnel based on the differences in monitoring data of each fire monitoring indicator at different monitoring points of the w-th reference tunnel at the same monitoring time. The fire spread characteristic value is then added to the multidimensional risk assessment characteristic value of the w-th reference tunnel to fully reflect the environmental characteristics of the w-th reference tunnel before the fire.

[0031] The steps for obtaining the fire spread characteristic value of the w-th reference tunnel are as follows: (1) Taking the cth fire monitoring index as an example, based on the difference in monitoring data of the cth fire monitoring index at different monitoring points of the wth reference tunnel at the same monitoring time, the abnormal propagation speed of the cth fire monitoring index is obtained.

[0032] Specifically: If the fire spreads quickly, the monitoring data at different monitoring points in the tunnel will become closer and closer as the fire spreads. Otherwise, the monitoring data at different monitoring points will continue to have large differences. Therefore, for any monitoring time, the standard deviation of the monitoring data of the cth fire monitoring index at all monitoring points of the wth reference tunnel at any monitoring time is calculated and denoted as the data difference degree of the cth fire monitoring index at any monitoring time. Similarly, following the method for obtaining the abnormal propagation speed of the c-th fire monitoring indicator, the data difference degree of the c-th fire monitoring indicator at each monitoring time is obtained to obtain a data difference degree sequence. Furthermore, the changing trend of the data difference degree at continuous monitoring times is used to characterize the propagation speed of the fire in the tunnel, that is, the absolute value of the difference between each two adjacent data in the data difference degree sequence is calculated, and all the absolute values ​​of the difference are accumulated to obtain the abnormal propagation speed of the c-th fire monitoring indicator of the w-th reference tunnel.

[0033] In one embodiment, the formula for calculating the abnormal propagation speed of the c-th fire monitoring index in the w-th reference tunnel is: in, This represents the abnormal propagation speed of the c-th fire monitoring indicator in the w-th reference tunnel, where m represents the number of monitoring moments. This represents the standard deviation of the monitoring data of the c-th fire monitoring index at all monitoring points of the w-th reference tunnel at the i-th monitoring time, which is also the data difference of the c-th fire monitoring index of the w-th reference tunnel at the i-th monitoring time. This represents the data difference of the c-th fire monitoring index of the w-th reference tunnel at the (i+1)-th monitoring time. Represents the absolute value symbol.

[0034] It should be noted that, since the monitoring frequency in this embodiment of the invention is 0.2Hz and the monitoring duration is 5 minutes, m=60; The larger the value, the faster the standard deviation of the monitoring data at different monitoring points changes at adjacent monitoring times. This means the monitoring data of the c-th fire monitoring indicator at each monitoring point within the w-th tunnel converge rapidly. The larger the value, the faster the abnormal propagation speed of the c-th fire monitoring indicator in the w-th reference tunnel, that is, the faster the fire spreads in the w-th tunnel.

[0035] (2) Obtain the abnormal propagation velocity of each fire monitoring indicator under the w-th reference tunnel. Since a fire will affect multiple fire monitoring indicators simultaneously, calculate the average value of the abnormal propagation velocity of all fire monitoring indicators as the fire spread velocity of the w-th reference tunnel. Then, perform linear normalization on the fire spread velocity to obtain the fire spread characteristic value of the w-th reference tunnel, denoted as . .

[0036] Thus, the fire spread characteristic value of the w-th reference tunnel was obtained.

[0037] Considering that wind speed is also a key factor in the rate of fire spread, wind speed sensors are used to obtain the wind speed at each monitoring point in the w-th tunnel at each monitoring time. The average wind speed of all monitoring points at all monitoring times is linearly normalized to obtain the wind speed characteristic value of the w-th reference tunnel. This wind speed characteristic value is also added to the multidimensional risk assessment characteristic value of the w-th reference tunnel, denoted as […]. .

[0038] Thus, the acquisition of the multidimensional risk assessment feature values ​​for the w-th reference tunnel has been completed. These feature values ​​include: temperature change trend feature values. Trend of smoke concentration , Concentration trend value , Concentration trend value , Concentration trend value Wind speed characteristic value and fire spread characteristic values This fully reflects the environmental characteristics of the w-th reference tunnel before the fire occurred.

[0039] Step S103: Based on the altitude and fire accident score of each reference tunnel, and the multidimensional risk assessment feature value of each reference tunnel, obtain the relationship curve between altitude and the importance of each dimension of risk assessment feature value. Based on the altitude of the target tunnel and the relationship curve between altitude and the importance of each dimension of risk assessment feature value, obtain the importance of each dimension of risk assessment feature value of the target tunnel.

[0040] Following the method for obtaining the multidimensional risk assessment feature value of the w-th reference tunnel in step S102, obtain the multidimensional risk assessment feature values ​​of all reference tunnels, as well as the multidimensional risk assessment feature value of the tunnel to be assessed in the current time period (5 minutes before the current time).

[0041] Since the occurrence and severity of tunnel fires are mainly related to the temperature and oxygen concentration inside the tunnel, the generation of toxic gases after the fire starts, the speed of fire spread, and the speed of personnel evacuation, the oxygen, air pressure, and climate conditions in the environment will change significantly with increasing altitude. For example, the low-oxygen environment at high altitudes may lead to incomplete combustion during a fire, which may produce more toxic gases. The complexity of the climate at high altitudes may also affect ventilation and fire spread in the tunnel. That is, when a fire occurs, the multidimensional risk assessment characteristic values ​​of tunnels at different altitudes will be significantly different. At the same time, even if the multidimensional risk assessment characteristic values ​​of tunnels at different altitudes are the same, the degree of fire risk of tunnels at different altitudes may also be different.

[0042] Therefore, in this embodiment of the invention, firstly, based on the damage status of each reference tunnel after a fire accident, relevant professionals score the fire accident of each reference tunnel to obtain a score value for each reference tunnel. To facilitate subsequent analysis, the score values ​​and altitude values ​​of all reference tunnels are linearly normalized to obtain the altitude and fire accident score for each reference tunnel. Then, the differences between the multidimensional risk assessment feature values ​​and fire accident scores of reference tunnels at different altitudes are compared to analyze the influence of altitude changes on the importance of each dimension of risk assessment feature value. The relationship curve between altitude and the importance of each dimension of risk assessment feature value is obtained to reflect the contribution of each dimension of risk assessment feature value to the fire accident score at different altitudes. Then, based on the altitude of the tunnel to be assessed and the corresponding multidimensional risk assessment feature value of the tunnel to be assessed in the current time period, the fire risk of the tunnel to be assessed in the current time period is assessed.

[0043] The steps for obtaining the relationship curve between altitude and the importance of each dimension of risk assessment feature value are as follows: (1) Based on the elevation similarity between each reference tunnel, all reference tunnels are divided into at least three elevation similarity groups.

[0044] Specifically: the elevation of each reference tunnel is mapped onto a number axis. For any data point on the number axis, the reciprocal of the distance between the data point and other data points on the number axis is calculated. All reciprocals are summed to obtain the density of the position of the data point on the number axis. Obtain the density corresponding to each data point on the number axis. Among all the densities, the data point corresponding to the maximum value is recorded as the target data point. On the number axis, the target number of data points closest to the target data point and the reference tunnel corresponding to the target data point are combined into an altitude similarity group. The target number is set to 10, which is not limited here. The implementer can set it according to the scenario. The data points corresponding to the altitude similarity groups are removed from the number axis to obtain a new number axis. This new number axis is then used as the new number axis. The process of obtaining the density of each data point on the number axis is repeated to obtain the altitude similarity groups corresponding to the number axis. This continues until the number of data points on the number axis is less than the target number, i.e., less than 10 data points. To avoid random interference caused by insufficient data, when the number of data points on the number axis is less than 10, the data points are discarded, resulting in all altitude similarity groups. It is assumed that the reference tunnels in the same altitude similarity group have similar altitudes, meaning that the contribution of each risk assessment feature value to the fire accident score is similar within the same altitude similarity group. It should be noted that to avoid insufficient data in subsequent analysis, the number of altitude similarity groups cannot be less than 3. Since the number of reference tunnels in each altitude similarity group is 10, the number of reference tunnels obtained in step S101 cannot be less than 30.

[0045] (2) Taking the t-th dimension risk assessment feature value as an example, based on the correlation between the t-th dimension risk assessment feature value of the reference tunnel in each altitude similarity group and the fire accident score, the importance of the t-th dimension risk assessment feature value to the fire accident score in each altitude similarity group is obtained respectively.

[0046] Specifically: Taking the qth altitude similarity group as an example, the t-th dimension risk assessment feature values ​​of all reference tunnels in the qth altitude similarity group are combined into a feature value sequence, and the fire accident scores of all reference tunnels in the qth altitude similarity group are combined into an accident score sequence. In the feature value sequence and the accident score sequence, the reference tunnels corresponding to two data points at the same location are the same. Calculate the absolute value of the Pearson correlation coefficient between the feature value sequence and the accident score sequence to obtain the importance of the t-th dimension risk assessment feature value to the fire accident score under the q-th altitude similarity group, denoted as . ; Similarly, we can determine the importance of the t-th dimension risk assessment feature value for fire accident scoring in each altitude similarity group.

[0047] (3) Based on the importance of the t-th dimension risk assessment feature value to the fire accident score under each similar altitude group, construct the relationship curve between altitude and the importance of the t-th dimension risk assessment feature value.

[0048] Specifically: the average elevation of all reference tunnels in each elevation similarity group is obtained and recorded as the average elevation of each elevation similarity group. Based on the average elevation of each elevation similarity group and the importance of the t-th dimension risk assessment feature value to the fire accident score in each elevation similarity group, a scatter plot is constructed. The horizontal axis of the scatter plot represents the average elevation, and the vertical axis represents the importance of the t-th dimension risk assessment feature value to the fire accident score. Each data point in the scatter plot corresponds to an elevation similarity group. Since the altitude similarity groups are obtained sequentially according to the density of the altitude of each reference tunnel on the number axis, the altitude difference between the reference tunnels in the later obtained altitude similarity groups may be larger than the altitude difference between the reference tunnels in the earlier obtained altitude similarity groups. That is, the altitude difference between the reference tunnels in the later obtained altitude similarity groups is larger, which may lead to the inaccuracy and low credibility of the importance of the t-th dimension risk assessment feature value corresponding to the average altitude in the fire accident score. Therefore, it is also necessary to obtain the credibility of the importance of the t-th dimension risk assessment feature value in the fire accident score based on the altitude difference between the reference tunnels in each altitude similarity group. For any data point in the scatter plot, the altitude similarity group corresponding to the data point is recorded as the target altitude similarity group. Referring to step (1), the density of the position of each reference tunnel in the target altitude similarity group on the number axis is obtained. The average value of the density corresponding to all reference tunnels in the target altitude similarity group is linearly normalized to obtain the weight of any data point in the scatter plot. The larger the average value, the closer the altitudes of the reference tunnels in the target altitude similarity group are, and the higher the credibility of the importance of the t-th dimension risk assessment feature value under the target altitude similarity group. Based on the weight of each data point in the scatter plot, the weighted least squares method is used to fit the data points in the scatter plot, obtaining the relationship curve between altitude and the importance of the t-th dimension risk assessment feature value. The weighted least squares method is existing technology and will not be elaborated upon here.

[0049] Thus, the relationship curve between altitude and the importance of the t-th dimension risk assessment feature value was obtained. Similarly, the relationship curve between altitude and the importance of each dimension risk assessment feature value was obtained.

[0050] Furthermore, any tunnel to be evaluated is designated as the target tunnel, and the altitude of the target tunnel (the value after linear normalization of the altitude value of the target tunnel) is obtained. The altitude of the target tunnel is then substituted into the relationship curve between altitude and the importance of each dimension of risk assessment feature value to obtain the importance of each dimension of risk assessment feature value of the target tunnel.

[0051] Step S104: Based on the multi-dimensional risk assessment feature value of the target tunnel in the current time period, the importance of each dimension of the risk assessment feature value under the target tunnel, and the fire accident score of each reference tunnel, obtain the fire risk score of the current tunnel in the current time period.

[0052] Following the steps in step S102 to obtain the multidimensional risk assessment feature value of the w-th reference tunnel, the multidimensional risk assessment feature value of the target tunnel in the current time period is obtained. Then, the multidimensional risk assessment feature value of the target tunnel in the current time period, as well as the multidimensional risk assessment feature value of each reference tunnel, are mapped to a multidimensional coordinate system. Each coordinate axis of the multidimensional coordinate system represents the risk assessment feature value of each dimension, so as to use the K-means clustering algorithm to cluster the data points in the multidimensional coordinate system and obtain the cluster to which the target tunnel belongs.

[0053] In the process of clustering data points using the K-means clustering algorithm based on the distance between each pair of data points in the multidimensional coordinate system, the importance of each risk assessment feature value of the target tunnel obtained in step S103 is used as a weight to obtain the weighted distance between each pair of data points in the multidimensional coordinate system. For example, assuming the multidimensional coordinate system is a three-dimensional coordinate system, where the x-axis, y-axis, and z-axis represent each risk assessment feature value, and the distance between the j-th data point and the k-th data point in the three-dimensional coordinate system is denoted as... ,but ,in, This represents the coordinates of the j-th data point on the x-axis. This represents the coordinates of the k-th data point on the x-axis. This represents the y-coordinate of the j-th data point. This represents the y-coordinate of the k-th data point. This represents the coordinates of the j-th data point on the z-axis. Let x represent the coordinates of the k-th data point on the z-axis. Assume the importance of the risk assessment feature value corresponding to the x-axis under the target tunnel is... The importance of the risk assessment feature value corresponding to the y-axis of the target tunnel is: The importance of the risk assessment feature value corresponding to the z-axis of the target tunnel is: Then the weighted distance between the j-th data point and the k-th data point in the three-dimensional coordinate system is .

[0054] Based on the weighted distance between every two data points in the multidimensional coordinate system, the data points in the multidimensional coordinate system are clustered to obtain the target cluster to which the target tunnel belongs. The K-means clustering algorithm is an existing technology and will not be described in detail here.

[0055] Furthermore, based on the weighted distance between the corresponding data points of each reference tunnel in the target cluster and the target tunnel, and the fire accident scores of all reference tunnels in the target cluster, the fire risk score of the target tunnel in the current time period is obtained, specifically: The reciprocal of the weighted distance between each reference tunnel in the target cluster and the corresponding data point of the target tunnel is linearly normalized to obtain the reference degree of each reference tunnel in the target cluster to the target tunnel. If the weighted distance between a certain reference tunnel in the target cluster and the target tunnel is smaller, it indicates that the environmental characteristics of the reference tunnel before the fire are more similar to the environmental characteristics of the target tunnel in the current time period. That is, the risk of the target tunnel being burned in the current time period is greater, and the fire accident score of the reference tunnel is more meaningful for the fire risk assessment of the target tunnel. Conversely, the risk of the target tunnel being burned in the current time period is smaller, and the reference degree of the reference tunnel to the target tunnel approaches 0. Using the degree of reference of each reference tunnel in the target cluster to the target tunnel as a weight, the weighted average of the fire accident scores of all reference tunnels in the target cluster is calculated as the fire risk score of the target tunnel in the current time period.

[0056] It should be noted that the greater the weighted distance between the corresponding data points of each reference tunnel in the target cluster and the target tunnel, the greater the difference between the environmental characteristics of the target tunnel in the current time period and the environmental characteristics of the reference tunnel before the fire. In other words, the lower the fire risk of the target tunnel in the current time period, and the lower the fire risk score of the target tunnel in the current time period. Conversely, the greater the fire risk of the target tunnel in the current time period, the closer the fire risk score of the target tunnel in the current time period is to the average fire accident score of all reference tunnels in the target cluster. In other words, the higher the fire risk score of the target tunnel in the current time period.

[0057] Furthermore, based on the fire risk score of the target tunnel during the current time period, the fire risk of the target tunnel during the current time period is assessed to facilitate timely fire response by relevant personnel. The method of assessing the fire risk of the target tunnel during the current time period based on its fire risk score is existing technology and will not be elaborated upon here.

[0058] In summary, this invention acquires a predetermined number of reference tunnels at different altitudes. For any reference tunnel, during the historical period of a fire accident in that reference tunnel, it acquires monitoring data for each fire monitoring indicator at each monitoring point of that reference tunnel at each monitoring time, resulting in a monitoring data sequence for each fire monitoring indicator at each monitoring point. Based on the data change trend in each monitoring data sequence and the differences in monitoring data for each fire monitoring indicator at different monitoring points at the same monitoring time, it acquires a multidimensional risk assessment feature value for that reference tunnel. Based on the altitude and fire accident score of each reference tunnel, and the multidimensional risk assessment feature value of each reference tunnel, it acquires a relationship curve between altitude and the importance of each dimension of risk assessment feature value. Based on the altitude of the target tunnel and the relationship curve between altitude and the importance of each dimension of risk assessment feature value, it acquires the importance of each dimension of risk assessment feature value of the target tunnel. Based on the multidimensional risk assessment feature value of the target tunnel in the current time period, the importance of each dimension of risk assessment feature value under the target tunnel, and the fire accident score of each reference tunnel, it acquires the fire risk score of the current tunnel in the current time period. Specifically, based on the altitude, fire accident score, and multi-dimensional risk assessment feature value of each reference tunnel, a relationship curve between altitude and the importance of each risk assessment feature value is obtained. This curve reflects the influence of altitude changes on the importance of different risk assessment feature values. Then, based on the altitude of the target tunnel, the importance of each risk assessment feature value of the target tunnel is obtained. This is used to adjust the weight of each risk assessment feature value when conducting risk assessment on the target tunnel, and obtain the fire risk score of the target tunnel in the current time period. This effectively improves the accuracy of fire risk assessment results for high-altitude tunnels.

[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for intelligent assessment of fire risk in high-altitude tunnels, characterized in that, The intelligent assessment method for fire risk in high-altitude tunnels includes: A preset number of reference tunnels at different altitudes are obtained. For any reference tunnel, during the historical period when a fire accident occurs in any reference tunnel, the monitoring data of each fire monitoring index at each monitoring point of any reference tunnel at each monitoring time are obtained, and the monitoring data sequence of each fire monitoring index at each monitoring point is obtained. Based on the data change trends in each monitoring data sequence and the differences in monitoring data of each fire monitoring indicator at different monitoring points at the same monitoring time, the multidimensional risk assessment characteristic value of any reference tunnel is obtained. Based on the altitude and fire accident score of each reference tunnel, as well as the multidimensional risk assessment feature value of each reference tunnel, the relationship curve between altitude and the importance of each dimension of risk assessment feature value is obtained. Based on the altitude of the target tunnel and the relationship curve between altitude and the importance of each dimension of risk assessment feature value, the importance of each dimension of risk assessment feature value of the target tunnel is obtained. Based on the multi-dimensional risk assessment characteristic value of the target tunnel in the current time period, the importance of each risk assessment characteristic value of the target tunnel, and the fire accident score of each reference tunnel, the fire risk score of the current tunnel in the current time period is obtained.

2. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 1, characterized in that, The process of obtaining the multidimensional risk assessment feature value of any reference tunnel based on the data change trend in each monitoring data sequence and the monitoring data differences of each fire monitoring indicator at different monitoring points at the same monitoring time includes: For any fire monitoring indicator, the monitoring data sequence of the fire monitoring indicator at each monitoring point of the reference tunnel is recorded as the target monitoring data sequence. Linear fitting is performed on the data in each target monitoring data sequence to obtain the fitting line of each target monitoring data sequence. The slope of each fitting line is obtained. Among all slopes, the maximum value is taken as the trend value of the fire monitoring indicator. The trend value is linearly normalized to obtain the change trend characteristic value of the fire monitoring indicator. Based on the differences in monitoring data of any fire monitoring index at different monitoring points of any reference tunnel at the same monitoring time, the abnormal propagation speed of any fire monitoring index is obtained. Obtain the abnormal propagation speed of each fire monitoring index under any reference tunnel, calculate the average value of the abnormal propagation speed of all fire monitoring indices as the fire spread speed of any reference tunnel, and linearly normalize the fire spread speed to obtain the fire spread characteristic value of any reference tunnel. Obtain the trend characteristic value of each fire monitoring indicator, and combine the trend characteristic values ​​of all fire monitoring indicators with the fire spread characteristic value to form the multidimensional risk assessment characteristic value of any reference tunnel.

3. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 2, characterized in that, The step of obtaining the abnormal propagation speed of any fire monitoring indicator based on the differences in monitoring data of any fire monitoring indicator at different monitoring points of any reference tunnel at the same monitoring time includes: For any given monitoring time, calculate the standard deviation of the monitoring data of any fire monitoring index at all monitoring points at that given monitoring time, and record it as the data difference degree of any fire monitoring index at that given monitoring time; The data difference degree of each fire monitoring indicator at each monitoring time is obtained to obtain a data difference degree sequence. The absolute value of the difference between each two adjacent data in the data difference degree sequence is calculated. All the absolute values ​​of the difference are accumulated to obtain the abnormal propagation speed of each fire monitoring indicator.

4. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 1, characterized in that, The process of obtaining the relationship curve between altitude and the importance of each dimension of risk assessment feature value based on the altitude and fire accident score of each reference tunnel, as well as the multidimensional risk assessment feature value of each reference tunnel, includes: Based on the elevation similarity between each reference tunnel, all reference tunnels are divided into at least three elevation similarity groups; For any one-dimensional risk assessment feature value, based on the correlation between the risk assessment feature value of the reference tunnel in each altitude similarity group and the fire accident score, the importance of the risk assessment feature value of the reference tunnel in each altitude similarity group to the fire accident score is obtained respectively. The average elevation of all reference tunnels in each similar elevation group is obtained and recorded as the average elevation of each similar elevation group. Based on the average elevation of each similar elevation group and the importance of any one-dimensional risk assessment feature value in the fire accident score under each similar elevation group, a scatter plot is constructed. The horizontal axis of the scatter plot represents the average elevation, and the vertical axis represents the importance of any one-dimensional risk assessment feature value in the fire accident score. Obtain the weight of each data point in the scatter plot. Based on the weight of each data point in the scatter plot, use the weighted least squares method to fit the data points in the scatter plot to obtain the relationship curve between altitude and the importance of any one-dimensional risk assessment feature value.

5. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 4, characterized in that, Based on the elevation similarity between each reference tunnel, all reference tunnels are divided into at least three elevation similarity groups, including: The elevation of each reference tunnel is mapped onto a number axis. For any data point on the number axis, the reciprocal of the distance between the data point and other data points on the number axis is calculated. All reciprocals are summed to obtain the density of the position of the data point on the number axis. Obtain the density corresponding to each data point on the number axis, and record the data point corresponding to the maximum value among all densities as the target data point. On the number axis, form an altitude similarity group by the number of target data points closest to the target data point and the reference tunnel corresponding to the target data point. Remove the data points corresponding to the altitude similarity group from the number axis to obtain a new number axis. Use the new number axis as the number axis and repeat the steps of obtaining the density of each data point on the number axis to obtain the altitude similarity group corresponding to the number axis until the number of data points on the number axis is less than the target number, and obtain at least three altitude similarity groups.

6. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 4, characterized in that, The step of obtaining the importance of any one-dimensional risk assessment feature value to the fire accident score for each altitude similarity group based on the correlation between the reference tunnel's risk assessment feature value and the fire accident score includes: For any altitude similarity group, the risk assessment feature values ​​of any dimension of all reference tunnels in the altitude similarity group are combined into a feature value sequence, and the fire accident scores of all reference tunnels in the altitude similarity group are combined into an accident score sequence. In the feature value sequence and the accident score sequence, the reference tunnels corresponding to two data points at the same location are the same. Calculate the absolute value of the Pearson correlation coefficient between the feature value sequence and the accident score sequence to obtain the importance of any dimension of risk assessment feature value to the fire accident score under any altitude similarity group.

7. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 5, characterized in that, The step of obtaining the weight of each data point in the scatter plot includes: For any data point in the scatter plot, the altitude similarity group corresponding to the data point is recorded as the target altitude similarity group. The density of each reference tunnel in the target altitude similarity group is obtained in the number axis. The average density of all reference tunnels in the target altitude similarity group is linearly normalized to obtain the weight of any data point in the scatter plot.

8. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 1, characterized in that, The process of obtaining the fire risk score of the current tunnel in the current time period based on the multi-dimensional risk assessment feature value corresponding to the target tunnel in the current time period, the importance of each dimension of the risk assessment feature value under the target tunnel, and the fire accident score of each reference tunnel includes: The multidimensional risk assessment feature value of the target tunnel in the current time period and the multidimensional risk assessment feature value of each reference tunnel are mapped to a multidimensional coordinate system, where each coordinate axis of the multidimensional coordinate system represents the risk assessment feature value of each dimension. In the process of clustering based on the distance between every two data points in the multidimensional coordinate system, the importance of each risk assessment feature value under the target tunnel is used as a weight to obtain the weighted distance between every two data points in the multidimensional coordinate system. Based on the weighted distance between every two data points in the multidimensional coordinate system, the data points in the multidimensional coordinate system are clustered to obtain the target cluster to which the target tunnel belongs. Based on the weighted distance between each reference tunnel in the target cluster and the corresponding data point of the target tunnel, and the fire accident score of all reference tunnels in the target cluster, the fire risk score of the target tunnel in the current time period is obtained.

9. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 8, characterized in that, The step of obtaining the fire risk score of the target tunnel in the current time period based on the weighted distance between the corresponding data points of each reference tunnel in the target cluster and the target tunnel, and the fire accident scores of all reference tunnels in the target cluster, includes: The reciprocal of the weighted distance between each reference tunnel in the target cluster and the corresponding data point of the target tunnel is linearly normalized to obtain the reference degree of each reference tunnel in the target cluster to the target tunnel. Using the degree of reference of each reference tunnel in the target cluster to the target tunnel as a weight, the weighted average of the fire accident scores of all reference tunnels in the target cluster is calculated as the fire risk score of the target tunnel in the current time period.

10. The intelligent assessment method for fire risk in high-altitude tunnels according to claim 2, characterized in that, The multidimensional risk assessment feature values ​​also include wind speed feature values. Therefore, obtaining the wind speed feature value corresponding to any reference tunnel includes: Obtain the wind speed monitoring data sequence at each monitoring point of any reference tunnel, calculate the average value of all data in all monitoring data sequences corresponding to the wind speed, and linearly normalize the average value to obtain the wind speed characteristic value of any reference tunnel.

Images