Method for monitoring a confined space work environment

Abstract

The application discloses a limited space operation environment monitoring method, and belongs to the technical field of gas monitoring. The steps include: time series alignment of gas concentration data of preset monitoring points in a limited space to obtain a plurality of time series concentration sequences; calculation of concentration change rates of the plurality of time series concentration sequences in a current time window, and spatial interpolation of the concentration change rates to obtain a gas concentration distribution gradient map; identification of an abnormal diffusion area with a gas concentration gradient exceeding a preset gradient threshold in the limited space, and determination of a diffusion direction vector of the abnormal diffusion area; generation of a targeted risk avoidance instruction for the limited space according to the diffusion direction vector and a relative positional relationship between the preset monitoring points and a current position of an operator in the limited space; reception of feedback information after execution of the instruction, and packaging of the feedback information into an operation environment monitoring report. The application realizes accurate identification of abnormal diffusion and dynamic risk avoidance through time series alignment, spatial interpolation and gradient analysis of gas concentration at a plurality of points.

Description

Technical Field

[0001] This invention belongs to the field of gas monitoring technology, and particularly relates to a method for monitoring confined space working environments. Background Technology

[0002] Confined space work refers to work activities conducted in enclosed or partially enclosed spaces with restricted access, poor natural ventilation, and a high risk of toxic, harmful, flammable, or explosive gas accumulation. Typical scenarios include underground utility tunnels, sewage wells, storage tanks, reaction vessels, cable trenches, and tunnels. Due to poor ventilation and limited gas diffusion in these spaces, oxygen-deficient environments can easily form, or toxic, harmful, flammable, and explosive gases such as hydrogen sulfide, carbon monoxide, and methane can accumulate, posing a fatal threat to workers. Therefore, strict environmental monitoring, especially real-time monitoring of gas concentrations, must be conducted before and during confined space work to ensure a safe working environment and prevent accidents such as poisoning, asphyxiation, and explosions.

[0003] However, existing monitoring technologies mostly rely on single-point data collection at fixed locations, failing to achieve temporal synchronization and spatial correlation of data from multiple points. This results in monitoring results that are difficult to reflect the true gas distribution status of the overall environment, leading to significant lag and bias in environmental risk identification. Furthermore, existing technologies only perform simple numerical comparisons of gas concentration changes, without analyzing the rate of concentration change and spatial diffusion patterns. This makes it impossible to accurately locate abnormal diffusion areas and trends, resulting in risk warnings lacking specificity and timeliness. In addition, risk control measures are mostly based on fixed threshold triggers, failing to dynamically match abnormal diffusion directions with the real-time location of workers. The generated avoidance instructions are highly generalized but poorly adaptable, and lack closed-loop feedback and effect evaluation mechanisms after instruction execution. Overall, the efficiency of operational environment monitoring and safety assurance capabilities are insufficient to meet actual operational needs. Summary of the Invention

[0004] Based on the above problems, the purpose of this invention is to provide a method for monitoring confined space working environments, which achieves accurate identification of abnormal diffusion and dynamic risk avoidance through multi-point gas concentration time-series alignment, spatial interpolation and gradient analysis.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The method for monitoring the working environment in confined spaces includes the following steps: Step 1: Perform time-series alignment on the gas concentration data of preset monitoring points within the confined space to obtain a multi-channel time-series concentration sequence for the confined space; Step 2: Calculate the concentration change rate of the multi-channel time series within the current time window, and perform spatial interpolation on the concentration change rate according to the spatial distribution of the preset monitoring points to obtain a gas concentration distribution gradient map in a limited space. Step 3: Based on the gas concentration distribution gradient map, identify the abnormal diffusion region in the limited space where the gas concentration gradient exceeds the preset gradient threshold, and determine the diffusion direction vector of the abnormal diffusion region. Step 4: Based on the diffusion direction vector and the relative orientation between the preset monitoring points and the current position of the workers in the confined space, generate targeted risk avoidance instructions for the confined space. Step 5: After executing the targeted risk avoidance instructions, receive the feedback information after the instructions are executed, and encapsulate the feedback information in a structured manner to obtain a monitoring report of the working environment in the confined space.

[0006] Preferably, in step 1, the process of obtaining the multi-channel time-series concentration sequence in a finite space is as follows: The gas concentration data at preset monitoring points within a confined space is obtained, and outliers are removed from the gas concentration data to obtain the effective concentration data for the confined space. The effective concentration data is timestamped according to a unified time base to obtain the standard concentration data of the preset monitoring points in a limited space at the same sampling time. The standard concentration data are sorted according to the spatial distribution order of the preset monitoring points to form a concentration distribution vector in a limited space; Arrange the concentration distribution vectors in chronological order to obtain a multi-channel time-series concentration sequence in a finite space.

[0007] Preferably, in step 2, the process of calculating the concentration change rate of the multi-time series concentration sequence within the current time window is as follows: From multiple time-series concentration sequences, the instantaneous concentration sampling values ​​of preset monitoring points within the current time window are extracted, and the instantaneous concentration sampling values ​​are decomposed into time series to extract the trend concentration component of the preset monitoring points within the current time window. With fluctuation concentration component ; The trend concentration component and the fluctuation concentration component are respectively differentiated within the current time window, and the first derivative of the trend concentration component with respect to time is taken as the rate of change of the trend. The first derivative of the fluctuation concentration component with respect to time is taken as the fluctuation rate of change. The formula for calculating the rate of change of the trend is as follows: ; In the formula, This represents the trend concentration component at time t, where t represents time. By traversing the connection paths between monitoring points and adjacent monitoring points in the preset monitoring point spatial topology diagram, the diffusion attenuation factor of the preset monitoring point relative to the risk source in the limited space is obtained. Based on the trend change rate, fluctuation change rate, and diffusion decay factor, the concentration change rate of multiple time series concentration sequences within the current time window is calculated.

[0008] Preferably, the formula for calculating the concentration change rate is: ; In the formula, This represents the rate of concentration change at the i-th monitoring point within the current time window. This represents the preset trend change rate weighting coefficient. This represents the preset weighting coefficient for the rate of change of volatility, where e represents the natural constant. This represents the diffusion attenuation factor of the i-th monitoring point relative to the risk source within a finite space. This represents the preset diffusion attenuation reference factor. This represents the preset diffusion response sensitivity coefficient.

[0009] Preferably, in step 2, the process of obtaining the gas concentration distribution gradient map of the finite space is as follows: Based on the spatial distribution of preset monitoring points, a spatial coordinate mapping grid is constructed in a limited space, and the concentration change rate of the preset monitoring points is anchored to the grid nodes in the spatial coordinate mapping grid to obtain the node rate value of the grid nodes. Based on the spatial topological adjacency relationship between adjacent anchored grid nodes in the spatial coordinate mapping grid, the rate value of the blank grid nodes without anchored rate values ​​is inherited through the connected path to obtain the derived rate value of the blank grid nodes. According to the grid spatial arrangement order, the gradient direction solution is performed on the nodal rate values ​​of the grid nodes in the spatial coordinate mapping grid to obtain the transverse gradient field and the longitudinal gradient field of the spatial coordinate mapping grid. By fusing the transverse and longitudinal gradient fields into vector fields, a gradient vector distribution map of a finite space is obtained. The gradient vector distribution map is then visualized and rendered to obtain a gas concentration distribution gradient map of a finite space.

[0010] Preferably, in step 3, abnormal diffusion regions where the gas concentration gradient in a confined space exceeds a preset gradient threshold are identified, and the diffusion direction vector of the abnormal diffusion regions is determined. The process is as follows: Based on the gas concentration distribution gradient map, the gradient amplitude is traversed for spatial points in a limited space, and spatial points whose gradient amplitude exceeds the preset gradient threshold are marked as abnormal points. Spatial connectivity clustering is performed on the abnormal locations to obtain the abnormal diffusion region in a limited space; Curvature analysis is performed on the spatial boundary contour of the anomalous diffusion region to locate the diffusion front boundary and diffusion wake boundary of the anomalous diffusion region. By vector synthesis of the gradient vectors at the boundary points on the diffusion front boundary, the main vector of the diffusion direction in the anomalous diffusion region is obtained; Spatial orientation calibration of the main diffusion direction vector yields the diffusion direction vector of the abnormal diffusion region.

[0011] Preferably, the process of locating the diffusion front boundary and diffusion wake boundary of the abnormal diffusion region is as follows: Discretize the spatial boundary contour of the abnormal diffusion region to decompose the spatial boundary contour into a continuous sequence of boundary micro-segments, and assign a corresponding set of boundary points to the sequence of boundary micro-segments. The curvature of the boundary point set in the spatial coordinate system is determined by the orientation offset. Boundary micro-segments with curvature reaching a preset curvature threshold are marked as high curvature micro-segments, and boundary micro-segments with curvature below the preset curvature threshold are marked as low curvature micro-segments. By performing connected clustering on continuously distributed and spatially adjacent high-curvature micro-segments and low-curvature micro-segments, the high-curvature boundary segments and low-curvature boundary segments of the anomalous diffusion region are obtained. Based on the overall diffusion trend of the abnormal diffusion region, the high curvature boundary segment located in the diffusion forward direction is positioned as the diffusion front boundary, and the low curvature boundary segment located in the diffusion source direction is positioned as the diffusion wake boundary.

[0012] Preferably, in step 4, the process of generating targeted risk avoidance instructions for a limited space is as follows: Obtain the spatial coordinates of the current position of the worker within a confined space, and compare the spatial coordinates with the spatial distribution of the preset monitoring points to obtain the relative positional relationship between the current position of the worker and the preset monitoring points; The diffusion direction vector is coupled with the relative orientation relationship to determine whether the diffusion direction vector points to the current position of the worker, thus obtaining the risk orientation judgment result in a limited space. Based on the risk orientation determination results, the risk sources in the confined space are classified into risk levels, and the risk sources whose risk levels reach the preset warning threshold are labeled to obtain a list of risk sources to be avoided in the confined space. Based on the list of risk sources to be avoided and the diffusion direction vector, a risk avoidance movement direction indication corresponding to the current position of the operator is generated. The risk avoidance movement direction indication and the list of risk sources to be avoided are then fused to obtain targeted risk avoidance instructions for a limited space.

[0013] Preferably, the process of obtaining a list of risk sources to be avoided in a limited space based on the risk orientation determination results is as follows: From the risk orientation determination results, the orientation association information of the risk source relative to the current position of the operator is extracted, and the orientation association information is matched and bound with the inherent attribute parameters of the risk source to obtain the risk situation association mapping of the risk source; Based on the risk situation correlation mapping, risk sources are classified according to the degree of risk urgency. Risk sources that point to the current location of the worker and whose pointing intensity reaches the preset intensity threshold are classified into the high-risk level, risk sources that point to the current location of the worker but whose pointing intensity does not reach the preset intensity threshold are classified into the medium-risk level, and risk sources that do not point to the current location of the worker are classified into the low-risk level, thus obtaining the risk level classification of the risk sources. Risk sources classified as high-risk and medium-risk are compared with preset warning trigger conditions to filter out risk sources that meet the warning trigger conditions, resulting in a set of risk sources to be marked in a limited space. Warning labels are attached to the risk sources in the set of risk sources to be marked, and the risk sources with attached warning labels are organized into items according to risk level to obtain a list of risk sources to be avoided in a limited space.

[0014] Preferably, in step 5, the process of obtaining the working environment monitoring report for the confined space is as follows: After the targeted risk avoidance command is executed, the gas concentration data of the preset monitoring points in the confined space is captured, and the captured gas concentration data is compared with the gas concentration data before the command is executed to obtain the concentration change status of the preset monitoring points after the command is executed. After the command is executed, the state of concentration change is checked for consistency in order to identify effective monitoring points whose concentration change trend matches the expected change direction of the targeted risk avoidance command, and the feedback concentration sequence of the effective monitoring points is extracted to obtain the effective feedback information after the command is executed. Based on effective feedback information, the recovery degree of the working environment situation in the confined space is assessed, the residual risk parameters of the abnormal diffusion area after the execution of the instructions are obtained, and the residual risk parameters are time-series aligned with the execution records of targeted risk avoidance instructions to obtain a set of elements for assessing the working environment situation in the confined space. The set of elements for assessing the operational environment, along with the spatial structure information of the confined space and the deployment information of the preset monitoring points, are arranged in an itemized manner according to a preset report template to obtain a monitoring report of the operational environment of the confined space.

[0015] The beneficial effects of this invention are: This invention generates a concentration distribution gradient map by performing time-series alignment, concentration change rate calculation, and spatial interpolation on gas concentration data from multiple points in a confined space. This allows for the accurate identification of abnormal diffusion areas and diffusion directions. Combined with the location of the workers, it generates appropriate risk avoidance instructions and completes closed-loop feedback and report encapsulation after instruction execution. This significantly improves the accuracy, real-time performance, and risk control efficiency of confined space work environment monitoring, achieving efficient control of the entire process from data monitoring and risk identification to instruction execution and effect evaluation, effectively ensuring work safety. Attached Figure Description

[0016] Figure 1 This is a schematic flowchart of the method of the present invention; Figure 2 This is a time-series variation curve of the multi-feature rate in Embodiment 1 of the present invention. Detailed Implementation

[0017] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0018] Example 1: As Figure 1 As shown, the method for monitoring the working environment in a confined space includes the following steps: Step 1: Perform time-series alignment on the gas concentration data of preset monitoring points within the confined space to obtain a multi-channel time-series concentration sequence for the confined space; Step 2: Calculate the concentration change rate of the multi-channel time series within the current time window, and perform spatial interpolation on the concentration change rate according to the spatial distribution of the preset monitoring points to obtain a gas concentration distribution gradient map in a limited space. Step 3: Based on the gas concentration distribution gradient map, identify the abnormal diffusion region in the limited space where the gas concentration gradient exceeds the preset gradient threshold, and determine the diffusion direction vector of the abnormal diffusion region. Step 4: Based on the diffusion direction vector and the relative orientation between the preset monitoring points and the current position of the workers in the confined space, generate targeted risk avoidance instructions for the confined space. Step 5: After executing the targeted risk avoidance instructions, receive the feedback information after the instructions are executed, and encapsulate the feedback information in a structured manner to obtain a monitoring report of the working environment in the confined space.

[0019] In step 1, a multi-channel time-series concentration sequence in a finite space is obtained, as follows: The gas concentration data at preset monitoring points within a confined space is obtained, and outliers are removed from the gas concentration data to obtain the effective concentration data for the confined space. The effective concentration data is timestamped according to a unified time base to obtain the standard concentration data of the preset monitoring points in a limited space at the same sampling time. The standard concentration data are sorted according to the spatial distribution order of the preset monitoring points to form a concentration distribution vector in a limited space; Arrange the concentration distribution vectors in chronological order to obtain a multi-channel time-series concentration sequence in a finite space.

[0020] Gas concentration data at preset monitoring points within a confined space is acquired, and outlier removal is performed to obtain the effective concentration data for the confined space. Concentration values ​​and corresponding timestamps are collected from each preset monitoring point. These data are processed using an outlier removal method based on median and absolute deviation. For each monitoring point's data sequence, the median of the sequence is calculated. Then, the median of the absolute value of the difference between each data point and the median is calculated as the absolute deviation. A threshold is set at the median plus or minus three times the absolute deviation. Data points outside the threshold range are identified as outliers and removed from the sequence. The remaining data points after outlier removal constitute the effective concentration data for that point. This process is repeated for all monitoring points to obtain the effective concentration data for the entire confined space.

[0021] Effective concentration data are timestamped according to a unified time base to obtain standard concentration data for preset monitoring points within a limited space at the same sampling time. The unified time base is defined as a sequence of identical sampling times divided at fixed time intervals from the earliest to the latest timestamp of all effective concentration data. For the effective concentration data of each monitoring point, a linear interpolation method is used to map the data to these identical sampling times. For each identical sampling time, the two closest effective data points before and after that time are found. Based on the time and concentration values ​​of these two data points, the concentration value at that time is calculated linearly. This interpolation operation is performed on all monitoring points so that each point has a corresponding concentration value at the same sampling time. These concentration values ​​are the standard concentration data.

[0022] The standard concentration data are sorted according to the spatial distribution order of the preset monitoring points to form a concentration distribution vector in a finite space. The preset monitoring points have a predefined spatial distribution order in the finite space, such as from the entrance to the exit or arranged according to grid coordinates. For each same sampling time, the standard concentration value of each point is extracted sequentially according to the spatial distribution order of the points, and these values ​​are arranged into an ordered list. This list is the concentration distribution vector at that sampling time. The order of the elements in the vector corresponds to the spatial order of the points, and the length of the vector is equal to the number of monitoring points.

[0023] Arranging the concentration distribution vectors in chronological order yields a multi-channel time-series concentration sequence in a finite space. The concentration distribution vectors at all sampling times are sorted chronologically from the earliest to the latest time, forming a sequence where each element is a concentration distribution vector corresponding to a sampling time. This sequence is the multi-channel time-series concentration sequence, containing complete information about the temporal and spatial variations of concentration within a finite space.

[0024] In step 2, the concentration change rate of the multi-series concentration series within the current time window is calculated, as follows: From multiple time-series concentration sequences, the instantaneous concentration sampling values ​​of preset monitoring points within the current time window are extracted, and the instantaneous concentration sampling values ​​are decomposed into time series to extract the trend concentration component of the preset monitoring points within the current time window. With fluctuation concentration component ; The trend concentration component and the fluctuation concentration component are respectively differentiated within the current time window, and the first derivative of the trend concentration component with respect to time is taken as the rate of change of the trend. The first derivative of the fluctuation concentration component with respect to time is taken as the fluctuation rate of change. The formula for calculating the rate of change of the trend is as follows: ; In the formula, This represents the trend concentration component at time t, where t represents time. By traversing the connection paths between monitoring points and adjacent monitoring points in the preset monitoring point spatial topology diagram, the diffusion attenuation factor of the preset monitoring point relative to the risk source in the limited space is obtained. Based on the trend change rate, fluctuation change rate, and diffusion decay factor, the concentration change rate of multiple time series concentration sequences within the current time window is calculated.

[0025] The formula for calculating the rate of concentration change is: ; In the formula, This represents the rate of concentration change at the i-th monitoring point within the current time window. This represents the rate of change of the trend of the i-th monitoring point within the current time window. This represents the rate of change of the i-th monitoring point within the current time window. This represents the preset trend change rate weighting coefficient. This represents the preset weighting coefficient for the rate of change of volatility, where e represents the natural constant. This represents the diffusion attenuation factor of the i-th monitoring point relative to the risk source within a finite space. This represents the preset diffusion attenuation reference factor. This represents the preset diffusion response sensitivity coefficient.

[0026] Figure 2 This is an exemplary multi-feature rate time-series variation curve, where the horizontal axis represents the monitoring time t, and the vertical axis represents the rate. The blue solid line, red dashed line, and green dotted line correspond to... , , According to the method in this embodiment, the original time-series concentration sequence is first decomposed to obtain trend concentration components and fluctuation concentration components, and the derivatives of each are then calculated to obtain the same dimension. , Then, by using the above formula, we can obtain... Sigmoid Modifier The value range is always (0, 1), which has an attenuation suppression effect. Therefore, after correction by the diffusion attenuation factor... The overall magnitude is lower than , As can be seen from the time-series variation curve, The overall benchmark level is higher, which characterizes the fundamental trend of long-term concentration accumulation within a limited space; The large fluctuations reflect short-term concentration disturbances caused by on-site airflow disturbances and ventilation fluctuations. By integrating trend and fluctuation characteristics and superimposing diffusion attenuation constraints, the curve changes smoothly, which can objectively characterize the speed of real concentration change at monitoring points after being affected by the diffusion of risk sources.

[0027] In step 2, the gas concentration distribution gradient map of the finite space is obtained, and the process is as follows: Based on the spatial distribution of preset monitoring points, a spatial coordinate mapping grid is constructed in a limited space, and the concentration change rate of the preset monitoring points is anchored to the grid nodes in the spatial coordinate mapping grid to obtain the node rate value of the grid nodes. Based on the spatial topological adjacency relationship between adjacent anchored grid nodes in the spatial coordinate mapping grid, the rate value of the blank grid nodes without anchored rate values ​​is inherited through the connected path to obtain the derived rate value of the blank grid nodes. According to the grid spatial arrangement order, the gradient direction solution is performed on the nodal rate values ​​of the grid nodes in the spatial coordinate mapping grid to obtain the transverse gradient field and the longitudinal gradient field of the spatial coordinate mapping grid. By fusing the transverse and longitudinal gradient fields into vector fields, a gradient vector distribution map of a finite space is obtained. The gradient vector distribution map is then visualized and rendered to obtain a gas concentration distribution gradient map of a finite space.

[0028] From multiple time-series concentration sequences, the instantaneous concentration sampling values ​​of preset monitoring points within the current time window are extracted. These sampling values ​​are the raw concentration data obtained from the monitoring point at each sampling time within the time window. The instantaneous concentration sampling values ​​of the monitoring point are decomposed into a time series using the moving average method. Specifically, each data point in the sequence, along with its two preceding and following data points (a total of five data points), is taken, and the arithmetic mean of these five data points is calculated. This average is used as the trend concentration component value at that data point. This process is repeated for each data point from the beginning of the sequence until the end, yielding the trend concentration component sequence of the monitoring point within the entire time window.

[0029] Subtracting the trend concentration component value at the corresponding position from each original data point in the instantaneous concentration sampling value sequence yields the difference sequence, which is the fluctuating concentration component sequence of that monitoring point within the current time window.

[0030] In practice, since concentration data is collected at discrete sampling times, it is impossible to directly differentiate a continuous function. Therefore, a numerical differentiation method is used for approximate calculation. The specific operation of this numerical differentiation method is as follows: Take the trend concentration component value corresponding to the start and end sampling times of the current time window, calculate the difference between them, and then divide it by the time length between the start and end sampling times. The quotient obtained is the trend change rate, which is numerically equal to the average rate of change of the trend concentration component within that time window, and is an approximate estimate of the derivative. Similarly, take the fluctuation concentration component value corresponding to the start and end sampling times of the current time window, calculate the difference between them, and then divide it by the same time length. The quotient obtained is the fluctuation change rate, which is an approximate estimate of the derivative of the fluctuation concentration component.

[0031] Traverse the preset monitoring point spatial topology diagram, which is a network consisting of monitoring points as nodes and direct connections between adjacent monitoring points as edges. Starting from the node corresponding to the location of the risk source within the specified limited space in the diagram, a breadth-first search method is used. That is, starting from the risk source node, expand outward layer by layer, first visiting all adjacent nodes directly connected to the risk source node, then visiting the adjacent nodes of these nodes, until the node where the preset monitoring point is located is found. Record the number of edges traversed from the risk source node to the monitoring point node, and then divide 1 by this number to obtain the diffusion attenuation factor of the monitoring point relative to the risk source.

[0032] Based on the trend change rate and fluctuation change rate of the monitoring point, the trend change rate is first multiplied by a preset trend change rate weighting coefficient to obtain a weighted trend value. Then, the fluctuation change rate is multiplied by a preset fluctuation change rate weighting coefficient to obtain a weighted fluctuation value. Finally, the weighted trend value and the weighted fluctuation value are added together to obtain a comprehensive rate value. Next, based on the diffusion attenuation factor of the monitoring point, the diffusion attenuation factor is compared with a preset diffusion attenuation benchmark factor. If the diffusion attenuation factor is greater than the benchmark factor, the influence of the diffusion attenuation factor amplifies the comprehensive rate value according to a preset diffusion response sensitivity. If the diffusion attenuation factor is less than the benchmark factor, the influence of the diffusion attenuation factor reduces the comprehensive rate value according to a preset diffusion response sensitivity. If the two are equal, the comprehensive rate value remains unchanged. The final value obtained after this adjustment is the concentration change rate of the monitoring point within the current time window. The above complete process from extracting instantaneous concentration sampling values ​​to calculating the concentration change rate is performed for each preset monitoring point within a limited space, thereby obtaining the concentration change rate of all monitoring points within the current time window.

[0033] The trend change rate weighting coefficient is a preset fixed value used to adjust the proportion of the trend change rate in the concentration change rate. This value is preset and stored before implementation based on the ventilation characteristics of the limited space and the statistical results of historical data.

[0034] The fluctuation rate weighting coefficient is a preset fixed value used to adjust the proportion of the fluctuation rate in the concentration change rate. This value is preset and stored before implementation based on the typical range of gas fluctuation amplitude and sensor noise level in a confined space.

[0035] The diffusion attenuation factor is derived from the value obtained after traversing the spatial topology of the monitoring point. Specifically, it is the diffusion attenuation factor of the monitoring point relative to the risk source in the limited space. Its value is equal to the number 1 divided by the number of edges traversed from the risk source node to the monitoring point node. The fewer the number of edges, the closer the value of the factor is to 1, and the more the number of edges, the closer the factor is to 0.

[0036] The diffusion attenuation benchmark factor is a preset fixed threshold used to determine whether the diffusion attenuation factor of the current monitoring point is too large or too small relative to the benchmark. This threshold is preset and stored before implementation based on the spatial size of the limited space and the deployment density of the monitoring points.

[0037] The diffusion response sensitivity coefficient is a preset fixed value used to control the severity of the influence of the deviation between the diffusion attenuation factor and the reference factor on the concentration change rate. This coefficient is preset and stored based on the measured data of the gas diffusion response rate in a confined space before implementation.

[0038] The significance of the concentration change rate calculation formula lies in integrating the three core factors affecting concentration change into a comprehensive index. The trend change rate reflects the rate of increase or decrease of gas concentration in the main direction within the time window, the fluctuation change rate reflects the speed at which the concentration fluctuates around the trend line, and the diffusion attenuation factor reflects the weakening effect of the distance between the monitoring point and the risk source on the efficiency of concentration change transmission. The three factors jointly determine the concentration change rate at the current moment through weight allocation and sensitivity adjustment, thus providing a numerical basis for subsequent spatial interpolation and gradient map generation.

[0039] The concentration change rate calculation formula shows that the concentration change rate increases with the increase of the trend change rate and the fluctuation change rate, and decreases with the decrease of these two change rates. At the same time, the diffusion attenuation factor plays a corrective role in the rate. When the diffusion attenuation factor of the monitoring point is greater than the preset diffusion attenuation benchmark factor, it means that the point is close to the risk source and the diffusion path is short. At this time, the concentration change rate will be reduced according to the intensity determined by the diffusion response sensitivity coefficient. When the diffusion attenuation factor is less than the benchmark factor, it means that the point is far from the risk source and the diffusion path is long. At this time, the concentration change rate will be amplified. When the two are equal, the rate remains unchanged. This correction mechanism makes the points closer to the risk source more sensitive to concentration changes and the values ​​more stable. Points far from the risk source compensate for the signal attenuation caused by long-distance transmission through the amplification effect, thereby ensuring that the concentration change rate of each point in the entire limited space is comparable and continuous in spatial distribution.

[0040] The formula for calculating the rate of concentration change generally uses the trend rate of change and the fluctuation rate of change to form the basic value of the rate. The deviation between the diffusion attenuation factor and the benchmark factor is dynamically scaled by the diffusion response sensitivity coefficient. The scaling intensity is controlled by the sensitivity coefficient. The larger the sensitivity coefficient, the more drastic the scaling effect, and the smaller the sensitivity coefficient, the smoother the scaling effect. The final output rate of concentration change reflects both the temporal variation characteristics of the concentration at the point and the location characteristics of the point relative to the risk source in the spatial topology, providing accurate numerical basis for subsequent identification of abnormal diffusion areas.

[0041] Based on the spatial distribution of preset monitoring points, a spatial coordinate mapping grid is constructed in a limited space. The concentration change rate of the preset monitoring points is anchored to the grid nodes in the spatial coordinate mapping grid, obtaining the node rate value of the grid node. According to the spatial coordinates of all preset monitoring points, a rectangular area covering all points is determined. This area is divided into uniform squares at fixed intervals in the horizontal and vertical directions, forming the spatial coordinate mapping grid. For each preset monitoring point, the nearest grid node is found, and the concentration change rate value of that point is directly assigned to this grid node; the assigned value is called the node rate value of that grid node. A grid node may carry the rate values ​​of multiple nearby monitoring points; in this case, the average of these rate values ​​is taken as the final node rate value of the node.

[0042] Based on the spatial topological adjacency relationships between adjacent anchored grid nodes in the spatial coordinate mapping grid, rate values ​​are inherited for blank grid nodes without anchored rate values ​​in the spatial coordinate mapping grid through connected paths, resulting in derived rate values ​​for the blank grid nodes. For any blank grid node, all its directly adjacent grid nodes in the spatial coordinate mapping grid that already possess node rate values ​​are identified. The average of the node rate values ​​of these adjacent nodes is calculated, and this average is then directly assigned to the blank grid node; this assigned value is called the derived rate value of the blank grid node. This operation is repeated for all blank nodes in the grid until all grid nodes possess rate values.

[0043] Following the grid spatial arrangement order, gradient direction calculations are performed on the nodal velocity values ​​of the grid nodes in the spatial coordinate mapping grid to obtain the transverse and longitudinal gradient fields of the spatial coordinate mapping grid. Starting from the second column of the first row of the grid, the difference in nodal velocity value between each grid node and its left-side neighbor is calculated sequentially. Subtracting the value of the left-side neighbor from the current node's value constitutes the transverse gradient field. Starting from the first column of the second row of the grid, the difference in nodal velocity value between each grid node and its upper-side neighbor is calculated sequentially. Subtracting the value of the upper-side neighbor from the current node's value constitutes the longitudinal gradient field. Each value in the gradient field represents the change in velocity along that direction.

[0044] The transverse and longitudinal gradient fields are fused using vector field fusion to obtain a gradient vector distribution map in a finite space. This gradient vector distribution map is then visualized and rendered to obtain a gas concentration distribution gradient map in the finite space. For each internal node in the grid, its corresponding transverse gradient value is taken as the horizontal component of the vector, and its corresponding longitudinal gradient value is taken as the vertical component of the vector. These two components together constitute a two-dimensional vector. At each node position in the grid, an arrow represents this vector, pointing in the direction of the fastest increase in the rate of concentration change. The length of the arrow is proportional to the magnitude of the vector. Finally, this grid map with vectors is plotted, and the color intensity of the arrows can also be used to represent the magnitude of the vector. The resulting visualization is the gas concentration distribution gradient map in the finite space.

[0045] In step 3, abnormal diffusion regions where the gas concentration gradient exceeds a preset gradient threshold within a confined space are identified, and the diffusion direction vector of the abnormal diffusion region is determined. The process is as follows: Based on the gas concentration distribution gradient map, the gradient amplitude is traversed for spatial points in a limited space, and spatial points whose gradient amplitude exceeds the preset gradient threshold are marked as abnormal points. Spatial connectivity clustering is performed on the abnormal locations to obtain the abnormal diffusion region in a limited space; Curvature analysis is performed on the spatial boundary contour of the anomalous diffusion region to locate the diffusion front boundary and diffusion wake boundary of the anomalous diffusion region: Discretize the spatial boundary contour of the abnormal diffusion region to decompose the spatial boundary contour into a continuous sequence of boundary micro-segments, and assign a corresponding set of boundary points to the sequence of boundary micro-segments. The curvature of the boundary point set in the spatial coordinate system is determined by the orientation offset. Boundary micro-segments with curvature reaching a preset curvature threshold are marked as high curvature micro-segments, and boundary micro-segments with curvature below the preset curvature threshold are marked as low curvature micro-segments. By performing connected clustering on continuously distributed and spatially adjacent high-curvature micro-segments and low-curvature micro-segments, the high-curvature boundary segments and low-curvature boundary segments of the anomalous diffusion region are obtained. Based on the overall diffusion trend of the abnormal diffusion area, the high curvature boundary section located in the diffusion forward direction is positioned as the diffusion front boundary, and the low curvature boundary section located in the diffusion source direction is positioned as the diffusion wake boundary. By vector synthesis of the gradient vectors at the boundary points on the diffusion front boundary, the main vector of the diffusion direction in the anomalous diffusion region is obtained; Spatial orientation calibration of the main diffusion direction vector yields the diffusion direction vector of the abnormal diffusion region.

[0046] Based on the gas concentration distribution gradient map, the gradient magnitude of spatial points within a finite space is traversed, and spatial points with gradient magnitudes exceeding a preset gradient threshold are marked as outliers. On the gas concentration distribution gradient map, for each node constituting the grid, the length of the gradient vector at that node is calculated as the gradient magnitude. This is done by summing the squares of the horizontal and vertical components of the vector and then taking the square root of this sum. The calculated gradient magnitude of each node is compared with a preset gradient threshold value. If the gradient magnitude of a node is greater than this preset threshold, the spatial point represented by that node is marked as an outlier. This comparison and marking operation is performed on all nodes in the grid.

[0047] Spatial connectivity clustering is performed on outliers to obtain anomaly diffusion regions in a finite space. In the spatial coordinate mapping grid, all nodes marked as outliers are examined. A region growing method is used for clustering. Specifically, an outlier not yet classified into any region is selected as a seed point and placed into a new region. Then, the four directly adjacent nodes (up, down, left, and right) of this seed point in the grid are checked. If these adjacent nodes are also outliers and not yet classified into any region, they are also added to the current region. Next, using these newly added points as new seed points, the process of checking their adjacent nodes is repeated until no new adjacent outliers can be added. At this point, all the outliers connected by adjacency constitute an anomaly diffusion region. The above process is repeated for the remaining unclassified outliers in the grid until all outliers are classified, ultimately resulting in several anomaly diffusion regions.

[0048] The spatial boundary contour of the anomalous diffusion region is discretized to decompose it into a continuous sequence of boundary micro-segments, and a corresponding set of boundary points is assigned to each micro-segment. The boundary of an anomalous diffusion region is determined by connecting all anomalous points within the region that are adjacent to non-anomaly points outside the region. Boundary points are selected sequentially along this closed boundary contour at fixed arc length intervals, where the arc length interval is a pre-defined fixed distance value. Adjacent boundary points are connected by straight lines; this line segment is a boundary micro-segment. Starting from the initial point, all sequentially selected boundary points are connected to form a series of consecutive straight line segments, which constitute the boundary micro-segment sequence. Each boundary micro-segment is defined by its two endpoints, which are the set of boundary points assigned to that micro-segment. The endpoint sets of all boundary micro-segments in the sequence together constitute the discretized boundary point set of the entire contour.

[0049] The curvature of the boundary point set in the spatial coordinate system is determined by assessing its orientation offset. Boundary micro-segments with curvature reaching a preset curvature threshold are marked as high-curvature micro-segments, while those with curvature below the threshold are marked as low-curvature micro-segments. For each boundary micro-segment in the sequence, its two adjacent micro-segments are obtained. The orientation angle of these three consecutive micro-segments is calculated, i.e., the angle between each micro-segment and the positive direction of the horizontal axis of the spatial coordinate system. Then, the absolute value of the difference between the orientation angle of the middle micro-segment and the two adjacent micro-segments is calculated, and the sum of these two absolute values ​​is used as the quantified curvature value of the middle boundary micro-segment. This quantified curvature value is compared with a preset curvature threshold. If the value is greater than or equal to the preset curvature threshold, the middle boundary micro-segment is marked as a high-curvature micro-segment. If the value is less than the threshold, it is marked as a low-curvature micro-segment. The same calculation and judgment are performed on all boundary micro-segments in the sequence.

[0050] Connectivity clustering is performed on continuously distributed and spatially adjacent high-curvature and low-curvature micro-segments to obtain high-curvature and low-curvature boundary segments of the anomalous diffusion region. In the boundary segment sequence, all labeled micro-segments are examined. Starting from any high-curvature micro-segment, a search is performed along the boundary contour in both forward and backward directions, merging all directly adjacent micro-segments that are also labeled as high-curvature micro-segments together. This search and merging process continues until a non-high-curvature micro-segment is encountered. The resulting set of continuous high-curvature micro-segments constitutes a high-curvature boundary segment. This process is repeated for all high-curvature micro-segments, aggregating the originally scattered high-curvature micro-segments into several continuous high-curvature boundary segments. Low-curvature micro-segments are processed in the same way: starting from any low-curvature micro-segment, a search is performed along the contour to merge all continuously adjacent low-curvature micro-segments, forming several low-curvature boundary segments.

[0051] Based on the overall diffusion trend of the anomalous diffusion region, the high-curvature boundary segment located in the diffusion forward direction is positioned as the diffusion front boundary, and the low-curvature boundary segment located in the diffusion source direction is positioned as the diffusion wake boundary. First, the overall diffusion trend of the anomalous diffusion region is determined, and the average direction of the gradient vectors of all anomalous points within the region is calculated. This average direction indicates the overall diffusion forward direction. Then, for each high-curvature boundary segment, the average position coordinates of all boundary points on that segment are calculated. From a known, presumed, or highest-concentration suspected diffusion source location within the region, a line is drawn to the average position point of this high-curvature boundary segment. The direction of this line is calculated and compared with the overall diffusion forward direction. If the angle between these two directions is less than 90 degrees, the high-curvature boundary segment is considered to be located in the diffusion forward direction and is positioned as the diffusion front boundary of the anomalous diffusion region. For each low-curvature boundary segment, the direction of the line connecting its average position point and the suspected diffusion source point is also calculated. If the angle between the direction of this line and the overall diffusion direction is greater than 90 degrees, then the low curvature boundary section is considered to be located in the direction of the diffusion source or the opposite direction, and is positioned as the diffusion wake boundary of the abnormal diffusion region.

[0052] The gradient vectors of the boundary points on the diffusion front boundary are vector synthesized to obtain the principal vector of the diffusion direction in the anomalous diffusion region. The coordinates of all boundary points on the high-curvature boundary segment located as the diffusion front boundary are obtained. For each such boundary point, the gradient vector corresponding to its grid node position is found in the previously generated gas concentration distribution gradient map. All these gradient vectors are synthesized by adding the horizontal components of all vectors to obtain a total horizontal component sum, and simultaneously adding the vertical components of all vectors to obtain a total vertical component sum. This total horizontal component sum and the total vertical component sum are used to construct a new two-dimensional vector, which is the principal vector of the diffusion direction in the anomalous diffusion region.

[0053] The diffusion direction vector of the anomalous diffusion region is obtained by spatially calibrating the principal vector of diffusion direction. The average position of all boundary points on the diffusion front boundary is used as the reference origin for spatial calibration. Starting from this reference origin, the due east direction or the positive X-axis of the spatial coordinate system is used as the reference azimuth. The angle between the principal vector of diffusion direction and this reference azimuth is calculated; this angle defines the specific direction of the principal vector. The diffusion direction vector is characterized by the coordinates of this reference origin combined with the calculated pointing angle, ultimately expressed as a directed line segment originating from the reference origin and pointing in the direction determined by this angle. This directed line segment is the diffusion direction vector of the anomalous diffusion region.

[0054] In step 4, targeted risk avoidance instructions for a limited space are generated, as follows: Obtain the spatial coordinates of the current position of the worker within a confined space, and compare the spatial coordinates with the spatial distribution of the preset monitoring points to obtain the relative positional relationship between the current position of the worker and the preset monitoring points; The diffusion direction vector is coupled with the relative orientation relationship to determine whether the diffusion direction vector points to the current position of the worker, thus obtaining the risk orientation judgment result in a limited space. Based on the risk orientation assessment results, risk sources within the confined space are classified into risk levels, and risk sources whose risk levels reach a preset warning threshold are labeled, resulting in a list of risk sources to be avoided in the confined space: From the risk orientation determination results, the orientation association information of the risk source relative to the current position of the operator is extracted, and the orientation association information is matched and bound with the inherent attribute parameters of the risk source to obtain the risk situation association mapping of the risk source; Based on the risk situation correlation mapping, risk sources are classified according to the degree of risk urgency. Risk sources that point to the current location of the worker and whose pointing intensity reaches the preset intensity threshold are classified into the high-risk level, risk sources that point to the current location of the worker but whose pointing intensity does not reach the preset intensity threshold are classified into the medium-risk level, and risk sources that do not point to the current location of the worker are classified into the low-risk level, thus obtaining the risk level classification of the risk sources. Risk sources classified as high-risk and medium-risk are compared with preset warning trigger conditions to filter out risk sources that meet the warning trigger conditions, resulting in a set of risk sources to be marked in a limited space. Warning labels are attached to the risk sources in the set of risk sources to be marked, and the risk sources with attached warning labels are organized into items according to risk level to obtain a list of risk sources to be avoided in a limited space; Based on the list of risk sources to be avoided and the diffusion direction vector, a risk avoidance movement direction indication corresponding to the current position of the operator is generated. The risk avoidance movement direction indication and the list of risk sources to be avoided are then fused to obtain targeted risk avoidance instructions for a limited space.

[0055] The spatial coordinates of the worker's current position within a confined space are obtained, and these coordinates are compared with the spatial distribution of preset monitoring points to determine the relative positional relationship between the worker's current position and the preset monitoring points. The worker's current planar coordinates are acquired in real-time using a positioning device. The known spatial coordinates of the preset monitoring points and the worker's coordinates are placed in the same coordinate system. The direction of the line connecting the worker's coordinates to each monitoring point is calculated sequentially. This direction is obtained by calculating the angle between the line connecting the two points and the positive X-axis of the coordinate system. Simultaneously, the straight-line distance between the two points is calculated. For each monitoring point, its relative positional relationship with the worker is defined by this directional angle and straight-line distance.

[0056] The diffusion direction vector is coupled directionally with the relative orientation to determine whether the diffusion direction vector points to the worker's current position, thus obtaining a risk orientation determination result for a limited space. The pointing angle of the diffusion direction vector is taken and compared with the angle of the line connecting the monitoring point corresponding to the risk source to the worker's coordinates, obtained in the previous step. The absolute value of the difference between these two angles is calculated. If the absolute value of this difference is less than 45 degrees, the diffusion direction vector is determined to point to the worker's current position, and the determination result is marked as pointing. If the absolute value of this difference is greater than or equal to 45 degrees, it is determined as not pointing. This pointing or not pointing result is the risk orientation determination result.

[0057] From the risk orientation determination result, the orientation association information of the risk source relative to the current position of the operator is extracted. This orientation association information is then matched and bound with the inherent attribute parameters of the risk source to obtain the risk situation association mapping of the risk source. The orientation association information includes whether the risk orientation determination result is pointing or not pointing; if pointing, it also includes the absolute value of the angle difference calculated in the previous step. The inherent attribute parameters of the risk source include the risk source type code, the current average concentration value, and its location coordinates. The angle difference or orientation marker in the orientation association information is bound one-to-one with the risk source type code, concentration value, and location coordinates to form an association record containing all information. This record is the risk situation association mapping of the risk source.

[0058] Based on the risk situation correlation mapping, risk sources are stratified according to their risk urgency. Risk sources that point to the current location of the worker and whose pointing intensity reaches a preset intensity threshold are classified into the high-risk level; those that point to the current location but whose pointing intensity does not reach the preset intensity threshold are classified into the medium-risk level; and those that do not point to the current location of the worker are classified into the low-risk level. This results in the risk level classification of each risk source. The pointing intensity is quantified by the current average concentration value of the risk source recorded in the risk situation correlation mapping. The preset intensity threshold is a specific concentration value. For risk sources whose pointing determination result is pointing, their current average concentration value is compared with the preset intensity threshold. If the concentration value is greater than or equal to the preset intensity threshold, the risk source is classified into the high-risk level; if the concentration value is less than the preset intensity threshold, it is classified into the medium-risk level; and for risk sources whose pointing determination result is not pointing, regardless of their concentration value, they are uniformly classified into the low-risk level. Each risk source is ultimately assigned a label of high-risk, medium-risk, or low-risk level, which is its risk level classification.

[0059] Risk sources classified as high-risk or medium-risk are compared with preset warning trigger conditions to filter out those that meet the conditions, resulting in a set of risk sources to be marked in a limited space. The preset warning trigger condition is that the straight-line distance between the risk source and the worker's current position is less than a preset safe distance threshold. From all high-risk and medium-risk risk sources, the straight-line distance between their location coordinates and the worker's coordinates is calculated one by one and compared with the preset safe distance threshold. If the distance is less than the safe distance threshold, the risk source is considered to meet the warning trigger condition. All risk sources that meet this condition are collected together to form the set of risk sources to be marked.

[0060] Warning labels are attached to the risk sources in the set of risk sources to be marked. These tagged risk sources are then organized into entries according to their risk level, resulting in a limited list of risk sources to be avoided. A warning label is generated for each risk source to be marked. The label content includes at least the risk source's type code, its risk level, and its location relative to the workers' orientation and distance. This label is associated with all the information for that risk source. Then, the risk sources are sorted according to their risk level, with all high-risk level risk sources listed first, followed by all medium-risk level risk sources. Within the same level, risk sources are sorted from closest to furthest from the workers. This sorted list, complete with information and warning labels, constitutes the list of risk sources to be avoided.

[0061] Based on the list of risks to be avoided and the diffusion direction vectors, a hazard avoidance movement direction indication corresponding to the current position of the worker is generated. This hazard avoidance movement direction indication is then fused with the list of risks to be avoided to obtain a targeted risk avoidance instruction for a limited space. The hazard avoidance movement direction indication is calculated by taking the opposite direction of the diffusion direction vector corresponding to each risk source in the list, and then synthesizing all these opposite direction vectors. The synthesis method involves calculating the average of the horizontal components of all vectors as the horizontal component of the synthesized vector, and calculating the average of the vertical components of all vectors as the vertical component of the synthesized vector. The direction defined by these two average components is the suggested hazard avoidance movement direction, generating a clear directional description, such as moving in a direction of 30 degrees north of east. This directional description is then merged with the list of risks to be avoided to form a complete text instruction. This instruction first lists all entries in the list of risks to be avoided, and then explicitly states at the end that the worker should move in the specified direction. This merged text is the targeted risk avoidance instruction.

[0062] In step 5, a monitoring report of the working environment in the confined space is obtained, as follows: After the targeted risk avoidance command is executed, the gas concentration data of the preset monitoring points in the confined space is captured, and the captured gas concentration data is compared with the gas concentration data before the command is executed to obtain the concentration change status of the preset monitoring points after the command is executed. After the command is executed, the state of concentration change is checked for consistency in order to identify effective monitoring points whose concentration change trend matches the expected change direction of the targeted risk avoidance command, and the feedback concentration sequence of the effective monitoring points is extracted to obtain the effective feedback information after the command is executed. Based on effective feedback information, the recovery degree of the working environment situation in the confined space is assessed, the residual risk parameters of the abnormal diffusion area after the execution of the instructions are obtained, and the residual risk parameters are time-series aligned with the execution records of targeted risk avoidance instructions to obtain a set of elements for assessing the working environment situation in the confined space. The set of elements for assessing the operational environment, along with the spatial structure information of the confined space and the deployment information of the preset monitoring points, are arranged in an itemized manner according to a preset report template to obtain a monitoring report of the operational environment of the confined space.

[0063] After the targeted risk avoidance command is executed, the gas concentration data at the preset monitoring points within the confined space is captured, and the captured gas concentration data is compared with the gas concentration data before the command execution to obtain the concentration change status of the preset monitoring points after the command execution. Immediately after the command execution, gas concentration data for a new time window is collected from all preset monitoring points as the post-command data. The pre-command data refers to the historical concentration data at the corresponding time point on which the targeted risk avoidance command was based. For each monitoring point, the average value of its post-command data within the corresponding time window is calculated, along with the average value within the corresponding pre-command time window. The difference between the post-command average and the pre-command average is the concentration change at that point. This concentration change is then divided by the pre-command average and multiplied by 100% to obtain a percentage value. This percentage value, along with its positive or negative sign, defines the concentration change status of that point after the command execution.

[0064] A consistency check is performed on the concentration changes after command execution to identify valid monitoring points whose concentration change trends match the expected direction of the targeted risk avoidance command. Feedback concentration sequences from these valid monitoring points are then extracted to obtain the valid feedback information after command execution. The expected direction of the targeted risk avoidance command is to reduce the concentration at monitoring points along the risk source and its diffusion path. Therefore, the consistency check method involves checking whether the percentage value of the concentration change after command execution at each monitoring point is negative. If the value is negative, the concentration decrease trend at that point matches the expected direction of the command, and the point is identified as a valid monitoring point. For all identified valid monitoring points, the original concentration data sequence captured within the entire time window after command execution is extracted; this sequence is the feedback concentration sequence for that point. All valid monitoring points and their corresponding feedback concentration sequences together constitute the valid feedback information after command execution.

[0065] Based on effective feedback information, the recovery level of the working environment in a confined space is assessed to obtain residual risk parameters of the abnormal diffusion area after the execution of instructions. These residual risk parameters are then time-sequentially aligned with the execution records of targeted risk avoidance instructions to obtain a set of elements for assessing the working environment situation in the confined space. The recovery level assessment is achieved by calculating residual risk parameters. First, effective monitoring points that originally belonged to the abnormal diffusion area are located from the effective feedback information. The average value of the feedback concentration sequence at these points after instruction execution is calculated. Then, this average value is divided by a predefined safe concentration limit standard for that gas type, and multiplied by 100% to obtain a percentage value. This percentage value represents the degree to which the abnormal area remains above the safety standard after instruction execution; this is the residual risk parameter. The calculated residual risk parameter is then arranged and linked chronologically with the detailed text of the targeted risk avoidance instructions, the specific time the instructions were issued, and the specific time the instructions were completed. This linked set, including residual risk parameters, instruction text, instruction issuance time, and instruction completion time, constitutes the set of elements for assessing the working environment situation.

[0066] The work environment situation assessment elements, along with the spatial structure information of the confined space and the deployment information of pre-set monitoring points, are arranged according to a pre-set report template to generate a confined space work environment monitoring report. The pre-set report template is a document framework that specifies the order of chapters, titles, and content. Spatial structure information includes a description of the confined space's dimensions, shape, main facilities, and entrance / exit locations. The deployment information of pre-set monitoring points includes a unique number for each point, a description of its specific physical location, and its node coordinates in the spatial coordinate mapping grid. During arrangement, the first part of the report first includes a descriptive text of the spatial structure information. The second part lists the deployment information of the monitoring points in tabular form. The third part fully incorporates the work environment situation assessment elements, including the specific values ​​and units of residual risk parameters, the full text of risk avoidance instructions, and the precise times of instruction issuance and completion. The fourth part includes a list of concentration changes at all points after instruction execution. Finally, a complete document containing all the above parts is generated; this document is the confined space work environment monitoring report.

[0067] Example 2: A confined space work environment monitoring device, comprising: One or more processors; Memory, used to store one or more computer programs; When one or more programs are executed by one or more processors, the one or more processors execute the method in Example 1.

[0068] Example 3: A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the method in Example 1.

[0069] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can make equivalent substitutions or modifications based on the technical solution and concept of the present invention within the scope of the technology disclosed in the present invention, and such modifications should also be considered to fall within the scope of protection of the present invention.

Claims

1. A method for monitoring the working environment in a confined space, characterized by the following steps: include: Step 1: Perform time-series alignment on the gas concentration data of preset monitoring points within the confined space to obtain a multi-channel time-series concentration sequence for the confined space; Step 2: Calculate the concentration change rate of the multi-channel time series within the current time window, and perform spatial interpolation on the concentration change rate according to the spatial distribution of the preset monitoring points to obtain a gas concentration distribution gradient map in a limited space. The process of calculating the rate of concentration change of multiple time-series concentration sequences within the current time window is as follows: From multiple time-series concentration sequences, the instantaneous concentration sampling values ​​of preset monitoring points within the current time window are extracted, and the instantaneous concentration sampling values ​​are decomposed into time series to extract the trend concentration component of the preset monitoring points within the current time window. With fluctuation concentration component ; Taking the trend derivatives of both the trend concentration component and the fluctuation concentration component within the current time window, and using the first derivative of the trend concentration component with respect to time as the rate of change of the trend. The first derivative of the fluctuation concentration component with respect to time is taken as the fluctuation rate of change. The formula for calculating the rate of change of the trend is as follows: ； In the formula, This represents the trend concentration component at time t, where t represents time. By traversing the connection paths between monitoring points and adjacent monitoring points in the preset monitoring point spatial topology diagram, the diffusion attenuation factor of the preset monitoring point relative to the risk source in the limited space is obtained. Based on the trend rate of change, fluctuation rate of change, and diffusion decay factor, the concentration change rate of multiple time-series concentration sequences within the current time window is calculated using the following formula: ； In the formula, This represents the rate of concentration change at the i-th monitoring point within the current time window. This represents the preset trend change rate weighting coefficient. This represents the preset weighting coefficient for the rate of change of volatility, where e represents the natural constant. This represents the diffusion attenuation factor of the i-th monitoring point relative to the risk source within a finite space. This represents the preset diffusion attenuation reference factor. This represents the preset diffusion response sensitivity coefficient; Step 3: Based on the gas concentration distribution gradient map, identify the abnormal diffusion region in the limited space where the gas concentration gradient exceeds the preset gradient threshold, and determine the diffusion direction vector of the abnormal diffusion region. Step 4: Based on the diffusion direction vector and the relative orientation between the preset monitoring points and the current position of the workers in the confined space, generate targeted risk avoidance instructions for the confined space. Step 5: After executing the targeted risk avoidance instructions, receive the feedback information after the instructions are executed, and encapsulate the feedback information in a structured manner to obtain a monitoring report of the working environment in the confined space.

2. The confined space work environment monitoring method as described in claim 1, characterized in that, In step 1, the process of obtaining the multi-channel time-series concentration sequence in a finite space is as follows: The gas concentration data at preset monitoring points within a confined space is obtained, and outliers are removed from the gas concentration data to obtain the effective concentration data for the confined space. The effective concentration data is timestamped according to a unified time base to obtain the standard concentration data of the preset monitoring points in a limited space at the same sampling time. The standard concentration data are sorted according to the spatial distribution order of the preset monitoring points to form a concentration distribution vector in a limited space; Arrange the concentration distribution vectors in chronological order to obtain a multi-channel time-series concentration sequence in a finite space.

3. The confined space work environment monitoring method as described in claim 1, characterized in that, In step 2, the process of obtaining the gas concentration distribution gradient map of the finite space is as follows: Based on the spatial distribution of preset monitoring points, a spatial coordinate mapping grid is constructed in a limited space, and the concentration change rate of the preset monitoring points is anchored to the grid nodes in the spatial coordinate mapping grid to obtain the node rate value of the grid nodes. Based on the spatial topological adjacency relationship between adjacent anchored grid nodes in the spatial coordinate mapping grid, the rate value of the blank grid nodes without anchored rate values ​​is inherited through the connected path to obtain the derived rate value of the blank grid nodes. According to the grid spatial arrangement order, the gradient direction solution is performed on the nodal rate values ​​of the grid nodes in the spatial coordinate mapping grid to obtain the transverse gradient field and the longitudinal gradient field of the spatial coordinate mapping grid. By fusing the transverse and longitudinal gradient fields into vector fields, a gradient vector distribution map of a finite space is obtained. The gradient vector distribution map is then visualized and rendered to obtain a gas concentration distribution gradient map of a finite space.

4. The method for monitoring confined space working environment as described in claim 1, characterized in that, In step 3, abnormal diffusion regions where the gas concentration gradient exceeds a preset gradient threshold within a confined space are identified, and the diffusion direction vector of the abnormal diffusion regions is determined. The process is as follows: Based on the gas concentration distribution gradient map, the gradient amplitude is traversed for spatial points in a limited space, and spatial points whose gradient amplitude exceeds the preset gradient threshold are marked as abnormal points. Spatial connectivity clustering is performed on the abnormal locations to obtain the abnormal diffusion area in a limited space; Curvature analysis is performed on the spatial boundary contour of the anomalous diffusion region to locate the diffusion front boundary and diffusion wake boundary of the anomalous diffusion region. By vector synthesis of the gradient vectors at the boundary points on the diffusion front boundary, the main vector of the diffusion direction in the anomalous diffusion region is obtained; Spatial orientation calibration of the main diffusion direction vector yields the diffusion direction vector of the abnormal diffusion region.

5. The confined space work environment monitoring method as described in claim 4, characterized in that, The process of locating the diffusion front boundary and diffusion wake boundary of the abnormal diffusion region is as follows: Discretize the spatial boundary contour of the abnormal diffusion region to decompose the spatial boundary contour into a continuous sequence of boundary micro-segments, and assign a corresponding set of boundary points to the sequence of boundary micro-segments. The curvature of the boundary point set in the spatial coordinate system is determined by the orientation offset. Boundary micro-segments with curvature reaching a preset curvature threshold are marked as high curvature micro-segments, and boundary micro-segments with curvature below the preset curvature threshold are marked as low curvature micro-segments. By performing connected clustering on continuously distributed and spatially adjacent high-curvature micro-segments and low-curvature micro-segments, the high-curvature boundary segments and low-curvature boundary segments of the anomalous diffusion region are obtained. Based on the overall diffusion trend of the abnormal diffusion region, the high curvature boundary segment located in the diffusion forward direction is positioned as the diffusion front boundary, and the low curvature boundary segment located in the diffusion source direction is positioned as the diffusion wake boundary.

6. The method for monitoring confined space working environments as described in claim 1, characterized in that, In step 4, the process of generating targeted risk avoidance instructions for a limited space is as follows: Obtain the spatial coordinates of the current position of the worker within a confined space, and compare the spatial coordinates with the spatial distribution of the preset monitoring points to obtain the relative positional relationship between the current position of the worker and the preset monitoring points; The diffusion direction vector is coupled with the relative orientation relationship to determine whether the diffusion direction vector points to the current position of the worker, thus obtaining the risk orientation judgment result in a limited space. Based on the risk orientation determination results, the risk sources in the confined space are classified into risk levels, and the risk sources whose risk levels reach the preset warning threshold are labeled to obtain a list of risk sources to be avoided in the confined space. Based on the list of risk sources to be avoided and the diffusion direction vector, a risk avoidance movement direction indication corresponding to the current position of the operator is generated. The risk avoidance movement direction indication and the list of risk sources to be avoided are then fused to obtain targeted risk avoidance instructions for a limited space.

7. The confined space work environment monitoring method as described in claim 6, characterized in that, Based on the risk orientation determination results, the process of obtaining a list of risk sources to be avoided in a limited space is as follows: From the risk orientation determination results, the orientation association information of the risk source relative to the current position of the operator is extracted, and the orientation association information is matched and bound with the inherent attribute parameters of the risk source to obtain the risk situation association mapping of the risk source; Based on the risk situation correlation mapping, risk sources are classified according to the degree of risk urgency. Risk sources that point to the current location of the worker and whose pointing intensity reaches the preset intensity threshold are classified into the high-risk level, risk sources that point to the current location of the worker but whose pointing intensity does not reach the preset intensity threshold are classified into the medium-risk level, and risk sources that do not point to the current location of the worker are classified into the low-risk level, thus obtaining the risk level classification of the risk sources. Risk sources classified as high-risk and medium-risk are compared with preset warning trigger conditions to filter out risk sources that meet the warning trigger conditions, resulting in a set of risk sources to be marked in a limited space. Warning labels are attached to the risk sources in the set of risk sources to be marked, and the risk sources with attached warning labels are organized into items according to risk level to obtain a list of risk sources to be avoided in a limited space.

8. The method for monitoring confined space working environment as described in claim 1, characterized in that, In step 5, the process of obtaining the working environment monitoring report for the confined space is as follows: After the targeted risk avoidance command is executed, the gas concentration data of the preset monitoring points in the confined space is captured, and the captured gas concentration data is compared with the gas concentration data before the command is executed to obtain the concentration change status of the preset monitoring points after the command is executed. After the command is executed, the state of concentration change is checked for consistency in order to identify effective monitoring points whose concentration change trend matches the expected change direction of the targeted risk avoidance command, and the feedback concentration sequence of the effective monitoring points is extracted to obtain the effective feedback information after the command is executed. Based on effective feedback information, the recovery degree of the working environment situation in the confined space is assessed, the residual risk parameters of the abnormal diffusion area after the execution of the instructions are obtained, and the residual risk parameters are time-series aligned with the execution records of targeted risk avoidance instructions to obtain a set of elements for assessing the working environment situation in the confined space. The set of elements for assessing the operational environment, along with the spatial structure information of the confined space and the deployment information of the preset monitoring points, are arranged in an itemized manner according to a preset report template to obtain a monitoring report of the operational environment of the confined space.

Images