A warehouse environment risk real-time monitoring and intelligent prevention and control system

Abstract

The application discloses a warehouse environment risk real-time monitoring and intelligent prevention and control system and relates to the technical field of warehouse safety, which comprises a multi-source heterogeneous sensor network, a data transmission module and a central processing unit arranged in a warehouse area; the multi-source heterogeneous sensor network and the central processing unit realize data interaction through the data transmission module; and the central processing unit generates initial continuous distribution of all physical fields through a normalized general potential conduction interpolation algorithm. Through the normalized general potential conduction interpolation algorithm, the application fuses spatial anisotropy and multi-physical field coupling effects, accurately reconstructs discrete sensor data into a full-space continuous physical field, and eliminates a monitoring blind area; based on physical field quantitative risk evolution, early hidden danger identification and hierarchical intelligent prevention and control are realized, and the accuracy of warehouse environment risk monitoring and the timeliness of prevention and control are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of warehouse safety technology, and in particular to a real-time monitoring and intelligent control system for warehouse environmental risks. Background Technology

[0002] With the rapid development of the modern logistics industry and the continuous improvement of the supply chain system, warehousing, as a core link in the logistics chain, is becoming increasingly larger and more complex. Modern warehousing not only undertakes the basic function of goods storage but is also gradually transforming into comprehensive logistics centers integrating storage, sorting, distribution, and processing. The types of goods stored have also expanded from ordinary daily necessities to special categories such as flammable and explosive materials, chemicals, fresh food, and pharmaceuticals. These special goods have strict requirements for parameters such as temperature, humidity, gas concentration, dust content, and light intensity in the warehousing environment. Abnormalities in any of these parameters can lead to spoilage, damage, or even major safety accidents such as fires, explosions, and poisoning, causing huge economic losses and casualties.

[0003] To ensure warehouse safety and reduce environmental risks, various warehouse environmental monitoring systems have emerged in existing technologies. These systems generally employ sensor network technology, deploying devices such as temperature and humidity sensors, smoke sensors, gas sensors, and dust sensors within the warehouse area to collect environmental parameters at discrete measurement points and transmit the data to a monitoring center for display and alarm functions. However, due to limitations in cost, installation conditions, and the number of devices, sensors can only be deployed at a limited number of discrete locations, failing to cover every corner of the warehouse. To obtain the environmental status of the entire space, it is necessary to reconstruct a continuous environmental physical field from discrete measurement point data using data interpolation methods. This is the core foundation for achieving full-space risk monitoring and control.

[0004] Currently, commonly used interpolation methods in the field of warehouse environment monitoring mainly include inverse distance weighting, Kriging interpolation, and linear interpolation. However, warehouse environments contain numerous obstacles such as shelves, stacks of goods, walls, and columns. These structures have a significant anisotropic impact on the propagation of temperature, humidity, gases, and dust—for example, the diffusion rate of gases in shelf aisles is much faster than that inside dense stacks of goods. Most existing interpolation methods assume that the space is homogeneous and isotropic, calculating interpolation weights solely based on the Euclidean distance between measuring points, completely ignoring the obstructive and guiding effects of spatial structures on the transmission of physical quantities. This leads to significant errors in the reconstructed physical field within sensor blind zones such as inside stacks of goods and behind shelves, failing to accurately reflect the true environmental conditions.

[0005] Due to the fundamental limitations of the aforementioned interpolation techniques, existing warehouse environment monitoring systems cannot accurately reconstruct the continuous physical field of the entire space. They can only rely on threshold alarms from a limited number of discrete measuring points, resulting in numerous monitoring blind spots. They are unable to identify early, potential risks and hazards, nor can they provide reliable spatial data support for precise intelligent prevention and control. Therefore, there is an urgent need to develop a high-precision interpolation method that can integrate multi-source heterogeneous data and consider the anisotropy of spatial structures and the coupling effects of multiple physical fields. This method would enable accurate reconstruction of the physical field of the warehouse environment, laying a solid foundation for real-time monitoring and intelligent prevention and control of warehouse environmental risks. Summary of the Invention

[0006] This invention provides a real-time monitoring and intelligent control system for warehouse environment risks, characterized by comprising: a multi-source heterogeneous sensor network deployed within the warehouse area, a data transmission module, and a central processing unit; the multi-source heterogeneous sensor network and the central processing unit interact via the data transmission module; the central processing unit specifically includes: The physical field initialization module is used to divide the entire storage area into multiple continuous three-dimensional grid cells. When the system starts, it initializes the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid cell, resulting in five mutually coupled continuous physical fields: temperature field, humidity field, gas concentration field, dust concentration field and risk potential energy field. The physical field update module is used to synchronously update four continuous physical fields—temperature field, humidity field, gas concentration field, and dust concentration field—using discrete observation data collected in real time by a multi-source heterogeneous sensor network. The risk potential energy field update module is used to continuously update the risk potential energy field based on the real-time updated temperature field, humidity field, gas concentration field, and dust concentration field. The risk potential energy field analysis module is used to perform real-time analysis of the risk potential energy field and implement prevention and control strategies based on the analysis results.

[0007] The aforementioned real-time monitoring and intelligent control system for warehouse environmental risks initializes the temperature, humidity, gas concentration, dust concentration, and risk potential energy value of each grid cell, specifically including: Collect sensor data and calculate reliability weights; The collected sensor data is uniformly mapped into a dimensionless normalized universal potential vector. Define a uniform propagation tensor for the center point of each grid cell; The unified equivalent transmission distance between the sensor and each grid cell is calculated based on the unified transmission tensor field. The temperature, humidity, gas concentration, and dust concentration of each grid cell are calculated by combining sensor reliability weights, normalized universal potential vectors, and unified equivalent conduction distance. The initial risk potential value of each grid cell is calculated based on the hazard level of the stored goods, storage time, and historical risk records.

[0008] The aforementioned real-time monitoring and intelligent control system for warehouse environmental risks utilizes discrete observation data collected in real time by a multi-source heterogeneous sensor network to simultaneously update four continuous physical fields: temperature field, humidity field, gas concentration field, and dust concentration field. Specifically, this includes: The discrete observation data acquired in real time is mapped to the observation universal potential field; Map the physical field of the previous cycle to the background universal potential field; The uniform equivalent transmission distance between the sensor and each grid cell at the current moment is calculated based on the background general potential field. The general potential field is constructed and analyzed by combining the observed general potential field and the unified equivalent transmission distance between the sensor and each grid cell at the current moment. The analysis of the universal potential field is reversed and mapped back to the real physical quantities to generate an updated continuous physical field.

[0009] The aforementioned real-time monitoring and intelligent control system for warehouse environment risks continuously updates the risk potential energy field based on real-time updates of the temperature field, humidity field, gas concentration field, and dust concentration field, specifically including: Based on the analysis of the current cycle, the general potential field is used to calculate the environmental risk source terms for each grid cell; A partial differential equation for the evolution of the risk potential field is established based on the environmental risk source terms of each grid cell; The risk potential field of the previous cycle is updated based on the established partial differential equation.

[0010] This invention also provides a method for real-time monitoring and intelligent control of risks in the warehouse environment, characterized by comprising: S10: Divide the entire storage area into multiple continuous three-dimensional grid cells, and initialize the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid cell when the system starts up, to obtain five mutually coupled continuous physical fields. S20: Utilizes discrete observation data collected in real time by a multi-source heterogeneous sensor network to synchronously update four continuous physical fields: temperature field, humidity field, gas concentration field, and dust concentration field. S30: The risk potential energy field is continuously updated based on the real-time updated temperature field, humidity field, gas concentration field, and dust concentration field; S40: Perform real-time analysis of the risk potential field and implement prevention and control strategies based on the analysis results.

[0011] The beneficial effects achieved by this invention are as follows: By using a normalized universal potential conduction interpolation algorithm, the spatial anisotropy and multi-physics coupling effects are integrated to accurately reconstruct discrete sensor data into a continuous physical field in the whole space, eliminating monitoring blind spots; based on the physical field quantification of risk evolution, early hidden danger identification and hierarchical intelligent prevention and control are realized, significantly improving the accuracy of warehouse environment risk monitoring and the timeliness of prevention and control. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0013] Figure 1 This is a schematic diagram of a real-time monitoring and intelligent control system for warehouse environment risks provided in Embodiment 1 of this application; Figure 2 This is a flowchart of a method for real-time monitoring and intelligent control of risks in the storage environment provided in Embodiment 2 of this application. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Example 1 like Figure 1 As shown, Embodiment 1 of this application provides a real-time monitoring and intelligent control system for warehouse environment risks, including: a multi-source heterogeneous sensor network 1 deployed in the warehouse area, a data transmission module 2, and a central processing unit 3; the multi-source heterogeneous sensor network 1 and the central processing unit 3 achieve data interaction through the data transmission module 2; wherein the multi-source heterogeneous sensor network 1 is used to collect discrete observation data in the warehouse area; the central processing unit 3 is used to process the data collected by the multi-source heterogeneous sensor network and execute control strategies based on the data processing results; the central processing unit 3 specifically includes: (1) Physical field initialization module 31 is used to divide the entire storage area into multiple continuous three-dimensional grid units, and initialize the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid unit when the system starts, so as to obtain five mutually coupled continuous physical fields: temperature field, humidity field, gas concentration field, dust concentration field and risk potential energy field. First, acquire 3D building information model data or 3D laser scanning point cloud data of the storage area, and extract the spatial boundary coordinates of fixed enclosure structures such as walls, columns, partitions, doors and windows; acquire the 3D geometric dimensions, arrangement, number of layers and precise location of each layer of the shelves; acquire the goods stacking distribution data, including the type, quantity, material, stacking geometry and stacking compactness of goods stored in each storage location; Then, based on the extracted spatial boundaries and cargo distribution data, a three-dimensional subdivision is performed using a Cartesian grid or an adaptive octree grid; a fine grid is used in cargo-dense areas and sensor deployment areas, while a coarser grid is used in open passage areas, and each grid cell is assigned a unique spatial index.

[0016] This step proposes a normalized universal potential conduction interpolation algorithm, which treats temperature, humidity, gas concentration, and dust concentration as a unified "universal potential." By constructing a unified conduction tensor that integrates spatial anisotropy and multi-physics coupling effects, the initial continuous distribution of all physical fields is generated in one step. The specific process is as follows: S101: Collect sensor data and calculate reliability weights; Upon system startup, a multi-source heterogeneous sensor network collects all discrete observation data for the current moment, including the three-dimensional spatial coordinates of each sensor measurement point i. Measured temperature value relative humidity value Gas concentration value and dust concentration value Simultaneously, the real-time signal-to-noise ratio output from the sensor's embedded self-diagnostic module is extracted. and zero-point drift accumulation The reliability weight of the sensor is calculated using the following formula. : in, This refers to the standard signal-to-noise ratio calibrated at the factory for this sensor model. The maximum permissible drift amount specified in the technical manual; The value range is (0,1], which is used to reduce the contribution of the degraded sensor to the interpolation result.

[0017] S102: Map the collected sensor data into a dimensionless normalized universal potential vector; Temperature general potential ,in For optimal storage temperature of goods, and These are the lower and upper limits of the temperature that are permissible for safe storage of goods, respectively, derived from the Material Safety Data Sheet (MSDS) provided by the goods supplier or industry warehousing standards; Humidity general potential ,in The optimal relative humidity for cargo storage is derived from the MSDS. The maximum allowable relative humidity limit deviation is determined by... It is obtained by adding the larger of the absolute values ​​of the allowable positive and negative deviations; this formula converts both deviations of humidity from the optimal value into an increase in potential energy, uniformly characterizing the risk of abnormal humidity. Universal potential of gas concentration ,in The concentration safety threshold for the gas is one percent of the occupational exposure limit or lower explosive limit of the gas, derived from the national standard "Occupational Exposure Limits for Hazardous Factors in the Workplace" or the Material Safety Data Sheet. General potential of dust concentration ,in The safe threshold for dust concentration is one-tenth of the lower limit of the dust explosion concentration, derived from the "Dust Explosion Prevention Safety Regulations" or experimental measurement data.

[0018] If the calculated universal potential is greater than 1, it is truncated to 1. The universal potential vector of sensor i is denoted as... .

[0019] S103: Define a uniform propagation tensor for the center point of each grid cell; Center point of each grid cell Define a Unified transmission tensor This tensor describes the propagation characteristics of a universal potential in three-dimensional space. Acting in spatial directions, it determines the relative speed of propagation of physical quantities in different directions. Its expression is: ,in, Indicates the center point of the grid cell. Based on the isotropic conduction term, , for The identity matrix; is the fundamental scalar conductivity coefficient of air. The calibration method is as follows: obtain the average temperature of the storage environment. and atmospheric pressure Query standard status (standard status reference temperature) Standard reference atmospheric pressure The geometric mean of the air thermal diffusivity and mass diffusivity under these conditions is taken as the reference value. Then calculate according to the following temperature and pressure correction formula: ; This is the structure-induced anisotropy term, used to describe the anisotropic influence of warehouse structures such as shelving and stacking on the transmission of the general potential. ,in The unit vector representing the main direction of the grid cell is determined by the following rules: if the cell is located in the aisle area, the unit vector tangent to the centerline of the aisle is taken; if the cell is located inside the stack of goods and there is obvious stratification, the unit vector parallel to the stratification plane and along the natural spreading direction of the material is taken; otherwise, the unit vector projected from the cell to the nearest aisle direction is taken. Represents the vector tensor product; The anisotropic strength coefficient is calibrated as follows: , Given the anisotropy ratio, if conditions permit, a tracer gas experiment would be conducted. Take the ratio of the measured equivalent diffusion coefficients in the direction of the shelf aisle to the vertical direction; if experimental conditions are unavailable, use an empirical formula. Estimate, The equivalent porosity of the grid cells. , This refers to the stacking density (the ratio of the mass of the goods to the volume they occupy). The bulk density of the cargo can be found in the materials handbook; This is a cross-physics coupling driving term used to describe the mutual driving effect between different physics fields. Its calculation process is as follows: For each physics field... (Representing temperature, humidity, gas concentration, and dust concentration, respectively), calculate its universal potential gradient vector at point x. A cohesion coupling coefficient is set for each physical field a. , Risk cross-coupling matrix The Middle Sum of all elements in the row; compute the coupling driving sub-items of the physical field: Perform matrix addition on the coupling driving terms of the four physics fields to obtain the cross-physics coupling driving terms of the mesh element. , It is The matrix is ​​such that each element is a tensor product of the gradient vectors.

[0020] The calibration was completed objectively based entirely on historical data, using the following method: 1. Collect a dataset of typical risk events involving simultaneous abnormal fluctuations of multiple sensors in the warehouse's history; 2. Select a time window before the event occurs, calculate the Granger causality or transfer entropy between the time series of each physical field, and determine the driving relationship and relative strength of field b on field a; 3. Set the reference coupling coefficient ,in This refers to the average temperature gradient magnitude observed in typical risk events. 4. Scaling based on the F-statistic of the Granger causality test: If the test is not significant, then , For the given significance threshold, this embodiment uses 3.07; When there is no historical risk event data in the warehouse Initialize as a diagonal matrix, setting off-diagonal elements to 0 and all diagonal elements to 0. .

[0021] When the initial continuous field is generated for the first time, the gradients of each physical field cannot be calculated, therefore The term is temporarily set to zero, and this coupling term will be enabled in the dynamic field update in the subsequent step S30.

[0022] S104: Calculate the unified equivalent transmission distance between the sensor and each grid cell based on the unified transmission tensor field; For the center point of the target grid cell With sensor points Define a unified equivalent transmission distance Let be the integral value along the path of least resistance in the conduction tensor field, i.e.: in For all connected sensor points Center point of the target grid cell A set of continuous paths in three-dimensional space. Let arc length be the parameter of the path, and lower limit of integration be... This indicates the upper limit of integration starting from the sensor position. This indicates the end point is at the center of the target grid cell. For any point on the path The unit tangent vector at point A, whose direction is along the tangent direction of the path, has a magnitude of 1. unit tangent vector transpose, The inverse matrix of the unified conduction tensor (i.e., the drag tensor) characterizes the spatial location of the universal potential. The magnitude of the conduction resistance at that point.

[0023] In practical solutions, complex continuous path integration is unnecessary; a discrete grid approximation method can yield sufficiently accurate results. On the partitioned 3D grid, the center point of each grid cell is used as a node in the graph. For any two adjacent grid cells, the edge weight between them is calculated: the drag tensor at the midpoint of the line connecting the two cell center points is taken, the drag per unit length in that direction is calculated, and then multiplied by the Euclidean distance between the two center points to obtain the edge weight. The Dijkstra shortest path algorithm is then run, starting from each sensor node, to calculate the shortest path length from it to all other grid cell nodes. This shortest path length represents the sensor node's path length. To the center point of the target grid cell Unified equivalent transmission distance .

[0024] S105: Combine sensor reliability weights, unified equivalent transmission distance, and normalized universal potential vector to calculate the temperature, humidity, gas concentration, and dust concentration of each grid cell; First, calculate the interpolation weight of each sensor for the target grid cell. Sensor i for the target grid cell... The interpolation weights are: in Let i be the reliability weight of sensor i. For bandwidth parameters, The calibration method is as follows: calculate the average value of the equivalent conduction distance between all sensors and their nearest neighbor sensors. ,Pick This ensures a smooth transition of weights between sensors.

[0025] Target Mesh Cell Universal potential vector The result is obtained in one step by normalizing the weighted average: ,in Let be the universal potential vector of sensor i; The normalized universal potential components of each grid cell are mapped back to the real physical quantities to obtain the initial continuous distribution of the entire space. The specific mapping formula is as follows: Temperature field: ,in To obtain the optimal storage temperature, For target mesh cells Temperature value at that location, and These are the lower and upper limits of the permissible temperature for safe storage of goods. For the target cell The general potential at the temperature is extracted from ; The determination rule is: if the distance to the target mesh cell Standardized equivalent conduction distance for the sensor's measured temperature value If, then take the positive sign. If , then take the negative sign.

[0026] Humidity field: ,in For target mesh cells The relative humidity value at that location The optimal relative humidity for storing goods; The maximum allowable relative humidity limit deviation, For target mesh cells Humidity general potential at the location, extracted from , The determination rule is: if the distance to the target mesh cell Standardized measured humidity values ​​from sensors with the shortest equivalent conduction distance If, then take the positive sign. If , then take the negative sign; Gas concentration field: ,in For target mesh cells The gas concentration value at that location, This is the safe concentration threshold for the gas. For target mesh cells The general potential of the gas concentration at that location, extracted from ; Dust concentration field: ,in For target mesh cells Dust concentration value at the location, The dust concentration safety threshold, For target mesh cells The general potential of dust concentration at the location is extracted from ; At this point, the initial continuous distribution of the temperature field, humidity field, gas concentration field, and dust concentration field has been generated.

[0027] S106: Calculate the initial risk potential value for each grid cell based on the hazard level, storage time, and historical risk records of the stored goods; Taking into account the inherent hazards of the goods, the cumulative effect of storage time, and historical risk records, the initial risk potential value is calculated for each grid cell. The specific steps are as follows: 1. Determine the hazard rating factor of the goods; Define a hazard level coefficient for each grid cell. This coefficient is determined based on the category of goods stored within the grid cell: a baseline value of 1.0 is used for non-toxic and non-flammable general goods; 2.5–4.0 for flammable solids; 5.0–7.0 for flammable liquids; and 8.0–10.0 for explosives and highly toxic substances. Specific values ​​can be found in a pre-set table of hazardous goods attributes, which is developed with reference to national standards such as the "Classification and Numbering of Dangerous Goods" and in conjunction with industry risk assessment models used by insurance companies.

[0028] 2. Calculate the cumulative risk factor over time. Extract the storage time of goods within each grid cell (In days), the time-cumulative risk factor is defined as: The time constant The timeframe is set based on the type of goods; 90 days is used for general goods, and 30 days for chemicals. This factor reflects the non-linear cumulative risk caused by prolonged storage, such as packaging aging and increased reactivity.

[0029] 3. Calculate the historical risk frequency factor Obtain historical risk records for each grid cell location and calculate the historical risk frequency factor: ,in This refers to the number of times that the grid cell has been identified as experiencing a risk event (such as abnormal temperature or leakage alarm) within a preset time period (e.g., one year). This represents the total number of historical monitoring records for this location. This is the historical risk gain coefficient, ranging from 0.5 to 2.0, selected based on the overall frequency of risk events in the warehouse. If this grid cell has no historical risk records, then... .

[0030] 4. Calculate the initial risk potential value of each grid cell by combining the cargo hazard level coefficient, time cumulative risk factor, and historical risk frequency factor; Center of each grid cell Initial risk potential value The calculation formula is: in, The total energy storage coefficient of the cargo is calculated based on the cargo's total calorific value or chemical energy release potential, and the unit is [unit missing]. The data comes from the Material Heat of Combustion Data Handbook; a value of 1.0 is used for ordinary non-toxic and non-flammable goods. ; This is the weighted balance coefficient between historical risk and time risk, with a default value of 1.0.

[0031] After traversing all grid cells and assigning values, the initial risk potential energy field is established.

[0032] (2) Physical field update module 32 is used to synchronously update four continuous physical fields: temperature field, humidity field, gas concentration field, and dust concentration field using discrete observation data collected in real time by a multi-source heterogeneous sensor network. This step is executed at the end of each sensor data acquisition cycle (e.g., every 30 seconds or 1 minute), using the discrete observation data reported by the sensors at the current moment to update the four continuous physical fields in real time. The update algorithm continues the normalized universal potential conduction interpolation framework, making full use of historical field information and new observation information to achieve dynamic correction while maintaining physical consistency. The specific implementation process is as follows: S201: Map the discrete observation data acquired in real time to the observation universal potential field; Using the same reliability weight formula as in step S101, the reliability weight of each sensor is calculated. If a sensor fails to report data in a given period, its reliability weight is reset to 0. If no data is reported for three consecutive acquisition cycles, an alarm is triggered to notify management personnel for maintenance. Using the same method as in step S102, the discrete observation data acquired in real time is converted into normalized universal potential vectors for each sensor to obtain the observation universal potential field. .

[0033] S202: Maps the physical field of the previous cycle to the background universal potential field; Using the temperature field, humidity field, gas concentration field, and dust concentration field updated in the previous acquisition cycle (or initially generated) as the background field, and employing the same mapping rule as in step S102, the physical quantity values ​​of the four background fields at the center point x of each grid cell are transformed into a background universal potential vector, thus obtaining the background universal potential field. .

[0034] S203: Calculate the unified equivalent transmission distance between the sensor and each grid cell at the current moment based on the background general potential field; During the initial field generation in step S10, the cross-physics coupling term in the unified conduction tensor The value is set to zero; in this real-time update step, the background universal potential field already has a complete spatial distribution, and the gradients of each field can be calculated, thereby enabling the coupling term and making the conduction tensor more accurately reflect the real physical process. The specific operation is as follows: 1. At the center point x of each grid cell, the gradient vectors of each component of the corresponding background universal potential vector are calculated using the central difference method; thus, four gradient vectors are obtained at the center point x of each grid cell: temperature gradient vector. Humidity gradient vector Gas concentration gradient vector and dust concentration gradient vector For boundary elements where central difference cannot be used, one-sided difference should be used instead.

[0035] 2. Recalculate the cross-physics coupling term for each mesh element. ; The cross-physics coupling term for each mesh element is calculated using the same rules as in step S103. .

[0036] 4. The newly calculated cross-physics coupling terms With the fundamental isotropic term Structure-induced anisotropy term Superposition yields the complete unified propagation tensor field at the current moment. ; Fundamental isotropic terms The calculation is repeated every m hours; in this embodiment, m is taken as 1. Structure-induced anisotropy term The calculation is recalculated when the location of goods in the warehouse changes.

[0037] 5. Recalculate the unified equivalent transmission distance between the sensor and each grid cell based on the current unified transmission tensor field; The uniform equivalent transmission distance between sensor i and each grid cell x is recalculated using the same algorithm as in step S104. , Let be the three-dimensional spatial coordinates of sensor i.

[0038] S204: Construct and analyze the general potential field by combining the observed general potential field and the unified equivalent transmission distance between the sensor and each grid cell at the current moment; The unified equivalent propagation distance at the current moment obtained in step S203 is used as the reference distance. Substitute the interpolation weights into the interpolation weight calculation formula in step S105; then combine the calculated interpolation weights with the observation universal potential vector of each effective sensor i at the current time. Substitute the normalized weighted average formula from step S105 to calculate the center point of each grid cell. Analysis of the general potential vector at the location The general potential field is obtained through analysis. .

[0039] Unlike the initial generation in step S10, the equivalent transmission distance used now has been updated with cross-physics coupling terms based on real-time gradient information in S203, which can more accurately reflect the spatial transmission characteristics under multi-physics coupling at the current moment, thus making the general potential field reconstructed by interpolation closer to the real physical state.

[0040] S205: The analysis general potential field is reverse-mapped back to the real physical quantities to generate an updated continuous physical field; Using the same reverse mapping rule as in step S105, the components of the general potential field are mapped back to physical quantities to obtain the updated temperature field, humidity field, gas concentration field, and dust concentration field.

[0041] (3) Risk potential energy field update module 33, used to continuously update the risk potential energy field based on the real-time updated temperature field, humidity field, gas concentration field and dust concentration field; This module analyzes the general potential field based on the real-time output of the physics field update module 32. The risk potential energy field is dynamically updated. Its evolution follows a partial differential equation of risk source generation, spatial diffusion, and natural decay, which is solved using the finite volume method. The update cycle is synchronized with the sensor acquisition cycle (e.g., 30 seconds or 1 minute). The specific process is as follows: S301: Calculate the environmental risk source terms for each grid cell based on the general potential field analysis of the current cycle; Extract the center of each grid cell from the general potential field of the current cycle. Analysis of the general potential vector at the location Then, the environmental risk source terms of the grid cell are obtained by weighted summation of each component. ,Right now ,in , , , The analysis of the universal potential vector is as follows. The general potentials for temperature, humidity, gas concentration, and dust concentration are included. , , , These are the risk weight coefficients for each component, all defaulting to 1.0. These can be adjusted based on the type of goods (e.g., for flammable liquids, the weight can be increased). and For precision instruments, it can increase ).

[0042] S302: Establish partial differential equations for the evolution of the risk potential energy field based on the environmental risk source terms of each grid cell; Based on the physical process of risk source generation, spatial diffusion, and natural decay, a partial differential equation for the evolution of the risk potential energy field is established: ,in: The time rate of change of risk potential energy; Let x be the risk potential energy value at the center x of the grid cell at time t; The environmental risk source term at the center x of the grid cell represents the gain of risk due to environmental parameter anomalies. This is the risk potential energy spatial diffusion term, which characterizes the propagation characteristics of risk in three-dimensional space; For divergence operators; The center of the grid cell calculated for the current sampling period The unified transport tensor at that location; Let be the gradient vector of the risk potential energy field at the center x of the grid cell at time t (calculated using the central difference method), which represents the rate of change and direction of the risk potential energy in three-dimensional space; The risk attenuation coefficient is determined based on warehouse ventilation conditions: 0.005~0.01 for warehouses with good natural ventilation. For mechanically ventilated warehouses, use 0.003~0.006. For sealed storage, take 0.001~0.003. .

[0043] S303: Update the risk potential energy field of the previous cycle based on the established partial differential equation; For different types of spatial boundaries within the warehouse, corresponding physical boundary conditions are applied to ensure the physical rationality of solving partial differential equations. For example, for solid structural enclosure boundaries, a zero-flux boundary condition is used, meaning that potential energy cannot penetrate the solid structure; for open boundaries such as doors, windows, and ventilation openings, a convection boundary condition is used, considering the carrying effect of ventilation airflow on potential energy. Those skilled in the art can set various boundary settings as needed; specific settings will not be elaborated here. The finite volume method is used to discretize and solve the established partial differential equations of risk potential energy field evolution. GPU parallel computing technology is combined to improve the computational efficiency of large-scale warehouse scenarios and obtain the preliminary distribution of risk potential energy of the entire warehouse three-dimensional grid at the current moment. The preliminary calculation results are subjected to non-negativity constraints, upper limit constraints, mutation anomaly corrections, and cargo dynamic update corrections to finally obtain the updated risk potential energy field.

[0044] (4) Risk potential energy field analysis module 34, used to perform real-time analysis of the risk potential energy field and execute prevention and control strategies based on the analysis results; This module receives the risk potential energy field output by the current period risk potential energy field update module 33. Based on the risk potential energy value at the center x of each grid cell, it classifies the cell into different risk levels: when the risk potential energy value is less than the first risk threshold, it is defined as a low-risk area; when the risk potential energy value is greater than or equal to the first risk threshold but less than the second risk threshold, it is defined as a medium-risk area; when the risk potential energy value is greater than or equal to the second risk threshold but less than the third risk threshold, it is defined as a high-risk area; and when the risk potential energy value is greater than or equal to the third risk threshold, it is defined as an extremely high-risk area. For low-risk areas, maintain the regular monitoring frequency without additional prevention and control measures; for medium-risk areas, increase the monitoring frequency to 30 seconds / time, activate the local ventilation system in the risk area, send a text warning to the central control room, and mark the risk location and level; for high-risk areas, increase the monitoring frequency to 10 seconds / time, activate the ventilation system in the risk area, shut off the non-fire-fighting power supply in the area, send a high-risk warning to the handheld terminals of on-site inspection personnel, lock the access control system in the risk area, and prohibit unauthorized personnel from entering; for extremely high-risk areas, activate the audible and visual alarms, trigger the fine water mist fire extinguishing system in the corresponding area, shut off the main power supply and gas valve of the warehouse, link the access control system to block the entrances and exits of the area, and send evacuation alarms to all handheld terminals.

[0045] The first risk threshold, the second risk threshold, and the third risk threshold are set as needed. In this embodiment, the first risk threshold is 250, the second risk threshold is 500, and the third risk threshold is 1000.

[0046] Example 2 like Figure 2 As shown in Embodiment 2 of this application, a method for real-time monitoring and intelligent control of risks in the warehousing environment is provided, including: Step S10: Divide the entire storage area into multiple continuous three-dimensional grid cells, and initialize the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid cell when the system starts up, to obtain five mutually coupled continuous physical fields. Step S20: Utilize the discrete observation data collected in real time by the multi-source heterogeneous sensor network to synchronously update the four continuous physical fields: temperature field, humidity field, gas concentration field, and dust concentration field. Step S30: Continuously update the risk potential energy field based on the real-time updated temperature field, humidity field, gas concentration field, and dust concentration field; Step S40: Perform real-time analysis of the risk potential field and implement prevention and control strategies based on the analysis results.

[0047] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of the present invention should be included within the scope of protection of the present invention.

Claims

1. A real-time monitoring and intelligent control system for warehouse environment risks, characterized in that, include: A multi-source heterogeneous sensor network, data transmission module, and central processing unit deployed within the warehouse area; The multi-source heterogeneous sensor network and the central processing unit interact with each other via a data transmission module; the central processing unit specifically includes: The physical field initialization module is used to divide the entire storage area into multiple continuous three-dimensional grid cells. When the system starts, it initializes the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid cell, resulting in five mutually coupled continuous physical fields: temperature field, humidity field, gas concentration field, dust concentration field and risk potential energy field. The physical field update module is used to synchronously update four continuous physical fields—temperature field, humidity field, gas concentration field, and dust concentration field—using discrete observation data collected in real time by a multi-source heterogeneous sensor network. The risk potential energy field update module is used to continuously update the risk potential energy field based on the real-time updated temperature field, humidity field, gas concentration field, and dust concentration field. The risk potential energy field analysis module is used to perform real-time analysis of the risk potential energy field and implement prevention and control strategies based on the analysis results.

2. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 1, characterized in that, When the system starts up, it initializes the temperature, humidity, gas concentration, dust concentration, and risk potential energy value of each grid cell, specifically including: Collect sensor data and calculate reliability weights; The collected sensor data is uniformly mapped into a dimensionless normalized universal potential vector. Define a uniform propagation tensor for the center point of each grid cell; The unified equivalent transmission distance between the sensor and each grid cell is calculated based on the unified transmission tensor field. The temperature, humidity, gas concentration, and dust concentration of each grid cell are calculated by combining sensor reliability weights, normalized universal potential vectors, and unified equivalent conduction distance. The initial risk potential value of each grid cell is calculated based on the hazard level of the stored goods, storage time, and historical risk records.

3. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 2, characterized in that, The unified conduction tensor is composed of three superpositions: the basic isotropic conduction term, the structure-induced anisotropic term, and the cross-physics field coupling driving term.

4. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 3, characterized in that, The calculation process for cross-physics coupling driving terms is as follows: For each physics field, calculate its universal potential gradient vector at the center of the grid cell; set a cohesive coupling coefficient for each physics field; calculate the coupling driving sub-terms of the physics field based on the universal potential gradient vector and the cohesive coupling coefficient; perform matrix addition on the coupling driving sub-terms of each physics field to obtain the cross-physics coupling driving terms of the grid cell.

5. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 2, characterized in that, The temperature, humidity, gas concentration, and dust concentration of each grid cell are calculated by combining sensor reliability weights, a unified equivalent conduction distance, and a normalized universal potential vector. Specifically, this includes: The interpolation weight of each sensor to the target grid cell is calculated based on the sensor reliability weight and the unified equivalent transmission distance. The general potential vector of each grid cell is calculated based on the interpolation weights and the sensor normalized general potential vector; The universal potential vector of each grid cell is mapped back to the real physical quantity.

6. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 1, characterized in that, Using discrete observation data acquired in real time by a multi-source heterogeneous sensor network, four continuous physical fields—temperature field, humidity field, gas concentration field, and dust concentration field—are updated synchronously, specifically including: The discrete observation data acquired in real time is mapped to the observation universal potential field; Map the physical field of the previous cycle to the background universal potential field; The uniform equivalent transmission distance between the sensor and each grid cell at the current moment is calculated based on the background general potential field. The general potential field is constructed and analyzed by combining the observed general potential field and the unified equivalent transmission distance between the sensor and each grid cell at the current moment. The analysis of the universal potential field is reversed and mapped back to the real physical quantities to generate an updated continuous physical field.

7. The real-time monitoring and intelligent control system for warehouse environment risks according to claim 1, characterized in that, The risk potential energy field is continuously updated based on real-time updates of the temperature field, humidity field, gas concentration field, and dust concentration field, specifically including: Based on the analysis of the current cycle, the general potential field is used to calculate the environmental risk source terms for each grid cell; A partial differential equation for the evolution of the risk potential field is established based on the environmental risk source terms of each grid cell; The risk potential field of the previous cycle is updated based on the established partial differential equation.

8. A method for real-time monitoring and intelligent control of risks in the warehousing environment, characterized in that, include: S10: Divide the entire storage area into multiple continuous three-dimensional grid cells, and initialize the temperature, humidity, gas concentration, dust concentration and risk potential energy value of each grid cell when the system starts up, to obtain five mutually coupled continuous physical fields. S20: Utilizes discrete observation data collected in real time by a multi-source heterogeneous sensor network to synchronously update four continuous physical fields: temperature field, humidity field, gas concentration field, and dust concentration field. S30: The risk potential energy field is continuously updated based on the real-time updated temperature field, humidity field, gas concentration field, and dust concentration field; S40: Perform real-time analysis of the risk potential field and implement prevention and control strategies based on the analysis results.

Images