A warehouse environment humidity closed-loop control system, method and medium

Abstract

The application relates to a warehouse environment humidity closed-loop control system, method and medium, and relates to the technical field of warehouse management.The warehouse environment humidity closed-loop control system comprises a perception collection layer, which is used for setting differentiated monitoring points in the global space of a warehouse, collecting and cleaning warehouse temperature and humidity and material stacking point clouds; an edge calculation layer, which is used for spatio-temporal fusion of the warehouse temperature and humidity and the material stacking point clouds, generation of stacking environment data groups, division of humidity control priorities in combination with material moisture absorption parameters; and a linkage execution layer, which is used for linkage of warehouse dehumidification equipment according to the humidity control priorities and humidity deviation amounts, generation of differentiated control instructions; the differentiated strategy of "intensive + sparse" is used for monitoring moisture-prone areas in the warehouse; meanwhile, non-contact three-dimensional scanning of material stacking is performed to realize real-time acquisition of spatial structure data such as stacking height, gap and density, to generate a dynamic thermal diagram and perform partitioned differentiated regulation and control, improve the regulation and control accuracy of the environment temperature and humidity, and effectively reduce the closed-loop regulation and control energy consumption.

Description

Technical Field

[0001] This application relates to the field of warehouse management technology, and in particular to a closed-loop control system, method and medium for warehouse environment humidity. Background Technology

[0002] With the rapid upgrading of modern logistics and intelligent manufacturing industries, automated warehouses (AS / RS), as core infrastructure for high-density storage and efficient turnover of goods, are seeing their application scale continuously expand in e-commerce retail, auto parts, and pharmaceutical cold chain industries. Their operational efficiency, storage security, and operational stability have become key factors affecting the efficiency of supply chain circulation. Especially in scenarios involving centralized management of multiple categories of goods and high-volume storage and retrieval, AS / RS need to support higher storage density, meet the stringent requirements of pharmaceuticals and food for temperature, humidity, and harmful gases, and cope with the reliability demands of high-load operation of stacker cranes, conveyors, and elevators. Traditional manual inspection and single-point monitoring models are no longer suitable for the real-time and precise control of intelligent AS / RS, making the construction of multi-dimensional intelligent monitoring systems an inevitable trend in the industry.

[0003] Bagged and ton-sized materials are generally highly hygroscopic. During storage, they are easily affected by factors such as ambient humidity, stacking and storage patterns, differences in regional ventilation, and dead corners in the warehouse. These factors can easily lead to problems such as softening due to moisture, clumping, mold growth, and deterioration, which directly affect the quality of material storage and its performance.

[0004] Existing patents disclose an IoT-driven multi-dimensional warehouse environment monitoring platform, comprising a multi-dimensional environmental data acquisition unit, a wireless data transmission unit, a multi-dimensional heterogeneous data fusion and processing unit, an environmental data prediction unit, an intelligent decision-making and control command generation unit, and an actuator driving unit. It collects multi-dimensional parameters of the warehouse environment, such as temperature, humidity, and gas concentration, in real time through a distributed sensor array. After adaptive anti-interference wireless transmission, it uses a multi-dimensional heterogeneous environmental parameter analysis model to complete data fusion and feature extraction. Based on an optimized long short-time memory network, it predicts environmental data and generates control commands by combining preset rules to drive the actuator to adjust the warehouse environment. This platform achieves comprehensive monitoring, accurate prediction, and intelligent control of multi-dimensional parameters of the warehouse environment, effectively improving warehouse management efficiency and cargo storage security.

[0005] The existing technical solutions mentioned above have the following drawbacks: 1. Traditional warehousing often adopts a unified dehumidification and overall constant temperature control mode, which relies on a single temperature and humidity sensor to collect environmental data. The risk assessment of moisture is limited in scope and the accuracy of prediction is low. It often produces local high humidity dead zones that are difficult to manage effectively. 2. The structural ventilation characteristics, such as the stacking height, interlayer gap, and stacking density of materials, were never taken into account, and the differences in moisture content in different areas were not distinguished. This easily leads to a large number of equipment running idle for a long time, resulting in high energy consumption and serious waste of resources. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this application is to provide a closed-loop control system, method, and medium for warehouse environment humidity. This system monitors moisture-prone areas in the warehouse using a differentiated "dense + sparse" strategy. Simultaneously, it employs non-contact 3D scanning of material stacks to acquire spatial structure data such as stack height, gaps, and density in real time. This data is used to generate dynamic heat maps and execute differentiated zone control, thereby improving the accuracy of environmental temperature and humidity control and effectively reducing the energy consumption of closed-loop control.

[0007] This was achieved using the following technical solutions: In a first aspect, this application provides a closed-loop control system for warehouse environment humidity, comprising: The sensing and acquisition layer is used to set up differentiated monitoring points throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; The edge computing layer is used to spatiotemporally fuse warehouse temperature and humidity data with material stacking point clouds, generate stacking environment data sets, and combine them with material moisture absorption parameters to prioritize humidity control. The linkage execution layer is used to link the warehouse dehumidification equipment and generate differentiated control commands based on the humidity control priority and humidity deviation.

[0008] By adopting the above technical solution, based on differentiated monitoring point deployment and spatiotemporal fusion algorithm, the warehouse temperature and humidity and stacking point cloud data are combined with material moisture absorption parameters to dynamically divide the humidity control priority, and the deviation-driven linkage dehumidification equipment generates differentiated instructions, which significantly improves the accuracy of warehouse environmental control and energy utilization efficiency.

[0009] Furthermore, the sensing and acquisition layer includes: The stacking sensing module is used to detect the combined stacking of materials and collect the initial stacking shape point cloud. The morphology layering module is used to decouple the initial stacking morphology point cloud according to the material stacking structure, divide the stacking material distribution area, and determine the risk level of the stacking area. The monitoring deployment module is used to match material monitoring strategies according to the risk level of the stockpiling area and set up differentiated monitoring points throughout the warehouse space. The environmental monitoring module is used to collect data on the stacked fabric area based on differentiated monitoring points to obtain the initial stacked temperature and humidity. The data cleaning module is used to perform time-series alignment and filtering and noise reduction on the initial stacking temperature and humidity based on the collection timestamp to obtain the warehouse temperature and humidity; at the same time, it performs filtering and noise reduction, outlier removal and voxel downsampling on the initial stacking morphology point cloud to obtain the material stacking point cloud.

[0010] By adopting the above technical solution, the stacking and distribution area is decoupled and the risk level is determined through the decoupling of the material stacking point cloud. Based on this, a differentiated monitoring strategy is matched to set up full-domain monitoring points. Then, high-precision warehouse temperature, humidity and point cloud data are obtained through time-series alignment and filtering and cleaning, which significantly improves the pertinence and reliability of environmental monitoring.

[0011] Furthermore, the morphological layering module includes: The hierarchical segmentation unit is used to divide the initial stacking shape point cloud into layers according to the material stacking structure, obtain the stacking layer point cloud, and determine the center coordinates of the point cloud. The clustering separation unit is used to calculate the distance between stacked point clouds based on the center coordinates of the point cloud, obtain the interlayer gap distance, and combine it with the inter-stack distance threshold to segment independent stacked point cloud clusters; Boundary generation unit is used to perform region growth on independent stacking point cloud clusters to construct stacked material distribution areas; The morphology extraction unit is used to perform feature analysis on independent stacked point cloud clusters and extract stacked structure feature parameters; the stacked structure feature parameters include the absolute height of the stack, the thickness of the gap between stacked layers, the local surface curvature, and the material stacking density. The adjacency association unit is used to map and fuse the stacking fabric area with the stacking structure feature parameters to generate a stacking space adjacency topology graph. The spatial determination unit is used to divide and quantify the stacking space adjacency topology based on the warehouse structural attributes, and to construct a spatial coupling feature vector. The weighted scoring unit is used to inspect the stacked fabric area based on the spatial coupling feature vector, calculate the interlayer gap rate, average humidity and humidity rise rate, and weight and fuse them into a comprehensive stacking score. The risk grading unit is used to match the overall score of the stacking area according to the risk threshold range of the stacking area and mark the risk level of the stacking and placement area; If the overall stacking score is within the first risk threshold range, the current stacking and placement area is determined to be high-risk. If the overall stacking score is within the second risk threshold range, the current stacking and placement area is determined to be of medium risk. If the overall stacking score is within the third risk threshold range, the current stacking area is determined to be low risk.

[0012] By adopting the above technical solution, based on the hierarchical clustering and morphological feature extraction of stacked point clouds, spatial adjacency topology is generated and stacking structure parameters are quantified. Combined with humidity and rise rate weighted fusion into a comprehensive score, multi-level classification judgment of stacking risk is realized, which significantly improves the refined management and moisture prevention early warning capabilities of warehouse environment.

[0013] Furthermore, the edge computing layer includes: The multi-source fusion module is used to perform timestamp calibration and spatial coordinate matching on warehouse temperature and humidity and material stacking point clouds to generate stacking environment data sets; The material scenario module is used to correlate the critical relative humidity, moisture absorption rate and agglomeration critical parameters of different types of materials, generate a set of material moisture absorption parameters, and construct a raw material mixed storage scenario. The risk assessment module is used to construct a dynamic moisture risk weight function based on the stacking environment data set and the material's hygroscopicity level, and generate a humidity heat map of the entire storage area. The humidity control sorting module is used to classify and sort the stacked fabric areas according to the mixed storage scenario of raw materials based on the humidity heat map of the entire storage area, and mark the humidity control priority.

[0014] By adopting the above technical solution, and by fusing temperature, humidity and point cloud data through timestamp calibration and spatial matching, and by constructing a dynamic risk weight function by combining the material moisture absorption rate and agglomeration critical parameters, a global humidity heat map is generated and humidity control priorities are divided. This achieves accurate assessment and differential decoupling of storage environment risks, and significantly improves the initiative of moisture prevention early warning and the scientific nature of humidity control resource allocation.

[0015] Furthermore, the material scenario module includes: The parameter input unit is used to correlate the critical relative humidity, moisture absorption rate and agglomeration critical parameters according to the material type to obtain the material moisture absorption parameter set; The coupling correlation unit is used to fit the material moisture absorption parameter set according to the temperature sequence, determine the effective critical humidity threshold, calculate the risk accumulation index, and construct the agglomeration risk accumulation model. The humidity grading unit is used to classify materials for compatibility based on the effective critical humidity threshold and mark the humidity compatibility level of the materials. The storage constraint unit is used to perform cross-checks between materials based on the material humidity compatibility level and the material volatility, and generate a stacking isolation rule library. The scenario evolution unit is used to simulate and combine material stacking based on the clumping risk accumulation model and the stacking isolation rule library to construct a raw material mixed storage scenario.

[0016] By adopting the above technical solutions, a clumping risk accumulation model is constructed based on the coupling of material moisture absorption parameters and temperature. Based on critical humidity, compatibility is classified and a stacking isolation rule library is generated. By simulating and combining raw material mixed storage scenarios, scientific isolation and risk pre-control of material stacking are achieved, significantly improving the accuracy of storage environment control and material storage safety.

[0017] Furthermore, the risk assessment module includes: Spatial partitioning unit is used to divide the entire warehouse space into grids, assign spatial index numbers, and record the grid center coordinates and warehouse functional areas. The data binding unit is used to map the stacking environment data group to the storage functional area based on the grid center coordinates, and to construct a complete functional area; An independent quantization unit is used to transform the stacking environment data set, calculate the humidity excess factor f1, ventilation gap factor f2, and stack level airtightness factor f3; and map the material hygroscopicity level to obtain the hygroscopicity level factor f4. The function building unit is used to perform nonlinear weighting operations on each factor to generate the dynamic moisture risk weight function Rcell: Rcell=w1*f1+w2*f2+w3*f3+w4*f4+w12*f1*f2+w14*f1*f4, Among them, w1, w2, w3 and w4 are the main effect weights of the corresponding factors, w12*f1*f2 is the humidity-ventilation coupling term, and w14*f1*f4 is the humidity-material coupling term; The risk calculation unit is used to perform comprehensive calculations on the complete functional area based on the dynamic moisture risk weighting function to obtain the regional moisture risk coefficient. The pixel mapping unit is used to smooth the transition grid between complete functional areas according to the spatial index number and the regional moisture risk coefficient, forming a moisture risk field. The information overlay unit is used to mark information on the moisture risk field based on the stacking structure characteristics and generate a humidity heat map of the entire storage area.

[0018] By adopting the above technical solution, a dynamic moisture risk function is constructed based on grid division and multi-factor nonlinear weighting (including humidity-ventilation and humidity-material coupling terms). Regional risk coefficients are generated and smoothed into a continuous risk field through spatial interpolation. Combined with stacking structure annotation, a warehouse-wide humidity heat map is formed, realizing refined dynamic assessment and visualization of moisture risk, which significantly improves the accuracy and decision-making efficiency of warehouse environment management.

[0019] Furthermore, the linkage execution layer includes: The regional aggregation module is used to perform grid aggregation of the entire warehouse space according to the humidity control priority, obtain the same-level control zones with different risk levels, and calculate the humidity deviation by combining the humidity heat map of the entire warehouse. The task sorting module is used to sort the control zones of the same level in descending order according to the humidity control priority to obtain the preliminary humidity control execution sequence, and to make internal corrections according to the humidity deviation to obtain the final humidity control execution sequence. The equipment detection module is used to query dehumidification equipment in the same control zone at different levels, establish an equipment space coverage table, and calculate the overlap of dehumidification spaces. The resource allocation module is used to allocate dehumidification equipment according to the overlap of dehumidification spaces based on the equipment space coverage table. If a single dehumidifier cannot provide complete coverage, multiple dehumidifiers will be deployed to work together, and corresponding grid responsibility areas will be defined. If different priority areas compete for the same dehumidifier, it will be assigned to the grid responsibility area with the larger humidity deviation or the greater risk of dampness in the area. The other grid responsibility area will be assigned the second-best coverage device or wait in line for the current device to be released. The instruction generation module is used to pre-regulate the operation mode of the dehumidification equipment based on the risk level of the stockpile area and the final humidity control execution sequence, generate equipment operating parameters, and convert them into differentiated control instructions.

[0020] By adopting the above technical solution, based on grid aggregation and priority sorting algorithms, the execution sequence is corrected by humidity deviation, and dehumidification equipment is allocated using spatial coverage tables and overlap calculations. This enables multi-device collaboration and conflict resolution, ultimately generating differentiated control commands, which significantly improves the accuracy and energy efficiency of warehouse environmental control.

[0021] Secondly, this application also provides a closed-loop control method for humidity in a storage environment, which adopts the following technical solution; A closed-loop control method for humidity in a warehouse environment includes: Differentiated monitoring points are set up throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; The warehouse temperature and humidity data and material stacking point cloud data are spatiotemporally fused to generate stacking environment data sets, and humidity control priorities are determined by combining material moisture absorption parameters. Based on the humidity control priority and humidity deviation, the warehouse dehumidification equipment is linked to generate differentiated control commands.

[0022] By adopting the above technical solution, based on the spatiotemporal fusion of multi-source sensors and modeling of material moisture absorption parameters, the priority of stacking humidity control is dynamically divided, and combined with the humidity deviation to drive the dehumidification equipment linkage, generating differentiated control commands, thus realizing precise energy-saving control of the warehouse environment.

[0023] Thirdly, this application also provides a storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the closed-loop control method for warehouse environment humidity as described above.

[0024] In summary, the beneficial technical effects of this application are as follows: A low-power LoRa temperature and humidity sensor network is deployed using a differentiated strategy of "dense + sparse" to focus on monitoring areas prone to moisture.

[0025] By using non-contact 3D scanning of stacks, spatial structure data such as stack height, gaps, and density can be acquired in real time, generating dynamic heat maps and implementing zoned differentiated control, which improves the accuracy of environmental temperature and humidity control and effectively reduces control energy consumption. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the closed-loop humidity control system for the storage environment in this application; Figure 2 This is a schematic diagram of the morphological layering module in this application; Figure 3 This is a schematic diagram of the risk assessment module in this application; Figure 4 This is a flowchart illustrating the closed-loop humidity control method for the storage environment in this application. Detailed Implementation

[0027] The present application will be further described in detail below with reference to the accompanying drawings.

[0028] Reference Figure 1 This application discloses a closed-loop control system for warehouse environment humidity, comprising: The sensing and acquisition layer is used to set up differentiated monitoring points throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; The edge computing layer is used to spatiotemporally fuse warehouse temperature and humidity data with material stacking point clouds, generate stacking environment data sets, and combine them with material moisture absorption parameters to prioritize humidity control. The linkage execution layer is used to link the warehouse dehumidification equipment and generate differentiated control commands based on the humidity control priority and humidity deviation.

[0029] In this embodiment, the sensing and acquisition layer includes a temperature and humidity sensing module and a stacking shape sensing module, which are used to carry out differentiated point deployment throughout the warehouse area to collect environmental temperature and humidity data and stacking three-dimensional shape data.

[0030] The temperature and humidity sensing module uses a low-power LoRa wireless temperature and humidity sensor array. It is densely deployed at 1-2 meter intervals in high-risk areas prone to moisture, such as the bottom stacks of materials, areas near walls, and enclosed corners. In conventional areas such as warehouse aisles and high-rise ventilation areas, it is sparsely deployed at 5-10 meter intervals, thus forming a differentiated gridded temperature and humidity sensing system to monitor temperature and humidity changes in different zones.

[0031] The stacking morphology perception module deploys a combination of ToF depth cameras and millimeter-wave radar at key points on warehouse beams, top supports, and aisles to form a complementary detection mechanism. This enables non-contact acquisition of stacking morphology and identification of information such as stacking height, interlayer gaps, stacking density, surface curvature, and stacking spacing. It generates 3D point cloud data of the stacking to capture dynamic changes such as material entry and exit, stacking displacement, disassembly and reassembly in real time, and continuously outputs warehouse space structure data.

[0032] The edge computing layer receives data from the sensing and acquisition layer, performs multi-source data fusion and dynamic risk assessment locally, and generates dynamic risk heatmaps and humidity control priorities. It includes: The multi-source data fusion module internally deploys a lightweight spatiotemporal fusion model to perform timestamp calibration and spatial coordinate matching on time-series environmental data from LoRa temperature and humidity sensors and three-dimensional spatial stacking data from ToF cameras / radars, thereby completing the feature alignment and fusion of heterogeneous data and eliminating data errors caused by acquisition delays and positional deviations.

[0033] The material parameter import module includes a pre-configured database of chemical material moisture absorption parameters. It allows users to input prior parameters such as critical relative humidity, moisture absorption rate, and clumping critical conditions for different types of chemical raw materials. The module also supports on-site customization of adding and modifying material parameters, adapting to mixed storage scenarios involving multiple types of raw materials. Furthermore, the system has a built-in automatic learning function for material moisture absorption characteristics, which can automatically optimize parameters by analyzing historical data.

[0034] The dynamic risk assessment module combines real-time environmental humidity data, stacking ventilation gaps, stacking level airtightness, and the material's own hygroscopicity level to construct a dynamic moisture risk weight function, quantifying the moisture risk coefficient of each grid area. Based on the real-time updated risk weight data, it automatically generates a heat map of humidity-sensitive areas across the entire warehouse, and through hierarchical visual labeling, divides the area into severely moisture-risk zones, moderately risk zones, and generally safe zones.

[0035] The priority determination module uses the heat map as the core basis for judgment, automatically sorts the humidity control priority of each area, and locks in vulnerable areas such as enclosed dead corners, low-level stacking, and poorly ventilated areas as key control targets.

[0036] The linkage execution layer is used to coordinate with the existing dehumidification equipment in the warehouse and execute differentiated control commands based on the humidity control priority and real-time humidity deviation output by the edge computing layer.

[0037] For high-risk, moisture-prone areas, increase the dehumidification load and enhance air convection to quickly reduce local humidity and keep it within the critical storage standard for materials. For medium-risk areas, low-power intermittent regulation is used to maintain environmental stability; For secure areas, equipment is put into sleep or standby mode to minimize energy consumption in the warehouse.

[0038] By replacing synchronous operation across the entire area with localized targeted regulation, precise closed-loop humidity control is achieved, effectively reducing energy consumption for warehouse operation and maintenance while ensuring anti-caking effects.

[0039] Preferably, the sensing and acquisition layer includes: The stacking sensing module is used to detect the combined stacking of materials and collect the initial stacking shape point cloud. Morphological layering module (refer to) Figure 2 This is used to decouple the initial stacking morphology point cloud based on the material stacking structure, divide the stacking and placement areas, and determine the risk level of the stacking areas, including: The hierarchical segmentation unit is used to divide the initial stacking shape point cloud into layers according to the material stacking structure, obtain the stacking layer point cloud, and determine the center coordinates of the point cloud. The clustering separation unit is used to calculate the distance between stacked point clouds based on the center coordinates of the point cloud, obtain the interlayer gap distance, and combine it with the inter-stack distance threshold to segment independent stacked point cloud clusters; Boundary generation unit is used to perform region growth on independent stacking point cloud clusters to construct stacked material distribution areas; The morphology extraction unit is used to perform feature analysis on independent stacked point cloud clusters and extract stacked structure feature parameters; the stacked structure feature parameters include the absolute height of the stack, the thickness of the gap between stacked layers, the local surface curvature, and the material stacking density. The adjacency association unit is used to map and fuse the stacking fabric area with the stacking structure feature parameters to generate a stacking space adjacency topology graph. The spatial determination unit is used to divide and quantify the stacking space adjacency topology based on the warehouse structural attributes, and to construct a spatial coupling feature vector. The weighted scoring unit is used to inspect the stacked fabric area based on the spatial coupling feature vector, calculate the interlayer gap rate, average humidity and humidity rise rate, and weight and fuse them into a comprehensive stacking score. The risk grading unit is used to match the overall score of the stacking area according to the risk threshold range of the stacking area and mark the risk level of the stacking and placement area; If the overall stacking score is within the first risk threshold range, the current stacking and placement area is determined to be high-risk. If the overall stacking score is within the second risk threshold range, the current stacking and placement area is determined to be of medium risk. If the overall stacking score is within the third risk threshold range, the current stacking area is determined to be low risk.

[0040] The monitoring deployment module is used to match material monitoring strategies according to the risk level of the stockpiling area and set up differentiated monitoring points throughout the warehouse space. The environmental monitoring module is used to collect data on the stacked fabric area based on differentiated monitoring points to obtain the initial stacked temperature and humidity. The data cleaning module is used to perform time-series alignment and filtering and noise reduction on the initial stacking temperature and humidity based on the collection timestamp to obtain the warehouse temperature and humidity; at the same time, it performs filtering and noise reduction, outlier removal and voxel downsampling on the initial stacking morphology point cloud to obtain the material stacking point cloud.

[0041] In this embodiment, the initial 3D point cloud generated by the ToF depth camera and the millimeter-wave radar is fused and denoised to construct a complete digital substrate for the warehouse space.

[0042] The high-precision surface point cloud provided by the ToF camera and the penetrating depth information provided by the millimeter-wave radar are synchronized in time and registered in space. An iterative nearest-point algorithm is used to unify the two types of point clouds into the global coordinate system of the warehouse, forming a complementary dense point cloud.

[0043] Based on the static CAD model of the warehouse or historical empty warehouse point cloud data, the fixed structures such as walls, floors, columns, and beams are removed by background subtraction, and only the foreground point cloud related to material stacking is retained.

[0044] A statistical outlier removal algorithm is applied to filter out isolated scattered points caused by sensor noise or dust, ensuring the accuracy of subsequent segmentation.

[0045] From a continuous foreground cloud, different stacking batches or independent stacking individuals are divided into unconnected logical units.

[0046] Clustering analysis is performed based on the 3D Euclidean distance between points in the point cloud. A clustering distance threshold is set, which is slightly larger than the gap between stack layers but smaller than the width of the passage between stacks, to automatically segment spatially separated stacked individuals into independent point cloud clusters.

[0047] For each clustered independent stack, a directed bounding box is calculated, recording its three-dimensional dimensions (length, width, and height), the coordinates of its base center, and the direction of its principal axes. This bounding box serves as the initial spatial boundary of the stacking material area.

[0048] Continuous ground areas not covered by any enclosures and with sufficient width for equipment passage are automatically identified as warehouse aisles. Narrow gaps adjacent to walls are identified as wall-to-wall gap areas.

[0049] Deep analysis was performed on each independent stacking point cloud cluster to extract structural features that affect the distribution of temperature and humidity.

[0050] Using the ground plane as a reference, calculate the vertical distance between the highest point in the stacking point cloud and the ground, and use this distance as the absolute height of the stack.

[0051] The point cloud is analyzed by layering and slicing along the stacking height. The location of the point density attenuation valley between adjacent slice layers is detected, and the thickness and continuity of the gaps between stacking layers are identified.

[0052] Surface fitting is performed on the point clouds of the top and sides of the stack to calculate the local surface curvature and identify points at risk of water accumulation in depressions. Simultaneously, the point cloud volume occupancy rate within the stack's bounding box is calculated as a quantitative indicator of stacking density.

[0053] Record the shortest distance between each stack and the wall, other stacks, and aisles to generate a stack space adjacency topology graph.

[0054] Based on the sensor deployment coordinates, each LoRa temperature and humidity node is mapped to its nearest spatial entity. If a sensor is located within or adjacent to a stack enclosure, it is assigned to that stack; if it is located in a passageway or a gap against a wall, it is marked as the corresponding area type.

[0055] Extract the current readings and recent historical trends of all temperature and humidity sensors associated with each stack, and calculate the average humidity, humidity gradient, and fluctuation range of the stack area.

[0056] A coupled feature vector is generated for each stack, including: stack height, stack density, distance from the wall, interlayer gap ratio, surface curvature, average humidity of the area, and rate of increase in humidity. These features directly reflect the moisture risk at that stack location.

[0057] Based on the coupled feature vector, each stacked fabric area is graded according to a preset risk classification standard: High-risk areas: simultaneously meet the following conditions: densely deployed areas, ventilation dead zones (<0.5 meters from the wall and >8 meters from the passageway), and the average humidity of the area is continuously higher than the warning value (e.g., >65%RH) or the rate of humidity increase is abnormal.

[0058] Medium-risk areas: those that meet the criteria of dense stacking but poor ventilation, or those located near walls but with normal humidity, or those sparsely populated areas but with locally high humidity.

[0059] Low-risk area: located on both sides of well-ventilated passages, with low stacking height, sparse stacking density, and the area humidity always within a safe range.

[0060] A weighted scoring model is used to score the coupling characteristics of each stack, resulting in a region score: If the score is greater than or equal to 0.7, it is designated as a high-risk stockpiling area and requires immediate attention.

[0061] If 0.3 ≤ Score < 0.7, it is designated as a medium-risk stockpile area, and monitoring needs to be strengthened.

[0062] If the score is less than 0.3, it is designated as a low-risk stockpile area, and routine inspections are maintained.

[0063] Preferably, the edge computing layer includes: The multi-source fusion module is used to perform timestamp calibration and spatial coordinate matching on warehouse temperature and humidity and material stacking point clouds to generate stacking environment data sets; The material scenario module is used to correlate the critical relative humidity, moisture absorption rate, and clumping critical parameters of different types of materials, generate a set of material moisture absorption parameters, and construct a raw material mixed storage scenario, including: The parameter input unit is used to correlate the critical relative humidity, moisture absorption rate and agglomeration critical parameters according to the material type to obtain the material moisture absorption parameter set; The coupling correlation unit is used to fit the material moisture absorption parameter set according to the temperature sequence, determine the effective critical humidity threshold, calculate the risk accumulation index, and construct the agglomeration risk accumulation model. The humidity grading unit is used to classify materials for compatibility based on the effective critical humidity threshold and mark the humidity compatibility level of the materials. The storage constraint unit is used to perform cross-checks between materials based on the material humidity compatibility level and the material volatility, and generate a stacking isolation rule library. The scenario evolution unit is used to simulate and combine material stacking based on the clumping risk accumulation model and the stacking isolation rule library to construct a raw material mixed storage scenario. Risk assessment module (refer to) Figure 3 This is used to construct a dynamic moisture risk weighting function based on the stacking environment data set and the material's hygroscopicity level, and generate a humidity heat map of the entire warehouse area, including: Spatial partitioning unit is used to divide the entire warehouse space into grids, assign spatial index numbers, and record the grid center coordinates and warehouse functional areas. The data binding unit is used to map the stacking environment data group to the storage functional area based on the grid center coordinates, and to construct a complete functional area; An independent quantization unit is used to transform the stacking environment data set, calculate the humidity excess factor f1, ventilation gap factor f2, and stack level airtightness factor f3; and map the material hygroscopicity level to obtain the hygroscopicity level factor f4. The function building unit is used to perform nonlinear weighting operations on each factor to generate the dynamic moisture risk weight function Rcell: Rcell=w1*f1+w2*f2+w3*f3+w4*f4+w12*f1*f2+w14*f1*f4, Among them, w1, w2, w3 and w4 are the main effect weights of the corresponding factors, w12*f1*f2 is the humidity-ventilation coupling term, and w14*f1*f4 is the humidity-material coupling term; The risk calculation unit is used to perform comprehensive calculations on the complete functional area based on the dynamic moisture risk weighting function to obtain the regional moisture risk coefficient. The pixel mapping unit is used to smooth the transition grid between complete functional areas according to the spatial index number and the regional moisture risk coefficient, forming a moisture risk field. The information overlay unit is used to mark information on the moisture risk field according to the stacking structure characteristics and generate a humidity heat map of the entire warehouse area. The humidity control sorting module is used to classify and sort the stacked fabric areas according to the mixed storage scenario of raw materials based on the humidity heat map of the entire storage area, and mark the humidity control priority.

[0064] In this embodiment, the core parameter set required to describe the moisture absorption behavior of a material category is defined as the smallest unit, forming the basic fields of the parameter library.

[0065] Set a relative humidity threshold for each material at which it begins to rapidly absorb moisture at a specific temperature. This value is typically derived from laboratory measurements or supplier technical data and is expressed as a percentage. The critical relative humidity for the same material at different temperatures can be stored as a curve or a segmented table.

[0066] Define a characteristic parameter for how the moisture increment of a material changes over time after exceeding the critical relative humidity. The initial moisture absorption rate and the rate of approaching saturation can be described using a linear rate constant or a time-based rate curve.

[0067] Set multidimensional critical thresholds that can trigger material agglomeration, including: upper limit of moisture content, duration of continuous high humidity exposure, and compressive strength threshold. This parameter is used to predict whether materials have entered a state of risk of agglomeration during storage.

[0068] Supplementing auxiliary parameters related to moisture absorption behavior, such as bulk density, specific surface area, and packaging sealing level, provides a basis for subsequent correlation analysis.

[0069] Based on the critical relative humidity values ​​measured at different temperature points, the CRH curve of the material as a function of temperature is fitted using a polynomial regression or exponential model, enabling the system to dynamically determine the current effective critical humidity threshold at any ambient temperature.

[0070] Establish a functional relationship between the moisture absorption rate and the difference between the ambient relative humidity and the CRH (Closing Humidity Level). The larger the humidity difference, the higher the moisture absorption rate; the two are usually positively correlated non-linearly. Store this relationship in a parameter library in the form of a piecewise linear table or a power function model.

[0071] The clumping critical condition is transformed into a quantifiable risk accumulation index. This index integrates factors such as real-time moisture content, duration exceeding the critical resistance level (CRH), and stacking compressibility, calculating the cumulative risk value in the form of a time integral. When the cumulative value exceeds a preset threshold, the system outputs a clumping warning.

[0072] Based on the critical relative humidity (CRH) levels of each material, they are classified into three humidity compatibility categories: high sensitivity, medium sensitivity, and low sensitivity. Materials with similar CRH values ​​are considered humidity compatible and can be stored together; materials with significantly different CRH values ​​are marked as incompatible and should be stored in separate areas.

[0073] For materials with high volatile organic compound (VOC) content or those prone to generating dust, the risk of cross-contamination with materials that have strong adsorption properties is marked. This mark is stored in a parameter library in the form of a prohibition list and is used to automatically alert for conflicts during mixed storage planning.

[0074] Based on the above rules, a stacking suggestion code is generated for each pair of materials: mixed stacking is allowed, spaced stacking is required, zoned stacking is required, and joint storage is strictly prohibited. This rule base is the core basis for mixed storage space planning.

[0075] Design a relational data table structure, with the main table storing basic material parameters and the related tables storing temperature dependence curves, moisture absorption rate model coefficients, agglomeration accumulation model parameters, and material compatibility matrices.

[0076] Import known categories of chemical raw materials in batches according to the above structure to serve as standard material templates. Assign a unique code to each material, aligning it with the material master data in the warehouse management system.

[0077] The user interface provides functions for adding, modifying, and deleting parameters. Operators can select the material category via drop-down menus, directly modify threshold parameters, or automatically update correlation curves by importing experimental data files. Modification records are automatically logged, and version rollback is supported.

[0078] Temperature and humidity time-series data, material inspection records for the corresponding time periods, clumping event reports, and quality sampling results are extracted from historical databases to form a tagged sample set. The tags represent the actual state of the materials, characterized by ambient temperature and humidity and exposure duration.

[0079] By analyzing cases where humidity repeatedly approached the critical level without any abnormal moisture absorption, and cases where humidity repeatedly fell below the original CRH level but slight moisture absorption occurred, the CRH threshold was fine-tuned towards a more precise direction using Bayesian inference or confidence interval adjustment methods.

[0080] Based on feedback from actual clumping events, the weight coefficients and trigger thresholds in the clumping risk accumulation model are adjusted using online gradient descent or incremental learning methods to make the prediction results more closely match the actual stacking and packaging conditions on site. After self-learning is complete, the optimized parameters are stored as a new version in the parameter library. After review and confirmation by process personnel, they are synchronized to the monitoring and control module of the production environment. Historical versions are retained for future reference, and parameter rollback is supported.

[0081] Based on the current temperature and humidity readings of each area, the monitoring system automatically indexes the parameter database records of the materials stored in that area, calculates the cumulative risk index of caking for each area, and generates a dynamic risk distribution map.

[0082] When a clumping warning is triggered in a certain area, the system automatically checks the list of all materials stored in that area and their moisture absorption parameters, sorts them by risk level, and outputs a priority list of affected materials to guide risk mitigation operations.

[0083] When new materials are received into the warehouse, the system automatically recommends the best storage area based on their moisture absorption parameters and compatibility rules with existing materials, avoiding the stacking of highly moisture-sensitive materials with materials that emit moisture, thus reducing the risk of mixed storage from the source.

[0084] Based on the warehouse floor level, the entire area is divided into square grids with fixed side lengths (e.g., 1 meter × 1 meter). Each grid is assigned a unique spatial index number, recording its center coordinates and the functional area it belongs to (stacking area, aisle, gap against the wall).

[0085] Based on the spatial distance between the grid coordinates and the locations of the temperature and humidity sensors, an inverse distance weighted interpolation method is used to calculate the real-time ambient temperature and relative humidity values ​​for each grid. The grid humidity RHcell is obtained by weighted averaging of readings from its neighboring sensors.

[0086] The 3D point cloud and structural parameters generated by the stacking morphology perception module are mapped to a mesh. For each mesh, the stacking height H, distance from the wall Dwall, interlayer gap ratio G, and stacking density ρ of the area it covers are extracted. If the mesh is located in a channel, the structural parameters are set to zero or the default value.

[0087] Based on the material storage location recorded by the warehouse management system, the stacking area grid is associated with the specific material category, and the critical relative humidity (CRH) and moisture absorption rate level (Kmat) of the material are read from the moisture absorption parameter library. Calculate the extent to which the current humidity of the grid exceeds the critical relative humidity of the material. If no material is stored, use the warehouse's general warning humidity RHwarn as the baseline.

[0088] ΔRH=max(0,RHcell-CRH), The difference is normalized so that f1∈[0,1]. When ΔRH=0, f1=0 (no risk); when ΔRH reaches the preset danger limit, f1=1 (extremely high risk).

[0089] Ventilation conditions are characterized by a combination of the distance from the wall (Dwall) and the inter-story gap ratio (G). Poorer ventilation results in higher risks.

[0090] f2=α*e -k1*Dwall +β*(1−G), Where α and β are weighting coefficients, and k1 is the attenuation constant. When the stack is close to the wall and there are no gaps between layers, f2→1; when the distance from the wall is sufficient and there are ample gaps, f2→0.

[0091] The greater the height and density of the stack, the weaker the air circulation inside the stack, and the easier it is for moisture to accumulate.

[0092] f3 = γ*H / Hmax + δ*ρ, where Hmax is the maximum allowable stacking height in the warehouse, and γ and δ are weighting coefficients. Both terms are normalized and then summed using a weighted average.

[0093] The hygroscopicity level of the material is directly read from the parameter library and mapped to a risk factor f4 of 0–1. Higher values ​​are assigned to materials with high hygroscopicity, and lower values ​​to materials with low hygroscopicity. This factor is an inherent property of the material and does not change with the environment.

[0094] The weights are dynamically adjusted based on the overall humidity level of the warehouse and seasonal patterns. During the rainy season, the weights of w1 and its coupling terms are appropriately increased; during the dry season, the influence weight of the f2 ventilation factor is increased to more sensitively detect local blind spots.

[0095] The Rcell calculated for each grid cell is the moisture hazard factor for that spatial unit, with a value range of [0,1]. The higher the value, the greater the probability and severity of material moisture absorption at that location under the current environmental and stacking conditions.

[0096] Using the center coordinates of each grid as the positioning point, the grid risk coefficient Rcell is used as the grayscale intensity value of that pixel region. Bilinear interpolation or Gaussian smoothing is applied to the blank areas between grids to form a continuous risk field.

[0097] The warehouse layout outline, stacking boundary boxes, sensor location icons, and real-time humidity values ​​are overlaid on the heat map. High-risk areas are automatically highlighted with a bright, flashing border, along with the material category and current key risk factors for that area.

[0098] Preferably, the linkage execution layer includes: The regional aggregation module is used to perform grid aggregation of the entire warehouse space according to the humidity control priority, obtain the same-level control zones with different risk levels, and calculate the humidity deviation by combining the humidity heat map of the entire warehouse. The task sorting module is used to sort the control zones of the same level in descending order according to the humidity control priority to obtain the preliminary humidity control execution sequence, and to make internal corrections according to the humidity deviation to obtain the final humidity control execution sequence. The equipment detection module is used to query dehumidification equipment in the same control zone at different levels, establish an equipment space coverage table, and calculate the overlap of dehumidification spaces. The resource allocation module is used to allocate dehumidification equipment according to the overlap of dehumidification spaces based on the equipment space coverage table. If a single dehumidifier cannot provide complete coverage, multiple dehumidifiers will be deployed to work together, and corresponding grid responsibility areas will be defined. If different priority areas compete for the same dehumidifier, it will be assigned to the grid responsibility area with the larger humidity deviation or the greater risk of dampness in the area. The other grid responsibility area will be assigned the second-best coverage device or wait in line for the current device to be released. The instruction generation module is used to pre-regulate the operation mode of the dehumidification equipment based on the risk level of the stockpile area and the final humidity control execution sequence, generate equipment operating parameters, and convert them into differentiated control instructions.

[0099] In this embodiment, the warehouse full-domain gridded data packet pushed by the edge computing layer is received, which includes the spatial index of each grid, the current humidity value, the target humidity value, the humidity control priority, and the real-time humidity deviation.

[0100] Adjacent grids with the same humidity control priority are aggregated into a unified control area to eliminate fragmented commands and prevent the same dehumidifier from receiving multiple conflicting commands.

[0101] Controlled humidity zones are arranged in descending order of priority. High-risk zones are at the top of the list, followed by medium-risk zones, and safe zones are at the bottom. Within the same risk level, zones are sorted by humidity deviation from largest to smallest, ensuring that the zone with the most severe deviation receives equipment resources first.

[0102] Find the installation location, rated coverage radius, adjustable airflow range, and current operating status of the warehouse dehumidification equipment. Establish an equipment-space coverage table to clarify the grid area that each device can effectively serve.

[0103] For each controlled area, the dehumidifier with the highest overlap in coverage area is selected as the main control device. If a single device cannot provide complete coverage, multiple devices are assigned to work together, and their respective responsible sub-areas are clearly defined.

[0104] If two high-priority areas compete for the same device, the area with the greater humidity deviation or higher risk level will be allocated priority. The other area will be assigned a device with secondary coverage, or will wait in a queue for the device to be released. Ensure that each device executes only one set of explicit control instructions at any given time.

[0105] Based on the risk level and humidity deviation of the area, specific equipment operating parameters are generated according to a preset strategy.

[0106] High-risk area control command generation: Operating mode setting: Dehumidifier set to continuous full-power mode. Commands include: compressor full-load operation, fan speed adjusted to the highest setting.

[0107] If the area is equipped with auxiliary circulating fans, they will be activated simultaneously and directed to blow airflow towards ventilation dead spots such as the bottom of the stack and gaps against the wall. Instructions include: fan activation and airflow direction / angle parameters.

[0108] The target humidity is set as the critical relative humidity of the stored materials in the area minus a safety margin. The system continuously calculates the real-time humidity deviation and uses an incremental PID controller to dynamically fine-tune the equipment power, ensuring that it quickly drops below the target line while avoiding overshoot.

[0109] Command data structure: {Device ID, Mode: Continuous Full Power, Target Humidity: Target Value, Fan Speed: Highest Level, Auxiliary Fan: Start}; Medium-risk area control command generation: Operation mode setting: The dehumidifier is set to intermittent operation mode. The command includes: operation cycle duration and shutdown cycle duration, and the duty cycle is dynamically adjusted according to the humidity deviation—the greater the deviation, the higher the operating percentage.

[0110] The compressor operates at a low to medium power setting, and the fan speed is set to medium. Energy consumption is minimized while maintaining environmental stability.

[0111] Set the allowable range for humidity fluctuations. When the humidity fluctuates within the range, maintain the current duty cycle; if the humidity continues to approach the critical value, automatically increase the operating duty cycle.

[0112] Instruction data structure: {Device ID, Mode: Intermittent operation, Runtime: t_on, Stoptime: t_off, Fan speed: Medium, Target humidity range: [Lower limit, Upper limit]}; Safety zone control command generation: Operating mode setting: The dehumidifier is set to sleep / standby mode. The compressor is turned off, and only the humidity monitoring function is retained.

[0113] Wake-up Condition Setting: Set a wake-up threshold for this area. When the humidity in the safe area spontaneously rises to near the material's critical standard, the equipment will automatically wake up and operate in medium-risk mode.

[0114] Command data structure: {Device ID, Mode: Sleep, Wake-up threshold: Wake-up humidity value}.

[0115] Reference Figure 4 This application discloses a closed-loop humidity control method for a storage environment, comprising: S1: Set up differentiated monitoring points throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; S2: Spatiotemporal fusion of warehouse temperature and humidity and material stacking point cloud to generate stacking environment data set, and combined with material moisture absorption parameters to classify humidity control priorities; S3: Based on the humidity control priority and humidity deviation, the warehouse dehumidification equipment is linked to generate differentiated control commands.

[0116] In this embodiment, in a large chemical raw material warehouse, for stacked ton bags and 25kg bagged hygroscopic raw materials (such as polyacrylamide and potassium chloride), the system first deploys 120 temperature and humidity sensors and 8 three-dimensional LiDARs on the roof truss, shelf panels and ground aisles according to the warehouse's three-dimensional spatial layout and material stacking height. The system collects and cleans the temperature and humidity data of each point and the point cloud outline of the material stack every 5 minutes, eliminating sensor drift and radar reflection noise.

[0117] By aligning the temperature and humidity fields at different times and locations with the stacking geometry model using a spatiotemporal fusion algorithm, a "surface-center" environmental data set for each stacking unit is generated. Then, combined with the standard moisture absorption isotherm curve for each material (such as potassium chloride with a critical relative humidity of 75%), the real-time moisture absorption risk level of each stack position is calculated, and moisture control priorities are divided into three levels: red, yellow, and green, based on the risk from high to low.

[0118] Meanwhile, the system uses the deviation between the actual humidity and the set target value (e.g., 65%RH) as input, and coordinates with the six dehumidifier units arranged in different zones within the warehouse. It issues "high-speed dehumidification, continuous operation" commands to the red zone stacks, "medium-speed intermittent operation" commands to the yellow zone, and "standby monitoring" commands to the green zone, thus forming a closed-loop control. Whenever the humidity at a certain stack drops to the target range or the point cloud detects a change in stack height causing airflow obstruction, the system dynamically adjusts the priority and regenerates control commands, ultimately stabilizing the humidity fluctuation across the entire warehouse within ±3%RH, effectively preventing bagged materials from clumping and deteriorating due to localized deliquescence.

[0119] This application discloses a storage medium storing at least one instruction, at least one program, code set, or instruction set. The at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the closed-loop control method for warehouse environment humidity as described above.

[0120] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A closed-loop control system for warehouse environment humidity, characterized in that, include: The sensing and acquisition layer is used to set up differentiated monitoring points throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; The edge computing layer is used to spatiotemporally fuse warehouse temperature and humidity data with material stacking point clouds, generate stacking environment data sets, and combine them with material moisture absorption parameters to prioritize humidity control. The linkage execution layer is used to link the warehouse dehumidification equipment and generate differentiated control commands based on the humidity control priority and humidity deviation.

2. The closed-loop control system for warehouse environment humidity according to claim 1, characterized in that, The sensing and acquisition layer includes: The stacking sensing module is used to detect the combined stacking of materials and collect the initial stacking shape point cloud. The morphology layering module is used to decouple the initial stacking morphology point cloud according to the material stacking structure, divide the stacking material distribution area, and determine the risk level of the stacking area; The monitoring deployment module is used to match material monitoring strategies according to the risk level of the stockpiling area and set up differentiated monitoring points throughout the warehouse space. The environmental monitoring module is used to collect data on the stacked fabric area based on differentiated monitoring points to obtain the initial stacked temperature and humidity. The data cleaning module is used to perform time-series alignment and filtering and noise reduction on the initial stacking temperature and humidity based on the collection timestamp to obtain the warehouse temperature and humidity; at the same time, it performs filtering and noise reduction, outlier removal and voxel downsampling on the initial stacking morphology point cloud to obtain the material stacking point cloud.

3. The closed-loop control system for warehouse environment humidity according to claim 2, characterized in that, The morphological layering module includes: The hierarchical segmentation unit is used to divide the initial stacking shape point cloud into layers according to the material stacking structure, obtain the stacking layer point cloud, and determine the center coordinates of the point cloud. The clustering separation unit is used to calculate the distance between stacked point clouds based on the center coordinates of the point cloud, obtain the interlayer gap distance, and combine it with the inter-stack distance threshold to segment independent stacked point cloud clusters; Boundary generation unit is used to perform region growth on independent stacking point cloud clusters to construct stacked material distribution areas; The morphology extraction unit is used to perform feature analysis on independent stacked point cloud clusters and extract stacked structure feature parameters; the stacked structure feature parameters include the absolute height of the stack, the thickness of the gap between stacked layers, the local surface curvature, and the material stacking density. The adjacency association unit is used to map and fuse the stacking fabric area with the stacking structure feature parameters to generate a stacking space adjacency topology graph. The spatial determination unit is used to divide and quantify the stacking space adjacency topology based on the warehouse structural attributes, and to construct a spatial coupling feature vector. The weighted scoring unit is used to inspect the stacked fabric area based on the spatial coupling feature vector, calculate the interlayer gap rate, average humidity and humidity rise rate, and weight and fuse them into a comprehensive stacking score. The risk grading unit is used to match the overall score of the stacking area according to the risk threshold range of the stacking area and mark the risk level of the stacking and placement area; If the overall stacking score is within the first risk threshold range, the current stacking and placement area is determined to be high-risk. If the overall stacking score is within the second risk threshold range, the current stacking and placement area is determined to be of medium risk. If the overall stacking score is within the third risk threshold range, the current stacking area is determined to be low risk.

4. The closed-loop control system for warehouse environment humidity according to claim 1, characterized in that, The edge computing layer includes: The multi-source fusion module is used to perform timestamp calibration and spatial coordinate matching on warehouse temperature and humidity and material stacking point clouds to generate stacking environment data sets; The material scenario module is used to correlate the critical relative humidity, moisture absorption rate and agglomeration critical parameters of different types of materials, generate a set of material moisture absorption parameters, and construct a raw material mixed storage scenario. The risk assessment module is used to construct a dynamic moisture risk weight function based on the stacking environment data set and the material's hygroscopicity level, and generate a humidity heat map of the entire storage area. The humidity control sorting module is used to classify and sort the stacked fabric areas according to the mixed storage scenario of raw materials based on the humidity heat map of the entire storage area, and mark the humidity control priority.

5. The closed-loop control system for warehouse environment humidity according to claim 4, characterized in that, The material scenario module includes: The parameter input unit is used to correlate the critical relative humidity, moisture absorption rate and agglomeration critical parameters according to the material type to obtain the material moisture absorption parameter set; The coupling correlation unit is used to fit the material moisture absorption parameter set according to the temperature sequence, determine the effective critical humidity threshold, calculate the risk accumulation index, and construct the agglomeration risk accumulation model. The humidity grading unit is used to classify materials for compatibility based on the effective critical humidity threshold and mark the humidity compatibility level of the materials. The storage constraint unit is used to perform cross-checks between materials based on the material humidity compatibility level and the material volatility, and generate a stacking isolation rule library. The scenario evolution unit is used to simulate and combine material stacking based on the clumping risk accumulation model and the stacking isolation rule library to construct a raw material mixed storage scenario.

6. The closed-loop control system for warehouse environment humidity according to claim 4, characterized in that, The risk assessment module includes: Spatial partitioning unit is used to divide the entire warehouse space into grids, assign spatial index numbers, and record the grid center coordinates and warehouse functional areas. The data binding unit is used to map the stacking environment data group to the storage functional area based on the grid center coordinates, and to construct a complete functional area; An independent quantization unit is used to transform the stacking environment data set, calculate the humidity excess factor f1, ventilation gap factor f2, and stack level airtightness factor f3; and map the material hygroscopicity level to obtain the hygroscopicity level factor f4. The function building unit is used to perform nonlinear weighting operations on each factor to generate the dynamic moisture risk weight function Rcell: Rcell=w1*f1+w2*f2+w3*f3+w4*f4+w12*f1*f2+w14*f1*f4, Among them, w1, w2, w3 and w4 are the main effect weights of the corresponding factors, w12*f1*f2 is the humidity-ventilation coupling term, and w14*f1*f4 is the humidity-material coupling term; The risk calculation unit is used to perform comprehensive calculations on the complete functional area based on the dynamic moisture risk weighting function to obtain the regional moisture risk coefficient. The pixel mapping unit is used to smooth the transition grid between complete functional areas according to the spatial index number and the regional moisture risk coefficient, forming a moisture risk field. The information overlay unit is used to mark information on the moisture risk field based on the stacking structure characteristics and generate a humidity heat map of the entire storage area.

7. The closed-loop control system for warehouse environment humidity according to claim 1, characterized in that, The linked execution layer includes: The regional aggregation module is used to perform grid aggregation of the entire warehouse space according to the humidity control priority, obtain the same-level control zones with different risk levels, and calculate the humidity deviation by combining the humidity heat map of the entire warehouse. The task sorting module is used to sort the control zones of the same level in descending order according to the humidity control priority to obtain the preliminary humidity control execution sequence, and to make internal corrections according to the humidity deviation to obtain the final humidity control execution sequence. The equipment detection module is used to query dehumidification equipment in the same control zone at different levels, establish an equipment space coverage table, and calculate the overlap of dehumidification spaces. The resource allocation module is used to allocate dehumidification equipment according to the overlap of dehumidification spaces based on the equipment space coverage table. If a single dehumidifier cannot provide complete coverage, multiple dehumidifiers will be deployed to work together, and corresponding grid responsibility areas will be defined. If different areas with the same priority compete for the same dehumidifier, it will be assigned to the grid responsibility area with the larger humidity deviation or the greater risk of dampness in the area. The other grid responsibility area will be assigned the second-best coverage device or wait in line for the current device to be released. The instruction generation module is used to pre-regulate the operation mode of the dehumidification equipment based on the risk level of the stockpile area and the final humidity control execution sequence, generate equipment operating parameters, and convert them into differentiated control instructions.

8. A closed-loop humidity control method for a storage environment, applied to the system described in any one of claims 1-7, characterized in that, include: Differentiated monitoring points are set up throughout the warehouse space to collect and clean warehouse temperature, humidity and material stacking point clouds; The warehouse temperature and humidity data and material stacking point cloud data are spatiotemporally fused to generate stacking environment data sets, and humidity control priorities are determined by combining material moisture absorption parameters. Based on the humidity control priority and humidity deviation, the warehouse dehumidification equipment is linked to generate differentiated control commands.

9. A storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the closed-loop humidity control method for warehouse environment as described in claim 8.

Images