Dust monitoring and analysis control system based on electronic greenhouse of port bulk cargo yard

Abstract

The present application relates to the field of environmental monitoring and automatic control technology, and specifically provides a dust monitoring and analysis control system based on a port bulk cargo yard electronic greenhouse. The system includes an environmental multi-dimensional perception module, a dust field dynamic reconstruction module, a dust suppression demand prediction module, and a precise execution control module. Through multi-dimensional data acquisition, dynamic three-dimensional dust field modeling, future dust suppression demand prediction, and spatial and timing optimization control of the spraying resource, the system realizes precise monitoring, prediction, and on-demand deployment of dust suppression resources of dust concentration, effectively improving the timeliness, foresight, and operation efficiency of control.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring and automatic control technology, specifically relating to a dust monitoring, analysis and control system based on an electronic shed in a port bulk cargo yard. Background Technology

[0002] In the field of environmental management and pollution control technology for port bulk cargo yards, dust pollution is a key issue affecting operational safety, the surrounding environment, and personnel health. With increasingly stringent environmental protection requirements and the promotion of the green port concept, effective monitoring and control of dust in storage yards has become a core requirement for industry development. Electronic storage sheds, as a widely used enclosed storage facility in recent years, are a key area of ​​current technological research and development, with the aim of achieving effective dust suppression and refined management through automation.

[0003] Existing technologies typically collect concentration data by deploying a network of dust sensors inside an electronic greenhouse and triggering a dust suppression spray system based on preset thresholds. However, existing systems have significant limitations: the sensor deployment strategy lacks adaptability to dynamic factors such as airflow patterns and material stacking patterns within the greenhouse, resulting in insufficient representativeness of the monitoring data.

[0004] Traditional fixed threshold control modes cannot respond to the coupled changes of multiple variables such as dust generation rate and ambient humidity, which can easily lead to waste of dust suppression resources or control lag. In addition, the system lacks the ability to deeply mine historical monitoring data and predict trends, making it difficult to achieve the upgrade of control from passive response to active prevention.

[0005] The aforementioned problems result in unstable dust control performance and low operational efficiency in electronic storage sheds, failing to meet increasingly sophisticated environmental management requirements. Therefore, there is an urgent need to develop a dust monitoring and analysis control system that deeply integrates real-time monitoring, intelligent analysis, and precise control functions to improve the environmental governance level and automated operation efficiency of electronic storage sheds in port bulk cargo yards. Summary of the Invention

[0006] The technical problem this invention aims to solve is to overcome the shortcomings of existing dust control systems for electronic sheds in port bulk cargo yards, such as insufficient representativeness of monitoring data, lagging control response, and lack of predictive management capabilities. This invention provides a dust monitoring, analysis, and control system based on electronic sheds in port bulk cargo yards. This system constructs a closed-loop control architecture integrating multi-dimensional sensing, dynamic modeling, intelligent prediction, and precise execution, enabling high-precision monitoring of the dust concentration field inside the shed, accurate prediction of its evolution trend, and precise on-demand allocation of dust suppression resources.

[0007] The technical solution of this invention is to construct a dust monitoring, analysis, and control system based on an electronic shed in a port bulk cargo yard. This system includes an environmental multi-dimensional sensing module, a dust field dynamic reconstruction module, a dust suppression demand prediction module, and a precise execution control module. The environmental multi-dimensional sensing module is responsible for collecting raw environmental status data inside and at the boundaries of the electronic shed.

[0008] The dust field dynamic reconstruction module receives raw data from the environmental multi-dimensional sensing module and constructs a three-dimensional dynamic distribution model of the dust concentration field inside the greenhouse. The dust suppression demand prediction module is connected to the dust field dynamic reconstruction module and is used to predict the dust concentration evolution trend and dust suppression demand intensity within a specific future time window based on the current and historical dust field dynamic distribution models.

[0009] The precision execution control module is connected to the dust suppression demand prediction module and the environmental multi-dimensional perception module, respectively. It is used to generate a sequence of control commands for the spray execution unit based on the predicted dust suppression demand intensity and the real-time feedback of local environmental parameters, so as to achieve precise control of dust suppression resources in the spatial and temporal dimensions.

[0010] Furthermore, the environmental multi-dimensional sensing module specifically includes a dust concentration sensing network, a meteorological element monitoring unit, and a material pile shape scanning unit. The dust concentration sensing network consists of a set of high-precision laser dust sensors deployed at key locations inside the electronic greenhouse, used to measure the dust mass concentration at different spatial points in real time.

[0011] The meteorological element monitoring unit includes miniature weather stations installed on the top and side walls of the greenhouse, which continuously collect data on wind speed, wind direction, temperature, and relative humidity inside the greenhouse. The material pile shape scanning unit uses a 3D laser scanner installed on the top of the greenhouse structure to periodically scan the 3D morphology of the material surface in the stockpile, obtaining data on changes in the outline height and surface area of ​​the material pile.

[0012] Furthermore, the core of the dust field dynamic reconstruction module lies in executing a numerical simulation process based on computational fluid dynamics and particulate transport theory. This module first establishes a steady-state calculation model of the airflow field inside the electronic greenhouse based on the three-dimensional terrain data provided by the material pile shape scanning unit and the boundary conditions provided by the meteorological element monitoring unit.

[0013] Subsequently, the module uses discrete-point concentration measurements obtained from the dust concentration sensor network as observation data, embedding them into the aforementioned airflow field model through a data assimilation algorithm. This allows for the solution of the particulate transport control equations describing the convection, diffusion, settling, and lifting processes of dust particles within the greenhouse. The final output is a time-series three-dimensional dust concentration distribution field covering the entire computational grid inside the greenhouse.

[0014] Furthermore, the workflow of the dust suppression demand prediction module includes two levels: trend extrapolation and event triggering. At the trend extrapolation level, the module extracts spatiotemporal features from the historical time series concentration field data output by the dust field dynamic reconstruction module and trains a time series prediction model to extrapolate the dust concentration change curves of the entire greenhouse and key areas within the next 1 to 2 hours.

[0015] At the event triggering level, the module integrates real-time collected positioning information of operating machinery and material loading and unloading flow data. When a high-intensity loading and unloading operation event is identified in a specific area, the module dynamically corrects the trend extrapolation results based on the current meteorological conditions of the area and calculates a quantitative dust suppression demand intensity index. This index comprehensively reflects the urgency of dust suppression actions and the scale of resource input.

[0016] Furthermore, the precision execution control module determines the specific control strategy based on the received dust suppression demand intensity index. This module internally maintains a database of the spatial layout and performance parameters of the spray execution units. The decision-making process begins with spatial targeting, that is, matching the optimal combination of spray execution units to cover the high-concentration areas or areas about to become high-concentration areas based on the coordinates output by the dust suppression demand prediction module.

[0017] Secondly, dosage and timing optimization are performed. The module combines the local humidity data fed back in real time by the environmental multi-dimensional sensing module to avoid over-humidification or insufficient dust suppression. The optimization algorithm calculates the optimal start-up time, water pressure and atomization particle size parameters for each activated spray execution unit, and generates a control command sequence containing start-up and stop times and working parameters, which is then sent to the corresponding spray execution unit.

[0018] Furthermore, the spray actuator employs independently controllable intelligent spray valve assemblies, with each assembly driving one or more spray heads. These intelligent spray valve assemblies receive digital commands from the precision execution control module, accurately adjusting the water circuit switch status and flow rate. Some advanced configurations can also receive pressure setting commands to adjust the spray atomization effect.

[0019] Furthermore, the system operates within a hierarchical decision-making framework, which includes a bottom-level second-level real-time control loop, a middle-level minute-level situation assessment loop, and a top-level hour-level strategy optimization loop. The second-level real-time control loop consists of a precise execution control module and a multi-dimensional environmental perception module, responsible for rapid response to sudden dust events.

[0020] The core of the minute-level situation assessment loop is the dust field dynamic reconstruction module, which periodically updates the panoramic view of dust distribution within the greenhouse. The hour-level strategy optimization loop is led by the dust suppression demand prediction module, which adaptively adjusts the prediction model parameters and macroscopically plans the dust suppression strategy based on data from longer time periods.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0022] 1. This invention achieves a leap from discrete point measurement to three-dimensional continuous field perception of dust concentration inside an electronic greenhouse through collaborative data acquisition of an environmental multi-dimensional sensing module and numerical simulation of a dust field dynamic reconstruction module. This fundamentally solves the problem of insufficient data representativeness caused by traditional sensor placement strategies that ignore dynamic changes in airflow and pile shape, providing a high-fidelity environmental situation map for precise control.

[0023] 2. This invention introduces a dust suppression demand prediction module, upgrading the dust control mode from passive threshold triggering to proactive trend prediction and event response. The system can predict the evolution path of dust concentration and key risk areas in advance, enabling dust suppression actions to be deployed in advance before the pollution concentration reaches the critical value. This effectively overcomes the lag of traditional control methods and significantly improves the timeliness and foresight of control.

[0024] 3. This invention achieves spatial targeting and on-demand delivery of dust suppression resources through a precise execution control module. Based on predicted demand and real-time feedback, the system performs refined scheduling and parameter optimization of the spray units, avoiding the waste of water and electricity in traditional uniform spraying or fixed-area spraying modes. While ensuring the dust suppression effect, it significantly improves the system's operational efficiency and economy, meeting the sustainable development requirements of green ports. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall technical solution architecture of the dust monitoring, analysis and control system based on the electronic shed of the port bulk cargo yard proposed in this invention.

[0026] Figure 2 This is a schematic diagram of the core principle framework of the dust field dynamic reconstruction module in this invention;

[0027] Figure 3 This is a logical flow diagram of the dust suppression demand prediction module in this invention;

[0028] Figure 4 This is a logical flow diagram of the precision execution control module in this invention;

[0029] Figure 5 This is a schematic diagram of the multi-level interaction relationships and data flow of the hierarchical decision-making framework of the system in this invention;

[0030] Figure 6 This is a schematic diagram of the interaction relationship and data flow between the real-time control loop of the multi-dimensional environmental perception module and the precise execution control module in this invention. Detailed Implementation

[0031] This invention provides a dust monitoring, analysis, and control system based on an electronic enclosure in a port bulk cargo yard. Please refer to the appendix. Figures 1 to 6The system's overall architecture consists of an environmental multi-dimensional sensing module, a dust field dynamic reconstruction module, a dust suppression demand prediction module, and a precise execution control module. These modules are connected to an industrial Ethernet network via a pre-set high-speed data bus, forming a data closed loop.

[0032] The system is deployed inside electronic sheds in port bulk cargo yards. These sheds are fully or semi-enclosed steel structures used for storing bulk materials such as coal and ore. The core objective of the system is to achieve three-dimensional dynamic perception of the dust concentration field inside the shed, accurate prediction of its future evolution trend, and precise spatial and temporal control of the dust suppression execution units.

[0033] The multi-dimensional environmental sensing module is responsible for collecting raw environmental data inside and around the electronic greenhouse. This module specifically includes a dust concentration sensing network, a meteorological element monitoring unit, and a material pile shape scanning unit. The dust concentration sensing network consists of a set of high-precision laser dust sensors deployed at key locations within the electronic greenhouse.

[0034] The placement strategy for these sensors is determined based on computational fluid dynamics pre-simulations. They are typically located upwind and downwind of the material pile, in corner areas, and in the roof space of the greenhouse, with a total deployment of between 32 and 64 sensors. Each high-precision laser dust sensor has a built-in miniature air pump that draws in ambient air at a constant flow rate. It uses the principle of laser scattering to measure the mass concentration of suspended particulate matter in the air in real time, with a measurement accuracy of ±3% and a range covering 0 to 1000 mg / m³. The sensor output is a 4 to 20 mA analog signal or a Modbus RTU digital signal, which is transmitted to the data acquisition box via shielded cable.

[0035] The data acquisition box incorporates signal conditioning circuitry and an analog-to-digital converter to filter, amplify, and digitize the signal, with a sampling frequency set to 1 Hz. The digitized concentration data is then timestamped and labeled with spatial coordinates, encapsulated into standard data frames, and transmitted to the central processing server via an industrial Ethernet switch.

[0036] The meteorological element monitoring unit consists of five miniature weather stations installed at the center of the greenhouse roof and the middle of the four side walls. The roof miniature weather station measures wind speed, wind direction, temperature, and relative humidity at the highest point inside the greenhouse. The side wall miniature weather stations monitor the meteorological conditions of the greenhouse boundary layer. Wind speed measurement uses ultrasonic principles, has no moving parts, a range of 0 to 30 meters per second, and an accuracy of ±0.2 meters per second.

[0037] Wind direction is measured using an electronic compass with an accuracy of ±3 degrees Celsius. Temperature is measured using a PT100 platinum resistance thermometer with an accuracy of ±0.2 degrees Celsius. Relative humidity is measured using a capacitive polymer sensor with an accuracy of ±2%. All meteorological data is sampled at a frequency of 1 Hz, collected via an RS485 bus to a meteorological data concentrator, and then transmitted to a central processing server via Ethernet. The data concentrator has a built-in real-time clock to ensure that all data are synchronized.

[0038] The material stacking scanning unit employs a 3D laser scanner installed on the top of the greenhouse structure, typically with 2 to 4 units symmetrically arranged on the main beam of the greenhouse. The 3D laser scanner uses either the time-of-flight method or the phase comparison method, emitting near-infrared laser pulses to scan the entire surface of the material in the stockpile. The scanner's horizontal scanning angle is 0 to 360 degrees, and its vertical scanning angle is -30 to +90 degrees, with an angular resolution better than 0.01 degrees.

[0039] The scanner can acquire hundreds of thousands of 3D point cloud data points per second. The scanning cycle is set according to the port's operational intensity: once per hour during periods of no operation and once every 10 minutes during periods of intensive loading and unloading. The acquired point cloud data is first preprocessed, including noise filtering, point cloud registration, and coordinate transformation, converting the point cloud from the scanner's coordinate system to a global coordinate system with one corner of the shed as the origin. The preprocessed point cloud is then used to reconstruct a 3D digital elevation model of the material pile surface using a triangulation algorithm.

[0040] The model includes the height value of each grid vertex, and uses this to calculate the surface area, volume, and slope distribution of the material pile. This terrain data is also timestamped and transmitted via gigabit Ethernet to a central processing server for storage.

[0041] The dust field dynamic reconstruction module receives all the raw data from the environmental multidimensional perception module and constructs a three-dimensional dynamic distribution model of the dust concentration field inside the greenhouse based on this data.

[0042] Please refer to the attached document. Figure 2 The core of this module lies in executing numerical simulations based on computational fluid dynamics and particulate transport theory. The module runs on a high-performance computing node of a central processing server, equipped with a multi-core CPU, high-speed memory, and a dedicated graphics processing unit (GPU) to accelerate computation.

[0043] The dust field dynamic reconstruction module first performs geometric modeling and mesh generation. The module reads the latest 3D digital elevation model data provided by the material pile shape scanning unit, using it as the lower boundary of the computational domain. The upper boundary of the computational domain is defined as the inner surface of the roof of the electronic greenhouse, and the side boundaries are the inner surfaces of the greenhouse walls. The module uses unstructured mesh generation technology to discretize the computational domain, generating millions of tetrahedral or hexahedral mesh elements. The mesh is locally refined near the material surface, spray devices, and sensor locations to ensure computational accuracy in critical areas.

[0044] Subsequently, the module performs airflow field calculations. It uses real-time wind speed, wind direction, temperature, and relative humidity data provided by the meteorological monitoring unit as the inlet boundary conditions and initial conditions for the flow field calculation. The module solves the Navier-Stokes equations describing airflow, including the continuity equation, momentum equation, and energy equation. The solution process employs a pressure-based coupled algorithm or a discrete solver, performing parallel computation on a graphics processor to obtain a stable airflow velocity and pressure field inside the electronic greenhouse. The convergence criterion is a residual reduction to a certain value. the following.

[0045] After obtaining a stable airflow field, the module initiates a dust transport simulation. It uses discrete-point concentration measurements acquired by a dust concentration sensor network as observation data and embeds them into the aforementioned airflow field model through a data assimilation algorithm. The data assimilation process employs either ensemble Kalman filtering or a three-dimensional variational method.

[0046] This method optimally fuses sensor observations with model predictions to correct the model's state variables, namely the dust concentration field. Specifically, the system maintains a set of model states, with each member representing a possible concentration field distribution. In each assimilation cycle, the module compares the sensor observations with the predictions of all set members and calculates the covariance matrix between the observed and predicted values.

[0047] Subsequently, the concentration fields of all set members are updated according to the Kalman gain formula, so that the updated set mean is closest to the true state.

[0048] ;

[0049] in, This represents the state vector after analysis, i.e., the assimilated concentration field; This represents the predicted state vector, i.e., the concentration field before assimilation; This represents the observation vector, i.e., the concentration value measured by the sensor; It is an observation operator that maps the model state space to the observation space; The Kalman gain matrix determines the weight of the observations in the analysis results. The calculation of the Kalman gain matrix depends on the model prediction error covariance matrix and the observation error covariance matrix.

[0050] After data assimilation, the module solves the set of governing equations for particulate transport, describing the convection, diffusion, settling, and uplift processes of dust particles within the greenhouse. This set of equations includes the dust mass conservation equation. The convection term is driven by a known air velocity field, the diffusion term is described by the turbulent diffusion coefficient, the settling term uses the Stokes settling velocity model, and the uplift term is parameterized based on empirical formulas for wind speed at the material pile surface and material properties. The equations are solved discretly on a pre-defined computational grid using the finite volume method.

[0051] The solution process is also executed in parallel on the graphics processing unit (GPU). Ultimately, the module outputs a time-series three-dimensional dust concentration distribution field covering all computational grid nodes within the entire greenhouse. This concentration field data is stored in scientific data formats such as NetCDF or HDF5, with a spatial resolution down to the meter level and a temporal resolution consistent with the simulation step size, typically 1 to 10 seconds. Each concentration value represents the average dust mass concentration within that grid cell, expressed in milligrams per cubic meter.

[0052] The dust suppression demand prediction module is connected to the dust field dynamic reconstruction module. Based on current and historical dust field dynamic distribution models, it predicts the evolution trend of dust concentration and the intensity of dust suppression demand within a specific future time window. Please refer to the appendix. Figure 3 The workflow of this module includes two levels: trend extrapolation and event triggering, and it runs on the data analysis node of the central processing server.

[0053] At the trend extrapolation level, the module first performs spatiotemporal feature extraction. It extracts key features from the historical time-series concentration field data output by the dust field dynamic reconstruction module.

[0054] These features include, but are not limited to: the spatial average concentration of the entire greenhouse, the maximum spatial gradient of the concentration field, the percentage of grid volumes where the concentration exceeds a preset threshold of 50 mg / m³, and the top 5 principal modes and their time coefficients obtained from principal component analysis of the concentration field. Feature extraction is performed once per minute.

[0055] Subsequently, the module trains and applies a time-series prediction model. The system employs a Long Short-Term Memory network or a gated recurrent unit as the core prediction model. The model's input is the feature time series from the past 60 minutes, and the output is the predicted value of the corresponding feature for the next 120 minutes.

[0056] Long Short-Term Memory (LSTM) networks, through their internal gating mechanisms including input gates, forget gates, and output gates, are able to learn long-term dependencies in time series data. The model is trained offline using historical data from the past 30 days, employing mean squared error as the loss function and the Adam optimizer for parameter updates.

[0057] During online forecasting, the model executes every 5 minutes, providing rolling predictions of the concentration field evolution trend for the next 2 hours. The prediction results include the overall concentration change curve of the greenhouse at each future time point, as well as the predicted concentration values ​​for key areas, such as the downwind area and the vicinity of the work area.

[0058] At the event triggering level, the module integrates external real-time data streams. This data includes the positioning information of operating machinery from the port scheduling system, such as the real-time 3D coordinates of ship unloaders and stacker-reclaimers; and material loading and unloading flow data, such as the instantaneous flow recorded by belt scales. The module has pre-defined event recognition rules.

[0059] When a specific area is identified, such as within 50 meters of a certain unloading point, where a ship unloader operates continuously for more than 2 minutes and the instantaneous flow rate exceeds 1,000 tons per hour, it is determined that a high-intensity loading and unloading operation event has occurred in that area.

[0060] Once the event is triggered, the module initiates a dynamic correction process. It combines the current meteorological conditions of the area, primarily wind speed and direction data from meteorological monitoring units, to calculate the dust uplift potential. The uplift potential is an empirical index, directly proportional to the 2.5th power of the wind speed and inversely proportional to the surface moisture content of the material.

[0061] The module uses the rising potential index to dynamically correct the future concentration curve of the region predicted by trend extrapolation. The correction method is to superimpose a concentration increment function based on event intensity and duration onto the predicted concentration.

[0062] Finally, the module calculates a quantified dust suppression demand intensity index. This index is a dimensionless value between 0 and 100. Its calculation formula takes into account the predicted concentration value, the rate of concentration increase, and the urgency of the event triggering.

[0063] Specifically, the base value of the index is determined by the ratio of the predicted maximum concentration to the threshold of 80 mg / m³. Then, it is weighted according to the rate of increase in concentration per unit time; the faster the rate of increase, the higher the weighting coefficient, up to a maximum of 1.5 times the base value. If an event triggers, an additional fixed increment of 20 is added. The calculated dust suppression demand intensity index is updated every minute, along with its corresponding spatial location information, i.e., the coordinates of high-concentration areas or areas that are about to become high-concentration areas.

[0064] The precision execution control module is connected to the dust suppression demand prediction module and the environmental multi-dimensional perception module, respectively, and is used to generate a sequence of control commands for the spray execution unit based on the predicted dust suppression demand intensity index and the real-time feedback of local environmental parameters.

[0065] Please refer to the attached document. Figure 4 This module runs on the real-time control node of the central processing server. This node has a hard real-time operating system to ensure the timely transmission of control commands.

[0066] The precision execution control module maintains a database containing the spatial layout and performance parameters of the spray actuators. This database records the three-dimensional installation coordinates, spray coverage radius, rated flow rate, working pressure range, atomized particle size range, and valve group number of each spray actuator in the greenhouse's global coordinate system.

[0067] The module's decision-making process begins with spatial targeting. It receives the dust suppression demand intensity index and its corresponding spatial location information from the dust suppression demand prediction module. Based on the coordinates of the demand area, the module performs a spatial query in the sprinkler execution unit database. The query algorithm, based on the geometric coverage principle, identifies all sprinkler execution units whose sprinkler coverage areas intersect with the demand area.

[0068] Subsequently, a greedy algorithm or integer programming is used to select the optimal combination of sprinkler execution units from the candidate units that can cover the target area with the minimum number of units and the maximum coverage efficiency. For example, for a circular demand area, sprinkler units near the center of the circle are selected first, ensuring that the coverage rate reaches more than 95%.

[0069] Next, dosage and timing optimization were performed. The module combined real-time local humidity data from the meteorological element monitoring unit in the environmental multi-dimensional sensing module with real-time dust concentration data from the dust concentration sensor network. The optimization goal was to ensure dust suppression effectiveness while avoiding excessive humidification that could lead to material adhesion or water waste, and also to avoid insufficient dust suppression. The optimization process was completed using a rule-based and model-based predictive control algorithm.

[0070] ;

[0071] in, It is the objective function that needs to be minimized; This is the number of activated spray execution units; It is the predicted number Dust concentration in the area after control by each unit; The target concentration is set at 30 milligrams per cubic meter. It is the predicted number The relative humidity of the area is controlled by each unit; This is the maximum allowable humidity, set to 85% to prevent over-humidification; It is the first Water consumption of each spray unit;

[0072] These are weighting coefficients, with values ​​of 1.0, 0.5, and 0.1, used to balance dust suppression, humidity control, and water conservation goals. The module solves this optimization problem to calculate the optimal operating time, operating water pressure, and target atomized particle size parameters for each activated spray unit. The operating time is accurate to the second, the water pressure adjustment accuracy is 0.1 MPa, and the atomized particle size is indirectly controlled by adjusting the water pressure and nozzle type.

[0073] Finally, the module generates a sequence of control commands containing start / stop times and operating parameters. The command sequence is encapsulated into data packets using a specific industrial communication protocol, such as Modbus TCP or OPC UA. Each data packet contains the address code of the target spray execution unit, the command type, parameter values, and an execution timestamp.

[0074] The command sequence is sent to the corresponding spray execution unit via an industrial Ethernet switch. The control cycle is synchronized with the dust suppression demand prediction module, typically 1 minute, but for a second-level real-time control loop, the response time can be shortened to less than 5 seconds.

[0075] The spray actuator employs independently controllable intelligent spray valve assemblies, with each assembly driving one or more spray heads. Essentially, the intelligent spray valve assembly is an electrically controlled water valve, its core being a ball valve or diaphragm valve driven by a stepper motor or servo motor. The valve assembly receives digital commands from the precision execution control module, which are transmitted via RS485 bus or industrial wireless network to the microcontroller built into the valve assembly.

[0076] The microcontroller interprets instructions and precisely controls the motor's rotation angle, thereby adjusting the valve opening and achieving precise control of the water circuit's on / off state and flow rate. The flow control accuracy can reach ±2% of the rated flow rate.

[0077] Some high-end configurations also integrate pressure sensors and pressure regulating valves, which can receive pressure setting commands and stabilize the water pressure at the set value through closed-loop control, thereby adjusting the spray atomization effect. The atomized particle size can be adjusted between 50 and 300 micrometers. The nozzles are made of wear-resistant ceramic material, and the spray angle can be selected according to the installation location, such as 90 degrees, 120 degrees, or 180 degrees.

[0078] The system operates within a hierarchical decision-making framework, which coordinates control tasks at different time scales. Please refer to the appendix. Figure 5 The framework includes a bottom-level second-level real-time control loop, a middle-level minute-level situation assessment loop, and a top-level hour-level strategy optimization loop.

[0079] The second-level real-time control loop consists of a precision execution control module and a multi-dimensional environmental sensing module. Please refer to the appendix. Figure 6 The typical interaction process of this loop is as follows: the dust concentration sensor network in the environmental multidimensional perception module continuously collects data at a frequency of 1 Hz.

[0080] When a sensor detects a sudden increase in dust concentration exceeding 100 milligrams per cubic meter within one second at its location, the event is marked as a sudden dust event. The data acquisition box immediately sends this abnormal data, along with a high-priority flag, to the central processing server. The real-time control node of the precision execution control module interrupts its normal control cycle and responds immediately. It first quickly queries the spray actuator database based on the coordinates of the event sensor to locate the nearest spray actuator that covers the sensor.

[0081] Subsequently, the module bypasses complex optimization calculations and directly generates a preset emergency control command. This command instructs the target sprinkler unit to immediately start at maximum flow for 10 seconds. The command is issued within 500 milliseconds of detecting the event.

[0082] Upon receiving the instruction, the spray execution unit's microcontroller immediately drives the valves to fully open, initiating powerful dust suppression. Simultaneously, this event is recorded and reported to the mid-level situation assessment loop, triggering an immediate dynamic reconstruction of the dust field.

[0083] The core of the mid-level minute-level situation assessment loop is the dust field dynamic reconstruction module. This loop runs at a fixed time period, typically set to initiate a complete dust field simulation every 5 minutes. Its triggering conditions can also be receiving an emergency event report from the second-level loop, or the dust suppression demand prediction module identifying a new operational event.

[0084] In this loop, the dust field dynamic reconstruction module integrates all new data collected by the environmental multi-dimensional sensing module within the past 5 minutes, including updated concentration measurements, meteorological data, and any latest material pile shape scan data. The module executes a complete process from geometric modeling to data assimilation and dust transport solution, outputting an updated panoramic view of the dust distribution within the greenhouse that reflects the current state.

[0085] This panoramic view serves as the most authoritative environmental situational map for the entire system. It is provided to the dust suppression demand prediction module and the precision execution control module to ensure that all decisions are based on consistent and up-to-date environmental information.

[0086] The top-level hourly strategy optimization loop is led by the dust suppression demand forecasting module. This loop has a long operating cycle, typically executing once per hour. Its main task is to perform in-depth analysis and strategy adjustments based on historical data over longer periods.

[0087] Specific tasks include: the dust suppression demand prediction module adaptively adjusting its internal time-series prediction model. The module collects all prediction results and actual observations from the past 24 hours and calculates the prediction error. If the mean absolute percentage error exceeds 15% for three consecutive periods, a model retraining process is triggered. Retraining uses data from the most recent seven days to update the weight parameters of the long short-term memory network to adapt to pattern changes caused by material properties or seasonal variations.

[0088] Furthermore, this loop also performs macro-level planning of dust suppression strategies. For example, it analyzes the dust suppression effect and water consumption data of the past day to evaluate performance under different control parameters. If it finds that certain areas are consistently insufficient in dust suppression under specific meteorological conditions, it can suggest adjusting the coverage radius parameter of the sprinkler unit in that area or optimizing the weight coefficients in the algorithm. After these strategy adjustment suggestions are confirmed, they will be updated in the database or algorithm parameters of the precision execution control module to achieve continuous self-optimization of the system.

[0089] Through the precise collaboration and hierarchical decision-making of the aforementioned modules, this system achieves intelligent management of the entire process of dust generation, diffusion, prediction, and suppression within the electronic enclosure of the port bulk cargo yard. Its core advantages lie in upgrading discrete point measurements to continuous three-dimensional field perception, transforming lagging threshold responses into proactive trend intervention, and optimizing extensive overall spraying into precise targeted control. This significantly improves water and energy utilization efficiency while ensuring dust suppression effectiveness.

[0090] This embodiment provides an implementation plan for a dust monitoring, analysis and control system based on an electronic shed in a port bulk cargo yard. Its core feature is that the algorithm models of the dust field dynamic reconstruction module and the dust suppression demand prediction module have been replaced, and a sensing network architecture based on wireless Internet of Things has been introduced. It is suitable for scenarios where existing electronic sheds are retrofitted at low cost or where there are strict restrictions on communication wiring.

[0091] In this embodiment, the architecture of the environmental multi-dimensional sensing module has been adjusted. The dust concentration sensing network uses wireless sensor nodes based on LoRa or NB-IoT technology.

[0092] Each node integrates a laser dust sensor, a micro battery or solar panel, and a LoRa communication module. The nodes are magnetically or mounted on the steel structure inside the greenhouse, eliminating the need for cabling. Operating in low-power mode, the nodes measure concentration every 10 seconds and upload the data to a cloud server or local edge computing gateway via a LoRa gateway.

Claims

1. A dust monitoring, analysis, and control system based on an electronic shed in a port bulk cargo yard, characterized in that, include: The multi-dimensional environmental sensing module is used to collect raw environmental data of the interior and boundaries of the electronic greenhouse; The dust field dynamic reconstruction module, connected to the environmental multi-dimensional sensing module, is used to receive raw environmental data and construct a three-dimensional dynamic distribution model of the dust concentration field inside the greenhouse. The dust suppression demand prediction module, connected to the dust field dynamic reconstruction module, is used to predict the dust concentration evolution trend and dust suppression demand intensity within a specific future time window based on the current and historical dust field dynamic distribution model. The workflow of the dust suppression demand prediction module includes two levels: trend extrapolation and event triggering. At the trend extrapolation level, the module extracts spatiotemporal features from the historical time series concentration field data output by the dust field dynamic reconstruction module and trains a time series prediction model to extrapolate the dust concentration change curves of the entire greenhouse and key areas within the next 1 to 2 hours. At the event triggering level, the module integrates real-time collected positioning information of operating machinery and material loading and unloading flow data. When a high-intensity loading and unloading operation event is identified in a specific area, the module dynamically corrects the trend extrapolation results based on the current meteorological conditions of the area and calculates a quantified dust suppression demand intensity index. The precision execution control module is connected to the dust suppression demand prediction module and the environmental multi-dimensional perception module, respectively, and is used to generate a sequence of control instructions for the spray execution unit based on the predicted dust suppression demand intensity and the real-time feedback of local environmental parameters. The decision-making process of the precision execution control module includes spatial targeting and dose and timing optimization; Spatial targeted positioning matches the optimal combination of spray execution units to cover areas with high concentrations or areas that are about to become high concentrations, based on the coordinates of high-concentration areas output by the dust suppression demand prediction module. Dosage and timing optimization combines real-time local humidity data from the environmental multi-dimensional sensing module to avoid over-humidification or insufficient dust suppression. The optimal activation duration, water pressure, and atomized particle size parameters for each activated spray execution unit are calculated using an optimization algorithm. The environmental multi-dimensional sensing module includes a dust concentration sensing network, a meteorological element monitoring unit, and a material pile shape scanning unit. The dust concentration sensing network consists of a set of high-precision laser dust sensors deployed at key locations inside the electronic greenhouse, used to measure the dust mass concentration at different spatial points in real time; the meteorological element monitoring unit includes miniature weather stations installed on the top and side walls of the greenhouse, continuously collecting data on wind speed, wind direction, temperature, and relative humidity inside the greenhouse; the material pile shape scanning unit uses a three-dimensional laser scanner installed on the top of the greenhouse structure to periodically scan the three-dimensional shape of the material surface in the stockpile, obtaining data on the changes in the outline height and surface area of ​​the material pile; The dust field dynamic reconstruction module performs a numerical simulation process based on computational fluid dynamics and particulate transport theory. This process includes: establishing a steady-state calculation model of the airflow field inside the electronic greenhouse based on the three-dimensional terrain data provided by the material pile shape scanning unit and the boundary conditions provided by the meteorological element monitoring unit; using the discrete point concentration measurements obtained by the dust concentration sensing network as observation data and embedding them into the airflow field model through a data assimilation algorithm; solving the particulate transport control equations describing the convection, diffusion, settling and lifting processes of dust particles inside the greenhouse, and outputting a time-series three-dimensional dust concentration distribution field covering the entire computational grid inside the greenhouse. The system operates within a hierarchical decision-making framework, which includes a bottom-level second-level real-time control loop, a middle-level minute-level situation assessment loop, and a top-level hour-level strategy optimization loop. The second-level real-time control loop consists of a precise execution control module and an environmental multi-dimensional perception module, and is responsible for responding quickly to sudden dust events. The core of the minute-level situation assessment loop is the dust field dynamic reconstruction module, which periodically updates the panoramic view of dust distribution inside the greenhouse; the hour-level strategy optimization loop is dominated by the dust suppression demand prediction module, which adaptively adjusts the prediction model parameters and macro-plans the dust suppression strategy based on data from a longer period. The calculation process of the dust suppression demand intensity index comprehensively considers the predicted concentration value, the rate of increase of concentration, and the urgency of the event triggering; the base value of the index is determined by the ratio of the predicted maximum concentration value to the threshold of 80 mg / m³; the index is weighted according to the rate of increase of concentration per unit time, and the faster the rate of increase, the higher the weighting coefficient, which can be up to 1.5 times the base value. When an event is triggered, an additional fixed increment of 20 is added. The calculated dust suppression demand intensity index is updated every minute, along with its corresponding spatial location information, namely the coordinates of high-concentration areas or areas that are about to become high-concentration areas. The objective function of the optimization algorithm is: ； in, It is the objective function that needs to be minimized; This is the number of activated spray execution units; It is the predicted number Dust concentration in the area after control by each unit; The target concentration is set at 30 milligrams per cubic meter. It is the predicted number The relative humidity of the area is controlled by each unit; This is the maximum allowable humidity, set to 85% to prevent over-humidification; It is the first Water consumption of each spray unit; These are weighting coefficients, with values ​​of 1.0, 0.5, and 0.1 respectively. The interaction process of the second-level real-time control loop is as follows: the dust concentration sensor network in the environmental multi-dimensional perception module continuously collects data at a frequency of 1 Hz; when the dust concentration is detected to rise sharply to more than 100 milligrams per cubic meter within 1 second, it is marked as a sudden dust event; the precision execution control module responds immediately, locates the nearest spray execution unit according to the coordinates of the event sensor; generates an emergency control command, ordering the target spray execution unit to immediately start at maximum flow rate for 10 seconds.

2. The dust monitoring, analysis, and control system based on an electronic shed in a port bulk cargo yard according to claim 1, characterized in that, The spraying execution unit adopts an independently controllable intelligent spraying valve group, and each valve group drives one or more spray heads; The intelligent spray valve assembly receives digital commands from the precision execution control module to precisely adjust the water circuit switch status and flow rate. Some advanced configurations of the valve assembly can also receive pressure setting commands to adjust the spray atomization effect.

Images