Port electronic greenhouse meteorological big data monitoring and joint operation method and system
By constructing a multi-source data resource pool and a dynamic dust diffusion prediction model in the electronic shed of the port yard, and combining it with CFD simulation correction factors, multi-device collaborative control rules were set, solving the predictability and accuracy problems of port dust control and realizing intelligent dust management.
Patent Information
- Application Number
- CN202511786248.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-01
- Publication Date
- 2026-03-06
AI Technical Summary
Existing port dust control technologies lack predictability, suffer from severe data silos, have insufficient model accuracy, employ simplistic control strategies, and lack closed-loop optimization, making it difficult to meet the demands of intelligent systems in terms of real-time performance, accuracy, and cost-effectiveness.
A sensing network is deployed in the electronic shed of the open port yard to build a multi-source data resource pool. A dynamic dust diffusion prediction model and correction factor are used to predict the dust concentration distribution. Multi-device collaborative control rules are set and risk scores are calculated to achieve adaptive linkage control of dust suppression equipment. Control parameters are adjusted and the model is incrementally trained through real-time monitoring.
It enables early prediction of dust diffusion trends, improves the real-time performance and accuracy of dust concentration prediction, reduces water consumption, realizes adaptive linkage control of multiple devices and dynamic adjustment of control parameters, and alleviates control imbalance.
Smart Images
Figure CN121615549A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent environmental protection technology for port bulk cargo yards, and in particular to a method and system for monitoring and jointly calculating meteorological big data in port electronic sheds. Background Technology
[0002] During the stacking, loading, unloading, and transshipment of bulk cargo such as coal, ore, and grain in ports, the disturbances caused by natural winds, mechanical operations, and the inherent characteristics of the materials themselves easily generate large amounts of suspended particulate matter. This not only pollutes the surrounding environment and threatens the health of workers, but also can lead to equipment malfunctions and affect the efficiency of yard operations due to dust accumulation. To control dust, existing technologies generally employ a "passive defense + local suppression" approach. For example, windbreak and dust suppression nets are used to reduce wind speed and thus reduce dust generation; spray systems are used to increase material humidity and suppress dust; and crusting agents are sprayed to form a solidified layer that seals the material surface. While these methods alleviate the dust problem to some extent, they still have limitations, making it difficult for existing methods to meet the needs of intelligent dust control in ports in terms of real-time performance, accuracy, and cost-effectiveness.
[0003] First, the methods are primarily passive and lack predictive capabilities: most rely on a delayed control logic of "monitoring-exceeding limits-triggering," failing to anticipate dust diffusion trends and resulting in delayed control effects, making it difficult to cope with sudden high-risk scenarios. Second, there are severe data silos and a lack of integration of multi-source information: meteorological data, pile shape data, dust concentration data, and equipment status data are collected from different systems or equipment without spatiotemporal alignment and fusion analysis, failing to form a global view of "stockyard-meteorology-equipment." Third, the models lack accuracy and physical constraints: some data-driven prediction models are trained only on historical concentration data, without coupling with physical mechanisms such as wind field and pile shape, leading to large prediction deviations under complex operating conditions. Fourth, the control strategies are simplistic and lack coordination: multiple devices are controlled independently, lacking dynamic prioritization and coordination logic (e.g., the spraying effect is offset by the wind field when the windbreak net is not raised first), and fixed thresholds cannot adapt to dynamic operating conditions such as material humidity and seasonal wind speed changes. Fifth, the lack of a closed-loop optimization mechanism: Control parameters are not adjusted based on actual dust suppression effects, which can easily lead to imbalances of "over-control" or "under-control" during long-term operation. These problems make it difficult for existing solutions to meet the needs of intelligent dust suppression in ports in terms of real-time performance, accuracy, and cost-effectiveness. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a method and system for monitoring and jointly computing meteorological big data for electronic greenhouses in ports.
[0005] To achieve the above objectives, in a first aspect, this invention provides a method for monitoring and jointly computing meteorological big data in electronic sheds at ports. The method includes the following steps: deploying a sensing network in an open port yard electronic shed, acquiring multi-source data in conjunction with a meteorological forecasting system, and constructing a multi-source data resource pool using the multi-source data; constructing a dynamic dust diffusion prediction model based on the multi-source data resource pool, and using the dynamic dust diffusion prediction model and correction factors to predict the dust concentration distribution in the open port yard electronic shed; setting multi-device collaborative control rules and calculating risk scores, and then performing adaptive linkage control of dust suppression equipment based on the current and predicted dust concentration distribution to achieve dust suppression; monitoring the dust suppression effect in real time, adjusting the control parameters of the dust suppression equipment, and periodically incrementally training the dynamic dust diffusion prediction model. This invention can predict dust diffusion trends in advance based on multi-source information such as meteorological data, achieve collaborative control of dust suppression equipment, and adjust control parameters according to the dust suppression effect to alleviate control imbalance and meet the needs of intelligent dust control in ports.
[0006] Optionally, the multi-source data includes meteorological data, site cloud data, and regional dust data in the electronic shed of the open port yard. The meteorological data includes real-time meteorological data and weather forecast data, both of which include temperature, humidity, rainfall, wind speed, and wind direction.
[0007] Optionally, the dust suppression equipment includes at least a windbreak net, a spraying device, and a crusting agent spraying device.
[0008] Optionally, after acquiring the multi-source data, it is also necessary to preprocess the multi-source data. The preprocessing includes data cleaning, standardization, spatiotemporal alignment, point cloud registration, point cloud noise reduction, point cloud segmentation, and point cloud clustering. The spatiotemporal alignment involves dividing the open port yard electronic shed into rectangular grids, each of which contains the corresponding meteorological data and regional dust data, thereby generating a continuous spatiotemporal distribution map.
[0009] Optionally, the step of constructing a dynamic dust diffusion prediction model based on the multi-source data resource pool, and using the dynamic dust diffusion prediction model and correction factors to predict the dust concentration distribution in the electronic shed of an open port yard, includes the following steps: The meteorological data of the rectangular grid is used as input and the dust concentration is used as output. A prediction network is constructed using ConvLSTM with attention mechanism. The prediction network is trained and validated using meteorological data and dust concentration data from the multi-source data resource pool to obtain the dynamic dust diffusion prediction model. The dynamic dust diffusion prediction model is used to obtain preliminary predicted dust concentration values for each of the rectangular grids. The dust concentration prediction value of the rectangular grid is multiplied by the corresponding correction factor to obtain the dust concentration correction value of the rectangular grid, thereby obtaining the dust concentration distribution in the electronic shed of the open port yard.
[0010] Optionally, a digital twin model of the open port yard electronic shed is constructed. Based on meteorological data, weather forecast data, site cloud data, and regional dust data in the open port yard electronic shed, reliable dust concentration values of the rectangular grid are periodically obtained through CFD simulation, and the ratio of the reliable dust concentration value to the preliminary dust concentration prediction value is used as the correction factor.
[0011] Optionally, the step of setting multi-device collaborative control rules and calculating risk scores, and then combining the current and predicted dust concentration distribution to perform adaptive linkage control of the dust suppression equipment to achieve dust suppression, includes the following steps: Set rules for multi-device collaborative control; The dust concentration and meteorological data in the rectangular grid are normalized, and the risk score of the rectangular grid is calculated by weighted summation. The current dust concentration distribution is determined using the regional dust data, and the dust suppression equipment is adaptively linked and controlled according to the multi-device collaborative control rules and the risk score to achieve real-time dust suppression. Based on the predicted dust concentration distribution, the multi-device collaborative control rules, and the risk score, the dust suppression equipment is adaptively linked for control, thereby achieving early dust suppression.
[0012] Optionally, the multi-device collaborative control rules include: When a rectangular grid with a risk score exceeding the risk threshold exists, the dust suppression equipment is controlled sequentially according to its priority, from highest to lowest: the windbreak net, the spraying equipment, and the crusting agent spraying equipment. The windbreak net is raised when the wind speed consistently exceeds a first wind speed threshold within a preset time window, and lowered when the wind speed consistently falls below a second wind speed threshold within the preset time window. When more than 50% of the area has a dust concentration exceeding the dust concentration threshold, the spraying angle is adjusted according to the wind direction for full coverage spraying; otherwise, only key areas are sprayed. The crusting agent spraying equipment is activated when the humidity before spraying is below the humidity threshold.
[0013] Optionally, the real-time monitoring of dust suppression effect, adjustment of the control parameters of the dust suppression equipment, and periodic incremental training of the dynamic dust diffusion prediction model include the following steps: Continuously monitor dust concentration and calculate the risk score. If the risk score or dust concentration is still higher than the corresponding threshold, generate a dust reduction suggestion and push the dust reduction suggestion to relevant personnel. Based on the monitoring results and the dust suppression recommendations, relevant personnel adjusted the control parameters of the dust suppression equipment. The multi-source data is continuously collected to expand the multi-source data resource pool, and the dynamic dust diffusion prediction model is incrementally trained periodically.
[0014] Secondly, the present invention provides a port electronic greenhouse meteorological big data monitoring and joint computing system, comprising: a data acquisition device, a data output device, a processor, and a storage device. The storage device includes a computer-readable storage medium storing a computer program. The computer program includes program instructions, which, when executed by the processor, cause the processor to implement the port electronic greenhouse meteorological big data monitoring and joint computing method provided by the present invention.
[0015] This method has at least the following beneficial effects: 1. This method constructs a dynamic dust diffusion prediction model, and then uses meteorological data to predict the dust concentration distribution in the electronic shed of an open port yard. The method also obtains a correction factor through CFD simulation to correct the prediction results of the dynamic dust diffusion prediction model, thereby improving the real-time performance and accuracy of the prediction.
[0016] 2. This method achieves multi-source data fusion prediction through "dynamic dust diffusion prediction model + CFD simulation", which further improves the accuracy of dust concentration prediction.
[0017] 3. This method divides the electronic sheds in open port yards into grids. By collecting and predicting the dust concentration in different grids, it is possible to select key areas for spraying or full-coverage spraying to reduce water consumption.
[0018] 4. This method sets out multi-device collaborative control rules, calculates the risk scores of different areas of the electronic shed in the open port yard, and then combines the current and predicted dust concentrations to achieve adaptive linkage control of the dust suppression equipment.
[0019] 5. This method generates dust suppression suggestions by calculating dust suppression efficiency and pushes these suggestions to relevant personnel to adjust the control parameters of the dust suppression equipment and alleviate control imbalance.
[0020] 6. A system adapted to this method is provided, which not only improves the practicality of this method, but also facilitates its promotion. Attached Figure Description
[0021] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0022] Figure 1 This is a flowchart illustrating a method for monitoring and jointly processing meteorological big data in port electronic greenhouses, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the framework of a port electronic greenhouse meteorological big data monitoring and joint computing system according to an embodiment of the present invention. Detailed Implementation
[0023] Specific embodiments of the present invention will now be described in detail. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other instances, well-known circuits, software, or methods have not been specifically described to avoid obscuring the invention.
[0024] Throughout this specification, references to "an embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale.
[0025] It should be noted in advance that, in one alternative embodiment, except for independent descriptions, the same symbols or letters appearing in all formulas have the same meaning and value.
[0026] In one optional embodiment, please refer to Figure 1 This invention provides a method for monitoring and jointly processing meteorological big data of electronic greenhouses in ports, the method comprising the following steps: S1. Deploy a sensing network in the electronic shed of the open port yard, combine it with the meteorological forecast system to obtain multi-source data, and use the multi-source data to construct a multi-source data resource pool.
[0027] Specifically, in this embodiment, the open-air port yard electronic enclosure is a virtual enclosure for the open-air yard constructed using distributed laser electronic fence technology. By arranging multi-channel lasers around and on top of the yard, a "multi-dimensional spatial three-dimensional laser monitoring system" is formed, achieving an electronic enclosure effect for the open-air yard. The sensing network deployed in the open-air port yard electronic enclosure includes ultrasonic anemometers, temperature and humidity sensors, lidar, and... X-ray dust monitor.
[0028] The multi-source data in the electronic enclosure of the open port yard includes meteorological data from different areas, site cloud data, and regional dust data. The meteorological data includes real-time meteorological data and weather forecast data, both of which include temperature, humidity, rainfall, wind speed, and wind direction. Specifically, wind speed and direction are collected in real-time using ultrasonic anemometers; temperature and humidity are collected in real-time using high-precision temperature and humidity sensors; rainfall is collected in real-time using rain gauges, typically assuming the same rainfall across different areas within the same open port yard electronic enclosure; site cloud data is obtained by scanning the area where the electronic enclosure is located using lidar; and... The X-ray dust monitor collects real-time dust data for the area, specifically the dust concentration in different areas; it also obtains weather forecast data for the next hour from the port's weather forecast system.
[0029] After collecting multi-source data from the electronic sheds in the open port yard, the collected data needs to be transmitted to a cloud platform via the Internet of Things (IoT) for preprocessing operations such as data cleaning, standardization, spatiotemporal alignment, point cloud registration, point cloud denoising, point cloud segmentation, and point cloud clustering. Data cleaning, standardization, and spatiotemporal alignment are mainly for meteorological data and regional dust data, while point cloud registration, point cloud denoising, point cloud segmentation, and point cloud clustering are mainly for site-specific point cloud data. Among these methods, missing value filling is achieved through linear interpolation; outlier identification is achieved through quartile range method; the open port yard electronic shed is divided into rectangular grids with a length and width of 4m, and the rectangular grids are assigned position codes according to their positions in rows and columns, such as (1,1) representing the rectangular grid in the first row and first column. Then, Kriging interpolation is used to spatially interpolate the multi-source data collected at the same time to generate a continuous spatiotemporal distribution map, realizing the spatiotemporal alignment of meteorological data and regional dust data; point clouds scanned by different lidars are unified into the same coordinate system to achieve point cloud registration; point cloud noise reduction is achieved through voxel grid filtering; point cloud segmentation is achieved through random sampling consensus algorithm; and point cloud clustering is achieved through Euclidean clustering method.
[0030] Finally, the preprocessed multi-source data is stored in a database to form a multi-source data resource pool. Meteorological data, dust concentration, and site cloud data of different rectangular grids are obtained through this multi-source data resource pool.
[0031] It should be noted that in this embodiment, the dust concentration within 1 meter above each grid surface is used as the dust concentration of that grid, so height is not considered. Furthermore, the grid surface here refers to the actual surface of material accumulation within the grid, rather than being fixed at ground level. In other alternative embodiments, a height attribute can also be assigned to the grid.
[0032] S2. Construct a dynamic dust diffusion prediction model based on the multi-source data resource pool, and use the dynamic dust diffusion prediction model and correction factor to predict the dust concentration distribution in the electronic shed of the open port yard.
[0033] Step S2 specifically includes the following steps: S21. Using the meteorological data of the rectangular grid as input and the dust concentration as output, a prediction network is constructed using ConvLSTM with attention mechanism.
[0034] Specifically, in this embodiment, a Convolutional Long Short-Term Memory (ConvLSTM) network with an attention mechanism is used to construct the prediction network. This prediction network consists of an input projection layer, two stacked ConvLSTM layers, a temporal attention layer connected to the last ConvLSTM layer, and an output prediction layer. Its input is a four-dimensional tensor with the shape (T, H, W, C), where T is the length of the input sequence, H is the number of rows in the rectangular grid, W is the number of columns in the rectangular grid, and C is the number of features in the rectangular grid. As shown in step S1, each rectangular grid has six features: temperature, humidity, rainfall, wind speed, wind direction, and dust concentration.
[0035] More specifically, the input projection layer of the prediction network uses a 3×3 convolutional kernel with a stride of 1 and padding of 0, mapping the number of feature channels in the input data from 6 to 64 dimensions to enhance the network's feature representation capability. Subsequently, the output of the input projection layer is fed into two stacked ConvLSTM layers to extract complex spatiotemporal features from the input sequence. Each ConvLSTM layer uses a 3×3 convolutional kernel with a stride of 1 and padding of 0, maintaining 64 hidden state channels. After the second ConvLSTM layer outputs the hidden states for all time steps, a temporal attention layer is used to calculate the importance weight of the hidden state at each time step. Then, the hidden states at all time steps are summed according to their weights to obtain a context vector representing the key information of the entire sequence. Finally, this context vector is output to the prediction layer, which consists of a 3×3 convolutional kernel with a stride of 1 and padding of 0, mapping the 64-channel feature map to a single channel, i.e., the dust concentration prediction result for each rectangular grid. The entire prediction network is trained by minimizing the mean squared error between the predicted dust concentration and the actual dust concentration.
[0036] S22. Using the meteorological data and dust concentration data in the multi-source data resource pool, the prediction network is trained and validated to obtain the dynamic dust diffusion prediction model.
[0037] Specifically, in this embodiment, the time step is set to 1 minute and the time series length is 90. Based on the set time step and time series length, meteorological data and dust concentration data from a multi-source data resource pool are used to construct training and validation sets, with a data ratio of 7:3. The constructed training and validation sets are used to train and validate the prediction network, resulting in a dynamic dust diffusion prediction model.
[0038] It should be noted that the multi-source data resource pool also includes weather forecast data, but it does not contain regional dust data corresponding to the weather forecast data. Therefore, weather forecast data cannot be used when constructing the training and validation sets; it will only be used when predicting dust concentrations.
[0039] Furthermore, since the acquired meteorological forecast data covers multiple future time steps, a rolling forecasting strategy can be employed to predict dust concentrations for multiple time steps within a relatively short period. For example, after predicting the dust concentration for the first future time step, the predicted dust concentration for that time step can be used as a known quantity to predict the dust concentration for the second future time step.
[0040] S23. Use the dynamic dust diffusion prediction model to obtain the preliminary predicted dust concentration values for each of the rectangular grids.
[0041] Specifically, in this embodiment, the real-time meteorological data and dust concentration of each rectangular grid over the last 90 time steps are input into the dynamic dust diffusion prediction model to obtain the dust concentration of each rectangular grid one minute later. By combining the meteorological forecast data and the rolling prediction strategy, the dust concentration of each rectangular grid in the future multiple time steps can be obtained, and the prediction result is recorded as the preliminary predicted value of dust concentration.
[0042] S24. Multiply the preliminary predicted dust concentration value of the rectangular grid with the corresponding correction factor to obtain the corrected dust concentration value of the rectangular grid, and then obtain the dust concentration distribution in the electronic shed of the open port yard.
[0043] Specifically, in this embodiment, the dynamic dust diffusion prediction model built on ConvLSTM is a purely data-driven model, adept at learning and predicting spatiotemporal sequence data. Once the model is trained, a forward prediction typically takes only a few seconds or even milliseconds, meeting real-time requirements. However, the dynamic dust diffusion prediction model does not understand physics; it merely fits the data. It may learn some spurious correlations in the data, thus making predictions that violate physical laws and accumulating errors during the prediction process. In contrast, computational fluid dynamics (CFD) simulations simulate the physical world by solving complex fluid dynamics and particulate motion equations, offering significant advantages in accuracy. However, its high computational cost and long processing time make it unsuitable for scenarios with high real-time requirements.
[0044] To alleviate this problem, this embodiment first constructs a digital twin model of the open port yard electronic enclosure. Then, based on meteorological data, weather forecast data, site cloud data, and regional dust data from the open port yard electronic enclosure, reliable dust concentration values for the rectangular grid are obtained every half month through CFD simulation. The ratio of the reliable dust concentration value of the rectangular grid at the same time to the preliminary dust concentration prediction value is used as a correction factor. Finally, the preliminary dust concentration prediction value of the rectangular grid is multiplied by the corresponding correction factor to obtain the corrected dust concentration value of the rectangular grid, thereby obtaining the dust concentration distribution in the open port yard electronic enclosure. The construction of the digital twin model and CFD simulation of the open port yard electronic enclosure are existing technologies.
[0045] This embodiment employs a hybrid prediction method combining a physical model and a data-driven model, enabling multi-source data fusion prediction. This not only improves the real-time performance and accuracy of dust concentration prediction but also reduces prediction costs to some extent. Furthermore, it allows for the acquisition of correction factors for rectangular grids at different time steps at the same moment, calculation of their average value, and use of this average value as the final correction factor for the rectangular grid, further enhancing prediction accuracy.
[0046] S3. Set multi-device collaborative control rules and calculate risk scores, and then combine the current and predicted dust concentration distribution to carry out adaptive linkage control of dust suppression equipment to achieve dust suppression.
[0047] The dust suppression equipment, also known as dust control equipment, includes at least windbreak nets, spraying equipment, and crust-forming agent spraying equipment. Specifically, the windbreak nets are liftable windbreak nets. Step S3 includes the following steps: S31. Set multi-device collaborative control rules.
[0048] Specifically, in this embodiment, the multi-device collaborative control rules are as follows: When there is a rectangular grid where the current or future risk score exceeds the risk threshold, the dust suppression equipment is controlled sequentially according to its priority, with the priority from high to low being the windbreak net, the spraying equipment, and the crusting agent spraying equipment; when the wind speed exceeds the first wind speed threshold within a preset time window, the windbreak net is raised; when the wind speed remains below the second wind speed threshold within the preset time window, the windbreak net is lowered; when the dust concentration exceeds the dust concentration threshold, the spraying angle is adjusted according to the wind direction for full coverage spraying; otherwise, only key areas are sprayed; when the humidity before spraying is lower than the humidity threshold, the crusting agent spraying equipment is activated.
[0049] More specifically, when a rectangular grid with a risk score exceeding 0.8 exists, dust suppression equipment is controlled sequentially according to equipment priority; the windbreak net is raised when the wind speed continuously exceeds 6 m / s for the next 10 minutes, and lowered when the wind speed continuously falls below 3 m / s for the next 30 minutes; when the dust concentration in more than 50% of the area exceeds... When the wind direction is high, adjust the spray angle to fully cover the stockpile; otherwise, only spray areas with dust concentrations exceeding a certain threshold. Spraying is carried out on the area containing the rectangular grid; when the humidity before spraying is below 50%, the crusting agent spraying equipment is started. Of course, relevant personnel can adjust parameters such as the time window, the first wind speed threshold, the second wind speed threshold, and the humidity threshold according to actual operational needs.
[0050] S32. Normalize the dust concentration and meteorological data in the rectangular grid, and calculate the risk score of the rectangular grid by weighted summation.
[0051] Specifically, in this embodiment, the maximum-minimum normalization method is used to normalize the dust concentration, temperature, reciprocal of humidity, reciprocal of rainfall, wind speed, and wind direction in the rectangular grid. Then, the current or future risk score of the rectangular grid is calculated according to the following formula: Where P is the risk score of the rectangular grid; and These represent the concentration weight and meteorological weight, respectively, with values of 0.7 and 0.3; c is the normalized dust concentration, used when calculating the current risk score of the rectangular grid, and the normalized dust concentration correction value when calculating the future risk score; t is the normalized temperature, used when calculating the current risk score of the rectangular grid, and the temperature in the meteorological forecast data when calculating the future risk score; h is the reciprocal of the normalized humidity, its value is determined by referring to the normalized temperature; V is the reciprocal of the normalized rainfall, its value is determined by referring to the normalized temperature; v is the normalized wind speed, its value is determined by referring to the normalized temperature; d is the normalized wind direction, its value is determined by referring to the normalized temperature. , , , and The weights for temperature, humidity, rainfall, wind speed, and wind direction are, in order. The Pearson correlation coefficients between temperature, humidity, rainfall, wind speed, and wind direction and dust concentration can be calculated using a sliding window, and then the Pearson correlation coefficients can be normalized. The sliding window size is 15 days.
[0052] S33. Determine the current dust concentration distribution using the regional dust data, and perform adaptive linkage control of the dust suppression equipment based on the multi-device collaborative control rules and the risk score to achieve real-time dust suppression.
[0053] S34. Based on the predicted dust concentration distribution, the multi-device collaborative control rules, and the risk score, the dust suppression equipment is subjected to adaptive linkage control to achieve early dust suppression.
[0054] S4. Monitor the dust suppression effect in real time, adjust the control parameters of the dust suppression equipment, and periodically perform incremental training on the dynamic dust diffusion prediction model.
[0055] Step S4 specifically includes the following steps: S41. Continuously monitor dust concentration and calculate the risk score. If the risk score or dust concentration is still higher than the corresponding threshold, generate a dust reduction suggestion and push the dust reduction suggestion to relevant personnel.
[0056] Specifically, in this embodiment, dust concentration is continuously monitored and the risk score is calculated. Ten minutes after dust suppression, if either the risk score or the dust concentration remains above the corresponding threshold, a dust suppression recommendation is generated in the cloud. This recommendation includes at least increasing the water spray volume, decreasing the first wind speed threshold, increasing the second wind speed threshold, and decreasing the humidity threshold.
[0057] S42. Based on the monitoring results and the dust reduction recommendations, relevant personnel adjust the control parameters of the dust reduction equipment.
[0058] Specifically, in this embodiment, relevant personnel adjust the control parameters of the dust suppression equipment based on monitoring results and the dust suppression recommendations, combined with their own experience. For example, they may increase the water spray volume, reduce the first wind speed threshold, increase the second wind speed threshold, reduce the humidity threshold, and increase the spray pressure.
[0059] S43. Continuously collect the multi-source data to expand the multi-source data resource pool, and periodically perform incremental training on the dynamic dust diffusion prediction model.
[0060] Specifically, in this embodiment, the multi-source data is continuously collected to expand the multi-source data resource pool, and then the training set and validation set are rebuilt every quarter to retrain and validate the dynamic dust diffusion prediction model, thereby improving the accuracy and reliability of the dynamic dust diffusion prediction model.
[0061] It should be noted that in some cases, the actions described in the specification can be performed in different orders and still achieve the desired results. In this embodiment, the order of steps is given only to make the embodiment clearer and easier to explain, and not to limit it.
[0062] In one optional embodiment, please refer to Figure 2 To improve the practicality and facilitate the promotion of this method, the present invention also provides a port electronic greenhouse meteorological big data monitoring and joint operation system. This system includes: a data acquisition device 1, a data output device 2, a processor 3, and a storage device 4. The storage device 4 includes a computer-readable storage medium storing a computer program. The computer program includes program instructions, which, when executed by the processor 3, enable the processor 3 to implement the port electronic greenhouse meteorological big data monitoring and joint operation method provided by the present invention.
[0063] In summary, this method offers at least the following advantages: First, it constructs a dynamic dust diffusion prediction model, then uses meteorological data to predict dust concentration distribution in electronic sheds of open port storage yards. Second, it uses CFD simulation to obtain correction factors to correct the prediction results of the dynamic dust diffusion prediction model, forming a hybrid prediction method combining a physical model and a data-driven model. This improves the real-time performance and accuracy of the predictions and reduces prediction costs to some extent. Third, this method achieves multi-source data fusion prediction by using a combination of obtaining preliminary dust concentration predictions from the dynamic dust diffusion prediction model and using CFD simulation to obtain correction factors to correct these preliminary predictions. Furthermore, it periodically performs incremental training on the dynamic dust diffusion prediction model, further improving the accuracy of dust concentration prediction. This method features several advantages: First, it divides the open port yard electronic shed into grids. By collecting and predicting dust concentrations in different grids, it allows for selective spraying of key areas or full coverage, reducing water consumption. Second, it establishes multi-device collaborative control rules, calculating risk scores for different areas of the open port yard electronic shed using dust concentration and meteorological data. Then, by combining the current and predicted dust concentrations in different areas, it achieves adaptive linkage control of the dust suppression equipment. Third, it generates dust suppression recommendations by calculating dust suppression efficiency and pushes these recommendations to relevant personnel to adjust the control parameters of the dust suppression equipment, mitigating control imbalances. Finally, it provides a system adapted to this method, which not only improves its practicality but also facilitates its widespread adoption.
[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.
Claims
1. A port electronic greenhouse weather big data monitoring and joint operation method, characterized in that, The method comprises the following steps: Deploy a sensing network in the open port yard electronic greenhouse, obtain multi-source data in combination with a weather forecasting system, and use the multi-source data to build a multi-source data resource pool; Build a dynamic dust diffusion prediction model based on the multi-source data resource pool, and use the dynamic dust diffusion prediction model and a correction factor to predict the dust concentration distribution in the open port yard electronic greenhouse; Set multi-device cooperative control rules and calculate a risk score, and then combine the current and predicted dust concentration distribution to perform adaptive linkage control on dust-settling devices to achieve dust settling; Real-time monitor the dust settling effect, and then adjust the control parameters of the dust-settling devices, and periodically perform incremental training on the dynamic dust diffusion prediction model.
2. The method according to claim 1, wherein the multi-source data comprises weather data, site point cloud data, and regional dust data in the open port yard electronic greenhouse, the weather data comprises real-time weather data and weather forecast data, and the real-time weather data and the weather forecast data each comprises temperature, humidity, rainfall, wind speed, and wind direction.
3. The method according to claim 2, wherein the dust-settling devices at least comprise a windproof net, a spraying device, and a crust agent spraying device.
4. The method according to claim 2, wherein after obtaining the multi-source data, the multi-source data needs to be preprocessed, the preprocessing comprises data cleaning, standardization, space-time alignment, point cloud registration, point cloud noise reduction, point cloud segmentation, and point cloud clustering; the space-time alignment is to divide the open port yard electronic greenhouse into rectangular grids, each rectangular grid contains corresponding weather data and regional dust data, and then a continuous space-time distribution map is generated. The method of building a dynamic dust diffusion prediction model based on the multi-source data resource pool and using the dynamic dust diffusion prediction model and a correction factor to predict the dust concentration distribution in the open port yard electronic greenhouse comprises the following steps: Taking the weather data of the rectangular grid as input and the dust concentration as output, a ConvLSTM with attention mechanism is used to build a prediction network; Using the weather data and dust concentration data in the multi-source data resource pool to complete the training and verification of the prediction network, the dynamic dust diffusion prediction model is obtained; 5.The port electronic greenhouse weather big data monitoring and joint operation method according to claim 4, characterized in that, Using the dynamic dust diffusion prediction model to obtain the preliminary prediction value of the dust concentration of each rectangular grid; Multiplying the preliminary prediction value of the dust concentration of the rectangular grid by the corresponding correction factor to obtain the dust concentration correction value of the rectangular grid, and then obtaining the dust concentration distribution in the open port yard electronic greenhouse.
6. The method according to claim 5, wherein The digital twin model of the open port yard electronic greenhouse is constructed, and according to meteorological data, meteorological forecast data, site cloud data and regional dust data in the open port yard electronic greenhouse, reliable values of dust concentration of the rectangular grid are obtained periodically through CFD simulation, and a ratio of the reliable values of dust concentration to the preliminary predicted values of dust concentration is taken as the correction factor. 7.The port electronic greenhouse weather big data monitoring and joint operation method according to claim 4, characterized in that, The setting of the multi-device cooperative control rule and the calculation of the risk score, and then the adaptive linkage control of the dust-settling device based on the current and predicted dust concentration distribution to realize dust settling, including the following steps: Setting a multi-device cooperative control rule; The dust concentration and meteorological data in the rectangular grid are normalized, and the risk score of the rectangular grid is calculated by weighted summation; The current dust concentration distribution is determined by using the regional dust data, and the adaptive linkage control of the dust-settling device is realized according to the multi-device cooperative control rule and the risk score to realize real-time dust settling; According to the predicted dust concentration distribution, the multi-device cooperative control rule and the risk score, the adaptive linkage control of the dust-settling device is realized to realize early dust settling. 8.The port electronic greenhouse weather big data monitoring and joint operation method according to claim 3, characterized in that, The multi-device cooperative control rule includes: When there is a rectangular grid with a risk score exceeding a risk threshold, the dust-settling devices are controlled in turn according to the device priority, and the device priority from high to low is the windbreak, the spraying device and the crust agent spraying device; when the wind speed continuously exceeds the first wind speed threshold within a preset time window, the windbreak is raised, and when the wind speed continuously falls below the second wind speed threshold within a preset time window, the windbreak is lowered; when more than 50% of the area has a dust concentration exceeding a dust concentration threshold, full-coverage spraying is performed by adjusting the spraying angle according to the wind direction, otherwise only key area spraying is performed; when the humidity before spraying is lower than a humidity threshold, the crust agent spraying device is started. 9.The port electronic greenhouse weather big data monitoring and joint operation method according to claim 1, characterized in that, The real-time monitoring of dust settling effect, the adjustment of the control parameters of the dust-settling device, and the incremental training of the dynamic dust diffusion prediction model at regular intervals, including the following steps: Continuously monitor the dust concentration and calculate the risk score, and if the risk score or dust concentration is still higher than the corresponding threshold, generate a dust-settling suggestion and push the dust-settling suggestion to relevant personnel; Based on the monitoring results and the dust-settling suggestion, the relevant personnel adjust the control parameters of the dust-settling device; Continuously collect the multi-source data to expand the multi-source data resource pool, and periodically perform incremental training on the dynamic dust diffusion prediction model.
10. A port electronic greenhouse weather big data monitoring and joint operation system, characterized in that, It includes: Data acquisition device, data output device, processor and storage, the storage includes a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program includes program instructions, the program instructions are executed by the processor to realize the method of port electronic greenhouse weather big data monitoring and joint operation as claimed in any one of claims 1-9.