Rainwater and atmospheric water collaborative collection and treatment method, system, device and medium

Abstract

This invention relates to the field of water resource collection and treatment technology, and provides a method, system, equipment, and medium for the coordinated collection and treatment of rainwater and atmospheric water. The method includes: acquiring environmental data, a time-series sequence of raindrop size distribution, and a raindrop velocity sequence; inputting the time-series sequence of raindrop size distribution and the raindrop velocity sequence into a regression neural network trained on micro-rainfall calibration data, and outputting an effective rainfall intensity correction value; inputting the effective rainfall intensity correction value and environmental data into a classification model, and outputting operation instructions, which include at least one of a rainwater priority mode, an atmospheric water production mode, or a synergistic enhancement mode. This invention can accurately quantify minute rainfall amounts, avoiding resource loss or energy waste due to misjudgment, and realizes intelligent coordinated collection and dynamic allocation of rainwater and atmospheric water based on real-time environmental perception, significantly improving the system's total water production rate and reducing energy consumption per unit of water production.

Description

Technical Field

[0001] This invention relates to the field of water resource collection and water treatment technology, and in particular to methods, systems, equipment and media for the coordinated collection and treatment of rainwater and atmospheric water. Background Technology

[0002] Currently, coastal saline-alkali areas, especially those reclaimed from the sea, generally face a severe shortage of freshwater resources. Traditional water supply methods, such as long-distance pipeline transportation and seawater desalination, suffer from high infrastructure construction costs, high energy consumption, and maintenance difficulties. Meanwhile, these areas often have relatively abundant rainfall and high air humidity; rainwater and atmospheric water vapor represent two unconventional freshwater resources with enormous potential but which have not yet been fully utilized. Therefore, how to efficiently collect rainwater and atmospheric water in a site-specific manner and treat it on-site into high-quality freshwater has become an important research direction for ensuring water resource security in coastal areas.

[0003] However, existing collection and processing methods cannot intelligently switch water source collection modes based on real-time environmental conditions. Rainwater and atmospheric water generation are highly unstable in time and space; periods without rain or fog may occur, or both may occur simultaneously. Current methods cannot automatically switch or combine operating modes based on real-time meteorological conditions (such as rainfall intensity, humidity, and wind speed). This leads to system idleness and waste when a single water source is scarce, and an inability to achieve complementary production capacity when both sources are abundant. Furthermore, the system cannot accurately determine the current environment when rain and fog are mixed; it may determine "no rain" and close the rainwater collection port, causing light rain to be lost; or it may determine "rain" and stop the efficient condensate collection equipment, resulting in energy waste and a decrease in water production rate. Therefore, how to achieve intelligent collaborative collection and dynamic allocation of rainwater and atmospheric water based on real-time environmental perception is a pressing technical problem that needs to be solved in this field. Summary of the Invention

[0004] This invention provides a method, system, device, and medium for the coordinated collection and treatment of rainwater and atmospheric water, in order to solve the technical problem that existing technologies cannot achieve intelligent coordinated collection and dynamic allocation of rainwater and atmospheric water based on real-time environmental perception.

[0005] This invention provides a method for the coordinated collection and treatment of rainwater and atmospheric water, characterized in that the method is applied to an integrated system, the integrated system including a rainwater collection unit, an atmospheric water condensation unit, a water quality treatment unit, and a central control module; the method includes: Acquire environmental data, time-series sequences of raindrop size distribution, and raindrop velocity sequences; The raindrop size distribution time series and raindrop falling velocity series are input into a regression neural network trained with micro-rainfall calibration data, and the effective rainfall intensity correction value is output. The effective rainfall intensity correction value and environmental data are input into the classification model, and an operation command is output. The operation command includes at least one of the following three modes: Rainwater priority mode: Enable rainwater collection unit; Atmospheric water generation mode: Rainwater collection unit is turned off; Synergistic Enhancement Mode: Simultaneously activate the rainwater harvesting unit and the atmospheric water condensation unit.

[0006] In some embodiments, the regression neural network is a long short-term memory network, whose input is the time sequence of raindrop size distribution and raindrop falling velocity within a continuous time window, and whose output is the effective rainfall intensity correction value at the current moment; the micro-rainfall calibration data includes: artificially simulated drizzle and micro-rainfall event data collected on site, with reference rainfall intensity measured by a high-precision weighing rain gauge as the training label.

[0007] In some embodiments, the classification model employs the XGBoost algorithm, and the decision function of the XGBoost algorithm is: , Where K is the number of regression trees. Let be the predicted score of the Kth regression tree for category c, calculated using the following formula: , in, Let be the total number of leaf nodes in the k-th regression tree; Let t be the weight value of the t-th leaf node in the k-th regression tree corresponding to category c; Let t be the feature space region corresponding to the t-th leaf node in the k-th regression tree; As an indicator function, when the input feature vector x Falling area The value is 1 if the condition is met, and 0 otherwise. Input feature vector , This is a correction value for effective rainfall intensity; RH represents air humidity. For ambient temperature, Here, WS represents the dew point temperature and wind speed. The input feature vector x is composed of the effective rainfall intensity correction value and environmental data; Output rain priority mode at the time. In real-time output atmospheric water production mode, Time-based output synergistic efficiency mode.

[0008] In some embodiments, the input features of the XGBoost algorithm further include: the turbidity (Turb) value and conductivity (EC) value of the rainwater collection unit inlet, and the estimated dust thickness (Dust) value of the condensation surface of the atmospheric water condensation unit. At this point, the input feature vector x is expanded to: .

[0009] In some embodiments, the preset weights of the classification model are obtained through offline supervised learning. The training samples are historical meteorological and water production records of the coastal area. Each sample includes an effective rainfall intensity correction value, air humidity, ambient temperature, dew point temperature, wind speed, and the corresponding optimal operating mode label. The operating mode that maximizes the total water production rate of the system or minimizes the unit energy consumption is selected as the optimal operating mode label. The training objective is to minimize the cross-entropy loss between the model output mode and the label.

[0010] In some embodiments, the method further includes: An initial runoff diversion device is installed at the inlet of the rainwater collection unit to automatically discharge the initial runoff of the first 10-15 mm based on the salinity detected by the chloride ion sensor. Apply an anti-salt spray corrosion coating to the condensation surface of the atmospheric water condensation unit and periodically activate the ultrasonic descaling device; Every two weeks, the water treatment unit undergoes combined air-water backwashing, and the backwash wastewater is either discharged externally or returned to the sedimentation area of ​​the rainwater collection unit.

[0011] In some embodiments, low-temperature rainwater collected by the rainwater harvesting unit is introduced into a heat exchanger to pre-cool the condensation surface of the atmospheric water condensation unit in order to improve condensation efficiency.

[0012] The present invention also provides a rainwater and atmospheric water co-collection and treatment system, characterized in that the system is used to execute the rainwater and atmospheric water co-collection and treatment method described in any of the above technical solutions, and the system includes a rainwater collection unit, an atmospheric water condensation unit, a water quality treatment unit and a central control module; Raindrop sensors are used to collect time-series sequences of raindrop size distribution and raindrop falling velocity. A meteorological sensor array is used to collect environmental data, including air humidity, ambient temperature, dew point temperature, and wind speed. The central control module integrates a regression neural network and a classification model; The central control module receives the raindrop size distribution time series and raindrop falling velocity series collected by the raindrop sensor, and inputs them into the regression neural network to output an effective rainfall intensity correction value; the central control module also inputs the effective rainfall intensity correction value and the environmental data collected by the meteorological sensor group into the classification model and outputs operation instructions.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the steps of any of the above-described methods for the coordinated collection and treatment of rainwater and atmospheric water.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the rainwater and atmospheric water co-collection and treatment method described in any of the above technical solutions.

[0015] The present invention provides a method, system, device, and medium for the coordinated collection and treatment of rainwater and atmospheric water. It utilizes raindrop sensors to collect time-series sequences of raindrop particle size distribution and falling velocity, inputting these sequences into a regression neural network trained with micro-rainfall calibration data. The network outputs an effective rainfall intensity correction value, thereby accurately quantifying the amount of light rainfall that traditional rain gauges cannot respond to under drizzle or mixed rain and fog conditions. This avoids the loss of light rain due to misjudging "no rain" and closing the rainwater collection port, or the energy waste caused by stopping the condensation equipment due to misjudging "rain." Secondly, the effective rainfall intensity correction value and environmental data are input into a classification model, outputting three operating commands: rainwater priority mode, atmospheric water production mode, or synergistic enhancement mode. This achieves intelligent coordinated collection and dynamic allocation of rainwater and atmospheric water based on real-time environmental perception, significantly improving the system's total water production rate and reducing energy consumption per unit of water production. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the method for the coordinated collection and treatment of rainwater and atmospheric water provided by the present invention. Figure 2 This is a schematic diagram of the framework of the rainwater and atmospheric water co-collection and treatment system provided by the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] Example 1 like Figure 1 As shown in the figure, this embodiment provides a method for the coordinated collection and treatment of rainwater and atmospheric water. This method is applied to an integrated system, which includes a rainwater collection unit, an atmospheric water condensation unit, a water quality treatment unit, and a central control module.

[0020] Specifically, the rainwater harvesting unit is responsible for collecting natural precipitation, and its inlet is connected to the water treatment unit via a pipeline; the atmospheric water condensation unit condenses water vapor in the air using a cooling surface, and its outlet is also connected to the water treatment unit; both the rainwater harvesting unit and the atmospheric water condensation unit are electrically connected to the central control module. It should be understood that although this embodiment describes physically separate units, in practical engineering applications, the rainwater harvesting unit and the atmospheric water condensation unit can also be designed as an integrated structure, for example, by installing a condensation component on the back of the rainwater collection surface to save space and utilize rainwater to assist in cooling the condensation surface.

[0021] The method specifically includes the following steps: Step S100: Obtain environmental data, raindrop size distribution time series, and raindrop falling velocity series.

[0022] Specifically, the acquisition of environmental data includes: using a temperature and humidity sensor to collect air humidity (RH) and ambient temperature, using an anemometer to collect wind speed (WS), using a dew point calculation module or dew point sensor to obtain dew point temperature, and using a raindrop spectrometer to record the number of raindrops, particle size and falling velocity in the sampling area, so as to obtain the time sequence of raindrop particle size distribution and the raindrop falling velocity sequence.

[0023] Step S200: Input the raindrop size distribution time series and raindrop falling velocity series into the regression neural network trained by micro-rainfall calibration data, and output the effective rainfall intensity correction value.

[0024] Specifically, the regression neural network is deployed in the processor of the central control module. Because raindrops in light drizzle (such as drizzle) are small in size, slow in velocity, and sparsely distributed, directly calculating rainfall intensity through simple volume accumulation often results in significant errors. The regression neural network, by learning from a large amount of drizzle calibration data, establishes a nonlinear mapping relationship between the time-series sequences of droplet size distribution and droplet velocity and the actual rainfall intensity. For example, when the input sequence shows a large number of small-diameter, low-velocity raindrops appearing in a short period, the regression neural network can identify this as a typical light drizzle event and output a corrected rainfall intensity value, thus preventing water loss caused by the system mistakenly closing rainwater collection inlets due to a misjudgment of "no rain."

[0025] Step S300: Input the effective rainfall intensity correction value and environmental data into the classification model, and output the operation command. The operation command includes at least one of the following three modes: Rainwater priority mode: enable the rainwater collection unit; Atmospheric water production mode: disable the rainwater collection unit; Synergistic effect mode: enable both the rainwater collection unit and the atmospheric water condensation unit.

[0026] Specifically, the switching logic for the three modes is as follows: Rainwater Priority Mode: When the effective rainfall intensity correction value is ≥2.0 mm / h, it indicates that the rainfall water production efficiency is high, and the system prioritizes the activation of the rainwater collection unit to collect rainwater at full speed. At this time, the atmospheric water condensation unit can be suspended to save energy.

[0027] Atmospheric water production mode: When the effective rainfall intensity correction value is <0.1mm / h, it indicates that there is no rain or only very weak rainfall, but the air moisture content is high. The system shuts down the rainwater collection unit to avoid invalid operation and runs the atmospheric water condensation unit at full capacity to produce water through condensation.

[0028] Synergistic Enhancement Mode: When the effective rainfall intensity correction value is between 0.1 mm / h and 2.0 mm / h, the water production efficiency of a single water source has not reached its peak. At this time, the system simultaneously activates the rainwater harvesting unit and the atmospheric water condensation unit. In this mode, not only is simultaneous water production from two water sources achieved, but the low-temperature rainwater collected by the rainwater harvesting unit can also serve as a cold source, helping to reduce the temperature of the condensation surface of the atmospheric water condensation unit, thereby improving the condensation efficiency.

[0029] By utilizing raindrop sensors to collect time-series sequences of raindrop size distribution and falling velocity, and inputting these sequences into a regression neural network trained with micro-rainfall calibration data, the system outputs an effective rainfall intensity correction value. This allows for the accurate quantification of minute rainfall amounts that traditional rain gauges cannot respond to under drizzle or mixed rain and fog conditions. This avoids the loss of minute rainfall due to the closure of rainwater collection ports caused by misjudging "no rain" or the energy waste caused by stopping condensation equipment due to misjudging "rain". Secondly, the effective rainfall intensity correction value and environmental data are input into a classification model, which outputs three operating commands: rainwater priority mode, atmospheric water production mode, or synergistic efficiency mode. This enables intelligent collaborative collection and dynamic allocation of rainwater and atmospheric water based on real-time environmental perception, significantly improving the system's total water production rate and reducing energy consumption per unit of water production.

[0030] Example 2 Based on Example 1, this example provides a detailed description of the specific structure of the regression neural network and the process of acquiring its training data. Specifically, the regression neural network in this example uses a Long Short-Term Memory (LSTM) network. The structure of an LSTM network mainly includes an input layer, a hidden layer, and an output layer.

[0031] For example, with a time window length of 10 minutes and a sampling frequency of 1 minute, the input layer receives a 10-dimensional feature sequence, where each dimension contains the statistical distribution of raindrop particle size and average falling velocity within that minute; the hidden layer contains several LSTM units, each of which controls the flow of information and the update of the state through forget gates, input gates, and output gates, thereby capturing the dynamic pattern of raindrop physical characteristics changing over time; the output layer is responsible for mapping the high-dimensional features extracted by the hidden layer to a specific numerical value, which serves as the effective rainfall intensity correction value for the current moment.

[0032] Drizzle may go through a cycle of onset, intensification, weakening, and cessation, with strong temporal correlations in the changes in raindrop size and velocity. LSTM networks can memorize the raindrop states at historical moments, thus inferring the true rainfall intensity based on historical trends even when the raindrop signal is extremely weak or even briefly interrupted at the current moment, avoiding misjudgments caused by the lack of instantaneous signals in traditional algorithms.

[0033] During calibration, the time-series sequences of raindrop size distribution and raindrop falling velocity collected by the raindrop sensor are used as input features, and the reference rainfall intensity measured by a high-precision weighing rain gauge within the same time period is used as training labels. By minimizing the mean square error between the network output value and the label value, the weight parameters of the LSTM network are continuously adjusted using the backpropagation algorithm. After training with a large number of samples, the network learns to extract effective information from the complex raindrop time-series features, thereby correcting the errors of traditional measurement methods. For example, when the raindrop sensor misses a raindrop due to its small size, the LSTM network can output a non-zero effective rainfall intensity correction value based on the raindrop trend at previous moments and implicit information such as the current ambient humidity, thus achieving accurate quantification of micro-rainfall.

[0034] Example 3 Based on the above embodiments, the classification model adopts the XGBoost algorithm. The decision function of the XGBoost algorithm is: , Where K is the number of regression trees. Let be the predicted score of the Kth regression tree for category c, calculated using the following formula: , in, Let be the total number of leaf nodes in the k-th regression tree; Let t be the weight value of the t-th leaf node in the k-th regression tree corresponding to category c; Let t be the feature space region corresponding to the t-th leaf node in the k-th regression tree; As an indicator function, when the input feature vector x Falling area The value is 1 if the condition is met, and 0 otherwise. Input feature vector , This is a correction value for effective rainfall intensity; RH represents air humidity. For ambient temperature, Here, WS represents the dew point temperature and wind speed. The input feature vector x is composed of the effective rainfall intensity correction value and environmental data; Output rain priority mode at the time. In real-time output atmospheric water production mode, Time-based output synergistic efficiency mode.

[0035] To further improve the accuracy of decision-making and make it more closely reflect actual operating conditions, the input features of the XGBoost algorithm can be further expanded. As a preferred implementation, the input features also include: the turbidity value and conductivity (EC) value of the rainwater collection unit inlet, and the estimated dust accumulation thickness on the condensation surface of the atmospheric water condensation unit. In this case, the input feature vector x is expanded to: .

[0036] Turbidity and conductivity can reflect rainwater quality in real time. If the initial rainwater turbidity is too high, the system can predict in advance that sedimentation or filtration efforts need to be increased. Dust accumulation thickness directly affects the heat exchange efficiency of the condensation surface. If dust accumulation is severe, condensation efficiency will decrease significantly. In this case, the model may reduce the weighting of the atmospheric water production mode and instead recommend a rainwater priority mode or trigger a cleaning command. It should be understood that the above features are merely examples, and those skilled in the art can introduce features such as air pressure and light intensity according to actual needs, all of which fall within the scope of protection of this invention.

[0037] The preset weights of the classification model are obtained through offline supervised learning. The training samples are historical meteorological and water production records of the coastal area. Each sample includes an effective rainfall intensity correction value, air humidity, ambient temperature, dew point temperature, wind speed, and the corresponding optimal operating mode label.

[0038] Among them, the operation mode that maximizes the total water production rate of the system or minimizes the unit energy consumption is selected as the optimal operation mode label; the training objective is to minimize the cross-entropy loss between the model output mode and the label, which can effectively measure the difference between the probability distribution predicted by the model and the true label distribution.

[0039] Considering that the present invention is mainly applied to coastal saline-alkali areas, where the environment is characterized by high salinity and high humidity, the method in this embodiment further includes: setting an initial runoff diversion device at the inlet of the rainwater collection unit to automatically discharge the initial runoff of the first 10-15 mm based on the salinity detected by the chloride ion sensor.

[0040] In coastal areas, due to salt deposition carried by sea breezes, rainwater collection surfaces (such as rooftops or collection pans) often accumulate large amounts of salt crystals and dust. Initial runoff during the early stages of rainfall washes these pollutants into the system, causing the collected rainwater salinity to exceed standards. This embodiment addresses this by connecting an initial runoff tank in series in the inlet pipe and installing a chloride ion sensor at the tank inlet. When rainfall begins, the central control module reads the salinity data detected by the chloride ion sensor in real time. If the salinity exceeds a preset threshold (e.g., 200 mg / L), the system identifies it as initial runoff and controls the solenoid valve to open, discharging this high-salinity wastewater into the sewage network or evaporation tank. As rainfall continues, the surface salt is washed away, the sensor reading drops below the threshold, and the system automatically closes the discharge valve, allowing subsequent clean rainwater to flow into the collection system. It should be understood that although this embodiment preferably discharges the first 10-15 mm of runoff, the actual discharge amount can be dynamically adjusted according to the salinity curve fed back by the sensor. For example, for the first rain after a long drought, more water may need to be discharged, while for continuous rainfall, the discharge amount can be reduced or even the runoff diversion step can be skipped, thereby maximizing the amount of rainwater collected while ensuring water quality.

[0041] For atmospheric water condensation units, the atmosphere in coastal areas contains high concentrations of salt spray aerosols. These salts readily adhere to the condensation surface, and the salts can react electrochemically with the metal material, leading to corrosion and perforation. Furthermore, the salt layer reduces the heat exchange efficiency of the condensation surface. This embodiment coats the surface of the condenser fins or plate with a layer of polytetrafluoroethylene (PTFE) or a nano-hydrophobic coating. This coating has excellent corrosion resistance and non-stick properties, effectively preventing salt from contacting the substrate. Simultaneously, an ultrasonic descaling device is periodically activated to remove the salt layer and dust adhering to the condensation surface.

[0042] In addition, this embodiment also includes a system of air-water combined backwashing of the water treatment unit every two weeks, with the backwash wastewater being discharged or returned to the sedimentation area of ​​the rainwater collection unit.

[0043] In terms of energy recovery, this embodiment introduces the low-temperature rainwater collected by the rainwater harvesting unit into the heat exchanger to pre-cool the condensation surface of the atmospheric water condensation unit in order to improve condensation efficiency.

[0044] like Figure 2 As shown, this embodiment provides a rainwater and atmospheric water co-collection and treatment system, which is used to execute the rainwater and atmospheric water co-collection and treatment method described in any of the above embodiments. The system includes a rainwater collection unit, an atmospheric water condensation unit, a water quality treatment unit, and a central control module.

[0045] Specifically, the rainwater harvesting unit includes a rainwater collection surface (such as a roof rainwater collection system or a ground rainwater collection tray), an initial diversion device, water conveyance pipelines, and a rainwater storage tank. The rainwater collection surface is responsible for receiving natural precipitation, and the rainwater is transported to the storage tank for temporary storage through the water conveyance pipelines.

[0046] The atmospheric water condensation unit includes a refrigeration compressor, a condenser (or condenser fins), a fan, and a water collection tray. Under the action of the fan, humid air flows over the condensation surface cooled by the refrigerant. Water vapor condenses into water droplets upon contact with the condenser and drips into the water collection tray under gravity, and then enters the water treatment unit through pipelines.

[0047] Water treatment units typically consist of multi-stage filtration devices, membrane modules, and disinfection devices, and are responsible for purifying collected rainwater and condensate to meet domestic or irrigation water standards.

[0048] Raindrop sensors are used to collect time-series sequences of raindrop size distribution and raindrop falling velocity sequences; specifically, raindrop sensors preferably employ raindrop spectrometers based on laser or photoelectric principles.

[0049] A meteorological sensor array is used to collect environmental data, including air humidity, ambient temperature, dew point temperature, and wind speed. The central control module integrates a regression neural network and a classification model; The central control module receives the raindrop size distribution time series and raindrop falling velocity series collected by the raindrop sensor, and inputs them into the regression neural network to output an effective rainfall intensity correction value; the central control module also inputs the effective rainfall intensity correction value and the environmental data collected by the meteorological sensor group into the classification model and outputs operation instructions.

[0050] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include a processor 110, a communication interface 120, a memory 130, and a communication bus 140, wherein the processor 110, the communication interface 120, and the memory 130 communicate with each other via the communication bus 140. The processor 110 can call logical instructions stored in the memory 130, which is the rainwater and atmospheric water co-collection and treatment method described in any of the above technical solutions.

[0051] Furthermore, the logical instructions in the aforementioned memory 130 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the rainwater and atmospheric water co-collection and treatment method provided by the above methods.

[0053] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rainwater and atmospheric water co-collection and treatment method provided by the above methods.

[0054] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0055] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0056] 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 of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for the coordinated collection and treatment of rainwater and atmospheric water, characterized in that, The method is applied to an integrated system, which includes a rainwater harvesting unit, an atmospheric water condensation unit, a water treatment unit, and a central control module; the method includes: Acquire environmental data, time-series sequences of raindrop size distribution, and raindrop velocity sequences; The raindrop size distribution time series and raindrop falling velocity series are input into a regression neural network trained with micro-rainfall calibration data, and the effective rainfall intensity correction value is output. The effective rainfall intensity correction value and environmental data are input into the classification model, and an operation command is output. The operation command includes at least one of the following three modes: Rainwater priority mode: Enable rainwater collection unit; Atmospheric water generation mode: Rainwater collection unit is turned off; Synergistic Enhancement Mode: Simultaneously activate the rainwater harvesting unit and the atmospheric water condensation unit.

2. The method for coordinated collection and treatment of rainwater and atmospheric water according to claim 1, characterized in that, The regression neural network is a long short-term memory network. Its input is the time sequence of raindrop size distribution and raindrop falling speed within a continuous time window, and its output is the effective rainfall intensity correction value at the current moment. The micro-rainfall calibration data includes: artificially simulated drizzle and micro-rainfall event data collected on-site, with reference rainfall intensity measured by a high-precision weighing rain gauge as the training label.

3. The method for coordinated collection and treatment of rainwater and atmospheric water according to claim 1, characterized in that, The classification model uses the XGBoost algorithm, and the decision function of the XGBoost algorithm is: ， Where K is the number of regression trees. Let be the predicted score of the Kth regression tree for category c, calculated using the following formula: ， in, Let be the total number of leaf nodes in the k-th regression tree; Let t be the weight value of the t-th leaf node in the k-th regression tree corresponding to category c; Let t be the feature space region corresponding to the t-th leaf node in the k-th regression tree; As an indicator function, when the input feature vector x Falling area The value is 1 if the condition is met, and 0 otherwise. Input feature vector , This is a correction value for effective rainfall intensity; RH represents air humidity. For ambient temperature, Here, WS represents the dew point temperature and wind speed. The input feature vector x is composed of the effective rainfall intensity correction value and environmental data; Output rain priority mode at the time. In real-time output atmospheric water production mode, Output collaborative efficiency mode.

4. The method for coordinated collection and treatment of rainwater and atmospheric water according to claim 3, characterized in that, The input features of the XGBoost algorithm also include: the turbidity (Turb) value and conductivity (EC) value of the rainwater collection unit inlet, and the estimated dust thickness (Dust) value of the condensation surface of the atmospheric water condensation unit. At this point, the input feature vector x is expanded to: 。 5. The method for coordinated collection and treatment of rainwater and atmospheric water according to claim 3, characterized in that, The preset weights of the classification model are obtained through offline supervised learning. The training samples are historical meteorological and water production records of the coastal area. Each sample includes an effective rainfall intensity correction value, air humidity, ambient temperature, dew point temperature, wind speed, and the corresponding optimal operating mode label. The operating mode that maximizes the total water production rate or minimizes the unit energy consumption is selected as the optimal operating mode label. The training objective is to minimize the cross-entropy loss between the model output mode and the label.

6. The method for coordinated collection and treatment of rainwater and atmospheric water according to claim 1, characterized in that, The method further includes: An initial runoff diversion device is installed at the inlet of the rainwater collection unit to automatically discharge the initial runoff of the first 10-15 mm based on the salinity detected by the chloride ion sensor. Apply an anti-salt spray corrosion coating to the condensation surface of the atmospheric water condensation unit and periodically activate the ultrasonic descaling device; Every two weeks, the water treatment unit undergoes combined air-water backwashing, and the backwash wastewater is either discharged externally or returned to the sedimentation area of ​​the rainwater collection unit.

7. The method for synergistic collection and treatment of rainwater and atmospheric water according to claim 1, characterized in that, The low-temperature rainwater collected by the rainwater harvesting unit is introduced into the heat exchanger to pre-cool the condensation surface of the atmospheric water condensation unit in order to improve the condensation efficiency.

8. A rainwater and atmospheric water co-collection and treatment system, characterized in that, The system is used to perform the rainwater and atmospheric water co-collection and treatment method according to any one of claims 1 to 7, and the system includes a rainwater collection unit, an atmospheric water condensation unit, a water quality treatment unit and a central control module; Raindrop sensors are used to collect time-series sequences of raindrop size distribution and raindrop falling velocity. A meteorological sensor array is used to collect environmental data, including air humidity, ambient temperature, dew point temperature, and wind speed. The central control module integrates a regression neural network and a classification model; The central control module receives the raindrop size distribution time series and raindrop falling velocity series collected by the raindrop sensor, and inputs them into the regression neural network to output an effective rainfall intensity correction value; the central control module also inputs the effective rainfall intensity correction value and the environmental data collected by the meteorological sensor group into the classification model and outputs operation instructions.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the rainwater and atmospheric water co-collection and treatment method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the rainwater and atmospheric water co-collection and treatment method as described in any one of claims 1 to 7.

Images