Bridge deck runoff purification method, device and system based on large model prediction and electro-catalysis-membrane coupling

Abstract

The application relates to the cross field of bridge deck runoff purification and artificial intelligence, and specifically discloses a bridge deck runoff purification method, device and system based on large model prediction and electrocatalysis-membrane coupling, the method comprising the following steps: receiving multi-dimensional real-time data sent by a sensor unit at a preset frequency, wherein the multi-dimensional real-time data comprises multi-dimensional water quality data, rainfall intensity forecast data and real-time traffic volume data; predicting pollution condition data of bridge deck runoff in a target time period by using an LSTM-Transformer hybrid model according to the multi-dimensional real-time data; the pollution condition data comprises a peak value of pollutant load in a rainfall initial washing stage; generating a control instruction for controlling a modular treatment unit to perform electrocatalysis-membrane coupling purification on the bridge deck runoff according to the pollution condition data, so that the modular treatment unit performs electrocatalysis-membrane coupling purification in response to the control instruction, and a closed-loop system mechanism of "prediction-electrocatalysis-membrane separation" is effectively realized, thereby achieving synergistic effect while reducing operation cost.

Description

Technical Field

[0001] This application relates to the intersection of bridge deck runoff purification and artificial intelligence, specifically to a bridge deck runoff purification method, device, and system based on large model prediction and electrocatalysis-membrane coupling. Background Technology

[0002] With the rapid expansion of my country's expressway and urban expressway networks, bridge runoff, as a typical intermittent and high-load non-point source pollution, has become a significant risk source threatening the ecological security of water bodies along the routes. Especially in sections crossing sensitive water bodies such as drinking water source protection areas, the "initial flushing effect" formed within the first 30 minutes of rainfall, carrying pollutants such as suspended solids, petroleum hydrocarbons, heavy metals, and chemical oxygen demand (COD), can cause pollutant concentrations to reach several times or even dozens of times the normal levels, resulting in instantaneous and impactful pollution of receiving water bodies. Statistics show that the concentration of suspended solids in expressway bridge runoff can reach 500-3000 mg / L, and the concentration of petroleum hydrocarbons can reach 10-50 mg / L, far exceeding surface water environmental quality standards.

[0003] Current bridge runoff treatment technologies mainly revolve around three levels: physical interception, ecological purification, and advanced treatment. However, all existing technologies have inherent limitations. For example, the core of physical interception schemes lies in collecting runoff through pipelines, using sedimentation tanks for natural settling, and setting up emergency pools to handle hazardous chemical leaks. Gravity sedimentation tanks, as recommended in the "Highway Drainage Design Code" (JTG / T D33-2012), offer advantages such as simple structure and low cost. However, their treatment efficiency relies entirely on physical sedimentation, resulting in poor removal of dissolved pollutants and fine particulate matter. Furthermore, they require a large footprint, lack real-time monitoring and adaptive capabilities, and cannot cope with dynamic changes in pollution load. Ecological purification schemes emphasize eco-friendliness, utilizing the synergistic effects of plants, soil, and microorganisms to purify runoff. For instance, the "integrated bridge runoff pollution treatment device" proposed by Beijing University of Technology integrates primary sedimentation, flocculation, and secondary sedimentation tanks, aiming to reduce the footprint. However, the treatment efficiency of such technologies is greatly affected by climate and season. During periods of heavy rainfall with a surge in hydraulic load, the purification effect drops sharply, and their ability to remove recalcitrant organic pollutants such as petroleum hydrocarbons is limited. Membrane separation deep treatment technology, with ultrafiltration and microfiltration membranes as its core, can efficiently remove suspended solids and some large molecular organic matter, resulting in stable effluent quality. However, its most prominent problem is severe membrane fouling, leading to frequent cleaning (e.g., chemical cleaning every 72 hours), high operating energy consumption, large chemical consumption, high operation and maintenance costs, and a lack of predictive adjustment for fluctuations in influent water quality.

[0004] Therefore, providing a predictable, low-cost, and highly effective method for purifying bridge surface runoff has become an urgent problem to be solved. Summary of the Invention

[0005] In view of the above-mentioned defects or deficiencies in the existing technology, it is desirable to provide a bridge surface runoff purification method, device and system based on large model prediction and electrocatalysis-membrane coupling, which effectively realizes the closed-loop system mechanism of "prediction-electrocatalysis-membrane separation", thereby reducing operating costs and synergistically increasing efficiency.

[0006] In a first aspect, embodiments of this application provide a bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling, comprising: The receiver receives multidimensional real-time data transmitted by the sensor unit at a preset frequency. The multidimensional real-time data includes multidimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. Based on the multidimensional real-time data, the pollution status of bridge deck runoff is predicted using the LSTM-Transformer hybrid model within the target time period; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall. Based on the pollution data, control commands are generated to control the modular treatment unit to perform electrocatalytic-membrane coupling purification of the bridge surface runoff, so that the modular treatment unit performs electrocatalytic-membrane coupling purification in response to the control commands.

[0007] In some embodiments, the pollution data includes the predicted peak value of petroleum pollutants during the initial flushing phase of rainfall. The step of generating control instructions based on the pollution data for controlling the modular treatment unit to perform electrocatalytic-membrane coupled purification of the bridge surface runoff includes: Based on the predicted petroleum data, the peak and baseline petroleum concentrations are predicted, and the impact factor of petroleum pollutants on the pollution baseline is determined. Based on the impact factor, determine the expected current density corresponding to the predicted peak value of the petroleum data; Based on the expected current density, determine the target current density for adjustment; Based on the target current density, control commands are generated to control the modular processing unit to electrocatalyze the bridge surface runoff.

[0008] In some embodiments, it also includes: Obtain the actual values ​​of the petroleum-related data for the target time period; The feedback current density is determined based on the difference between the actual value of the petroleum data and the predicted peak value of the petroleum data. Feedback control commands are generated based on the feedback current density to control the modular processing unit to electrocatalyze the bridge surface runoff.

[0009] In some embodiments, the pollution data includes the predicted peak value of suspended solids in the peak pollutant load during the initial flushing phase of rainfall. The step of generating control instructions based on the pollution data for controlling the modular treatment unit to perform electrocatalytic-membrane coupled purification of the bridge surface runoff includes: Based on the predicted peak value and threshold concentration of the suspended matter data, the flux decay rate is determined. Based on the flux decay rate, determine the target flux corresponding to the predicted peak value of the suspended solids data; Based on the target flux, control instructions are generated to control the modular processing unit to adjust the membrane flux of the bridge deck runoff.

[0010] In some embodiments, it also includes: Obtain the actual values ​​of suspended matter data related to the suspended matter data within the target time period; The membrane fouling rate is determined based on the actual values ​​related to the suspended matter data; The target backwash interval is determined based on the membrane fouling rate; Perform membrane flushing according to the target backflushing interval.

[0011] In some embodiments, the actual values ​​related to suspended solids data include actual suspended solids data values, actual operating flux, and transmembrane pressure differential. Determining the membrane fouling rate based on the actual values ​​related to suspended solids data includes: The rate of change of transmembrane pressure difference, influent suspended solids load, and flux load are determined based on the actual values ​​of the suspended solids data, the actual operating flux, and the transmembrane pressure difference, respectively. The membrane fouling rate is determined based on the rate of change of the transmembrane pressure difference, the influent suspended solids load, and the flux load.

[0012] Secondly, embodiments of this application provide a bridge surface runoff purification device based on large model prediction and electrocatalysis-membrane coupling, comprising: The receiving module is used to receive multi-dimensional real-time data sent by the sensor unit at a preset frequency. The multi-dimensional real-time data includes multi-dimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. The prediction module is used to predict the pollution status of bridge surface runoff within a target time period based on the multidimensional real-time data and using an LSTM-Transformer hybrid model; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall; The instruction module is used to generate control instructions for controlling the modular treatment unit to perform electrocatalytic-membrane coupling purification of the bridge surface runoff based on the pollution data, so that the modular treatment unit performs electrocatalytic-membrane coupling purification in response to the control instructions.

[0013] Thirdly, embodiments of this application provide a bridge surface runoff purification system based on large model prediction and electrocatalysis-membrane coupling, comprising: A sensor unit is used to collect multidimensional real-time data and send the multidimensional real-time data to a large model prediction unit. A modular processing unit is used to receive control commands sent by the large model prediction unit to perform electrocatalysis-membrane coupling purification. The large model prediction unit used to perform the method.

[0014] Fourthly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the embodiments of this application.

[0015] Fifthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in embodiments of this application.

[0016] In a sixth aspect, embodiments of this application provide a computer program product, including a computer program, characterized in that, when the computer program is executed by a processor, it implements the method described in embodiments of this application.

[0017] The bridge surface runoff purification method, apparatus, and system based on large model prediction and electrocatalysis-membrane coupling provided in this application embodiment receive multi-dimensional real-time data transmitted by sensor units at a preset frequency. The multi-dimensional real-time data includes multi-dimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. Based on the multi-dimensional real-time data, an LSTM-Transformer hybrid model is used to predict the pollution status of bridge surface runoff within a target time period. Based on the pollution status data, control commands are generated to control the modular processing unit to perform electrocatalysis-membrane coupling purification of bridge surface runoff. This enables the modular processing unit to respond to the control commands and perform electrocatalysis-membrane coupling purification, effectively realizing a closed-loop system mechanism of "prediction-electrocatalysis-membrane separation," thereby reducing operating costs and synergistically increasing efficiency.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1This paper shows an architecture diagram of a bridge surface runoff purification system based on large model prediction and electrocatalysis-membrane coupling provided in an embodiment of this application. Figure 2 A schematic flowchart of a bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling provided in an embodiment of this application is shown. Figure 3 This application shows a graph comparing the predicted rainstorm event with the actual pollution load according to a specific embodiment of the present application. Figure 4 This illustration shows a schematic diagram of a bridge surface runoff purification device based on large model prediction and electrocatalysis-membrane coupling according to an embodiment of this application; Figure 5 A schematic diagram of the structure of a computer system suitable for implementing an electronic device or server according to embodiments of this application is shown. Detailed Implementation

[0020] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] The structural schematic diagram of the bridge runoff purification system based on large model prediction and electrocatalysis-membrane coupling proposed in this application is shown below. Figure 1 . Figure 1 This paper illustrates the architecture of a bridge surface runoff purification system based on large model prediction and electrocatalysis-membrane coupling provided in an embodiment of this application.

[0023] like Figure 1 As shown, the architecture includes: a sensor unit 101, a large model prediction unit 102, and a modular processing unit 103.

[0024] Sensor unit 101 acts as the system's "sensory nerves," responsible for collecting various input information around the clock and at high density. Specifically, multifunctional monitoring nodes are deployed along the bridge deck at preset intervals (preferably 50 meters). Each node integrates suspended solids sensors, petroleum hydrocarbon sensors, chemical oxygen demand (COD) sensors, pH sensors, conductivity sensors, and ultrasonic flow meters. Sensor unit 101 collects water quality data such as suspended solids (SS), petroleum hydrocarbons, COD, pH value, conductivity, and real-time water flow at a preset frequency (e.g., once / second, delay <1 second). It also receives real-time rainfall intensity forecast data from regional high-precision weather radar and real-time traffic flow data from traffic management departments, providing comprehensive input dimensions for the prediction model.

[0025] The large-scale model prediction unit 102 is the "intelligent brain" of the system, responsible for processing information, predicting the future, and making decisions. Specifically, the large-scale model prediction unit 102 is used to execute the bridge surface runoff purification method based on large-scale model prediction and electrocatalysis-membrane coupling proposed in this application embodiment. By analyzing the pollution fluctuation patterns recorded in the past, and using the current sensor readings uploaded at a frequency of 1 time / second as historical and real-time water quality sequences, as well as future short-term rainfall forecasts (0-200 mm / h), real-time and predicted traffic flow (distinguishing between heavy-duty vehicles), seasonal and temperature factors, it predicts the most intense instantaneous concentrations of suspended solids (SS), petroleum, heavy metals, and chemical oxygen demand (COD) entering the system in the initial flushing within the next 15-30 minutes, i.e., the water-lipid peak, and predicts the maximum instantaneous flow rate (m³ / h) entering the treatment system within the next 15-30 minutes, i.e., the water peak. Based on the prediction results, it calculates the optimal control parameters and sends them to the module treatment unit, including the target current density of the electrocatalytic unit, the expected operating flux of the membrane unit, and the optimal backwashing timing.

[0026] The modular processing unit 103 acts as the system's "executor," precisely completing the purification task based on the control parameters output by the large model prediction unit 102. The modular processing unit 103 includes a Ti / RuO2-IrO2 electrode and a polyvinylidene fluoride hollow fiber ultrafiltration membrane. Specifically, the modular processing unit 103 uses a Ti / RuO2-IrO2 electrode to construct a micro-electric field catalytic oxidation system. Upon receiving control commands from the large model prediction unit 102, it steplessly adjusts the current density within the range of 10-15 mA / cm². It operates energy-efficiently under low load and intensifies oxidation in advance when the large model prediction unit 102 predicts the peak value of petroleum pollutants, decomposing recalcitrant organic matter into easily biodegradable or easily membrane-filterable small molecules, while simultaneously generating strong oxidants such as ·OH, which provides disinfection. The polyvinylidene fluoride hollow fiber ultrafiltration membrane is selected due to its excellent antifouling performance and mechanical strength, and its 0.01 μm pore size ensures efficient retention. The large model prediction unit 102 predicts the trend of pollutant concentration changes, dynamically adjusts the membrane operating pressure and flux (50-80L / m²·h), and adopts a combination mode of "gas-water pulse backwash + predictive chemical cleaning", which extends the average membrane chemical cleaning cycle from the traditional 72 hours to more than 120 hours and reduces reagent consumption by 40%.

[0027] The large model prediction unit 102 can be deployed on a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0028] The sensor unit 101, the large model prediction unit 102, and the modular processing unit 103 are directly or indirectly connected via wired or wireless communication. Optionally, the aforementioned wireless or wired network uses standard communication technologies and / or protocols. The network is typically the Internet, but can also be any network, including but not limited to any combination of Local Area Network (LAN), Metropolitan Area Network (MAN), Wide Area Network (WAN), mobile, wired or wireless networks, private networks, or virtual private networks.

[0029] The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling proposed in this application can be implemented by a bridge surface runoff purification device based on large model prediction and electrocatalysis-membrane coupling, which can be installed on terminal equipment or a server.

[0030] To further illustrate the technical solutions provided in the embodiments of this application, a detailed description is provided below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiments of this application provide method operation instruction steps as shown in the following embodiments or drawings, the method may include more or fewer operation instruction steps based on conventional or non-creative effort. In steps where there is no logically necessary causal relationship, the execution order of these steps is not limited to the execution order provided in the embodiments of this application. In actual processing or when the device executes the method, it may be executed sequentially or in parallel according to the method shown in the embodiments or drawings.

[0031] Please refer to Figure 2 , Figure 2 This illustration shows a schematic flowchart of a bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling, according to an embodiment of this application. Figure 2 As shown, the method includes: Step 201: Receive multi-dimensional real-time data sent by the sensor unit at a preset frequency. The multi-dimensional real-time data includes multi-dimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data.

[0032] The multidimensional water quality data includes suspended solids data collected by the suspended solids sensor integrated in the sensor unit, petroleum data collected by the petroleum sensor, chemical oxygen demand (COD) value collected by the chemical oxygen demand sensor, pH value collected by the pH sensor, conductivity data collected by the conductivity sensor, and real-time water volume data collected by the ultrasonic flow meter, etc. This application does not make specific limitations on these data, which can be determined according to the type of sensor integrated in the sensor unit.

[0033] Suspended solids (SS) are organic and inorganic particulate matter suspended in water that cannot pass through 0.45-micron filter paper or filters. Examples include insoluble silt, clay, organic matter, algae, and microorganisms, and they are one of the indicators for measuring the degree of water pollution. Petroleum sensors can be petroleum hydrocarbon sensors used to measure data on light oil, heavy oil, and other petroleum-related substances in water. Chemical oxygen demand (COD) sensors are online monitoring devices used to quickly determine the content of organic pollutants in water. Their core principle is based on ultraviolet absorption, calculating COD values ​​by measuring the degree of absorption of specific wavelengths of ultraviolet light by organic matter. pH sensors collect pH values, which are the negative logarithm of the hydrogen ion concentration in an aqueous solution, used to indicate the acidity or alkalinity of the solution. Electrical conductivity is the degree to which a material conducts current; the higher the value, the stronger the conductivity.

[0034] Based on the acquisition of multidimensional water quality data, further high-precision meteorological radar rainfall intensity forecast data for the bridge surface runoff area will be obtained, for example, by accessing rainfall intensity forecast data from meteorological departments. Real-time traffic flow data from traffic management departments will also be acquired, including data on the proportion of heavy-duty vehicles.

[0035] In some preferred embodiments, multidimensional real-time data sent by the sensor unit can be received at a preset frequency, such as once per second with a delay of less than 1 second. This application does not make specific limitations in this regard.

[0036] Step 202: Based on multidimensional real-time data, use the LSTM-Transformer hybrid model to predict the pollution status of bridge deck runoff within the target time period; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall.

[0037] It should be noted that LSTM networks excel at handling time-dependent relationships and capturing historical trends in water quality parameters. The Transformer's sub-attention mechanism can effectively correlate multi-source heterogeneous data (such as the relationship between rainfall intensity in the next 15 minutes and the baseline concentration of pollutants accumulated on the bridge surface before or during rainfall). Based on this, this application utilizes an LSTM-Transformer hybrid model to predict the pollution status of bridge surface runoff within a target time period based on multi-dimensional real-time data. The target time period can be a reasonable expectation of the time for modular treatment units to pre-treat the water quality, such as 15-30 minutes; this application does not impose a specific limitation on this.

[0038] Specifically, in this embodiment, the LSTM layer in the LSTM-Transformer is used to process the historical trends of multidimensional water quality data and capture long-term dependencies, such as the accumulation and scouring patterns of pollutants during rainy days. The Transformer layer automatically learns the nonlinear interactions between multi-source features through a self-attention mechanism. In this embodiment, the self-attention mechanism of the Transformer layer is used to learn the contribution weights of rainfall intensity and the proportion of heavily loaded vehicles to suspended solids, thereby achieving advanced prediction of pollutants (e.g., suspended solids).

[0039]

[0040]

[0041]

[0042]

[0043] Therefore, this application deeply applies the LSTM-Transformer hybrid model to bridge runoff scenarios, achieving accurate prediction of complex nonlinear pollution processes. Unlike existing early warning methods that rely solely on historical water quality data, this application deeply integrates meteorological forecast data with real-time traffic data—two key external driving factors—transforming prediction from "post-event retrospective" to "pre-event forecast."

[0044] Step 203: Based on the pollution data, generate control commands for controlling the modular treatment unit to perform electrocatalytic-membrane coupling purification of bridge surface runoff, so that the modular treatment unit responds to the control commands to perform electrocatalytic-membrane coupling purification.

[0045] It should be noted that the modular processing unit includes an adaptive electrocatalytic oxidation module. Specifically, the adaptive electrocatalytic oxidation module uses a Ti / RuO2-IrO2 electrode to construct a micro-field catalytic oxidation system. Petroleum-based contaminants lose electrons at the anode surface, undergoing oxidative decomposition. The anode generates strong oxidants such as hydroxyl radicals (·OH), which pre-oxidize recalcitrant organic matter (petroleum-based) into smaller molecules, while simultaneously altering the surface charge of the colloid, reducing subsequent membrane fouling. The current density controls the electron transfer rate at the electrode surface. Too low a current density results in insufficient hydroxyl radical generation, the oxidation reaction is controlled by mass transfer, large organic molecules cannot be fully pre-oxidized, and the risk of subsequent membrane fouling is high. Too high a current density, while increasing the oxidation rate, leads to increased oxygen evolution side reactions, decreased current efficiency, accelerated electrode overheating and passivation, and a significant increase in energy consumption. Therefore, determining a reasonable current density is crucial to solving the problem.

[0046] In one feasible embodiment, the pollution data includes the predicted peak value of petroleum pollutants during the initial flushing phase of rainfall. Based on the pollution data, control instructions are generated for controlling the modular treatment unit to perform electrocatalytic-membrane coupled purification of bridge runoff. These instructions include: determining the impact factor of petroleum pollutants on the pollution baseline based on the predicted peak value and the limiting petroleum concentration; determining the expected current density corresponding to the predicted peak value of petroleum pollutants based on the impact factor; determining the target current density for adjustment based on the expected current density; and generating control instructions for controlling the modular treatment unit to perform electrocatalytic purification of bridge runoff based on the target current density.

[0047] For example, the target current density can be determined using the following formula:

[0048]

[0049] As can be seen, the expression for the target current density reflects the core idea that the higher the pollution load, the stronger the electrocatalytic intensity should be. Introducing a safety margin coefficient can effectively avoid the risk of predictive plate smudging.

[0050] Furthermore, to avoid deviations in prediction results due to uncontrollable factors, such as abnormal vehicle petroleum pollutant emissions, this application proposes a feedback adjustment strategy. In one feasible embodiment, the actual values ​​of petroleum data for the target time period are obtained; the feedback current density is determined based on the difference between the actual petroleum data values ​​and the predicted peak values; and a feedback control command is generated based on the feedback current density to control the modular processing unit to electrocatalyze bridge runoff.

[0051] For example, the feedback current density can be determined using the following expression:

[0052]

[0053] Preferably, the feedback current density can be determined based on the sampling frequency of the sensor unit, and then feedback control commands can be generated.

[0054] Therefore, by using feedback control commands, it is possible to respond quickly when the actual measured value of petroleum pollutants does not match the predicted value, rapidly compensate for instantaneous changes, eliminate steady-state residual error, and prevent long-term under-oxidation.

[0055] It should also be noted that the modular processing unit includes a polyvinylidene fluoride hollow fiber ultrafiltration membrane. Using a polyvinylidene fluoride hollow fiber ultrafiltration membrane, a 0.01 μm pore size ensures high-efficiency retention. Its separation principle includes a sieving effect that retains suspended solids, colloids, and bacteria larger than the pore size; adsorption retention of some dissolved organic matter through hydrophobic or electrostatic interactions on the membrane surface; and the dynamic gel layer formed during operation further improving the retention rate.

[0056] In one feasible embodiment, the pollution data includes the predicted peak value of suspended solids data during the initial flushing phase of rainfall. Based on the pollution data, control instructions are generated for controlling the modular treatment unit to perform electrocatalytic-membrane coupled purification of bridge runoff. These instructions include: determining the flux decay rate based on the predicted peak value of suspended solids data and the threshold concentration of suspended solids data; determining the target flux corresponding to the predicted peak value of suspended solids data based on the flux decay rate; and generating control instructions for controlling the modular treatment unit to adjust the membrane flux of bridge runoff based on the target flux.

[0057] For example, the target flux can be determined using the following formula:

[0058]

[0059] Specifically, after obtaining the target flux, the target flux is written as a set value into the flux regulation control logic to generate an adjustment of the feed pump speed, thereby realizing membrane flux regulation.

[0060] For example, when the peak value of high suspended solids data is predicted, the target flux at the target time t decreases from 80 L / m·h to 50-34 L / m·h, and the water flow is automatically reduced by the membrane flux adjustment control command.

[0061] Therefore, by predicting suspended solids data, the lag in reducing flux after the traditional method increases membrane pressure differential is effectively solved, effectively avoiding irreversible membrane fouling, and proactively reducing operating load before the fouling peak is reached, thus slowing down the rate of fouling accumulation from the source.

[0062] Furthermore, excessively high concentrations of pollutants in bridge runoff can easily lead to irreversible membrane fouling. Therefore, to avoid accelerated membrane aging due to fluctuations in pollutant levels, this application proposes adaptive backwashing based on actual values ​​of suspended solids data. This avoids both excessive backwashing (wasting energy and water) and insufficient backwashing (accelerating membrane aging). By precisely controlling the timing of backwashing, unnecessary mechanical and chemical stresses are reduced, extending the average service life of the membrane module.

[0063] In one feasible embodiment, the actual values ​​of suspended solids data related to the target time period are obtained; the membrane fouling rate is determined based on the actual values ​​of suspended solids data; the target backwashing interval is determined based on the membrane fouling rate; and membrane flushing is performed according to the target backwashing interval.

[0064] The actual values ​​related to suspended solids data include actual suspended solids data, actual operating flux, and transmembrane pressure difference. Based on the actual values ​​related to suspended solids data, the membrane fouling rate is determined, including: determining the rate of change of transmembrane pressure difference, influent suspended solids load, and flux load based on the actual suspended solids data, actual operating flux, and transmembrane pressure difference, respectively; and determining the membrane fouling rate based on the rate of change of transmembrane pressure difference, influent suspended solids load, and flux load.

[0065] It should be noted that the membrane fouling rate is an indicator for evaluating the trend of membrane fouling and is used to reflect the current health status of the membrane.

[0066] For example, the membrane fouling rate can be determined using the following expression:

[0067]

[0068] Furthermore, the target backwash interval can be determined using the following expression:

[0069]

[0070] It should be noted that in the expression for membrane fouling rate, the rate of change of transmembrane pressure directly reflects the membrane fouling rate. As pollutants accumulate on the membrane, the pressure required to maintain a constant flux will continuously increase. The feed water suspended solids load reflects that the higher the suspended solids concentration, the faster the filter cake layer forms on the membrane surface. Correspondingly, the closer the operating flux is to the limiting flux, the more severe the concentration polarization, and the faster the particles migrate to the membrane surface.

[0071] It should be understood that, in the embodiments of this application, after the predicted peak data of suspended solids is obtained, the membrane flux is dynamically adjusted, for example, by reducing the membrane flux. The purpose of flux adjustment is to proactively reduce the operating load before the peak of fouling arrives, thereby reducing the rate of increase in membrane fouling. If the membrane fouling rate continues to rise, the backwashing interval is further shortened to prevent the filter cake on the membrane surface from thickening rapidly and causing irreversible fouling.

[0072] In some feasible embodiments, the actual values ​​of suspended solids data can be independently acquired in real time to determine the membrane fouling rate. When the membrane fouling rate is consistently higher than the membrane fouling rate threshold and backwashing cannot effectively reduce it, chemical cleaning can be triggered in advance to avoid membrane fiber breakage or irreversible fouling.

[0073] It should be noted that in the embodiments of this application, electrocatalysis is not only a pretreatment unit, but also an "intelligent pretreatment" unit that operates dynamically according to control commands. Its output water quality (biodegradability, colloidal stability) is matched in real time with the feed water requirements of the subsequent membrane unit. This transcends the single concept of "micro-electric field coupled membrane separation" and achieves a leap in the level of intelligent coupling. The use of this synergistic mechanism enables the electrocatalytic unit to maintain an average removal rate of petroleum hydrocarbons at 92-95%, greatly reducing the organic fouling load on the membrane. Under the precise control of the large model prediction unit, the fouling rate of the membrane system is reduced by 60%, the cleaning cycle is extended to 120 hours, and the energy consumption per ton of water is reduced to ≤1.2kWh. Compared with traditional membrane processes, the overall energy consumption is reduced by more than 30%, the chemical reagent consumption is reduced by 40%, and the annual operation and maintenance cost is reduced by 25%.

[0074] When the peak concentration of petroleum compounds is predicted, the large model prediction unit instructs the electrocatalytic unit to increase the current density to 15 mA / cm² 30 seconds in advance, pre-oxidizing large organic molecules into smaller molecules, while changing the surface charge of the colloid and reducing the risk of membrane pore blockage; the membrane unit simultaneously adjusts the backwashing frequency to achieve synergistic effect.

[0075] Therefore, the bridge runoff purification method based on large model prediction and electrocatalysis-membrane coupling provided in this application embodiment receives multi-dimensional real-time data sent by sensor units at a preset frequency. The multi-dimensional real-time data includes multi-dimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. Based on the multi-dimensional real-time data, the pollution status of bridge runoff within a target time period is predicted using an LSTM-Transformer hybrid model. Based on the pollution status data, control commands are generated to control the modular treatment unit to perform electrocatalysis-membrane coupling purification of bridge runoff. This effectively realizes a closed-loop system mechanism of "prediction-electrocatalysis-membrane separation", reducing operating costs while synergistically increasing efficiency.

[0076] In one specific embodiment, the G98 Ring Expressway at the Chitian Reservoir section is used.

[0077] Scenario: The bridge spans a secondary protection zone of a drinking water source. It is 898.2 meters long, with a peak daily traffic volume of 28,000 vehicles, of which about 25% are heavy-duty vehicles. Environmental protection requirements are stringent.

[0078] System configuration: 18 monitoring nodes, 1 AI edge computing control cabinet, 1 electrocatalysis module (processing capacity 50 m³ / h, electrode area 12 m²), and ultrafiltration membrane module (PVDF hollow fiber, effective area 200 m²).

[0079] Model Training: Rainfall data from all 2023 events (23 valid rainfall events) were collected as the training set. Each training sample included past and current water quality sequences, rainfall intensity, traffic flow, and environmental factors as input, with the corresponding peak SS / petroleum hydrocarbon concentration for the next 15-30 minutes as output. Features such as rainfall intensity, traffic flow, and water quality time series were extracted to construct an LSTM-Transformer prediction model. The model was trained for 500 epochs, and the validation set error converged to RMSE = 18.2 mg / L (SS prediction).

[0080] Operational results (actual measurement during the 2024 rainy season): Influent water quality (early stage of heavy rain): SS=2000mg / L, petroleum hydrocarbons=11.25mg / L.

[0081] Effluent water quality: SS=4.2mg / L, petroleum=0.3mg / L (stable and better than GB 3838-2002 Class III standard).

[0082] Intelligent performance: The AI ​​model achieved an accuracy rate of 92.3% in issuing pollution peak warnings for three major rainfall events, with an average response time of 2.5 minutes.

[0083] Economic indicators: The direct cost of treating 1 ton of water is 0.76 yuan (including 0.48 yuan of electricity consumption, 0.12 yuan of chemical reagents, and 0.16 yuan of membrane replacement amortization). The annual water saving is about 70,000 tons. The estimated carbon credit benefits and water saving benefits are about 80,000 yuan per year, which can shorten the investment payback period to 3-4 years.

[0084] In another specific embodiment, the test conditions were: simulated hourly rainfall of 95mm (historical extreme value), traffic flow of 3000 vehicles / hour, heavy-duty vehicles accounting for 35%, and continuous operation for 24 hours. System Response: The AI ​​model predicted 18 minutes in advance that the peak SS would reach 1800 mg / L and the peak petroleum hydrocarbons would reach 12 mg / L, and automatically generated an "enhanced operation mode" command: the electrocatalytic current density was increased from 10 mA / cm² to 15 mA / cm², the membrane system operating flux was reduced from 80 L / m·h to 60 L / m·h, and the backwashing frequency was adjusted from once every 30 minutes to once every 20 minutes; Test results: The system operated continuously and stably for 24 hours, with the effluent SS concentration consistently below 8 mg / L and petroleum hydrocarbons consistently below 0.5 mg / L. The transmembrane pressure differential (TMP) rise rate was 0.8 kPa / h, a 47% reduction compared to the mode without AI control (1.5 kPa / h). There were no membrane module fouling alarms, demonstrating the system's strong resistance to shock loads and the effectiveness of the AI ​​control strategy.

[0085]

[0086] In yet another specific embodiment, two sets of control experiments were set up under the same hardware configuration:

[0087] Experimental results show that large model prediction unit control can significantly improve processing efficiency, reduce energy consumption and extend membrane life, verifying the advanced nature of the method of this invention.

[0088] It should be noted that although the operation of the method of the present invention is described in a specific order in the accompanying drawings, this does not require or imply that the operations must be performed in that specific order, or that all the operations shown must be performed in order to achieve the desired result.

[0089] Figure 4 A schematic diagram of a bridge surface runoff purification device based on large model prediction and electrocatalysis-membrane coupling provided in an embodiment of this application is shown.

[0090] like Figure 4 As shown, the bridge surface runoff purification device 10 based on large model prediction and electrocatalysis-membrane coupling includes: The receiving module 11 is used to receive multidimensional real-time data sent by the sensor unit at a preset frequency. The multidimensional real-time data includes multidimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. Prediction module 12 is used to predict the pollution status data of bridge deck runoff within a target time period based on the multidimensional real-time data and using an LSTM-Transformer hybrid model; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall; The instruction module 13 is used to generate control instructions for controlling the modular treatment unit to perform electrocatalytic-membrane coupling purification of the bridge surface runoff based on the pollution data, so that the modular treatment unit performs electrocatalytic-membrane coupling purification in response to the control instructions.

[0091] In some embodiments, the pollution data includes the predicted peak value of petroleum pollutants during the initial flushing phase of rainfall. The instruction module 13 is specifically used for: Based on the predicted petroleum data, the peak and baseline petroleum concentrations are predicted, and the impact factor of petroleum pollutants on the pollution baseline is determined. Based on the impact factor, determine the expected current density corresponding to the predicted peak value of the petroleum data; Based on the expected current density, determine the target current density for adjustment; Based on the target current density, control commands are generated to control the modular processing unit to electrocatalyze the bridge surface runoff.

[0092] In some embodiments, instruction module 13 is further configured to: Obtain the actual values ​​of the petroleum-related data for the target time period; The feedback current density is determined based on the difference between the actual value of the petroleum data and the predicted peak value of the petroleum data. Feedback control commands are generated based on the feedback current density to control the modular processing unit to electrocatalyze the bridge surface runoff.

[0093] In some embodiments, the pollution data includes the predicted peak value of suspended solids data in the peak pollutant load during the initial flushing stage of rainfall. The instruction module 13 is specifically used for: Based on the predicted peak value and threshold concentration of the suspended matter data, the flux decay rate is determined. Based on the flux decay rate, determine the target flux corresponding to the predicted peak value of the suspended solids data; Based on the target flux, control instructions are generated to control the modular processing unit to adjust the membrane flux of the bridge deck runoff.

[0094] In some embodiments, instruction module 13 is further configured to: Obtain the actual values ​​of suspended matter data related to the suspended matter data within the target time period; The membrane fouling rate is determined based on the actual values ​​related to the suspended matter data; The target backwash interval is determined based on the membrane fouling rate; Perform membrane flushing according to the target backflushing interval.

[0095] In some embodiments, the actual values ​​related to the suspended matter data include the actual value of the suspended matter data, the actual operating flux, and the transmembrane pressure difference. The instruction module 13 is specifically used for: The rate of change of transmembrane pressure difference, influent suspended solids load, and flux load are determined based on the actual values ​​of the suspended solids data, the actual operating flux, and the transmembrane pressure difference, respectively. The membrane fouling rate is determined based on the rate of change of the transmembrane pressure difference, the influent suspended solids load, and the flux load.

[0096] It should be understood that the modules or modules described in the bridge surface runoff purification device 10 based on large model prediction and electrocatalysis-membrane coupling are similar to those in the reference model. Figure 2 The steps in the described method correspond accordingly. Therefore, the operations and features described above for the method are also applicable to the bridge surface runoff purification device 10 based on large model prediction and electrocatalysis-membrane coupling, and its included modules, and will not be repeated here. The bridge surface runoff purification device 10 based on large model prediction and electrocatalysis-membrane coupling can be pre-implemented in the browser or other secure applications of electronic devices, or can be loaded into the browser or other secure applications of electronic devices through download or other means. The corresponding modules in the bridge surface runoff purification device 10 based on large model prediction and electrocatalysis-membrane coupling can cooperate with the modules in the electronic device to implement the solution of the embodiments of this application.

[0097] The division of modules or units mentioned in the detailed description above is not mandatory. In fact, according to the embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0098] The following is for reference. Figure 5 , Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the embodiments of this application is shown. like Figure 5As shown, the computer system 500 includes a central processing unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 502 or programs loaded from storage section 508 into random access memory (RAM) 503. RAM 503 also stores various programs and data required for the system's operating instructions. CPU 501, ROM 502, and RAM 503 are interconnected via bus 504. Input / output (I / O) interface 505 is also connected to bus 504.

[0099] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. A removable medium 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 510 as needed so that computer programs read from it can be installed into storage section 508 as needed.

[0100] Specifically, according to embodiments of this application, the flowchart above refers to... Figure 2 The described process can be implemented as a computer software program. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such an embodiment, the computer program contains program code for performing the methods shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs the functions defined in the system of this application.

[0101] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0102] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operational instructions of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two connected blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified functions or operational instructions, or using a combination of dedicated hardware and computer instructions.

[0103] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be housed in a processor; for example, a processor can be described as including a receiving module, a prediction module, and an instruction module. The names of these units or modules do not necessarily limit the specific unit or module itself. For example, a receiving module can also be described as "receiving multidimensional real-time data transmitted by a sensor unit at a preset frequency, wherein the multidimensional real-time data includes multidimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data."

[0104] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments, or may exist independently and not assembled into the electronic device. The computer-readable storage medium stores one or more programs that, when used by one or more processors, execute the bridge runoff purification method based on large model prediction and electrocatalysis-membrane coupling described in this application.

[0105] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling, characterized in that, include: The receiver receives multidimensional real-time data transmitted by the sensor unit at a preset frequency. The multidimensional real-time data includes multidimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. Based on the multidimensional real-time data, the pollution status of bridge deck runoff is predicted using the LSTM-Transformer hybrid model within the target time period; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall. Based on the pollution data, control commands are generated to control the modular treatment unit to perform electrocatalytic-membrane coupling purification of the bridge surface runoff, so that the modular treatment unit performs electrocatalytic-membrane coupling purification in response to the control commands.

2. The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling according to claim 1, characterized in that, The pollution data includes the predicted peak value of petroleum pollutants during the initial flushing phase of rainfall. Based on this pollution data, control instructions are generated to control the modular treatment unit to perform electrocatalytic-membrane coupled purification of the bridge surface runoff, including: Based on the predicted petroleum data, the peak and baseline petroleum concentrations are predicted, and the impact factor of petroleum pollutants on the pollution baseline is determined. Based on the impact factor, determine the expected current density corresponding to the predicted peak value of the petroleum data; Based on the expected current density, determine the target current density for adjustment; Based on the target current density, control commands are generated to control the modular processing unit to electrocatalyze the bridge surface runoff.

3. The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling according to claim 2, characterized in that, Also includes: Obtain the actual values ​​of the petroleum-related data for the target time period; The feedback current density is determined based on the difference between the actual value of the petroleum data and the predicted peak value of the petroleum data. Feedback control commands are generated based on the feedback current density to control the modular processing unit to electrocatalyze the bridge surface runoff.

4. The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling according to claim 1, characterized in that, The pollution data includes the predicted peak value of suspended solids during the initial flushing phase of rainfall. Based on this pollution data, control instructions are generated for the modular treatment unit to perform electrocatalytic-membrane coupled purification of the bridge surface runoff, including: Based on the predicted peak value and threshold concentration of the suspended matter data, the flux decay rate is determined. Based on the flux decay rate, determine the target flux corresponding to the predicted peak value of the suspended solids data; Based on the target flux, control instructions are generated to control the modular processing unit to adjust the membrane flux of the bridge deck runoff.

5. The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling according to claim 4, characterized in that, Also includes: Obtain the actual values ​​of suspended matter data related to the suspended matter data within the target time period; The membrane fouling rate is determined based on the actual values ​​related to the suspended matter data; The target backwash interval is determined based on the membrane fouling rate; Perform membrane flushing according to the target backflushing interval.

6. The bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling according to claim 5, characterized in that, The actual values ​​related to the suspended solids data include actual suspended solids data values, actual operating flux, and transmembrane pressure difference. Determining the membrane fouling rate based on the actual values ​​related to the suspended solids data includes: The rate of change of transmembrane pressure difference, influent suspended solids load, and flux load are determined based on the actual values ​​of the suspended solids data, the actual operating flux, and the transmembrane pressure difference, respectively. The membrane fouling rate is determined based on the rate of change of the transmembrane pressure difference, the influent suspended solids load, and the flux load.

7. A bridge surface runoff purification device based on large model prediction and electrocatalysis-membrane coupling, characterized in that, include: The receiving module is used to receive multi-dimensional real-time data sent by the sensor unit at a preset frequency. The multi-dimensional real-time data includes multi-dimensional water quality data, rainfall intensity forecast data, and real-time traffic flow data. The prediction module is used to predict the pollution status of bridge surface runoff within a target time period based on the multidimensional real-time data and using an LSTM-Transformer hybrid model; the pollution status data includes the peak pollutant load during the initial flushing stage of rainfall; The instruction module is used to generate control instructions for controlling the modular treatment unit to perform electrocatalytic-membrane coupling purification of the bridge surface runoff based on the pollution data, so that the modular treatment unit performs electrocatalytic-membrane coupling purification in response to the control instructions.

8. A bridge surface runoff purification system based on large model prediction and electrocatalysis-membrane coupling, characterized in that, include: A sensor unit is used to collect multidimensional real-time data and send the multidimensional real-time data to a large model prediction unit. A modular processing unit is used to receive control commands sent by the large model prediction unit to perform electrocatalysis-membrane coupling purification. The large model prediction unit is used to execute the bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling as described in claims 1-6.

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 bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling as described in any one of claims 1-6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the bridge surface runoff purification method based on large model prediction and electrocatalysis-membrane coupling as described in any one of claims 1-6.

Images