Medication management method and device based on deep learning model and computer device
By using deep learning models and multi-parameter coupled analysis, the problem of identifying scaling or biofilm adsorption of chemicals in cooling systems was solved, enabling appropriate increases and precise management of chemical dosage, and ensuring system stability and environmentally friendly emissions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU INST OF TECH
- Filing Date
- 2026-03-02
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies struggle to accurately identify and quantify scaling or biofilm adsorption phenomena in cooling systems, leading to inaccurate dosage adjustments and potentially causing over-compression or excessive emissions.
By employing a deep learning model combined with multi-parameter coupling analysis, the adsorption phenomenon of the reagent is identified by comparing the theoretical and actual total emissions and combining the coupling degree threshold. Based on the difference, the dosage of cleaning and removal reagent for the next cycle is calculated, and the dosing management plan is updated.
It enables precise identification of reagent adsorption phenomena, ensuring that the cooling system maintains heat exchange performance and water quality safety while avoiding excessive emissions, thus reducing operating costs and environmental risks.
Smart Images
Figure CN122434084A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent analysis technology for environmental data, and in particular to a method, apparatus, computer equipment, and storage medium for pesticide application management based on a deep learning model. Background Technology
[0002] Currently, large-scale industrial cooling systems are widely used in industries such as petrochemicals, power, metallurgy, electronics manufacturing, and data centers. During continuous operation, these systems typically require the periodic addition of chemical agents with water quality stabilizing or cooling regulating effects to control scaling, biofilm formation, and corrosion, while ensuring heat exchange efficiency. Existing technologies generally calculate the total chemical discharge by monitoring the concentration of chemicals at the discharge outlet, the wastewater volume, and the dosage data online. They then use predictive models to extrapolate theoretical discharge volumes, thereby assessing the rationality of chemical dosing and the stability of system operation. While existing online monitoring and data modeling methods can provide basic data support for emission compliance and chemical dosing management, they still lack effective identification and quantitative analysis of the complex physicochemical process of chemicals being temporarily adsorbed by scaling or biofilm during circulation.
[0003] In existing chemical dosing management schemes, a common practice is to directly adjust the dosage based on the difference between the theoretical and actual total emissions. However, this simplistic method has significant limitations. First, changes in operating conditions such as dosing device malfunctions, abnormal valve adjustments, flow fluctuations, or temporary discharges can all cause the actual total emissions to be lower than the theoretical value. Simply relying on the difference to adjust the dosage makes it difficult to distinguish these interfering factors from the actual adsorption effect of the chemicals, easily leading to excessive reduction in the dosage. Second, most existing deep learning or predictive models only focus on the input-output relationship of the chemicals and do not fully consider key process parameters directly related to scaling or biofilm formation, such as the coupled changes in the pressure difference between the heat exchanger inlet and outlet and the heat transfer coefficient. Therefore, they cannot accurately determine the root cause of the difference. Summary of the Invention
[0004] Based on this, the purpose of this invention is to provide a dosing management method, device, computer equipment, and storage medium based on a deep learning model. By leveraging deep learning and multi-parameter coupling analysis, it accurately identifies the phenomenon of pesticide adsorption. By comparing theoretical and actual total emissions and combining this with a coupling degree threshold, it accurately identifies the amount of adsorbed pesticide that must be removed before the next maintenance. Based on this, the difference is mapped to the dosing management of the next cycle, enabling a moderate increase in the dosage, thus preventing excessive emissions while maintaining heat exchange performance and water quality safety in the target cooling system.
[0005] In a first aspect, embodiments of this application provide a drug administration management method based on a deep learning model, comprising the following steps:
[0006] Obtain index data for a specific maintenance cycle of the target cooling system, the actual total amount of reagent discharged, and the predicted total amount of reagent discharged. The index data includes the amount of reagent discharged at the outlet for several identical dosing operations, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. The index data is input into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. Based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree, it is determined whether there is agent adsorption phenomenon in the specific maintenance cycle; If chemical adsorption occurs, calculate the difference between the actual total chemical emission and the predicted total chemical emission, and use this difference as the cleaning and removal dosage for the next specific maintenance cycle. Update the current dosing management plan based on the cleaning and removal dosage for the next specific maintenance cycle to obtain the updated dosing management plan.
[0007] Secondly, embodiments of this application provide a drug administration management device based on a deep learning model, comprising: The data acquisition module is used to obtain the index data, actual total discharge of the agent, and predicted total discharge of the agent for a specific maintenance cycle of the target cooling system. The index data includes the discharge port dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger for several identical dosing operations. The coupling degree calculation module is used to input the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. The agent adsorption judgment module is used to determine whether agent adsorption occurs during the specific maintenance cycle based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree. The scheme update module is used to calculate the difference between the actual total emission of the agent and the predicted total emission of the agent if the agent adsorption phenomenon exists, and use this difference as the cleaning and removal agent dosage for the next specific maintenance cycle; and update the current dosing management scheme according to the cleaning and removal agent dosage for the next specific maintenance cycle to obtain the updated dosing management scheme.
[0008] Thirdly, embodiments of this application provide a computer device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the drug administration method based on a deep learning model as described in the first aspect.
[0009] Fourthly, embodiments of this application provide a storage medium storing a computer program that, when executed by a processor, implements the steps of the drug administration management method based on a deep learning model as described in the first aspect.
[0010] This application provides a dosing management method, apparatus, computer device, and storage medium based on a deep learning model. By leveraging deep learning and multi-parameter coupling analysis, it accurately identifies pesticide adsorption phenomena. By comparing theoretical and actual total emissions and combining this with a coupling degree threshold, it accurately identifies the amount of adsorbed pesticide that must be removed before the next maintenance. Based on this, the difference is mapped to the dosing management of the next cycle, allowing for a moderate increase in the dosage, ensuring that the target cooling system maintains heat exchange performance and water quality safety while avoiding excessive emissions.
[0011] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description
[0012] Figure 1 A flowchart illustrating a drug administration management method based on a deep learning model, provided as an embodiment of this application; Figure 2 A schematic diagram of the process of a drug administration management method based on a deep learning model provided in another embodiment of this application; Figure 3 This is a schematic diagram of step S2 in a drug administration management method based on a deep learning model provided in one embodiment of this application; Figure 4 This is a schematic diagram of step S4 in a drug administration management method based on a deep learning model provided in one embodiment of this application; Figure 5 A schematic diagram of a drug administration device based on a deep learning model provided in one embodiment of this application; Figure 6 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. Detailed Implementation
[0013] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0014] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.
[0015] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."
[0016] Please see Figure 1 , Figure 1 The flowchart illustrates a drug administration management method based on a deep learning model, as provided in one embodiment of this application. The method includes the following steps: S1: Obtain indicator data for a specific maintenance cycle of the target cooling system, the actual total amount of chemicals discharged, and the predicted total amount of chemicals discharged.
[0017] The execution entity of the drug administration management method based on the deep learning model is the management device of the drug administration management method based on the deep learning model (hereinafter referred to as the management device). In an optional embodiment, the management device may be a computer device, a server, or a server cluster composed of multiple computer devices.
[0018] In an optional embodiment, the management device can obtain the index data of a specific maintenance cycle of the target cooling system, the actual total discharge of chemicals, and the predicted total discharge of chemicals by querying a preset database. The specific maintenance cycle refers to a fixed operation and maintenance cycle preset under normal and stable operating conditions to maintain the heat exchange performance and water quality safety of the target cooling system. At the end of each specific maintenance cycle, the heat exchangers and pipelines of the target cooling system that come into contact with the cooling water are cleaned or descaled.
[0019] The database originates from monitoring, control, and maintenance records accumulated during the long-term operation of the target cooling system. This includes data such as flow rate, concentration, pressure, and temperature collected by online monitoring instruments; dosage and timing recorded by the dosing control system; cleaning time and operating parameters recorded by the maintenance system; and environmental and process condition information. Because the cooling system operates continuously year-round with numerous monitoring points and high data collection frequency, the resulting database often reaches massive scales, meeting the requirements of deep learning and big data analysis. Preferably, the database should include at least the following types of data: dosing operation data (dosage and time), discharge port chemical concentration or dosage, inlet and outlet pressures of the heat exchanger, inlet and outlet temperatures of the heat exchanger, and the calculated heat transfer coefficient, makeup water volume and wastewater discharge volume, cooling load and its changing trends, and environmental climate data.
[0020] The dosage of pesticides at the discharge outlet refers to the cumulative amount of pesticides in the discharged water during the discharge interval between the end of one pesticide application and the start of the next pesticide application.
[0021] In this embodiment, the target cooling system can be understood as a circulating cooling water system used to remove waste heat from processes or equipment in industrial production or large public facilities, such as large chemical plants, metallurgical production lines, power units, data center cooling facilities, or central air conditioning cooling water systems. Such systems typically include cooling towers, heat exchangers, circulating water pumps, pipelines, and associated water treatment and chemical dosing devices. During operation, it is continuously necessary to add chemical agents with water stabilizing or cooling regulating effects, such as slow-release dispersants, complexing stabilizers, organic amine regulators, or other water conditioners containing surface-active groups. Because the cooling system operates continuously for extended periods and has a high process load, the amount of data generated is extremely large, providing a solid application scenario and data foundation for subsequent intelligent mining of massive amounts of data.
[0022] In practical operation, chemical agents with water quality stabilization or cooling regulation functions are key targets for emission supervision in the environmental protection field. This is because these agents often contain organophosphorus, organonitrogen, or other chemical groups that may affect aquatic ecosystems. Direct discharge into the environment with circulating cooling water could lead to environmental risks such as eutrophication of surface water or increased chemical oxygen demand (COD). Therefore, various environmental emission standards typically require online monitoring, total quantity calculation, and compliance reporting of the emissions of these agents. This invention, when acquiring a massive operational database of the target cooling system, incorporates the dosage of chemicals at the discharge outlet as key monitoring data into the analysis. This not only meets the management requirements for emission compliance but also provides a foundational data guarantee for subsequent deep learning coupling analysis and positive dosage correction.
[0023] A specific maintenance cycle refers to a fixed operation and maintenance cycle preset for a target cooling system under normal and stable operating conditions to maintain the system's heat exchange performance and water quality safety. Preferably, this cycle can be one week, one month, or one quarter, with one month or one quarter being the most common in industries such as chemical, power, and metallurgy. This is because within a month or one quarter of operation, the accumulation of dissolved salts, the proliferation of microbial colonies, and the resulting scale, biofilm, and other deposits in the cooling water will reach a level requiring unified cleaning. A cycle that is too short will increase maintenance costs, while a cycle that is too long will increase the risk of scaling. A specific maintenance cycle begins after the completion of the previous cleaning operation and ends before the start of the next cleaning operation. Within a cycle, chemical dosing operations are usually performed at the same dosage and interval to maintain water quality stability and facilitate the evaluation of dosing effectiveness; therefore, "multiple identical chemical dosing operations" can be considered objectively present.
[0024] The database originates from monitoring, control, and maintenance records accumulated during the long-term operation of the target cooling system. This includes data such as flow rate, concentration, pressure, and temperature collected by online monitoring instruments; dosage and timing recorded by the dosing control system; cleaning time and operating parameters recorded by the maintenance system; and environmental and process condition information. Because the cooling system operates continuously year-round with numerous monitoring points and high data collection frequency, the resulting database often reaches massive scales, meeting the requirements of deep learning and big data analysis. Preferably, the database should include at least the following types of data: dosing operation data (dosage and time), discharge port chemical concentration or dosage, inlet and outlet pressures of the heat exchanger, inlet and outlet temperatures of the heat exchanger and the calculated heat transfer coefficient, makeup water and wastewater discharge, cooling load and its changing trends, and environmental climate data.
[0025] The discharge outlet dosage refers to the cumulative dosage of pesticide in the discharged water during the discharge interval between the end of one dosing and the start of the next. It reflects the combined results of pesticide retention, consumption, and discharge within the system after dosing. The inlet and outlet pressure difference of a heat exchanger is the difference between the inlet and outlet pressures, indirectly reflecting the scaling or biofilm buildup on the inner wall of the heat exchanger: as scaling or biofilm increases, water flow resistance rises, and the pressure difference increases accordingly. The heat exchanger heat transfer coefficient refers to the amount of heat transferred per unit heat transfer area and per unit temperature difference. It comprehensively reflects the heat transfer performance of the heat exchanger; when scaling or biofilm forms on the heat exchanger surface, the heat transfer coefficient decreases. These three factors together constitute important indicators for monitoring the formation and adsorption behavior of scaling or biofilm in a system.
[0026] The actual total discharge of the agent is the total discharge of the agent calculated by integrating the concentration and discharge volume of the agent through online monitoring at the discharge outlet between the start of the maintenance and the start of the next maintenance in the specific maintenance cycle; the predicted total discharge of the agent is the theoretical total discharge of the agent obtained by modeling and extrapolating the agent discharge process in the specific maintenance cycle based on the index data and a deep learning model under the condition of stable dosage.
[0027] The predicted total chemical emissions can also be obtained using existing deep learning technology. Deep learning models have been used in cooling water chemical dosing management to model and predict theoretical chemical emission trends based on various operating conditions such as dosage, dosing interval, water temperature, system load, circulation rate, evaporation rate, and wastewater discharge. In an optional embodiment, the management device can use historical dosing data and operating condition data as input, and utilize an established or trainable deep learning time series model to calculate the theoretical total chemical emissions under ideal conditions of stable dosing and no abnormal adsorption.
[0028] Please see Figure 2 , Figure 2 A schematic diagram of the flow of a drug administration management method based on a deep learning model provided for another embodiment of this application includes step S5, which, prior to step S2, specifically is as follows: S5: Preprocess the discharge port dosage, heat exchanger inlet and outlet pressure difference, and heat exchanger heat transfer coefficient in the index data to obtain preprocessed index data.
[0029] In this embodiment, the management device preprocesses the discharge outlet drug dosage, heat exchanger inlet and outlet pressure difference, and heat exchanger heat transfer coefficient in the index data to obtain preprocessed index data. The preprocessing includes noise reduction and smoothing.
[0030] Specifically, the management equipment can select algorithms such as Kalman filtering, moving average, and exponential smoothing based on the operating characteristics of the target cooling system to eliminate high-frequency noise and outliers caused by transient flow changes, sensor jitter, and sampling errors, while retaining the true trend of data changes. The processed data can more realistically reflect the concentration decay after reagent addition, the pressure difference increase caused by scaling or biofilm accumulation, and the true dynamic changes in heat transfer performance, providing a reliable foundation for subsequent pattern recognition.
[0031] S2: Input the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger.
[0032] In this embodiment, the management device inputs the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. The deep learning model has been used in the field of cooling water dosing management to model and predict the theoretical emission trend of the agent based on various operating conditions such as dosage, dosing interval, water temperature, system load, circulation volume, evaporation volume, and sewage discharge volume.
[0033] In the initial stage after drug administration, some of the drug is adsorbed by the continuously growing scale or biofilm structure, resulting in a continuous decrease in the drug dosage detected at the discharge port. When the scale or biofilm approaches saturation, the adsorption rate slows down, and the drug dosage at the discharge port tends to stabilize. Subsequently, local damage to the scale or biofilm or changes in the microenvironment can cause the release of adsorbed drug, leading to a resurgence of the drug dosage at the discharge port. Therefore, if the deep learning model identifies this specific pattern, it can be assumed that the disappeared drug is most likely due to adsorption by scale or biofilm.
[0034] Please see Figure 3 , Figure 3 The schematic diagram of step S2 in the drug administration management method based on a deep learning model provided in one embodiment of this application includes steps S21 to S23, as detailed below: S21: Perform time-series modeling and trend graph construction on the index data to obtain the trend curve of drug dosage at the discharge port, the trend curve of pressure difference between the inlet and outlet of the heat exchanger, and the trend curve of heat transfer coefficient of the heat exchanger.
[0035] In this embodiment, the management device performs time-series modeling and trend graph construction on the indicator data based on the deep learning model to obtain the external discharge outlet drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve.
[0036] S22: Perform pattern recognition based on the external discharge port drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve to obtain the pattern recognition result for the specific maintenance cycle.
[0037] In the initial stage after dosing, some of the agent is adsorbed by the continuously growing scale or biofilm structure, resulting in a continuous decrease in the amount of agent detected at the discharge port. When the scale or biofilm is close to saturation, the adsorption rate slows down, and the amount of agent at the discharge port tends to stabilize. Subsequently, local damage to the scale or biofilm or changes in the microenvironment will cause the adsorbed agent to be released, causing the amount of agent at the discharge port to rise again.
[0038] In this embodiment, the management device, based on the deep learning model, performs pattern recognition according to the discharge port dosage trend curve, the heat exchanger inlet / outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve to obtain the pattern recognition results for the specific maintenance cycle. This is used to analyze whether the missing agent is due to scaling or biofilm adsorption. The pattern recognition results include the discharge port dosage transitioning from a continuous decreasing phase in the early stage, to a stable plateau phase in the middle stage, and to a rising phase in the later stage. The heat exchanger inlet / outlet pressure difference pattern recognition results include the heat exchanger inlet / outlet pressure difference transitioning from a continuous rising phase to a plateau phase. The heat exchanger heat transfer coefficient pattern recognition results include the heat exchanger heat transfer coefficient transitioning from a continuous decreasing phase to a plateau phase.
[0039] S23: Based on the pattern recognition results, identify the corresponding transition nodes of each stage in the external discharge port drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve; calculate the coupling degree based on the corresponding transition nodes of each stage in the external discharge port drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve to obtain the coupling degree.
[0040] In this embodiment, the management device identifies the corresponding transition nodes in each stage of the external discharge port drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve based on the pattern recognition results.
[0041] The management equipment calculates the coupling degree based on the corresponding transition nodes in the external discharge outlet drug dosage trend curve, the heat exchanger inlet / outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve. Specifically, the coupling degree can be calculated using methods such as proximity calculation, correlation coefficient, lag correlation analysis, dynamic time warping (DTW) similarity, trend consistency rate, or change point overlap rate to obtain the calculated values for each transition node. These calculated values are then averaged to obtain the coupling degree. This method further improves the accuracy and anti-interference capability of the coupling degree calculation, providing strong data analysis support.
[0042] S3: Based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree, determine whether there is an agent adsorption phenomenon in the specific maintenance cycle.
[0043] Fluctuations in operating conditions or external factors, such as occasional flow fluctuations, abnormal valve adjustments, or temporary discharges during system operation, may cause a short-term decrease, plateau, and rebound in the dosage at the discharge port. It should be noted that, under normal operating conditions, the inlet and outlet pressure difference and the heat exchanger heat transfer coefficient typically exhibit a pattern of continuous increase followed by a plateau, or continuous decrease followed by a plateau, as scaling or biofilm accumulation occurs.
[0044] To further determine that the reduction in drug dosage during the specific maintenance cycle is mainly caused by adsorption by biofilm or scale, and to avoid misjudgments due to other interfering factors, in this embodiment, the management device determines whether drug adsorption exists during the specific maintenance cycle based on the actual total drug emission, the predicted total drug emission, and the coupling degree. Specifically, if the actual total drug emission is less than the predicted total drug emission, and the coupling degree is higher than a set threshold, the management device determines whether drug adsorption exists during the specific maintenance cycle.
[0045] On the one hand, the actual total emissions being lower than the theoretical value indicates that some of the reagents were temporarily retained during the process from administration to discharge. On the other hand, the high coupling of the three key indicators suggests that this retention behavior is closely related to scaling or biofilm formation. Since the coupling threshold setting excludes other interfering factors such as occasional operating condition fluctuations, abnormal valve adjustments, and unplanned emissions, it can be reliably determined that the difference is mainly caused by scaling or biofilm adsorption, rather than other reasons. This determination is not only based on existing monitoring and prediction technologies but also significantly improves the accuracy and reliability of identifying scaling or biofilm adsorption phenomena through multi-parameter deep coupling verification.
[0046] S4: If chemical adsorption occurs, calculate the difference between the actual total chemical emission and the predicted total chemical emission, and use it as the cleaning and removal dosage for the next specific maintenance cycle; update the current dosing management plan based on the cleaning and removal dosage for the next specific maintenance cycle to obtain the updated dosing management plan.
[0047] If chemical adsorption exists, meaning the reduction in total chemical emissions primarily stems from adsorption by scaling or biofilm, without other abnormal interfering factors, then based on this reliable premise, the amount of cleaning chemicals removed in the next specific maintenance cycle can be considered as the total amount of chemicals adsorbed by biofilm or scaling during that cycle. This adsorbed amount of chemicals will be completely removed during system cleaning and descaling at the end of the cycle, thus preventing delayed release or potential environmental emissions risks in the next cycle.
[0048] Therefore, in this embodiment, the management device calculates the difference between the actual total discharge of the agent and the predicted total discharge of the agent, and uses this difference as the cleaning and removal agent dosage for the next specific maintenance cycle, as a safe replenishment amount, thereby ensuring the water quality regulation effect and avoiding excessive discharge.
[0049] The management equipment updates the current dosing management plan based on the cleaning and removal dosage for the next specific maintenance cycle, thus obtaining the updated dosing management plan.
[0050] Please see Figure 4 , Figure 4 The schematic diagram of step S4 in the drug administration management method based on a deep learning model provided in one embodiment of this application includes steps S41 to S42, as follows: S41: The cleaning and removal dosage for the next specific maintenance cycle is allocated according to the time sequence and dosage stage of each dosing operation within the next specific maintenance cycle, so as to obtain the increased dosage for each dosing stage of each dosing operation.
[0051] In this embodiment, the management device allocates the cleaning and removal drug dosage for the next specific maintenance cycle according to the time sequence and drug administration stage of each drug administration operation within the next specific maintenance cycle, thereby obtaining the increased drug dosage for each drug administration stage of each drug administration operation.
[0052] Furthermore, regarding the allocation of the "timing sequence of dosing operations" for the next specific maintenance cycle and the method of allocating operations that may include multiple dosing stages, this embodiment does not limit the specific dosing strategy optimization method. Instead, it uses the corrected safe dosing replenishment amount as a compensation input that can be directly applied to the existing dosing control system. Existing dosing plans in the art are usually arranged based on a predetermined dosing cycle, load changes, dosing methods (such as continuous dosing, pulse dosing, or time-segmented dosing), and dosage settings for different stages. In this embodiment, the management device allocates the compensation amount in a weighted manner according to factors such as the basic dosing amount ratio of each dosing stage, the operating load distribution, or historical adsorption contribution, to obtain the increased dosage of each dosing stage in each dosing operation.
[0053] S42: Amplify and correct the increased drug dosage at each stage of each drug administration operation and the preset natural consumption coefficient of the target cooling system to obtain the corrected increased drug dosage at each stage of each drug administration operation; update the current drug administration management scheme based on the corrected increased drug dosage at each stage of each drug administration operation to obtain the updated drug administration management scheme.
[0054] Considering that the agent will still be consumed in the cooling water due to normal reaction and natural decay after dosing, in this embodiment, the management device amplifies and corrects the increased dosage of each dosing stage of each dosing operation based on the preset natural consumption coefficient of the target cooling system, and obtains the corrected increased dosage of each dosing stage of each dosing operation. The natural consumption coefficient of the target cooling system can be directly obtained using the aforementioned deep learning model for generating the predicted total emission of the agent. This deep learning model has integrated long-term data on natural decay factors such as the chemical stability of the agent, the temperature of the circulating water, and the flow rate.
[0055] The management equipment updates the current dosing management scheme based on the increased dosage at each stage of the corrected dosing operation, obtaining an updated dosing management scheme. In an optional embodiment, to improve the scientific nature of the allocation, the management equipment can set a weighting coefficient based on factors such as the dosage of each dosing operation, cooling load, or historical adsorption rate. This optimizes the compensation accuracy of each dosing while ensuring consistency in the total amount, updating the current dosing management scheme to achieve positive correction management of subsequent dosing amounts. This significantly reduces the risk of misjudgment due to factors such as device failure or abnormal flow. Furthermore, through natural consumption coefficient correction and weighted allocation strategies, it ensures that the increased dosage is fully utilized, allowing the system to maintain heat exchange performance and water quality safety while avoiding excessive discharge, reducing operating costs and improving system stability.
[0056] Please refer to Figure 5 , Figure 5 This is a schematic diagram of a drug administration management device based on a deep learning model, provided in one embodiment of this application. The device can be implemented entirely or partially through software, hardware, or a combination of both. The deep learning model-based drug administration management device 5 includes: The data acquisition module 51 is used to obtain the index data of a specific maintenance cycle of the target cooling system, the actual total discharge of the agent and the predicted total discharge of the agent, wherein the index data includes the discharge port dosage of the agent in several identical dosing operations, the pressure difference between the inlet and outlet of the heat exchanger and the heat transfer coefficient of the heat exchanger. The coupling degree calculation module 52 is used to input the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger and the heat transfer coefficient of the heat exchanger. The agent adsorption judgment module 53 is used to determine whether there is an agent adsorption phenomenon in the specific maintenance cycle based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree. The scheme update module 54 is used to calculate the difference between the actual total emission of the agent and the predicted total emission of the agent if the agent adsorption phenomenon exists, and use it as the cleaning and removal agent dosage for the next specific maintenance cycle; and update the current dosing management scheme according to the cleaning and removal agent dosage for the next specific maintenance cycle to obtain the updated dosing management scheme.
[0057] In this embodiment, a data acquisition module obtains indicator data, actual total chemical discharge, and predicted total chemical discharge for a specific maintenance cycle of the target cooling system. The indicator data includes the discharge port dosage, heat exchanger inlet / outlet pressure difference, and heat exchanger heat transfer coefficient for several identical dosing operations. A coupling degree calculation module inputs the indicator data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port dosage, heat exchanger inlet / outlet pressure difference, and heat exchanger heat transfer coefficient. A chemical adsorption judgment module determines whether chemical adsorption exists in the specific maintenance cycle based on the actual total chemical discharge, predicted total chemical discharge, and coupling degree. A scheme update module calculates the difference between the actual total chemical discharge and the predicted total chemical discharge if chemical adsorption exists, using this difference as the cleaning and removal chemical dosage for the next specific maintenance cycle. The current dosing management scheme is updated based on the cleaning and removal chemical dosage for the next specific maintenance cycle to obtain an updated dosing management scheme. By leveraging deep learning and multi-parameter coupling analysis, chemical adsorption phenomena are accurately identified. By comparing theoretical and actual total emissions and combining this with a coupling degree threshold, the dosage of adsorbed chemicals that must be removed before the next maintenance can be accurately determined. Based on this, the difference is mapped to the dosing management of the next cycle, allowing for a moderate increase in the dosage, thus preventing excessive emissions while maintaining heat exchange performance and water quality safety in the target cooling system.
[0058] Please refer to Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. The computer device 6 includes: a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61; the computer device can store multiple instructions, which are adapted to be loaded and executed by the processor 61. Figures 1 to 4 The method steps of the illustrated embodiment can be found in the following documentation for detailed execution. Figures 1 to 4 The specific details of the illustrated embodiments will not be elaborated here.
[0059] The processor 61 may include one or more processing cores. The processor 61 connects to various parts of the server using various interfaces and lines. It executes various functions and processes data of the deep learning model-based drug delivery management device 5 by running or executing instructions, programs, code sets, or instruction sets stored in memory 62, and by accessing data in memory 62. Optionally, the processor 61 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 61 may integrate one or a combination of several of the following: a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required to be displayed on the touch screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 61 and may be implemented as a separate chip.
[0060] The memory 62 may include random access memory (RAM) or read-only memory. Optionally, the memory 62 may include a non-transitory computer-readable storage medium. The memory 62 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 62 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch instructions), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data involved in the various method embodiments described above, etc. Optionally, the memory 62 may also be at least one storage device located remotely from the aforementioned processor 61.
[0061] This application embodiment also provides a storage medium that can store multiple instructions, which are adapted to be loaded and executed by a processor as described above. Figures 1 to 4 For details of the method steps described in the embodiment, please refer to [link / reference]. Figures 1 to 4 The specific details of the embodiments will not be elaborated here.
[0062] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0063] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0064] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the algorithm. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0065] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.
[0066] 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 units can be selected to achieve the purpose of this embodiment according to actual needs.
[0067] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0068] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms.
[0069] This invention is not limited to the above-described embodiments. If any modifications or variations to this invention do not depart from the spirit and scope of this invention, and if such modifications and variations fall within the scope of the claims and equivalent technologies of this invention, then this invention also intends to include such modifications and variations.
Claims
1. A drug administration management method based on a deep learning model, characterized in that, Includes the following steps: Obtain index data for a specific maintenance cycle of the target cooling system, the actual total amount of reagent discharged, and the predicted total amount of reagent discharged. The index data includes the amount of reagent discharged at the outlet for several identical dosing operations, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. The index data is input into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. Based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree, it is determined whether there is agent adsorption phenomenon in the specific maintenance cycle; If chemical adsorption occurs, calculate the difference between the actual total chemical emission and the predicted total chemical emission, and use this difference as the cleaning and removal dosage for the next specific maintenance cycle. Update the current dosing management plan based on the cleaning and removal dosage for the next specific maintenance cycle to obtain the updated dosing management plan.
2. The drug administration management method based on a deep learning model according to claim 1, characterized in that: The specific maintenance cycle refers to a fixed maintenance cycle preset under normal and stable operating conditions to maintain the heat exchange performance and water quality safety of the target cooling system. At the end of each specific maintenance cycle, the heat exchangers and pipelines of the target cooling system, as well as other components in contact with the cooling water, are cleaned or descaled.
3. The drug administration management method based on a deep learning model according to claim 1, characterized in that: The dosage of pesticides at the discharge outlet refers to the cumulative amount of pesticides in the discharged water during the discharge interval between the end of one pesticide application and the start of the next pesticide application.
4. The drug administration management method based on a deep learning model according to claim 1, characterized in that: Before inputting the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the inlet and outlet pressure difference of the heat exchanger, and the heat transfer coefficient of the heat exchanger, the following steps are included: The dosage of the drug at the outlet, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger in the index data are preprocessed to obtain the preprocessed index data.
5. The drug administration management method based on a deep learning model according to claim 4, characterized in that: The step of inputting the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the inlet and outlet pressure difference of the heat exchanger, and the heat transfer coefficient of the heat exchanger includes the following steps: Time series modeling and trend graph construction are performed on the aforementioned index data to obtain the trend curve of drug dosage at the discharge port, the trend curve of pressure difference between the inlet and outlet of the heat exchanger, and the trend curve of heat transfer coefficient of the heat exchanger. Pattern recognition is performed based on the trends of the discharge port dosage, the heat exchanger inlet / outlet pressure difference, and the heat exchanger heat transfer coefficient to obtain the pattern recognition results for the specific maintenance cycle. The pattern recognition results include the discharge port dosage transitioning from a continuously decreasing phase in the early stage, to a stable plateau phase in the middle stage, and then to an increasing phase in the later stage. The heat exchanger inlet / outlet pressure difference pattern recognition results include the heat exchanger inlet / outlet pressure difference transitioning from a continuously increasing phase to a plateau phase. The heat exchanger heat transfer coefficient pattern recognition results include the heat exchanger heat transfer coefficient transitioning from a continuously decreasing phase to a plateau phase. Based on the pattern recognition results, the corresponding transition nodes of each stage in the external discharge outlet drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve are identified. The coupling degree is calculated based on the corresponding conversion nodes of each stage in the external discharge port drug dosage trend curve, the heat exchanger inlet and outlet pressure difference trend curve, and the heat exchanger heat transfer coefficient trend curve.
6. The drug administration management method based on a deep learning model according to claim 5, characterized in that, The actual total discharge of the agent is the total discharge of the agent calculated by integrating the concentration and discharge volume of the agent through online monitoring at the discharge outlet between the start of the maintenance and the start of the next maintenance in the specific maintenance cycle; the predicted total discharge of the agent is the theoretical total discharge of the agent obtained by modeling and extrapolating the agent discharge process in the specific maintenance cycle based on the index data and a deep learning model under the condition of stable dosage.
7. The drug administration management method based on a deep learning model according to claim 6, characterized in that, The step of updating the current dosing management plan based on the cleaning and removal dosage for the next specific maintenance cycle to obtain the updated dosing management plan includes the following steps: The cleaning and removal dosage for the next specific maintenance cycle is allocated according to the time sequence and dosage stage of each dosing operation within the next specific maintenance cycle, so as to obtain the increased dosage for each dosing stage of each dosing operation. The increased drug dosage for each stage of each drug administration operation is amplified and corrected based on the preset target cooling system's natural consumption coefficient to obtain the corrected increased drug dosage for each stage of each drug administration operation. The current drug administration management plan is then updated based on the corrected increased drug dosage for each stage of each drug administration operation to obtain the updated drug administration management plan.
8. A drug administration management device based on a deep learning model, characterized in that, include: The data acquisition module is used to obtain the index data, actual total discharge of the agent, and predicted total discharge of the agent for a specific maintenance cycle of the target cooling system. The index data includes the discharge port dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger for several identical dosing operations. The coupling degree calculation module is used to input the index data into a preset deep learning model for pattern recognition and coupling degree analysis to obtain the coupling degree of the discharge port drug dosage, the pressure difference between the inlet and outlet of the heat exchanger, and the heat transfer coefficient of the heat exchanger. The agent adsorption judgment module is used to determine whether agent adsorption occurs during the specific maintenance cycle based on the actual total emission of the agent, the predicted total emission of the agent, and the coupling degree. The scheme update module is used to calculate the difference between the actual total emission of the agent and the predicted total emission of the agent if the agent adsorption phenomenon exists, and use this difference as the cleaning and removal agent dosage for the next specific maintenance cycle; and update the current dosing management scheme according to the cleaning and removal agent dosage for the next specific maintenance cycle to obtain the updated dosing management scheme.
9. A computer device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the drug administration method based on a deep learning model as described in any one of claims 1 to 7.
10. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the drug administration method based on a deep learning model as described in any one of claims 1 to 7.