Environmental monitoring data classification management method and system based on digital twinning

Abstract

The invention is suitable for the technical field of environmental monitoring data processing, and provides an environmental monitoring data classification management method and system based on digital twinning, and the method comprises the steps: obtaining the ammonia nitrogen concentration monitoring data of a target reservoir inlet within a specified time period after rainstorm, and determining an initial goodness of fit value corresponding to the monitoring data in the digital twin model, and obtaining historical archiving data of the target reservoir. According to the method, a double-correction-factor mechanism constructed based on the trend slope and the temperature disturbance frequency is introduced, dynamic correction of the initial goodness of fit value of the digital twin model is achieved, and the problem that in the prior art, response out-of-control and evaluation distortion are likely to happen to the model in a specific rainstorm scene is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental monitoring data processing, and in particular relates to a method and system for environmental monitoring data classification management based on digital twins. Background Art

[0002] In the current environmental monitoring system, more and more regions are beginning to use digital twin models to predict and classify water pollutants. This type of model builds a simulation system corresponding to the real environment and uses real-time or historical monitoring data for dynamic simulation. It can predict the trend of pollutant concentration changes and assist environmental protection departments in achieving early warning, pollution tracking, and water quality classification. In practical applications, digital twin technology has been widely used in the simulation management of water environments such as rivers, reservoirs, and lakes, especially in the monitoring and trend analysis of common pollutants such as ammonia nitrogen and total phosphorus. It has a certain modeling foundation and promotion value.

[0003] However, the digital twin models used in existing technologies mostly rely on established parameters and static rules for simulation and evaluation, and are unable to fully cope with the complex pollution changes caused by sudden hydrological events such as rainstorms. In rainstorm scenarios, surface runoff quickly flows into the water body, which may carry a large amount of pollutants and cause drastic fluctuations in water quality, accompanied by other coupling factors such as temperature disturbances. Under such nonlinear dynamic conditions, the fit values ​​generated by the original model often deviate due to insufficient response to changes in pollution trends or lack of consideration of the impact of disturbances, resulting in the model's classification of pollution characteristics inconsistent with reality, affecting the stability of the model and the accuracy of subsequent judgments. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for environmental monitoring data classification management based on digital twins, aiming to solve the problems raised in the background technology.

[0005] The present invention is implemented as follows: a method for classifying and managing environmental monitoring data based on digital twins, the method comprising: Obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after the rainstorm, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir; Perform several groups of parameter debugging actions on the digital twin model, record the changes in the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this; If it is determined that there is a systematic deviation, obtain some local historical archived data consistent with the current rainstorm intensity, and build a trend model of ammonia nitrogen concentration based on it, and set the first correction factor based on the trend model; Detect whether the drop in water temperature after the current rainstorm meets the condition of exceeding the preset threshold. If so, count the occurrence frequency of the corresponding situation in the above local historical archived data, and set the second correction factor according to the frequency; The first correction factor and the second correction factor are jointly applied to the initial goodness of fit value to generate a corrected goodness of fit value, which is used for classification label determination of subsequent ammonia nitrogen concentration monitoring data.

[0006] As a further limitation of the technical solution of the embodiment of the present invention, the steps of executing several groups of parameter debugging behaviors on the digital twin model, recording the change of the goodness of fit value after each group of debugging, and analyzing whether there is a systematic deviation in the initial goodness of fit value based on the change include: Restarting the processing program for generating the goodness-of-fit value for the ammonia nitrogen concentration monitoring data in the digital twin model several times, and performing a preset amplitude debugging operation on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; Record the test fit value after each group of debugging, and calculate the change range of each test fit value compared with the initial fit value, and determine whether the change range exceeds the preset range threshold; If the proportion of the number of debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset proportion threshold, it is determined that there is a systematic deviation in the initial fit value.

[0007] As a further limitation of the technical solution of the embodiment of the present invention, if it is determined that there is a systematic deviation, a number of local historical archived data consistent with the current rainstorm intensity are obtained, and an ammonia nitrogen concentration change trend model is constructed based on the data, and the step of setting the first correction factor based on the trend model includes: After determining that the initial fit value has systematic deviations, the historical archived data of the target reservoir are analyzed, and several local historical archived data that match the current rainstorm intensity are selected; Each selected local historical archive data is processed to extract the average historical ammonia nitrogen concentration values ​​within a specified period of time after the rainstorm, and arranged in chronological order to construct an ammonia nitrogen concentration change trend model; The average slope of the constructed ammonia nitrogen concentration change trend model was calculated and used as the first correction factor.

[0008] As a further limitation of the technical solution of the embodiment of the present invention, the step of detecting whether the drop in the temperature of the water body after the current rainstorm satisfies the condition of exceeding a preset threshold value, and if so, counting the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data, and setting the second correction factor according to the frequency includes: Based on historical archived data, the water temperature data of the target reservoir inlet in the preset time period before and after the current rainstorm is extracted, and it is determined whether the drop in water temperature after the rainstorm exceeds the preset threshold; If it is determined that the drop exceeds the preset threshold, each local historical archive data is analyzed to determine whether there is a situation in which the water temperature drops by more than the preset threshold in the preset time period before and after the rainstorm; The occurrence frequency of local historical archived data that meets the above conditions is counted, and the occurrence frequency is used as a basis to set the second correction factor.

[0009] As a further limitation of the technical solution of the embodiment of the present invention, the step of applying the first correction factor and the second correction factor together to the initial goodness of fit value to generate a corrected goodness of fit value for subsequent classification label determination of ammonia nitrogen concentration monitoring data includes: Calling a preset goodness of fit value correction calculation formula, substituting the first correction factor and the second correction factor into the goodness of fit value correction calculation formula to correct the initial goodness of fit value, and obtaining a corrected goodness of fit value; The classification label determination operation of the subsequent ammonia nitrogen concentration monitoring data is performed based on the corrected fit value.

[0010] As a further limitation of the technical solution of the embodiment of the present invention, the calculation formula for correcting the numerical value of the fit degree is: ,in Refers to the corrected goodness of fit value. Refers to the initial goodness of fit value, Refers to the first correction factor, that is, the average slope of the ammonia nitrogen concentration trend model. Refers to the adjustment weight corresponding to the first correction factor, Refers to the second correction factor, that is, the frequency of occurrence of local historical archived data that meets the conditions. Refers to the adjustment weight corresponding to the second correction factor; In the calculation formula for the numerical correction of the degree of fit ,in Refers to the total amount of local historical archived data. Refers to the trend model of ammonia nitrogen concentration change. The changing slope of the segment line; ,in Refers to the number of local historical archive data that meets the conditions.

[0011] A digital twin-based environmental monitoring data classification management system, the system comprising: a data acquisition module, a systematic deviation judgment module, a first correction factor determination module, a second correction factor determination module and a goodness of fit value correction module, wherein: The data acquisition module is used to obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after the rainstorm, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir; The systematic deviation judgment module is used to perform several groups of parameter debugging behaviors on the digital twin model, record the changes in the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this; A first correction factor determination module is used to obtain a number of local historical archived data consistent with the current rainstorm intensity if it is determined that there is a systematic deviation, and to construct an ammonia nitrogen concentration change trend model based on the data, and to set a first correction factor based on the trend model; The second correction factor determination module is used to detect whether the drop in water temperature after the current rainstorm meets the condition of exceeding a preset threshold. If so, the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data is counted, and the second correction factor is set according to the frequency; The goodness of fit value correction module is used to jointly apply the first correction factor and the second correction factor to the initial goodness of fit value to generate a corrected goodness of fit value for use in the classification label determination of subsequent ammonia nitrogen concentration monitoring data.

[0012] As a further limitation of the technical solution of the embodiment of the present invention, the systematic deviation judgment module specifically includes: A model debugging unit, used to restart the processing program for generating a goodness of fit value for ammonia nitrogen concentration monitoring data in the digital twin model several times, and to perform a debugging operation of a preset amplitude on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; A change range calculation unit is used to record the test fit value of each group after debugging, and calculate the change range of each test fit value compared with the initial fit value, and determine whether each change range exceeds a preset range threshold; The proportion comparison unit is used to determine that there is a systematic deviation in the initial fit value if the proportion of the debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset proportion threshold.

[0013] As a further limitation of the technical solution of the embodiment of the present invention, the first correction factor determination module specifically includes: A data screening unit is used to analyze the historical archived data of the target reservoir and screen out a number of local historical archived data that match the current rainstorm intensity after determining that the initial fit value has a systematic deviation; The trend model building unit is used to process each selected local historical archive data, extract the average historical ammonia nitrogen concentration value within a specified time period after the rainstorm, and arrange them in chronological order to build an ammonia nitrogen concentration change trend model; The average slope calculation unit is used to calculate the average slope of the constructed ammonia nitrogen concentration change trend model and use the average slope as the first correction factor.

[0014] As a further limitation of the technical solution of the embodiment of the present invention, the second correction factor determination module specifically includes: The water temperature analysis unit is used to extract the water temperature data of the target reservoir inlet within a preset time period before and after the current rainstorm based on the historical archived data, and determine whether the drop in water temperature after the rainstorm exceeds a preset threshold; A drop range judgment unit is used to analyze each local historical archive data to determine whether there is a situation in which the water body temperature drops by more than the preset threshold value in the preset time period before and after the rainstorm if it is determined that the drop range exceeds the preset threshold value; The occurrence frequency calculation unit is used to count the occurrence frequency of the local historical archived data that meets the above conditions, and use the occurrence frequency as a basis to set the second correction factor.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention achieves dynamic correction of the initial fit value of the digital twin model by introducing a dual correction factor mechanism based on trend slope and temperature disturbance frequency, solving the problem of uncontrolled response and evaluation distortion in the prior art model under specific rainstorm scenarios. The average slope is extracted by constructing an ammonia nitrogen concentration change trend model to capture the temporal characteristics of the pollution response. At the same time, the historical frequency of water temperature changes after rainstorms is combined to reflect the potential impact of environmental disturbances on simulation accuracy, so that the model correction process has both internal response and external drive dimensions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a method provided by an embodiment of the present invention; Figure 2 A flow chart of determining whether there is a systematic deviation in the initial goodness of fit value in the method provided in an embodiment of the present invention; Figure 3 A flow chart of generating a first correction factor in the method provided in an embodiment of the present invention; Figure 4 A flow chart of generating a second correction factor in the method provided in an embodiment of the present invention; Figure 5 A flow chart of generating a corrected fit value in the method provided in an embodiment of the present invention; Figure 6 An application architecture diagram of a system provided by an embodiment of the present invention; Figure 7 A structural block diagram of a systematic deviation judgment module in a system provided by an embodiment of the present invention; Figure 8 A structural block diagram of a first correction factor determination module in a system provided by an embodiment of the present invention; Fig. 9 This is a structural block diagram of a second correction factor determination module in a system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] Figure 1 A flow chart of a method provided by an embodiment of the present invention is shown.

[0019] Specifically, a method for classifying and managing environmental monitoring data based on digital twins comprises the following steps: Step S100, obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after the heavy rain occurs, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir.

[0020] In an embodiment of the present invention, the inlet of the target reservoir is selected as the location for collecting ammonia nitrogen concentration monitoring data, based on the fact that this location can most directly reflect the input of surface runoff caused by rainstorm events to the water quality of the reservoir. After a rainstorm, the surface runoff that flows into the reservoir usually carries nitrogen pollutants, especially ammonia nitrogen, which has sensitive and fast-responding characteristics and often shows a trend of short-term sudden increases. By monitoring the changes in ammonia nitrogen concentration at the inlet, the initial characteristics of pollution input can be captured in time before the water quality changes affect the entire water body, thereby improving the accuracy of pollution response and the timeliness of early warning.

[0021] The purpose of collecting ammonia nitrogen concentration monitoring data within a specified time period after a rainstorm is to align the time period with historical data and ensure consistency between trend analysis and model comparison. This time period is generally preset based on factors such as reservoir catchment area, flow rate, and water body renewal cycle. It can effectively cover the high-frequency change range of pollution impacts and is a key period for establishing concentration change trend models and identifying model deviations.

[0022] The acquisition of ammonia nitrogen concentration monitoring data can be achieved through the online water quality monitoring device deployed at the inlet of the target reservoir, which can use mature monitoring technologies such as spectroscopy, electrode method or automated colorimetry for real-time collection. The monitoring equipment is usually integrated with the reservoir digital twin system, uploaded to the model platform in real time through the hydrological information system, and automatically participates in data modeling and parameter evaluation.

[0023] The initial fit value refers to the degree of consistency index obtained by comparing the digital twin system with the actual collected monitoring data after simulating and predicting the ammonia nitrogen concentration based on the current model structure, input parameters and historical training logic. The fit value can be calculated through a variety of mathematical indicators, such as root mean square error (RMSE), mean absolute error (MAE), relative deviation rate, goodness of fit, etc. The purpose of generating the fit value is to measure the accuracy of the current digital twin model's response to actual water quality changes. In the present invention, the initial fit value is not only used as an evaluation indicator of water quality simulation accuracy, but also as a key reference parameter in classification label determination.

[0024] The historical archived data of the target reservoir can be retrieved from the environmental monitoring system, water information platform, historical data center or remote sensing database, usually including monitoring data of multiple time periods and their corresponding meteorological, hydrological and pollutant characteristic information. The historical archived data described in the present invention refers to the original monitoring records and related environmental parameter information collected and stored in the target reservoir in multiple historical time periods, usually including but not limited to: ammonia nitrogen concentration monitoring data at various stages before and after the rainstorm, used to restore the time series changes of ammonia nitrogen concentration under the condition of rainstorm of specific intensity; meteorological data records related to the intensity of the rainstorm, including cumulative rainfall, hourly rainfall intensity and rainfall distribution process; water temperature monitoring data in the corresponding time period, used to describe the change of water temperature after the rainstorm.

[0025] Furthermore, the digital twin-based environmental monitoring data classification management method also includes the following steps: Step S200, execute several groups of parameter debugging behaviors on the digital twin model, record the change of the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this.

[0026] Specifically, Figure 2 A flow chart for determining whether there is a systematic bias in the initial goodness of fit values ​​is shown.

[0027] Among them, several groups of parameter debugging behaviors are performed on the digital twin model, and the changes in the fit value after each group of debugging are recorded. Based on this, it is analyzed whether there is a systematic deviation in the initial fit value. The specific steps include: Step S201, restarting the processing program for generating a goodness-of-fit value for ammonia nitrogen concentration monitoring data in the digital twin model several times, and performing a debugging operation of a preset amplitude on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; Step S202, recording the test fit value of each group after debugging, and calculating the change range of each test fit value compared with the initial fit value, and judging whether each change range exceeds a preset range threshold; Step S203: If the proportion of the number of debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset ratio threshold, it is determined that there is a systematic deviation in the initial fit value.

[0028] In an embodiment of the present invention, several groups of parameter debugging behaviors are performed on the digital twin model, specifically by restarting the processing program in the model for generating the goodness-of-fit value for the ammonia nitrogen concentration monitoring data multiple times. The processing program is usually integrated into the digital twin system in a modular manner, and can be automatically called or batch triggered in the system, and each call runs independently. The number of restarts is comprehensively set according to the number of model control parameters, the accuracy requirements of the system response analysis, and the acceptable degree of model running time, and is preferably set to more than five times to ensure that the subsequent analysis is representative and statistically stable. The basis for setting the number of times usually includes the empirical value of parameter sensitivity analysis, the processing capacity of the simulation platform, and the fluctuation characteristics of the model output caused by the debugging behavior.

[0029] During each model startup, at least one model control parameter that is directly related to the predicted result of ammonia nitrogen concentration is debugged once with a set amplitude. Such model control parameters may include pollutant diffusion rate, water mixing coefficient, weight of water temperature on reaction rate, ammonia nitrogen release coefficient in bottom sediment, inlet boundary condition setting value, inlet flow rate correction term, and weight of historical trend fitting in the model. The amplitude of each debugging operation is set to a moderate deviation of the parameter within the allowable range of the model. It is usually set with reference to the disturbance range recommended in the model manual and combined with the results of historical sensitivity experiments on the parameters, such as increasing or decreasing the original value by a certain percentage to ensure that each disturbance will not cause the model structure to become unstable and significantly affect the simulation results.

[0030] After each set of debugging is completed, the system records the corresponding test fit value and compares it with the initial fit value, and calculates the change in the fit value after each debugging compared to the initial value. In actual implementation, the system's built-in data difference analysis tool can be used to output each change result according to the configured calculation logic and record it uniformly in the model response log. The system determines the amplitude of each change to determine whether it exceeds the preset amplitude threshold, which is determined by the model error tolerance range and is usually set to a fixed percentage range of the initial fit value.

[0031] After completing all debugging behaviors, count the number of times the change in the degree of fit value exceeds the preset amplitude threshold in all debugging, and calculate the proportion of such times to the total number of debugging times. When the statistical results show that the proportion of debugging times exceeding the preset amplitude threshold exceeds the set proportion threshold, the system will determine that there is a systematic deviation in the initial degree of fit value. This deviation phenomenon reflects that under the complex input background of the current rainstorm event, the simulation results of the digital twin model for ammonia nitrogen concentration are highly sensitive to the model control parameters, and the response of the degree of fit value to the debugging results is unstable, which can easily cause drastic fluctuations in the model output due to slight disturbances.

[0032] This phenomenon shows that when the digital twin model is used to evaluate the fit of the rainstorm-ammonia nitrogen response mechanism, there may be structural imbalance, simulation logic loss of control, or evaluation mechanism distortion due to insufficient scene adaptability, resulting in the original fit value being unable to truly reflect the degree of match between the current model and the actual monitoring data. Therefore, it is necessary to further introduce external objective factors corresponding to the current scene, perform model correction or evaluation value adjustment, in order to improve the accuracy of the classification judgment results and the adaptability of the model.

[0033] Furthermore, the digital twin-based environmental monitoring data classification management method also includes the following steps: Step S300: If it is determined that there is a systematic deviation, obtain a number of local historical archived data consistent with the current rainstorm intensity, and construct an ammonia nitrogen concentration change trend model based on the data, and set a first correction factor based on the trend model.

[0034] Specifically, Figure 3 A flow chart for generating a first correction factor is shown.

[0035] If it is determined that there is a systematic deviation, a number of local historical archived data consistent with the current rainstorm intensity are obtained, and a trend model of ammonia nitrogen concentration change is constructed based on the data. The first correction factor is set based on the trend model, which specifically includes the following steps: Step S301, after determining that there is a systematic deviation in the initial fit value, analyzing the historical archived data of the target reservoir, and screening out a number of local historical archived data that match the current rainstorm intensity; Step S302, processing each of the selected local historical archived data, extracting the average historical ammonia nitrogen concentration values ​​within a specified time period after the rainstorm, and arranging them in chronological order to construct an ammonia nitrogen concentration change trend model; Step S303, calculating the average slope of the constructed ammonia nitrogen concentration change trend model, and using the average slope as the first correction factor.

[0036] In the embodiment of the present invention, when the initial fit value is judged to have a systematic deviation, in order to further identify the performance characteristics of the deviation in a specific rainstorm scene and thus achieve targeted correction, it is necessary to perform model attribution and trend analysis based on historical scene data similar to the current monitoring scene. To this end, a number of local historical archived data matching the current rainstorm intensity are selected from the historical archived data of the target reservoir as the sample basis for the subsequent construction of the ammonia nitrogen concentration change trend model.

[0037] The screening of local historical archived data that matches the current rainstorm intensity is mainly to ensure that the construction basis of the trend model has situational consistency and environmental comparability. Rainstorm intensity is one of the important external driving factors that affect the change of ammonia nitrogen concentration. Rainstorms of different intensities will trigger completely different pollution impact processes and water quality response patterns. If historical data with a large difference from the current rainstorm intensity are directly used for trend modeling, non-representative patterns may be introduced, resulting in a deviation in the correction results, and the actual evolution characteristics of ammonia nitrogen concentration in the current scene cannot be effectively restored. Therefore, by matching the rainstorm intensity, it can be ensured that the selected sample data has sufficient scene similarity, providing relevant sample support for subsequent model calculations.

[0038] After completing the sample data screening, each local historical archived data is processed to extract the average historical ammonia nitrogen concentration values ​​within the specified time period after the rainstorm, and arranged in chronological order. This processing process can ensure the consistency of different historical data in the time dimension by setting a unified statistical window after the rainstorm, such as 24 hours or 48 hours. Subsequently, each historical data point is sequentially established as a sequence with the horizontal axis as the historical time and the vertical axis as the average ammonia nitrogen concentration value, forming a data structure of the concentration change trend. This trend data structure serves as the basis for modeling and can be used to extract change characteristics, evaluate the steepness of the trend, and identify systematic deviation patterns.

[0039] After the ammonia nitrogen concentration change trend model is built, the average slope of the model is calculated to obtain the overall change trend of ammonia nitrogen concentration in this type of rainstorm scenario. The average slope reflects the overall rate of change of ammonia nitrogen concentration over time in historical scenarios. It is a key parameter that can characterize the pollution load intensity, pollutant input rate and system response characteristics. The purpose of using the average slope as the first correction factor is that in the current scenario, if the digital twin model does not fully reflect similar concentration change intensity or trend characteristics, the correction factor can be used to compensate the trend of the initial fit value, so that it is closer to the simulation behavior that the model should have in real situations. By introducing the trend slope as a dynamic adjustment amount, the system's response adaptability to sudden environmental events such as "rainstorm-pollution" can be enhanced, and the reliability and actual representativeness of the model output in the classification management process can be improved.

[0040] Furthermore, the digital twin-based environmental monitoring data classification management method also includes the following steps: Step S400, detecting whether the drop in water temperature after the current rainstorm meets the condition of exceeding a preset threshold. If so, counting the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data, and setting a second correction factor based on the frequency.

[0041] Specifically, Figure 4 A flow chart for generating a second correction factor is shown.

[0042] Among them, detecting whether the drop in water temperature after the current rainstorm meets the condition of exceeding the preset threshold, if so, counting the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data, and setting the second correction factor according to the frequency specifically includes the following steps: Step S401, extracting the water temperature data of the target reservoir inlet within a preset time period before and after the current rainstorm based on the historical archived data, and determining whether the drop in water temperature after the rainstorm exceeds a preset threshold; Step S402: if it is determined that the drop exceeds the preset threshold, each local historical archived data is analyzed to determine whether there is a situation in which the drop in water temperature exceeds the preset threshold in the preset time period before and after the rainstorm; Step S403: Count the occurrence frequency of local historical archived data that meets the above conditions, and use the occurrence frequency as a basis to set the second correction factor.

[0043] In an embodiment of the present invention, judging whether the drop in water temperature after heavy rain exceeds a preset threshold is to identify whether there is a typical rapid water temperature change scenario, which often affects the diffusion behavior, reaction rate and model fitting ability of pollutants. In actual hydrological processes, a large amount of surface runoff brought by heavy rain is usually accompanied by low-temperature inflow. If the inflow temperature is significantly lower than the water temperature in the reservoir area, it may cause a sharp drop in local temperature. Water temperature is a key influencing factor in controlling the ammonia nitrogen conversion process (such as ammonia volatilization, nitrification reaction rate, etc.). Rapid changes in temperature will cause distortion of certain response parameters in the digital twin model, thereby affecting the model's simulation accuracy of changes in ammonia nitrogen concentration.

[0044] The drop in water temperature can be calculated by comparing the temperature values ​​before and after the rainstorm. The preferred method is to extract the average water temperature in the set time period before the rainstorm and the average water temperature in the same time period after the rainstorm. The difference between the two is the drop. If the drop exceeds the temperature threshold set in the system (for example, 2°C or 3°C), it is considered that there is a significant temperature disturbance, which may be one of the potential causes for the current model to deviate from reality.

[0045] After determining that the current temperature change exceeds the preset threshold, the selected local historical archived data are analyzed one by one to determine whether there is a similar situation where the water temperature drops by more than the preset threshold before and after the corresponding historical rainstorm event. This process is to find historical samples that are consistent with the current temperature change characteristics and confirm whether such scenes are representative or common in historical data. If a certain type of feature appears repeatedly in historical samples and fails to get a reasonable response in the model, it means that the feature is reasonable to be included in the correction system.

[0046] The number of historical data with significant temperature drop characteristics is further counted, and its proportion in all local historical archived data is calculated, which is the frequency of occurrence of this situation. This frequency reflects the prevalence of such temperature disturbance scenarios in historical data, and is an objective basis for measuring whether the current model needs targeted corrections. If the frequency is high, it means that such disturbances are common scenarios, but the model's response capability is insufficient and there are systematic adaptation problems; if the frequency is low, it indicates that the scenario is an occasional situation, and the correction intensity can be weakened accordingly. By using this frequency as the source parameter of the second correction factor, dynamic adjustment of the model evaluation value can be achieved, the sensitivity of the digital twin system to special hydrological and water quality coupling events can be enhanced, and the adaptability and accuracy of classification label determination under multi-source interference can be improved.

[0047] Furthermore, the digital twin-based environmental monitoring data classification management method also includes the following steps: Step S500, applying the first correction factor and the second correction factor together to the initial goodness of fit value to generate a corrected goodness of fit value for use in subsequent classification label determination of ammonia nitrogen concentration monitoring data.

[0048] Specifically, Figure 5 A flow chart for generating a revised goodness-of-fit value is shown.

[0049] The first correction factor and the second correction factor are jointly applied to the initial goodness of fit value to generate a corrected goodness of fit value, which is used for the classification label determination of subsequent ammonia nitrogen concentration monitoring data. Specifically, the following steps are included: Step S501, calling a preset goodness of fit value correction calculation formula, substituting a first correction factor and a second correction factor into the goodness of fit value correction calculation formula to correct the initial goodness of fit value, and obtaining a corrected goodness of fit value; Step S502, performing a classification label determination operation on subsequent ammonia nitrogen concentration monitoring data based on the corrected fit value.

[0050] The calculation formula for correcting the numerical value of the fit is: ,in Refers to the corrected goodness of fit value. Refers to the initial goodness of fit value, Refers to the first correction factor, that is, the average slope of the ammonia nitrogen concentration trend model. Refers to the adjustment weight corresponding to the first correction factor, Refers to the second correction factor, that is, the frequency of occurrence of local historical archived data that meets the conditions. Refers to the adjustment weight corresponding to the second correction factor; In the calculation formula for the numerical correction of the degree of fit ,in Refers to the total amount of local historical archived data. Refers to the trend model of ammonia nitrogen concentration change. The changing slope of the segment line; ,in Refers to the number of local historical archive data that meets the conditions.

[0051] In an embodiment of the present invention, the first correction factor and the second correction factor are selected to be jointly applied to the initial fit value, with the aim of realizing multi-dimensional dynamic correction of the output results of the digital twin model under a specific rainstorm scenario. The first correction factor is based on the average slope extracted from the ammonia nitrogen concentration trend model, which reflects the temporal evolution speed and intensity of the target pollutant in historically similar scenarios, and is the internal dynamic characteristic of the pollution response; the second correction factor is derived from the frequency of occurrence of water temperature changes, representing the external influencing factors of environmental conditions on the diffusion and reaction process of pollutants. These two types of correction factors start from the two dimensions of the result trend of pollution behavior and the control variables of the pollution process, complement each other, and can form a more complete correction logic system.

[0052] Combining these two types of correction factors can effectively improve the accuracy of identifying and adjusting systematic deviations. When the model has simulation structure or parameter adaptation problems in the current scenario, the single-dimensional correction method may cause an imbalance in the correction strength due to the one-sided reflection of a certain type of deviation factor; and the dual-factor joint adjustment mechanism can cover more types of deviation sources, and achieve systematic comprehensive compensation from model trend misfit to external disturbance insensitivity. This joint correction method not only improves the accuracy of classification label determination, but also has the ability to dynamically adapt to actual scene changes without increasing the complexity of the model structure. It is a lightweight technical path that can greatly optimize the application effect without changing the main body of the model.

[0053] The following are some examples: The current monitoring data of ammonia nitrogen concentration in the target reservoir within 48 hours after a rainstorm is compared with the predicted value of the digital twin model. The system calculates an initial consistency value of 0.83.

[0054] The system then starts several sets of parameter debugging programs to perturb the model control parameters by a preset amplitude. Statistical analysis shows that there are systematic deviations in the initial fit values.

[0055] According to the current rainstorm intensity, the system screened out five matching local historical archived data, extracted the average ammonia nitrogen concentration data series within 48 hours after the rainstorm, constructed an ammonia nitrogen concentration change trend model in chronological order, and calculated the average slope to be 0.14, which was used as the first correction factor.

[0056] At the same time, the system extracted water temperature data within 12 hours before and after the rainstorm. It was found that the water temperature dropped by more than the set threshold of 2°C. After further analysis of the historical data, it was found that there were 3 historical data that met the same temperature drop conditions, accounting for 60% of all historical data. Based on this, the occurrence frequency of 0.6 was used as the second correction factor.

[0057] The system calls the preset fit value correction calculation formula, substitutes the initial fit value, the first correction factor, the second correction factor and their respective weights, and performs the following calculation: The corrected goodness of fit value = initial goodness of fit value × (1-first correction factor × weight 1-second correction factor × weight 2); assuming that weight 1 is 0.5 and weight 2 is 0.3, substituting the values ​​into the formula, we get: corrected goodness of fit value = 0.83 × (1-0.14 × 0.5-0.6 × 0.3) = 0.83 × (1-0.07-0.18) = 0.83 × 0.75 = 0.6225.

[0058] The final system output corrected fit value is 0.6225, and based on the classification interval to which this value belongs, the subsequent ammonia nitrogen concentration monitoring data is assigned a classification label of "model deviation", and the subsequent data review or model adjustment operation is triggered. This correction process ensures that the digital twin model's response to monitoring data in actual abnormal scenarios is more reasonable and explainable.

[0059] Furthermore, Figure 6 The application architecture diagram of the system provided by the embodiment of the present invention is shown.

[0060] Among them, in another preferred embodiment provided by the present invention, an environmental monitoring data classification management system based on digital twins includes: The data acquisition module 100 is used to obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after a rainstorm, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir.

[0061] In an embodiment of the present invention, the inlet of the target reservoir is selected as the location for collecting ammonia nitrogen concentration monitoring data, based on the fact that this location can most directly reflect the input of surface runoff caused by rainstorm events to the water quality of the reservoir. After a rainstorm, the surface runoff that flows into the reservoir usually carries nitrogen pollutants, especially ammonia nitrogen, which has sensitive and fast-responding characteristics and often shows a trend of short-term sudden increases. By monitoring the changes in ammonia nitrogen concentration at the inlet, the initial characteristics of pollution input can be captured in time before the water quality changes affect the entire water body, thereby improving the accuracy of pollution response and the timeliness of early warning.

[0062] The purpose of collecting ammonia nitrogen concentration monitoring data within a specified time period after a rainstorm is to align the time period with historical data and ensure consistency between trend analysis and model comparison. This time period is generally preset based on factors such as reservoir catchment area, flow rate, and water body renewal cycle. It can effectively cover the high-frequency change range of pollution impacts and is a key period for establishing concentration change trend models and identifying model deviations.

[0063] The acquisition of ammonia nitrogen concentration monitoring data can be achieved through the online water quality monitoring device deployed at the inlet of the target reservoir, which can use mature monitoring technologies such as spectroscopy, electrode method or automated colorimetry for real-time collection. The monitoring equipment is usually integrated with the reservoir digital twin system, uploaded to the model platform in real time through the hydrological information system, and automatically participates in data modeling and parameter evaluation.

[0064] The initial fit value refers to the degree of consistency index obtained by comparing the digital twin system with the actual collected monitoring data after simulating and predicting the ammonia nitrogen concentration based on the current model structure, input parameters and historical training logic. The fit value can be calculated through a variety of mathematical indicators, such as root mean square error (RMSE), mean absolute error (MAE), relative deviation rate, goodness of fit, etc. The purpose of generating the fit value is to measure the accuracy of the current digital twin model's response to actual water quality changes. In the present invention, the initial fit value is not only used as an evaluation indicator of water quality simulation accuracy, but also as a key reference parameter in classification label determination.

[0065] The historical archived data of the target reservoir can be retrieved from the environmental monitoring system, water information platform, historical data center or remote sensing database, usually including monitoring data of multiple time periods and their corresponding meteorological, hydrological and pollutant characteristic information. The historical archived data described in the present invention refers to the original monitoring records and related environmental parameter information collected and stored in the target reservoir in multiple historical time periods, usually including but not limited to: ammonia nitrogen concentration monitoring data at various stages before and after the rainstorm, used to restore the time series changes of ammonia nitrogen concentration under the condition of rainstorm of specific intensity; meteorological data records related to the intensity of the rainstorm, including cumulative rainfall, hourly rainfall intensity and rainfall distribution process; water temperature monitoring data in the corresponding time period, used to describe the change of water temperature after the rainstorm.

[0066] Furthermore, the digital twin-based environmental monitoring data classification management system also includes: The systematic deviation judgment module 200 is used to perform several groups of parameter debugging behaviors on the digital twin model, record the changes in the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this.

[0067] Specifically, Figure 7 The structure block diagram of the systematic deviation judgment module 200 in the system provided by the embodiment of the present invention is shown.

[0068] Among them, in the preferred embodiment provided by the present invention, the systematic deviation judgment module 200 specifically includes: The model debugging unit 201 is used to restart the processing program for generating the goodness-of-fit value for the ammonia nitrogen concentration monitoring data in the digital twin model several times, and perform a debugging operation of a preset amplitude on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; The variation range calculation unit 202 is used to record the test fit value of each group after debugging, and calculate the variation range of each test fit value compared with the initial fit value, and determine whether each variation range exceeds a preset range threshold; The proportion comparison unit 203 is used to determine that there is a systematic deviation in the initial fit value if the proportion of the debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset proportion threshold.

[0069] In an embodiment of the present invention, several groups of parameter debugging behaviors are performed on the digital twin model, specifically by restarting the processing program in the model for generating the goodness-of-fit value for the ammonia nitrogen concentration monitoring data multiple times. The processing program is usually integrated into the digital twin system in a modular manner, and can be automatically called or batch triggered in the system, and each call runs independently. The number of restarts is comprehensively set according to the number of model control parameters, the accuracy requirements of the system response analysis, and the acceptable degree of model running time, and is preferably set to more than five times to ensure that the subsequent analysis is representative and statistically stable. The basis for setting the number of times usually includes the empirical value of parameter sensitivity analysis, the processing capacity of the simulation platform, and the fluctuation characteristics of the model output caused by the debugging behavior.

[0070] Furthermore, the digital twin-based environmental monitoring data classification management system also includes: The first correction factor determination module 300 is used to obtain a number of local historical archived data consistent with the current rainstorm intensity if it is determined that there is a systematic deviation, and to construct an ammonia nitrogen concentration change trend model based on the data, and to set the first correction factor based on the trend model.

[0071] Specifically, Figure 8 It shows a structural block diagram of the first correction factor determination module 300 in the system provided by an embodiment of the present invention.

[0072] In a preferred embodiment of the present invention, the first correction factor determination module 300 specifically includes: The data screening unit 301 is used to analyze the historical archived data of the target reservoir and screen out a number of local historical archived data that match the current rainstorm intensity after determining that the initial fit value has a systematic deviation; The trend model building unit 302 is used to process each of the selected local historical archived data, extract the average historical ammonia nitrogen concentration values ​​within a specified time period after the rainstorm, and arrange them in chronological order to build an ammonia nitrogen concentration change trend model; The average slope calculation unit 303 is used to calculate the average slope of the constructed ammonia nitrogen concentration change trend model and use the average slope as the first correction factor.

[0073] In the embodiment of the present invention, when the initial fit value is judged to have a systematic deviation, in order to further identify the performance characteristics of the deviation in a specific rainstorm scene and thus achieve targeted correction, it is necessary to perform model attribution and trend analysis based on historical scene data similar to the current monitoring scene. To this end, a number of local historical archived data matching the current rainstorm intensity are selected from the historical archived data of the target reservoir as the sample basis for the subsequent construction of the ammonia nitrogen concentration change trend model.

[0074] The screening of local historical archived data that matches the current rainstorm intensity is mainly to ensure that the construction basis of the trend model has situational consistency and environmental comparability. Rainstorm intensity is one of the important external driving factors that affect the change of ammonia nitrogen concentration. Rainstorms of different intensities will trigger completely different pollution impact processes and water quality response patterns. If historical data with a large difference from the current rainstorm intensity are directly used for trend modeling, non-representative patterns may be introduced, resulting in a deviation in the correction results, and the actual evolution characteristics of ammonia nitrogen concentration in the current scene cannot be effectively restored. Therefore, by matching the rainstorm intensity, it can be ensured that the selected sample data has sufficient scene similarity, providing relevant sample support for subsequent model calculations.

[0075] After completing the sample data screening, each local historical archived data is processed to extract the average historical ammonia nitrogen concentration values ​​within the specified time period after the rainstorm, and arranged in chronological order. This processing process can ensure the consistency of different historical data in the time dimension by setting a unified statistical window after the rainstorm, such as 24 hours or 48 hours. Subsequently, each historical data point is sequentially established as a sequence with the horizontal axis as the historical time and the vertical axis as the average ammonia nitrogen concentration value, forming a data structure of the concentration change trend. This trend data structure serves as the basis for modeling and can be used to extract change characteristics, evaluate the steepness of the trend, and identify systematic deviation patterns.

[0076] After the ammonia nitrogen concentration change trend model is built, the average slope of the model is calculated to obtain the overall change trend of ammonia nitrogen concentration in this type of rainstorm scenario. The average slope reflects the overall rate of change of ammonia nitrogen concentration over time in historical scenarios. It is a key parameter that can characterize the pollution load intensity, pollutant input rate and system response characteristics. The purpose of using the average slope as the first correction factor is that in the current scenario, if the digital twin model does not fully reflect similar concentration change intensity or trend characteristics, the correction factor can be used to compensate the trend of the initial fit value, so that it is closer to the simulation behavior that the model should have in real situations. By introducing the trend slope as a dynamic adjustment amount, the system's response adaptability to sudden environmental events such as "rainstorm-pollution" can be enhanced, and the reliability and actual representativeness of the model output in the classification management process can be improved.

[0077] Furthermore, the digital twin-based environmental monitoring data classification management system also includes: The second correction factor determination module 400 is used to detect whether the drop in water temperature after the current rainstorm meets the condition of exceeding a preset threshold. If so, the frequency of occurrence of the corresponding situation in the above-mentioned local historical archived data is counted, and the second correction factor is set according to the frequency.

[0078] Specifically, Fig. 9It shows a structural block diagram of the second correction factor determination module 400 in the system provided by an embodiment of the present invention.

[0079] In a preferred embodiment of the present invention, the second correction factor determination module 400 specifically includes: The water temperature analysis unit 401 is used to extract the water temperature data of the target reservoir inlet within a preset time period before and after the current rainstorm based on the historical archived data, and determine whether the drop in water temperature after the rainstorm exceeds a preset threshold; The drop range determination unit 402 is used to analyze each local historical archive data to determine whether there is a situation in which the drop range of water body temperature exceeds the preset threshold value in the preset time period before and after the rainstorm if it is determined that the drop range exceeds the preset threshold value; The occurrence frequency calculation unit 403 is used to count the occurrence frequency of the local historical archived data that meets the above conditions, and use the occurrence frequency as a basis to set the second correction factor.

[0080] In an embodiment of the present invention, judging whether the drop in water temperature after heavy rain exceeds a preset threshold is to identify whether there is a typical rapid water temperature change scenario, which often affects the diffusion behavior, reaction rate and model fitting ability of pollutants. In actual hydrological processes, a large amount of surface runoff brought by heavy rain is usually accompanied by low-temperature inflow. If the inflow temperature is significantly lower than the water temperature in the reservoir area, it may cause a sharp drop in local temperature. Water temperature is a key influencing factor in controlling the ammonia nitrogen conversion process (such as ammonia volatilization, nitrification reaction rate, etc.). Rapid changes in temperature will cause distortion of certain response parameters in the digital twin model, thereby affecting the model's simulation accuracy of changes in ammonia nitrogen concentration.

[0081] The drop in water temperature can be calculated by comparing the temperature values ​​before and after the rainstorm. The preferred method is to extract the average water temperature in the set time period before the rainstorm and the average water temperature in the same time period after the rainstorm. The difference between the two is the drop. If the drop exceeds the temperature threshold set in the system (for example, 2°C or 3°C), it is considered that there is a significant temperature disturbance, which may be one of the potential causes for the current model to deviate from reality.

[0082] After determining that the current temperature change exceeds the preset threshold, the selected local historical archived data are analyzed one by one to determine whether there is a similar situation where the water temperature drops by more than the preset threshold before and after the corresponding historical rainstorm event. This process is to find historical samples that are consistent with the current temperature change characteristics and confirm whether such scenes are representative or common in historical data. If a certain type of feature appears repeatedly in historical samples and fails to get a reasonable response in the model, it means that the feature is reasonable to be included in the correction system.

[0083] Furthermore, the digital twin-based environmental monitoring data classification management system also includes: The goodness-of-fit value correction module 500 is used to apply the first correction factor and the second correction factor jointly to the initial goodness-of-fit value to generate a corrected goodness-of-fit value for use in subsequent classification label determination of ammonia nitrogen concentration monitoring data.

[0084] In an embodiment of the present invention, the first correction factor and the second correction factor are selected to be jointly applied to the initial fit value, with the aim of realizing multi-dimensional dynamic correction of the output results of the digital twin model under a specific rainstorm scenario. The first correction factor is based on the average slope extracted from the ammonia nitrogen concentration trend model, which reflects the temporal evolution speed and intensity of the target pollutant in historically similar scenarios, and is the internal dynamic characteristic of the pollution response; the second correction factor is derived from the frequency of occurrence of water temperature changes, representing the external influencing factors of environmental conditions on the diffusion and reaction process of pollutants. These two types of correction factors start from the two dimensions of the result trend of pollution behavior and the control variables of the pollution process, complement each other, and can form a more complete correction logic system.

[0085] Combining these two types of correction factors can effectively improve the accuracy of identifying and adjusting systematic deviations. When the model has simulation structure or parameter adaptation problems in the current scenario, the single-dimensional correction method may cause an imbalance in the correction strength due to the one-sided reflection of a certain type of deviation factor; and the dual-factor joint adjustment mechanism can cover more types of deviation sources, and achieve systematic comprehensive compensation from model trend misfit to external disturbance insensitivity. This joint correction method not only improves the accuracy of classification label determination, but also has the ability to dynamically adapt to actual scene changes without increasing the complexity of the model structure. It is a lightweight technical path that can greatly optimize the application effect without changing the main body of the model.

[0086] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

Claims

1. A method for classifying and managing environmental monitoring data based on digital twins, characterized in that: The method comprises: Obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after the rainstorm, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir; Perform several groups of parameter debugging actions on the digital twin model, record the changes in the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this; If it is determined that there is a systematic deviation, obtain some local historical archived data consistent with the current rainstorm intensity, and build a trend model of ammonia nitrogen concentration based on it, and set the first correction factor based on the trend model; Detect whether the drop in water temperature after the current rainstorm meets the condition of exceeding the preset threshold. If so, count the occurrence frequency of the corresponding situation in the above local historical archived data, and set the second correction factor according to the frequency; The first correction factor and the second correction factor are jointly applied to the initial goodness of fit value to generate a corrected goodness of fit value, which is used for classification label determination of subsequent ammonia nitrogen concentration monitoring data.

2. The method for classifying and managing environmental monitoring data based on digital twins according to claim 1 is characterized in that: The steps of performing several groups of parameter debugging actions on the digital twin model, recording the change of the goodness of fit value after each group of debugging, and analyzing whether there is a systematic deviation in the initial goodness of fit value based on this include: Restarting the processing program for generating the goodness-of-fit value for the ammonia nitrogen concentration monitoring data in the digital twin model several times, and performing a preset amplitude debugging operation on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; Record the test fit value after each group of debugging, and calculate the change range of each test fit value compared with the initial fit value, and determine whether the change range exceeds the preset range threshold; If the proportion of the number of debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset proportion threshold, it is determined that there is a systematic deviation in the initial fit value.

3. The method for classifying and managing environmental monitoring data based on digital twins according to claim 1 is characterized in that: If it is determined that there is a systematic deviation, obtain some local historical archived data consistent with the current rainstorm intensity, and build an ammonia nitrogen concentration change trend model based on it. The steps of setting the first correction factor based on the trend model include: After determining that the initial fit value has systematic deviations, the historical archived data of the target reservoir are analyzed, and several local historical archived data that match the current rainstorm intensity are selected; Each selected local historical archive data is processed to extract the average historical ammonia nitrogen concentration values ​​within a specified period of time after the rainstorm, and arranged in chronological order to construct an ammonia nitrogen concentration change trend model; The average slope of the constructed ammonia nitrogen concentration change trend model was calculated and used as the first correction factor.

4. The method for classifying and managing environmental monitoring data based on digital twins according to claim 3 is characterized in that: The steps of detecting whether the drop in water temperature after the current rainstorm meets the condition of exceeding a preset threshold value, and if so, counting the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data, and setting the second correction factor according to the frequency include: Based on historical archived data, the water temperature data of the target reservoir inlet in the preset time period before and after the current rainstorm is extracted, and it is determined whether the drop in water temperature after the rainstorm exceeds the preset threshold; If it is determined that the drop exceeds the preset threshold, each local historical archive data is analyzed to determine whether there is a situation in which the water temperature drops by more than the preset threshold in the preset time period before and after the rainstorm; The occurrence frequency of local historical archived data that meets the above conditions is counted, and the occurrence frequency is used as a basis to set the second correction factor.

5. The method for classifying and managing environmental monitoring data based on digital twins according to claim 4 is characterized in that: The steps of applying the first correction factor and the second correction factor together to the initial goodness of fit value to generate a corrected goodness of fit value for use in subsequent classification label determination of ammonia nitrogen concentration monitoring data include: Calling a preset goodness of fit value correction calculation formula, substituting the first correction factor and the second correction factor into the goodness of fit value correction calculation formula to correct the initial goodness of fit value, and obtaining a corrected goodness of fit value; The classification label determination operation of the subsequent ammonia nitrogen concentration monitoring data is performed based on the corrected fit value.

6. The method for classifying and managing environmental monitoring data based on digital twins according to claim 5 is characterized in that: The calculation formula for correcting the numerical value of the fit is: ,in Refers to the corrected goodness of fit value. Refers to the initial goodness of fit value, Refers to the first correction factor, that is, the average slope of the ammonia nitrogen concentration trend model. Refers to the adjustment weight corresponding to the first correction factor, Refers to the second correction factor, that is, the frequency of occurrence of local historical archived data that meets the conditions. Refers to the adjustment weight corresponding to the second correction factor; In the calculation formula for the numerical correction of the degree of fit ,in Refers to the total amount of local historical archived data. Refers to the trend model of ammonia nitrogen concentration change. The changing slope of the segment line; ,in Refers to the number of local historical archive data that meets the conditions.

7. An environmental monitoring data classification management system based on digital twins, characterized in that: The system comprises: a data acquisition module, a systematic deviation judgment module, a first correction factor determination module, a second correction factor determination module and a goodness of fit value correction module, wherein: The data acquisition module is used to obtain the ammonia nitrogen concentration monitoring data at the inlet of the target reservoir within a specified time period after the rainstorm, determine the initial fit value corresponding to the monitoring data in the digital twin model, and obtain the historical archived data of the target reservoir; The systematic deviation judgment module is used to perform several groups of parameter debugging behaviors on the digital twin model, record the changes in the fit value after each group of debugging, and analyze whether there is a systematic deviation in the initial fit value based on this; A first correction factor determination module is used to obtain a number of local historical archived data consistent with the current rainstorm intensity if it is determined that there is a systematic deviation, and to construct an ammonia nitrogen concentration change trend model based on the data, and to set a first correction factor based on the trend model; The second correction factor determination module is used to detect whether the drop in water temperature after the current rainstorm meets the condition of exceeding a preset threshold. If so, the occurrence frequency of the corresponding situation in the above-mentioned local historical archived data is counted, and the second correction factor is set according to the frequency; The goodness of fit value correction module is used to jointly apply the first correction factor and the second correction factor to the initial goodness of fit value to generate a corrected goodness of fit value for use in the classification label determination of subsequent ammonia nitrogen concentration monitoring data.

8. The digital twin-based environmental monitoring data classification management system according to claim 7 is characterized in that: The systematic deviation judgment module specifically includes: A model debugging unit, used to restart the processing program for generating a goodness of fit value for ammonia nitrogen concentration monitoring data in the digital twin model several times, and to perform a debugging operation of a preset amplitude on at least one model control parameter affecting the ammonia nitrogen concentration simulation result during each startup process; A change range calculation unit is used to record the test fit value of each group after debugging, and calculate the change range of each test fit value compared with the initial fit value, and determine whether each change range exceeds a preset range threshold; The proportion comparison unit is used to determine that there is a systematic deviation in the initial fit value if the proportion of the debugging times with a change amplitude exceeding a preset amplitude threshold exceeds a preset proportion threshold.

9. The digital twin-based environmental monitoring data classification management system according to claim 8, characterized in that: The first correction factor determination module specifically includes: A data screening unit is used to analyze the historical archived data of the target reservoir and screen out a number of local historical archived data that match the current rainstorm intensity after determining that the initial fit value has a systematic deviation; The trend model building unit is used to process each selected local historical archive data, extract the average historical ammonia nitrogen concentration value within a specified time period after the rainstorm, and arrange them in chronological order to build an ammonia nitrogen concentration change trend model; The average slope calculation unit is used to calculate the average slope of the constructed ammonia nitrogen concentration change trend model and use the average slope as the first correction factor.

10. The digital twin-based environmental monitoring data classification management system according to claim 9, characterized in that: The second correction factor determination module specifically includes: The water temperature analysis unit is used to extract the water temperature data of the target reservoir inlet within a preset time period before and after the current rainstorm based on the historical archived data, and determine whether the drop in water temperature after the rainstorm exceeds a preset threshold; A drop range judgment unit is used to analyze each local historical archive data to determine whether there is a situation in which the water body temperature drops by more than the preset threshold value in the preset time period before and after the rainstorm if it is determined that the drop range exceeds the preset threshold value; The occurrence frequency calculation unit is used to count the occurrence frequency of the local historical archived data that meets the above conditions, and use the occurrence frequency as a basis to set the second correction factor.

Images