Ecological wetland denitrification data analysis method and operation and maintenance system based on environmental data

Abstract

The application discloses an ecological wetland denitrification data analysis method and operation and maintenance system based on environmental data, relates to the technical field of environmental data analysis, and comprises the following steps: continuously collecting environmental monitoring data of multiple different regions in an ecological wetland; evaluating the consistency of changes between dissolved oxygen sensor data of different regions to obtain an evaluation result; performing correlation verification on the dissolved oxygen sensor data and at least one of water temperature data, pH value data and nitrogen concentration data to obtain a verification result; analyzing the deviation degree of behavior characteristics presented by the dissolved oxygen sensor data from historical healthy behavior patterns to obtain an analysis result; and identifying abnormal behavior patterns of the environmental monitoring data according to the evaluation result, the verification result and the analysis result, and generating a diagnostic intervention instruction. The application can effectively identify and eliminate distorted dissolved oxygen data caused by biological membranes, improve denitrification efficiency, and guarantee stable and efficient operation of the system.

Description

Technical Field

[0001] This application relates to the field of environmental data analysis technology, and in particular to an ecological wetland denitrification data analysis method and operation and maintenance system based on environmental data. Background Technology

[0002] In the daily operation of denitrification processes in ecological wetlands, dissolved oxygen sensors are crucial devices for monitoring the activity of microorganisms in the water and supporting the assessment of denitrification efficiency. However, during long-term operation, suspended particles and microorganisms in the wetland environment adhere to the sensor's sensitive surface and form a biofilm. The growth of this biofilm is non-uniform, leading to slow, continuous, and asynchronous deviations in the readings of different dissolved oxygen sensors. Because this deviation is gradual and insidious, it is difficult to detect in a timely manner through routine short-term calibration.

[0003] Existing methods for predicting nitrogen removal efficiency rely on the correlation between historical physicochemical data and biological activity status. When dissolved oxygen data with the aforementioned asynchronous biases is input, the inherent logic of "extracting biological efficiency from physical parameters" is disrupted, leading to erroneous interpretations of process status based on inaccurate sensor readings. For example, it might overestimate nitrification efficiency when actual dissolved oxygen is insufficient but readings are high, or mislead operators into over-aeration when readings are low.

[0004] The failure of this forecasting method deprived operation management of reliable real-time guidance. To grasp the true state of the process, it became necessary to revert to and significantly increase the costly and time-consuming manual water sampling and testing, which had a lag of several hours to several days. The lag in decision-making led to missed opportunities for optimal intervention, and process optimization could only rely on rough experience, which not only increased operating costs and risks but also hindered further improvements in denitrification efficiency. Summary of the Invention

[0005] This application proposes an ecological wetland denitrification data analysis method and operation and maintenance system based on environmental data. It aims to solve the technical problem that in the ecological wetland denitrification process, the dissolved oxygen sensor data is asynchronized due to the adhesion of biofilm, which causes the existing denitrification efficiency prediction method to fail, and thus leads to operation and management difficulties, increased costs and decreased denitrification efficiency.

[0006] Firstly, this application provides a method for analyzing nitrogen removal data from ecological wetlands based on environmental data, comprising the following steps:

[0007] Continuous collection of environmental monitoring data from multiple different areas within the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency;

[0008] The consistency of changes in dissolved oxygen sensor data across different regions was evaluated to obtain the evaluation results.

[0009] The dissolved oxygen sensor data is correlated with at least one of the water temperature data, pH value data, and nitrogen concentration data to obtain the verification results;

[0010] The analysis results are obtained by analyzing the degree of deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical healthy behavior patterns.

[0011] Based on the assessment results, the verification results, and the analysis results, abnormal behavior patterns in the environmental monitoring data are identified, and diagnostic intervention instructions for the corresponding regions and / or projects are generated.

[0012] As some embodiments of this application, the multiple different areas are divided according to the process flow and functional zones of the ecological wetland, and include an inlet area, at least one aerobic area and an outlet area arranged sequentially along the water flow direction of the ecological wetland; the environmental monitoring data are collected by multiple sets of sensors correspondingly arranged in the multiple different areas.

[0013] As some embodiments of this application, the step of evaluating the consistency of changes in dissolved oxygen sensor data between different regions to obtain evaluation results includes:

[0014] Calculate the difference or ratio between dissolved oxygen sensor data in adjacent different regions;

[0015] Determine whether the difference or ratio exceeds the normal fluctuation range determined based on historical stable operating data to obtain an assessment result of whether the dissolved oxygen sensor data is abnormal or normal;

[0016] The historical stable operation data is obtained by selecting data segments that meet preset conditions within a preset continuous operating period of the ecological wetland in the past. The preset conditions include: a first condition that the denitrification efficiency is continuously higher than a preset compliance threshold, and a second condition that the fluctuation value of the environmental monitoring data is less than a preset fluctuation value or is reliably determined by on-site calibration.

[0017] As some embodiments of this application, the step of correlating and verifying the dissolved oxygen sensor data with at least one of the water temperature data, pH value data, and nitrogen concentration data to obtain the verification result includes:

[0018] Determine whether the dissolved oxygen sensor data meets the preset correlation verification conditions with the water temperature data, the pH value data, and the nitrogen concentration data, and obtain the verification result of whether the dissolved oxygen sensor data is normal or abnormal;

[0019] The correlation verification conditions include: when the water temperature data and pH value data are respectively within the preset suitable range for nitrification reaction, and the ammonia nitrogen removal efficiency reflected by the nitrogen concentration data is higher than the preset efficiency threshold, the dissolved oxygen sensor data is greater than or equal to the preset lower concentration limit.

[0020] As some embodiments of this application, the step of analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns to obtain analysis results includes:

[0021] Based on the historical stable operating data, the statistical range of the trend characteristics and fluctuation characteristics of the dissolved oxygen sensor data is calculated to determine historical health behavior patterns;

[0022] Calculate the behavioral characteristics presented by the dissolved oxygen sensor data and the degree of deviation from the historical health behavior pattern;

[0023] The system determines whether the deviation is greater than a preset deviation threshold to obtain an analysis result indicating whether the dissolved oxygen sensor data has deviated.

[0024] As some embodiments of this application, after the step of analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns to obtain the analysis results, the following steps are also included:

[0025] Extract the data sequence from the dissolved oxygen sensor data that was determined to have deviated;

[0026] A low-pass filter is used to separate the slow-changing trend component that characterizes the potential bias of dissolved oxygen sensor data, and a high-pass filter is used to separate the high-frequency fluctuation component that characterizes environmental or physical disturbances.

[0027] The average rate of change of the slow-changing trend component within a preset trend period and the standard deviation of the high-frequency fluctuation component within a preset fluctuation period are extracted as quantitative features to help identify abnormal behavior patterns in the environmental monitoring data.

[0028] As some embodiments of this application, the step of identifying abnormal behavior patterns in the environmental monitoring data based on the evaluation results, the verification results, and the analysis results, and generating diagnostic intervention instructions for the corresponding region and / or the corresponding project includes:

[0029] Based on preset evidence weighting rules, current probabilities are assigned to multiple preset potential anomaly causes according to the evaluation results, the verification results, the analysis results, and the quantitative characteristics; wherein, the potential anomaly causes include dissolved oxygen sensor inaccuracy, process anomalies in the local microenvironment where the dissolved oxygen sensor is located, and local physical disturbances; the evidence weighting rules define the degree of support of each input information for different potential anomaly causes;

[0030] Calculate the information entropy of the current diagnostic status based on the current probability of each of the potential causes of abnormality.

[0031] Based on a pre-defined operation-result correlation model, the expected probability of each potential abnormal cause is predicted after each of the multiple pre-defined diagnostic operations. The diagnostic operations include: on-site calibration of the dissolved oxygen sensor in the target area, collection of water samples from the target area for microbial community analysis, detection of the concentration of specific organic matter in the water of the target area, and measurement of the water flow velocity in the target area. The operation-result correlation model defines the result and probability obtained after each diagnostic operation, as well as the probability update relationship between each result and each potential abnormal cause.

[0032] Calculate the predicted information entropy based on the expected probability of each of the potential causes of anomalies;

[0033] Based on the information entropy of the current diagnostic status and the predicted information entropy, the expected reduction value of information entropy is obtained;

[0034] Select the diagnostic action that maximizes the expected reduction value as the generated diagnostic intervention instruction for the corresponding region and / or item.

[0035] As some embodiments of this application, the information entropy is calculated based on the following formula:

[0036] ;

[0037] Where n represents the total number of the plurality of preset potential abnormal causes, H represents the probability assigned to the i-th potential cause of the anomaly, and H represents the information entropy.

[0038] As some embodiments of this application, the method further includes the following steps:

[0039] Obtain the feedback results after executing the diagnostic intervention instructions;

[0040] Query the probability update relationship between the operation-result association model and the feedback result;

[0041] Based on the retrieved probability update relationship, update the current probability assigned to each of the potential abnormal causes;

[0042] Based on the updated probabilities, the information entropy of the current diagnostic status is recalculated;

[0043] If the recalculated information entropy is higher than the preset entropy threshold, then based on the updated current probability, the expected reduction value after performing each diagnostic operation is re-predicted, and the diagnostic operation that maximizes the expected reduction value is selected as the subsequently generated diagnostic intervention instruction.

[0044] Otherwise, generate the corresponding processing instructions based on the most probable potential cause of the anomaly.

[0045] Secondly, this application also provides an ecological wetland denitrification data analysis and operation and maintenance system based on environmental data, including:

[0046] The data acquisition module is used to continuously collect environmental monitoring data from multiple different areas in the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency;

[0047] The consistency assessment module is used to evaluate the consistency of changes in dissolved oxygen sensor data in different regions and obtain the assessment results.

[0048] The correlation verification module is used to perform correlation verification between the dissolved oxygen sensor data and at least one of the water temperature data, the pH value data, and the nitrogen concentration data to obtain the verification result;

[0049] The deviation analysis module is used to analyze the degree of deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns, and to obtain the analysis results;

[0050] The instruction generation module is used to identify abnormal behavior patterns in the environmental monitoring data based on the evaluation results, the verification results, and the analysis results, and to generate diagnostic intervention instructions for the corresponding region and / or the corresponding project.

[0051] The technical solution according to the embodiments of this application has at least the following beneficial effects:

[0052] In summary, this application, through multi-dimensional cross-validation, effectively identifies and eliminates distorted dissolved oxygen data caused by biofilms, reconstructing the logical foundation for "physical parameter-based biological efficiency estimation" and restoring accurate state assessment. It transforms expensive and delayed manual sampling into precise diagnosis triggered on demand by an intelligent system, reducing unnecessary sampling and testing frequency while achieving near real-time state judgment and decision-making, overcoming the drawbacks of relying on delayed test data. Based on accurate data and real-time diagnosis, it enables refined prediction and optimized control of wetland processes, thereby improving denitrification efficiency and ensuring stable and efficient system operation.

[0053] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0054] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0055] Figure 1 This is a flowchart illustrating an ecological wetland denitrification data analysis method based on environmental data, provided in an embodiment of this application.

[0056] Figure 2 This is a schematic diagram of the architecture of an ecological wetland denitrification data analysis and maintenance system based on environmental data, provided in an embodiment of this application. Detailed Implementation

[0057] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0058] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0059] In the denitrification process of ecological wetlands, the dissolved oxygen sensor readings exhibit slow and continuous deviations due to the non-uniform and asynchronous growth of the surface biofilm. This disrupts the inherent logic of "physical parameters estimating biological efficiency" relied upon by existing denitrification efficiency prediction methods, leading to inaccurate state assessments and process guidance. This forces operation and management to revert to a high-cost, highly delayed manual sampling and testing mode, which not only increases operating costs but also causes the best intervention opportunity to be missed due to decision-making delays. Ultimately, this restricts further improvement in denitrification efficiency and increases the risk of system operation.

[0060] Therefore, such as Figure 1 As shown, this application discloses a method for analyzing nitrogen removal data in ecological wetlands based on environmental data, including the following steps:

[0061] S110, continuously collects environmental monitoring data from multiple different areas in the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency;

[0062] S120, evaluate the consistency of changes in dissolved oxygen sensor data in different regions, and obtain the evaluation results;

[0063] S130, perform correlation verification between the dissolved oxygen sensor data and at least one of the water temperature data, pH value data and nitrogen concentration data to obtain the verification result;

[0064] S140, Analyze the degree of deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns to obtain the analysis results;

[0065] S150, based on the assessment results, the verification results, and the analysis results, identify abnormal behavior patterns in the environmental monitoring data and generate diagnostic intervention instructions for the corresponding region and / or the corresponding project.

[0066] To better understand the technical solution proposed in this application, some key terms involved will be explained first.

[0067] "Ecological wetlands" refer to artificial or natural wetland systems whose main function is to remove pollutants, especially nitrogen compounds, from water bodies through the synergistic action of plants, soil, and microorganisms. The denitrification process typically includes two main stages: nitrification and denitrification, each completed by different microbial communities.

[0068] Environmental monitoring data is the fundamental input of this method, encompassing several key parameters. Dissolved oxygen sensor data reflects the concentration of dissolved oxygen in the water, a crucial factor influencing nitrification and denitrification reaction rates. Water temperature data affects microbial activity and reaction kinetics. pH data is another important environmental factor affecting microbial enzyme activity. Nitrogen concentration data directly reflects denitrification efficiency and typically includes indicators such as ammonia nitrogen, nitrate nitrogen, and total nitrogen. These data provide comprehensive information for subsequent data analysis.

[0069] "Abnormal behavior patterns" refer to specific data characteristics in environmental monitoring data that are inconsistent with normal operating conditions and may indicate system failure or efficiency degradation. Identifying these abnormal patterns is one of the core objectives of this method.

[0070] "Diagnostic intervention instructions" are action recommendations output by this method, designed to guide operators in conducting targeted inspections, calibrations, or adjustments to the wetland system to resolve identified anomalies. These instructions may include specific sensor calibration recommendations, water sample collection and analysis recommendations, or other process parameter adjustment recommendations.

[0071] The core of the ecological wetland denitrification data analysis method based on environmental data proposed in this application lies in conducting multi-dimensional and in-depth analysis of environmental monitoring data collected during the operation of ecological wetlands, so as to accurately identify potential abnormal behavior patterns and provide targeted intervention instructions.

[0072] First, continuous environmental monitoring data is collected from multiple different areas within the ecological wetland. This data is fundamental to understanding the wetland's operational status. For example, continuous data collection can be achieved by deploying multiple dissolved oxygen sensors, water temperature sensors, pH sensors, and nitrogen concentration sensors within the wetland. These sensors can be connected to a central data acquisition system via wired or wireless means and transmit data to a data storage server at preset time intervals (e.g., every minute, every hour).

[0073] Secondly, the consistency of dissolved oxygen sensor data across different regions is evaluated to obtain assessment results. This evaluation aims to identify any unreasonable discrepancies or trend deviations between different sensor readings. If a significant inconsistency is found between the data variation trend of a particular dissolved oxygen sensor and surrounding area or historical data, it may indicate a problem with the sensor or an anomaly in the local environment of its location.

[0074] Furthermore, the dissolved oxygen sensor data is correlated with at least one of the following: water temperature, pH, and nitrogen concentration data, to obtain validation results. This validation aims to check whether the dissolved oxygen data maintains a reasonable biochemical correlation with other key environmental parameters. For example, in areas with vigorous nitrification, a certain negative correlation is often observed between dissolved oxygen concentration and ammonia nitrogen concentration; that is, high ammonia nitrogen removal efficiency is accompanied by high dissolved oxygen consumption. If the dissolved oxygen sensor data shows high dissolved oxygen, but the nitrogen concentration data shows low nitrogen removal efficiency, this may indicate a bias in the dissolved oxygen sensor data, or that the biological activity in the area is inhibited by other factors. This correlation validation can be achieved by establishing an empirical model or a statistical model trained based on historical data.

[0075] Furthermore, the analysis examines the degree of deviation between the behavioral characteristics of dissolved oxygen sensor data and historical health behavior patterns to obtain analytical results. Historical health behavior patterns refer to the typical trends, fluctuation ranges, and periodic characteristics exhibited by dissolved oxygen sensor data during normal and efficient operation of the wetland system. Significant deviations from this baseline pattern in dissolved oxygen sensor data, such as abnormal long-term drift, drastic fluctuations, or periodic loss, indicate potential sensor malfunctions or process anomalies. This degree of deviation analysis can be achieved by calculating the distance, similarity, or statistical significance between the current data and the historical baseline.

[0076] Finally, diagnostic intervention instructions are generated for the corresponding areas and / or projects. This step is the decision-making stage, integrating the aforementioned analyses. For example, if the assessment results show poor consistency of dissolved oxygen sensor data, validation results show abnormal correlations with other parameters, and analysis results show deviations from historical health patterns, the system can comprehensively determine that the dissolved oxygen sensor or its area is abnormal. Based on this determination, the system can automatically generate specific intervention instructions, such as "perform on-site calibration of the dissolved oxygen sensor in the influent area," "collect water samples from the aerobic area for microbial community analysis," or "inspect the aeration equipment in the effluent area." These instructions aim to guide operators to take precise actions to address the identified problems, thereby restoring the normal operation and denitrification efficiency of the wetland system.

[0077] This application addresses the asynchronous bias problem of dissolved oxygen sensors through a multi-dimensional, self-verifying intelligent diagnostic framework. Its core working principle is as follows: First, a complete view of the system status is established through multi-source data acquisition (dissolved oxygen, water temperature, pH, and nitrogen concentration). Second, a three-layer progressive analysis is performed: horizontally, dissolved oxygen data from different locations is compared to assess the consistency of changes, directly exposing asynchronous bias; vertically, dissolved oxygen data is correlated with other key parameters to verify logical consistency and utilize denitrification biochemical principles to test the data's rationality; in the time dimension, current dissolved oxygen data behavior is compared with historical health patterns to identify slow drift or abnormal fluctuations. Finally, the system integrates the results of these three layers of analysis, intelligently identifies abnormal behavior patterns, and generates precise diagnostic or treatment instructions targeting specific locations and items, forming a closed loop of "monitoring-analysis-decision-intervention".

[0078] In summary, this application, through multi-dimensional cross-validation, effectively identifies and eliminates distorted dissolved oxygen data caused by biofilms, reconstructing the logical foundation for "physical parameter-based biological efficiency estimation" and restoring accurate state assessment. It transforms expensive and delayed manual sampling into precise diagnosis triggered on demand by an intelligent system, reducing unnecessary sampling and testing frequency while achieving near real-time state judgment and decision-making, overcoming the drawbacks of relying on delayed test data. Based on accurate data and real-time diagnosis, it enables refined prediction and optimized control of wetland processes, thereby improving denitrification efficiency and ensuring stable and efficient system operation.

[0079] In a specific embodiment of this application, the multiple different areas are divided according to the process flow and functional zones of the ecological wetland, and preferably include an inlet area, at least one aerobic area and an outlet area arranged sequentially along the water flow direction of the ecological wetland; the environmental monitoring data are collected by multiple sets of sensors correspondingly arranged in the above-mentioned multiple different areas.

[0080] The process flow and functional zoning of ecological wetlands refers to dividing the wetland space into sub-regions with corresponding treatment tasks based on the different functions the wetland undertakes in the water treatment process. For example, the influent zone is mainly responsible for receiving raw water and performing preliminary treatment; the aerobic zone maintains a high dissolved oxygen level through aeration or other methods to promote nitrification; and the effluent zone is used to collect treated water and monitor it before final discharge. This zoning method ensures that data collection in each zone has a clear process background and functional orientation, thereby improving the relevance of data analysis. The influent zone, at least one aerobic zone, and effluent zone, arranged sequentially along the water flow direction of the ecological wetland, can be understood as simulating a typical wastewater treatment plant process, enabling data collection to cover the entire treatment path of water from entering to leaving the wetland. The influent zone is usually located upstream of the wetland, the aerobic zone can be one or more zones connected in series or parallel, and the effluent zone is located downstream. This sequential arrangement helps track changes in water quality and the evolution of denitrification efficiency at different treatment stages. In practical applications, environmental monitoring data is collected through multiple sensor groups strategically positioned across various zones. A sensor group is a collection of sensors that provide data on dissolved oxygen, water temperature, pH levels, and nitrogen concentration, among others. This strategic placement means that each functional zone is equipped with a dedicated sensor group. The location and number of these sensor groups are optimized based on the characteristics and monitoring needs of each zone to ensure comprehensive and accurate acquisition of environmental monitoring data. For example, in aerobic zones, the density of dissolved oxygen sensors may be higher to obtain more data on the aerobic environment.

[0081] The proposed solution divides the ecological wetland into functional zones such as an influent zone, an aerobic zone, and an effluent zone, and strategically deploys sensor arrays within these zones. This allows the collected environmental monitoring data to more accurately reflect the actual operational status of different treatment stages within the wetland. This refined zone division and data acquisition strategy enables subsequent consistency assessments, correlation verifications, and behavioral characteristic deviation analyses of dissolved oxygen sensor data to be based on more representative and targeted data, significantly improving the accuracy of abnormal behavior pattern identification and the effectiveness of diagnostic intervention commands. For example, when dissolved oxygen data in the aerobic zone is abnormal, it can quickly pinpoint a potential problem with nitrification in that area, rather than making a general judgment that the entire wetland system is abnormal.

[0082] In some embodiments of this application, the step of evaluating the consistency of changes in dissolved oxygen sensor data across different regions to obtain evaluation results preferably includes:

[0083] Calculate the difference or ratio between dissolved oxygen sensor data in adjacent different regions;

[0084] Determine whether the difference or ratio exceeds the normal fluctuation range determined based on historical stable operating data to obtain an assessment result of whether the dissolved oxygen sensor data is abnormal or normal;

[0085] The historical stable operation data is obtained by selecting data segments that meet preset conditions within a preset continuous operating period of the ecological wetland in the past. The preset conditions include: a first condition that the denitrification efficiency is continuously higher than a preset compliance threshold, and a second condition that the fluctuation value of the environmental monitoring data is less than a preset fluctuation value or is reliably determined by on-site calibration.

[0086] First, dissolved oxygen sensor data from adjacent areas needs to be acquired. These readings can be instantaneous values ​​or values ​​averaged over a certain time window. Then, the relationship between these adjacent readings is quantified by calculating the differences or ratios between them. For example, the reading of the next area can be subtracted from the reading of the previous area, or the ratio of the reading of the next area to the reading of the previous area can be calculated.

[0087] The normal fluctuation range refers to the range within which the difference or ratio of dissolved oxygen sensor data in adjacent areas should fall under the normal and healthy operating conditions of the ecological wetland. This range is not arbitrarily set but determined based on historical stable operating data. The historical stable operating data refers to data segments selected from the past operation of the ecological wetland that meet specific conditions. These conditions aim to ensure that the selected data segments accurately reflect the state of the wetland during healthy operation. Specifically, the preset conditions include two aspects: the first condition requires that within the selected data segment, the wetland's nitrogen removal efficiency consistently exceeds a preset threshold, ensuring that the wetland function is in good condition; the second condition requires that the fluctuation value of the environmental monitoring data is less than a preset fluctuation value, or that the data reliability is confirmed through on-site calibration, ensuring the stability and accuracy of the sensor data itself. By meeting these stringent conditions, a reliable historical healthy operating baseline can be constructed.

[0088] The difference or proportion of dissolved oxygen sensor data in adjacent areas is calculated and compared with the normal fluctuation range determined based on the aforementioned historical stable operating data. If the calculated difference or proportion exceeds this normal fluctuation range, the dissolved oxygen sensor data is considered abnormal, and a corresponding evaluation result is generated; otherwise, the reading is considered normal.

[0089] Through the above technical solution, this application provides a more objective, quantitative, and reliable method for assessing the consistency of dissolved oxygen sensor data. By introducing a normal fluctuation range based on historical stable operating data, this method significantly improves the accuracy and reliability of the assessment results, effectively avoiding false alarms or missed alarms that may occur in traditional assessment methods. Furthermore, rigorous screening of historical data ensures the scientific validity and representativeness of the assessment baseline, thus providing a solid foundation for subsequent identification of abnormal behavior patterns and the generation of diagnostic intervention instructions, thereby enhancing the intelligent management level and operational stability of the denitrification process in ecological wetlands.

[0090] The following is a specific example to illustrate this.

[0091] Suppose an ecological wetland is divided into an influent zone, aerobic zone 1, aerobic zone 2, and an effluent zone, with dissolved oxygen sensors deployed in each zone. To assess the consistency of dissolved oxygen sensor data between aerobic zone 1 and aerobic zone 2, the system continuously collects dissolved oxygen sensor data from both zones. For example, at a certain moment, the dissolved oxygen reading in aerobic zone 1 is 3.5 mg / L, and the dissolved oxygen reading in aerobic zone 2 is 2.8 mg / L. In this case, the difference between the two can be calculated as 0.7 mg / L.

[0092] Simultaneously, the system utilizes pre-defined historical stable operating data to establish a normal fluctuation range. This historical data is selected from data segments where wetland denitrification efficiency consistently exceeds, for example, a threshold of 80%, and where fluctuations in all environmental monitoring data (including dissolved oxygen, water temperature, pH, etc.) are less than preset fluctuation values ​​(e.g., dissolved oxygen fluctuation less than 0.2 mg / L / hour) or have been verified as reliable through on-site calibration. Statistical analysis of this historical data determines the normal fluctuation range for the difference in dissolved oxygen readings between adjacent aerobic zones under normal operating conditions, for example, 0.3 mg / L to 0.6 mg / L.

[0093] Subsequently, the real-time calculated difference of 0.7 mg / L was compared with the established normal fluctuation range (0.3 mg / L to 0.6 mg / L). Since 0.7 mg / L exceeded the upper limit of 0.6 mg / L, the system determined that there was an anomaly in the dissolved oxygen sensor data between aerobic zone 1 and aerobic zone 2, and generated a corresponding assessment result indicating possible sensor inaccuracy or localized process abnormalities. This assessment method, based on quantitative differences and historical health baselines, makes anomaly detection more accurate and objective.

[0094] In some embodiments of this application, the step of verifying the correlation between dissolved oxygen sensor data and at least one of water temperature data, pH value data, and nitrogen concentration data to obtain the verification result preferably includes:

[0095] Determine whether the dissolved oxygen sensor data meets the preset correlation verification conditions with the water temperature data, the pH value data, and the nitrogen concentration data, and obtain the verification result of whether the dissolved oxygen sensor data is normal or abnormal;

[0096] The correlation verification conditions include: when the water temperature data and pH value data are respectively within the preset suitable range for nitrification reaction, and the ammonia nitrogen removal efficiency reflected by the nitrogen concentration data is higher than the preset efficiency threshold, the dissolved oxygen sensor data is greater than or equal to the preset lower concentration limit.

[0097] The pre-defined correlation verification conditions aim to establish a logical relationship between dissolved oxygen sensor data and key environmental parameters of the denitrification process in ecological wetlands. The "pre-defined suitable nitrification range" refers to a specific range of water temperature and pH values ​​determined based on ecological wetland design and operational experience, where these values ​​are favorable for nitrifying bacteria activity. For example, water temperature should be between 15°C and 35°C, and pH between 7.0 and 8.5. When both water temperature and pH data fall within this range, it indicates that the wetland environment is suitable for nitrification. Simultaneously, "nitrogen concentration data reflecting ammonia nitrogen removal efficiency higher than a pre-defined efficiency threshold" means that by monitoring changes in ammonia nitrogen concentration, the wetland's ammonia nitrogen removal effect is assessed to determine whether it reaches the expected level, such as a removal efficiency higher than 80%. When both conditions are met, i.e., the wetland environment is suitable for nitrification and ammonia nitrogen removal efficiency is good, the "dissolved oxygen sensor data greater than or equal to a pre-defined lower concentration limit" is considered normal. This lower concentration limit is typically set based on the dissolved oxygen requirements of the nitrification reaction, for example, 2.0 mg / L. If dissolved oxygen data is below this lower limit under suitable nitrification conditions, it may indicate an anomaly in the dissolved oxygen sensor data.

[0098] The above technical solutions significantly improve the accuracy and reliability of dissolved oxygen sensor data verification, preventing inappropriate judgments of dissolved oxygen data when wetland process conditions are unsuitable or denitrification efficiency is low, thereby reducing false alarms and missed alarms. By logically correlating dissolved oxygen data with water temperature, pH, and nitrogen concentration data, the identification of anomalies in dissolved oxygen sensor data becomes more accurate, helping to quickly pinpoint whether the problem is sensor malfunction or wetland process abnormality. This provides a more solid data foundation for generating subsequent diagnostic intervention commands, thus improving the overall effectiveness of ecological wetland denitrification data analysis methods.

[0099] In a further embodiment of this application, the step of analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns to obtain the analysis results preferably includes:

[0100] Based on the historical stable operating data, the statistical range of the trend characteristics and fluctuation characteristics of the dissolved oxygen sensor data is calculated to determine historical health behavior patterns;

[0101] Calculate the behavioral characteristics presented by the dissolved oxygen sensor data and the degree of deviation from the historical health behavior pattern;

[0102] The system determines whether the deviation is greater than a preset deviation threshold to obtain an analysis result indicating whether the dissolved oxygen sensor data has deviated.

[0103] Historical stable operation data refers to data segments in which the ecological wetland has met the first condition of consistently exceeding a preset threshold for nitrogen removal efficiency within a predetermined continuous operating period, and the second condition of environmental monitoring data fluctuation values ​​being less than a preset fluctuation value or reliably determined through on-site calibration. Based on this reliable historical data, the statistical ranges of trend and fluctuation characteristics of dissolved oxygen sensor data can be calculated. Trend characteristics may include the long-term average, slope, and direction of change of the data, while fluctuation characteristics may include the standard deviation, variance, peak and trough values ​​of the data. By statistically analyzing the normal ranges of these characteristics, a dissolved oxygen behavior pattern representing the healthy operation state of the ecological wetland can be constructed.

[0104] Furthermore, during actual monitoring, the behavioral characteristics of the current dissolved oxygen sensor data are calculated in real time, such as its current trend and fluctuations. These real-time behavioral characteristics are then compared with pre-defined historical healthy behavior patterns to calculate the deviation. The deviation can be quantified using various statistical methods, such as Euclidean distance, Mahalanobis distance, and correlation coefficients, to measure the degree of difference between the current data and historical healthy patterns. A preset deviation threshold is set based on actual application requirements and system sensitivity. When the calculated deviation exceeds this preset threshold, the dissolved oxygen sensor data is considered to have deviated, indicating a possible anomaly; conversely, if the deviation is not greater than the threshold, the data is considered to be within the normal range.

[0105] This application establishes a multi-dimensional health behavior model based on historical stable operating data and quantifies the degree of deviation of current data from this model, effectively avoiding misjudgments or omissions caused by a lack of comprehensive understanding of the dynamic characteristics of data in traditional methods. This significantly improves the accuracy and sensitivity of anomaly identification, enabling potential problems in the denitrification process of ecological wetlands to be detected in a timely manner, thereby providing managers with more timely and accurate decision support and ensuring the stable achievement of wetland denitrification efficiency standards.

[0106] The following is a specific example to illustrate this.

[0107] Suppose that the stable operation data of a certain ecological wetland over the past year is confirmed as historical stable operation data. Based on this data, the system analyzes the dissolved oxygen sensor data, calculating the statistical range of its daily average, daily maximum, daily minimum, and daily standard deviation. For example, in a healthy mode, the daily average dissolved oxygen typically fluctuates between 4-6 mg / L, and the daily standard deviation is typically less than 0.5 mg / L. These statistical ranges collectively constitute the historical healthy behavior pattern. In subsequent daily monitoring, the system continuously collects dissolved oxygen sensor data. On a certain day, the system detects a daily average dissolved oxygen of 3.5 mg / L and a daily standard deviation of 0.8 mg / L. At this point, the system calculates the deviation of the current data from the historical healthy behavior pattern. For example, it can calculate the distance between the current daily average and the center of the healthy mode average range, and the distance between the current daily standard deviation and the center of the healthy mode standard deviation range, and then sum these distances using weighted averages to obtain the total deviation. If the preset deviation threshold is 1.5 (assuming a normalized distance is used), and the calculated total deviation is 2.1, the system will determine that the dissolved oxygen sensor data has deviated. This deviation will serve as a crucial basis for identifying abnormal behavior patterns in environmental monitoring data, and will be used in conjunction with other assessment and validation results to generate diagnostic intervention instructions. For example, if the deviation is primarily reflected in excessively low daily average values, it may indicate insufficient oxygen supply in the aerobic zone; if the deviation is primarily reflected in excessively large daily standard deviations, it may indicate localized water flow disturbances or sensor instability. In this way, potential problems can be identified more precisely.

[0108] Based on the above implementation scheme, after analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns to obtain the analysis results, the following steps are preferably also included:

[0109] Extract the data sequence from the dissolved oxygen sensor data that was determined to have deviated;

[0110] A low-pass filter is used to separate the slow-changing trend component that characterizes the potential bias of dissolved oxygen sensor data, and a high-pass filter is used to separate the high-frequency fluctuation component that characterizes environmental or physical disturbances.

[0111] The average rate of change of the slow-changing trend component within a preset trend period and the standard deviation of the high-frequency fluctuation component within a preset fluctuation period are extracted as quantitative features to help identify abnormal behavior patterns in the environmental monitoring data.

[0112] When dissolved oxygen sensor data is determined to have deviated, the first step is to extract the data sequence for the period in which the deviation occurred. This data sequence contains detailed information about how dissolved oxygen data changes over time. A low-pass filter is a signal processing tool that allows signals below a certain frequency to pass through while attenuating signals above that frequency. In this application, the low-pass filter is used to extract slowly changing trend components from the raw data sequence. These components typically reflect potential deviations such as sensor calibration drift, slow changes in biological activity, or long-term process parameter adjustments. A high-pass filter, on the other hand, allows signals above a certain frequency to pass through while attenuating signals below that frequency. In this application, the high-pass filter is used to separate high-frequency fluctuation components, which are typically associated with short-term environmental disturbances (such as instantaneous changes in water flow or localized aeration anomalies) or physical disturbances (such as the shedding of deposits from the sensor surface).

[0113] After separating the slow-changing trend component and the high-frequency fluctuation component, quantitative features are extracted from these components. Specifically, the average rate of change of the slow-changing trend component within a preset trend period can quantify the speed and direction of the trend change; for example, a continuous downward trend may indicate sensor malfunction or insufficient oxygen supply in the aerobic zone. The standard deviation of the high-frequency fluctuation component within a preset fluctuation period can quantify the severity of the fluctuation; for example, a significant increase in the standard deviation may indicate local hydraulic instability or intermittent disturbances. These average rates of change and standard deviations, as quantitative features, can provide more refined and discriminative information for subsequent identification of abnormal behavior patterns in environmental monitoring data.

[0114] This application, by separating trend and fluctuation components and extracting their quantitative characteristics, can distinguish different types of anomalies, such as drift caused by slow sensor misalignment or high-frequency fluctuations caused by environmental or physical disturbances. This refined analytical capability makes subsequent diagnostic intervention instructions more targeted, avoiding blind investigations and unnecessary interventions. This improves the diagnostic efficiency and accuracy of denitrification data analysis in ecological wetlands, helps to promptly identify and resolve potential problems in wetland operation, and ensures stable achievement of denitrification efficiency standards.

[0115] In further embodiments of this application, the steps of QuickSearch identifying abnormal behavior patterns in environmental monitoring data based on assessment results, verification results, and analysis results, and generating diagnostic intervention instructions for the corresponding region and / or project, include:

[0116] Based on preset evidence weighting rules, current probabilities are assigned to multiple preset potential anomaly causes according to the evaluation results, the verification results, the analysis results, and the quantitative characteristics; wherein, the potential anomaly causes include dissolved oxygen sensor inaccuracy, process anomalies in the local microenvironment where the dissolved oxygen sensor is located, and local physical disturbances; the evidence weighting rules define the degree of support of each input information for different potential anomaly causes;

[0117] Calculate the information entropy of the current diagnostic status based on the current probability of each of the potential causes of abnormality.

[0118] Based on a pre-defined operation-result correlation model, the expected probability of each potential abnormal cause is predicted after each of the multiple pre-defined diagnostic operations. The diagnostic operations include: on-site calibration of the dissolved oxygen sensor in the target area, collection of water samples from the target area for microbial community analysis, detection of the concentration of specific organic matter in the water of the target area, and measurement of the water flow velocity in the target area. The operation-result correlation model defines the result and probability obtained after each diagnostic operation, as well as the probability update relationship between each result and each potential abnormal cause.

[0119] Calculate the predicted information entropy based on the expected probability of each of the potential causes of anomalies;

[0120] Based on the information entropy of the current diagnostic status and the predicted information entropy, the expected reduction value of information entropy is obtained;

[0121] Select the diagnostic action that maximizes the expected reduction value as the generated diagnostic intervention instruction for the corresponding region and / or item.

[0122] First, a preliminary probability allocation is performed on several pre-defined potential causes of anomalies. These potential causes may include dissolved oxygen sensor inaccuracy, process anomalies in the local microenvironment where the dissolved oxygen sensor is located (e.g., obstruction of nitrification or denitrification processes), and local physical disturbances (e.g., changes in water flow or blockage). The evidence weighting rule aims to quantify the degree to which different input information (such as consistency assessment results of dissolved oxygen sensor data, correlation verification results with other environmental parameters, behavioral characteristic deviation analysis results, and trend and fluctuation quantification features extracted through filtering) supports or opposes each potential cause of anomaly. For example, if the correlation verification of dissolved oxygen sensor data with other parameters fails, the probability of sensor inaccuracy may be increased.

[0123] Secondly, based on the current probabilities of each assigned potential cause of anomaly, the information entropy of the current diagnostic state is calculated. Information entropy here quantifies the uncertainty of the current diagnostic state, i.e., the degree of ambiguity in determining the cause of the anomaly. A higher information entropy value indicates greater uncertainty in the current determination of the cause of the anomaly.

[0124] Furthermore, based on a pre-defined operation-outcome correlation model, multiple pre-defined diagnostic operations are simulated and predicted. These diagnostic operations aim to obtain more information to narrow down the diagnostic scope, such as on-site calibration of dissolved oxygen sensors in the target area to verify their accuracy, collecting water samples from the target area for microbial community analysis to assess biological activity, detecting the concentration of specific organic matter in the water of the target area to determine the presence of inhibitors, or measuring the water flow velocity in the target area to identify physical blockages. The operation-outcome correlation model defines the possible outcomes and probabilities after performing each diagnostic operation, and how these outcomes affect the probability updates of various potential causes of anomalies. Through this model, the expected probability of each potential cause of anomalies after performing each operation can be predicted.

[0125] Subsequently, the predicted information entropy is calculated based on the expected probabilities of each potential cause of anomaly. This predicted information entropy reflects the level of diagnostic uncertainty that may be reached after performing a specific diagnostic procedure and obtaining its results.

[0126] Furthermore, by comparing the information entropy of the current diagnostic state with the predicted information entropy, the expected reduction in information entropy can be obtained. This value quantifies the degree to which performing a diagnostic procedure is expected to reduce diagnostic uncertainty, i.e., the information value of that procedure.

[0127] Finally, the diagnostic operation that maximizes the expected reduction in information entropy is selected as the generated diagnostic intervention instruction for the corresponding region and / or item. This means the system will recommend the operation that most effectively reduces diagnostic uncertainty and is most likely to quickly pinpoint the cause of the anomaly.

[0128] This application's solution transforms the traditional anomaly identification and diagnosis process into an intelligent decision-making process by introducing concepts from probabilistic reasoning and information theory. When the aforementioned evaluation results, verification results, and analysis results indicate the existence of an abnormal behavior pattern, the system no longer simply outputs a preset intervention command. Instead, it first synthesizes multi-source information (evaluation results, verification results, analysis results, and quantitative characteristics) to probabilistically assess potential anomaly causes, thereby forming a preliminary understanding of the current problem state. By calculating the information entropy of the current diagnostic state, the degree of ambiguity of this understanding is quantified. To reduce this uncertainty, the system utilizes an operation-outcome correlation model to simulate the information gain that different diagnostic operations may bring, i.e., measuring the value of the operation by predicting the reduction in information entropy after execution. This decision-making mechanism based on the expected reduction in information entropy enables the system to intelligently select the intervention command with the highest diagnostic efficiency, thereby avoiding blind attempts and resource waste.

[0129] The following is a specific example to illustrate this.

[0130] Suppose that dissolved oxygen sensor data in an aerobic area of ​​an ecological wetland suddenly and continuously declines, and the aforementioned deviation analysis shows a significant deviation between its behavioral characteristics and historical healthy behavior patterns. In this case, the system will assign initial probabilities to potential causes of anomalies (e.g., dissolved oxygen sensor inaccuracy, local microenvironmental process anomalies, local physical disturbances) based on the aforementioned assessment results, verification results, analysis results, and quantitative characteristics, using pre-defined evidence weighting rules. For example, if the correlation verification results between dissolved oxygen sensor data and other water quality parameters (such as pH and water temperature) show inconsistencies, the current probability of sensor inaccuracy might be assigned as 0.5, process anomalies as 0.3, and physical disturbances as 0.2.

[0131] The system then calculates the information entropy of the current diagnostic status to quantify this uncertainty. Next, based on a pre-defined operation-outcome correlation model, the system simulates different diagnostic operations, such as "on-site calibration of the dissolved oxygen sensor in the target area," "collecting water samples from the target area for microbial community analysis," or "measuring the water flow velocity in the target area." For each operation, the model predicts the possible results and their probabilities, as well as how these results update the probability of potential anomalies. For example, if the "on-site calibration" operation is performed, there is an 80% probability that the sensor reading will deviate from the standard value (Result A), and a 20% probability that the sensor reading will be normal (Result B). If Result A occurs, the probability of sensor misalignment will increase significantly, while the probability of process anomalies and physical disturbances will decrease; if Result B occurs, the probability of sensor misalignment will decrease. The system calculates the expected probability of potential anomalies under each operation and further calculates the predicted information entropy.

[0132] By comparing the current information entropy with the predicted information entropy, the system obtains the expected reduction in information entropy for each diagnostic operation. Assuming the "on-site calibration" operation has the largest expected reduction, it indicates that this operation is most effective in reducing diagnostic uncertainty. In this case, the system will generate a diagnostic intervention instruction: "Perform on-site calibration of the dissolved oxygen sensor in the target area." In this way, the system can intelligently recommend the most effective diagnostic steps, thereby quickly locating the problem and guiding subsequent processing.

[0133] It should be noted that the information entropy is calculated based on the following formula:

[0134] ;

[0135] Where n represents the total number of the plurality of preset potential abnormal causes, H represents the probability assigned to the i-th potential cause of the anomaly, and H represents the information entropy.

[0136] This formula is the classic definition of Shannon's information entropy, used to quantify the uncertainty of random variables. H represents the information entropy of the current diagnostic state or the predicted diagnostic state; a larger value indicates higher diagnostic uncertainty regarding potential causes of anomalies, and vice versa. n refers to the number of all potential causes of anomalies considered during the diagnostic process, such as dissolved oxygen sensor malfunction, process anomalies in the local microenvironment where the dissolved oxygen sensor is located, and local physical disturbances. P_i represents the probability assigned to the i-th potential cause of anomaly, reflecting the likelihood of that cause occurring. By logarithmically weighted summing the probabilities of all potential causes of anomalies, a comprehensive measure of uncertainty can be obtained.

[0137] In the steps of identifying abnormal behavior patterns and generating diagnostic intervention instructions, this application first assigns current probabilities to multiple potential abnormal causes based on assessment results, verification results, analysis results, and quantitative characteristics, and calculates the information entropy of the current diagnostic state based on these probabilities. Subsequently, it predicts the expected probability after performing different diagnostic operations using an operation-outcome correlation model and calculates the predicted information entropy. By comparing the current information entropy with the predicted information entropy, the expected reduction in information entropy can be obtained. The introduction of this formula provides a clear mathematical basis and operability for the calculation of information entropy, thereby ensuring the accuracy and consistency of quantifying diagnostic uncertainty and providing a reliable basis for subsequently selecting the optimal diagnostic operation.

[0138] Based on the above embodiments, the method of this application preferably includes the following steps:

[0139] Obtain the feedback results after executing the diagnostic intervention instructions;

[0140] Query the probability update relationship between the operation-result association model and the feedback result;

[0141] Based on the retrieved probability update relationship, update the current probability assigned to each of the potential abnormal causes;

[0142] Based on the updated probabilities, the information entropy of the current diagnostic status is recalculated;

[0143] If the recalculated information entropy is higher than the preset entropy threshold, then based on the updated current probability, the expected reduction value after performing each diagnostic operation is re-predicted, and the diagnostic operation that maximizes the expected reduction value is selected as the subsequently generated diagnostic intervention instruction.

[0144] Otherwise, generate the corresponding processing instructions based on the most probable potential cause of the anomaly.

[0145] Obtaining feedback results after executing diagnostic intervention instructions refers to collecting data on the actual effects of the instructions after the system generates and executes the first round of diagnostic intervention instructions. For example, if the instruction is to perform on-site calibration of a dissolved oxygen sensor, the feedback result could be changes in the calibrated sensor readings, a calibration report, or manual confirmation. If the instruction is to collect water samples for microbial community analysis, the feedback result could be changes in the microbial community structure or abundance data of specific functional bacteria in the analysis report. The purpose is to obtain direct evidence of the impact of the intervention measures on the system's state.

[0146] The probability update relationship between the query operation-result association model and the feedback result can be understood as the system searching for rules or functions that match the current feedback result based on a pre-established operation-result association model. This model defines the various possible outcomes and their probabilities after performing a specific diagnostic operation, and how these outcomes affect the probability distribution of various potential causes of anomalies. For example, if the sensor reading returns to normal after calibration, the model will indicate a significant increase in the probability of the potential cause of anomaly: "dissolved oxygen sensor inaccuracy," while the probabilities of other causes will decrease accordingly. Its purpose is to provide a basis for subsequent probability updates.

[0147] Based on the retrieved probability update relationships, the current probabilities assigned to each potential cause of anomaly are updated. Specifically, this involves using Bayesian inference or other probability update algorithms to incorporate feedback information into the judgment of potential causes of anomalies. For example, if the feedback strongly supports a certain potential cause of anomaly, the probability of that cause will be increased, while the probabilities of other causes will be decreased accordingly, thus making the diagnosis more accurate. The goal is to dynamically adjust the confidence level for the causes of anomalies.

[0148] Recalculating the information entropy of the current diagnostic state based on the updated probabilities means recalculating the uncertainty of the current diagnostic state after the probability update. Information entropy is an indicator of system uncertainty; a higher entropy value indicates greater uncertainty and the need for more information to clarify the diagnosis. Its purpose is to assess the clarity of the current diagnosis.

[0149] If the recalculated information entropy exceeds the preset entropy threshold, the expected reduction after performing each diagnostic operation is re-predicted based on the updated current probability. The diagnostic operation that maximizes the expected reduction is then selected as the subsequent diagnostic intervention instruction. This indicates that if the system still has high uncertainty after receiving feedback (i.e., information entropy exceeds the threshold), a new round of diagnostic decision-making is required. The system will re-evaluate all possible diagnostic operations, predict the amount of information entropy reduction they can bring, and select the operation that minimizes uncertainty as the new intervention instruction. The preset entropy threshold can be set according to the actual application scenario and the requirements for diagnostic accuracy. For example, when the information entropy is below a certain value, the diagnosis is considered sufficiently clear. The purpose is to ensure that the system can continue to perform effective diagnoses even when uncertainty is high.

[0150] Otherwise, based on the most probable potential cause of the anomaly, corresponding processing instructions are generated. This means that if the recalculated information entropy is already below the preset entropy threshold, it indicates that the system's diagnosis of the cause of the anomaly is sufficiently clear. At this point, the system will directly generate targeted processing instructions based on the most probable potential cause of the anomaly. For example, if it is a sensor malfunction, an instruction to replace the sensor will be generated; if it is a process anomaly, an instruction to adjust the aeration rate or carbon source dosage will be generated. The purpose is to take corrective measures directly after the diagnosis is clear.

[0151] This application's solution effectively addresses the limitations of a single diagnostic command that may not fully resolve the problem or result in inaccurate diagnoses by introducing a feedback mechanism and an iterative diagnostic process. Specifically, after the system generates and executes a diagnostic intervention command, it can obtain feedback results to understand the actual effectiveness of the intervention measures in real time. These feedback results are used to update the probability of potential abnormal causes, making the system's judgment of the root cause of the anomaly closer to the actual situation. Subsequently, by recalculating the information entropy, the system can assess the uncertainty of the current diagnosis. If the diagnosis still has high uncertainty, it will again use the expected reduction value of the information entropy to select a new, better diagnostic operation, thus forming a closed, adaptive diagnostic optimization process. This iterative mechanism ensures continuous improvement and accuracy in diagnosis, avoiding blind intervention or diagnostic stagnation.

[0152] The following is a specific example to illustrate this.

[0153] Assuming that the dissolved oxygen sensor data in the aerobic zone of the ecological wetland remains low, the system, based on the evaluation results, verification results, and analysis results, initially judges that there is a high probability of "dissolved oxygen sensor inaccuracy" and "process abnormality in the local microenvironment", and generates the first round of diagnostic intervention instructions: "perform on-site calibration of the dissolved oxygen sensor in the aerobic zone".

[0154] After executing the instruction, the system obtains feedback results. For example, the calibration personnel report that the sensor reading has increased slightly after calibration, but is still below the normal range, and no obvious physical damage to the sensor was found during the calibration process. The system inputs this feedback result into the operation-outcome correlation model, which indicates that the probability of "dissolved oxygen sensor misalignment" has decreased slightly, while the probability of "process anomalies in the local microenvironment" has increased.

[0155] Based on the updated probabilities, the system recalculates the information entropy of the current diagnostic state. If the information entropy is still higher than the preset entropy threshold, it indicates that the diagnosis remains uncertain. The system will then re-predict the expected reduction in information entropy after performing other diagnostic operations (such as "collecting water samples from the target area for microbial community analysis" or "detecting the concentration of specific organic matter in the water body of the target area") based on the updated probabilities. If "collecting water samples from the target area for microbial community analysis" yields the largest reduction in information entropy, the system will select this operation as the subsequent diagnostic intervention instruction.

[0156] If, after executing the microbial community analysis command, the feedback results show that the abundance of nitrifying bacteria in the aerobic zone is significantly lower than normal, the system updates the probability again. At this point, the probability of "process anomaly in the local microenvironment" becomes extremely high, and the recalculated information entropy is lower than the preset entropy threshold. The system will then generate corresponding processing commands based on the highest probability potential cause of the anomaly (i.e., "process anomaly in the local microenvironment"), such as "adjusting the aeration rate in the aerobic zone and supplementing the carbon source," thereby achieving precise location and effective resolution of the anomaly.

[0157] like Figure 2 As shown, this application also discloses an ecological wetland denitrification data analysis and operation and maintenance system 200 based on environmental data, including:

[0158] The data acquisition module 210 is used to continuously collect environmental monitoring data from multiple different areas in the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency;

[0159] The consistency assessment module 220 is used to assess the consistency of changes in dissolved oxygen sensor data in different regions and obtain assessment results.

[0160] The correlation verification module 230 is used to perform correlation verification between the dissolved oxygen sensor data and at least one of the water temperature data, the pH value data and the nitrogen concentration data to obtain the verification result;

[0161] The deviation analysis module 240 is used to analyze the degree of deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and the historical health behavior pattern, and to obtain the analysis results;

[0162] The instruction generation module 250 is used to identify abnormal behavior patterns in the environmental monitoring data based on the evaluation results, the verification results, and the analysis results, and to generate diagnostic intervention instructions for the corresponding region and / or the corresponding project.

[0163] The data acquisition module 210 can consist of a series of physical sensors and data transmission units. For example, dissolved oxygen sensors, water temperature sensors, pH sensors, and nitrogen concentration sensors are deployed inside the wetland and transmit the collected data to the central processing unit in real time via wired or wireless communication networks (such as LoRa, NB-IoT, or Ethernet). Alternatively, the data acquisition module 210 can be a standalone hardware device integrating sensor interfaces, data storage, and communication functions; or it can be a software component running on an edge computing device, responsible for reading data from various sensor interfaces and performing preliminary processing.

[0164] The consistency assessment module 220 can be a software program running on a server or embedded controller. For example, the consistency assessment module 220 can receive dissolved oxygen sensor data streams from the data acquisition module 210 and perform real-time or periodic analysis on these data. As one implementation, the consistency assessment module 220 can employ simple threshold comparison logic, for example, considering inconsistency as a result when the difference in dissolved oxygen readings between adjacent areas exceeds a preset value; alternatively, it can measure the consistency between data based on statistical methods, such as calculating correlation coefficients or mean squared errors.

[0165] The correlation verification module 230 can also be a software component. Its inputs include dissolved oxygen sensor data and other environmental parameter data. For example, the correlation verification module 230 can pre-set a series of logical rules to determine whether a reasonable biochemical relationship exists between dissolved oxygen data and other parameters. As one implementation method, the correlation verification module 230 can make judgments based on a rule base built from expert experience; alternatively, it can use a simple linear regression model to analyze the correlation between dissolved oxygen and other parameters, and consider the verification to have failed when the correlation significantly deviates from expectations.

[0166] The deviation analysis module 240 can be a data analysis engine, the core of which is to establish and maintain a baseline of historical health behavior patterns. For example, the deviation analysis module 240 can store dissolved oxygen data during normal wetland operation over a past period and extract statistical characteristics such as average values ​​and fluctuation ranges. As one implementation method, the deviation analysis module 240 can use a sliding window averaging method to smooth the data and identify trends; alternatively, it can use a simple statistical control chart method, where a deviation is considered to have occurred when data points exceed control limits.

[0167] The instruction generation module 250 is the decision-making core of the operation and maintenance system, and can be a rule-based expert system or a machine learning model. For example, the instruction generation module 250 can receive outputs from the aforementioned evaluation, verification, and analysis modules, and make a comprehensive judgment based on preset priority or weight rules. As one implementation method, the instruction generation module 250 can preset a decision tree and output corresponding diagnostic intervention instructions based on different input combinations; or, it can use a simple lookup table to map preset abnormal patterns to predefined intervention measures.

[0168] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0169] The preferred embodiments of this application have been described in detail above, but this application is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application.

Claims

1. A method for analyzing nitrogen removal data in ecological wetlands based on environmental data, characterized in that, Includes the following steps: Continuous collection of environmental monitoring data from multiple different areas within the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency; The consistency of changes in dissolved oxygen sensor data across different regions was evaluated to obtain the evaluation results. The dissolved oxygen sensor data is correlated with at least one of the water temperature data, pH data, and nitrogen concentration data to obtain a verification result. This includes determining whether the dissolved oxygen sensor data meets preset correlation verification conditions with the water temperature data, pH data, and nitrogen concentration data, and obtaining a verification result indicating whether the dissolved oxygen sensor data is normal or abnormal. The correlation verification conditions include: when the water temperature data and pH value data are respectively within the preset suitable range for nitrification reaction, and the ammonia nitrogen removal efficiency reflected by the nitrogen concentration data is higher than the preset efficiency threshold, the dissolved oxygen sensor data is greater than or equal to the preset lower limit of concentration. The deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns is analyzed. Trend and fluctuation quantification features are extracted by filtering to obtain the analysis results. Behavioral characteristics refer to the current trend and fluctuation situation. Based on the assessment results, the verification results, and the analysis results, abnormal behavior patterns in the environmental monitoring data are identified, and diagnostic intervention instructions for the corresponding regions and / or projects are generated, including: Based on preset evidence weighting rules, current probabilities are assigned to multiple preset potential anomaly causes according to the evaluation results, the verification results, the analysis results, and the quantitative characteristics; wherein, the potential anomaly causes include dissolved oxygen sensor inaccuracy, process anomalies in the local microenvironment where the dissolved oxygen sensor is located, and local physical disturbances; the evidence weighting rules define the degree of support of each input information for different potential anomaly causes; Calculate the information entropy of the current diagnostic status based on the current probability of each of the potential causes of abnormality. Based on a pre-defined operation-result correlation model, the expected probability of each potential abnormal cause is predicted after each of the multiple pre-defined diagnostic operations. The diagnostic operations include: on-site calibration of the dissolved oxygen sensor in the target area, collection of water samples from the target area for microbial community analysis, detection of the concentration of specific organic matter in the water of the target area, and measurement of the water flow velocity in the target area. The operation-result correlation model defines the result and probability obtained after each diagnostic operation, as well as the probability update relationship between each result and each potential abnormal cause. Calculate the predicted information entropy based on the expected probability of each of the potential causes of anomalies; Based on the information entropy of the current diagnostic status and the predicted information entropy, the expected reduction value of information entropy is obtained; Select the diagnostic action that maximizes the expected reduction value as the generated diagnostic intervention instruction for the corresponding region and / or item.

2. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 1, characterized in that, The multiple different areas are divided according to the process flow and functional zones of the ecological wetland, and include an inlet area, at least one aerobic area and an outlet area arranged sequentially along the water flow direction of the ecological wetland; the environmental monitoring data are collected by multiple sets of sensors correspondingly arranged in the multiple different areas.

3. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 1, characterized in that, The steps for evaluating the consistency of dissolved oxygen sensor data across different regions and obtaining the evaluation results include: Calculate the difference or ratio between dissolved oxygen sensor data in adjacent different regions; Determine whether the difference or ratio exceeds the normal fluctuation range determined based on historical stable operating data to obtain an assessment result of whether the dissolved oxygen sensor data is abnormal or normal; The historical stable operation data is obtained by selecting data segments that meet preset conditions within a preset continuous operating period of the ecological wetland in the past. The preset conditions include: a first condition that the denitrification efficiency is continuously higher than a preset compliance threshold, and a second condition that the fluctuation value of the environmental monitoring data is less than a preset fluctuation value or is reliably determined by on-site calibration.

4. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 3, characterized in that, The steps for analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical healthy behavioral patterns, and obtaining the analysis results, include: Based on the historical stable operating data, the statistical range of the trend characteristics and fluctuation characteristics of the dissolved oxygen sensor data is calculated to determine historical health behavior patterns; Calculate the behavioral characteristics presented by the dissolved oxygen sensor data and the degree of deviation from the historical health behavior pattern; The system determines whether the deviation is greater than a preset deviation threshold to obtain an analysis result indicating whether the dissolved oxygen sensor data has deviated.

5. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 1, characterized in that, After the step of analyzing the deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical healthy behavioral patterns to obtain the analysis results, the following steps are also included: Extract the data sequence from the dissolved oxygen sensor data that was determined to have deviated; A low-pass filter is used to separate the slow-changing trend component that characterizes the potential bias of dissolved oxygen sensor data, and a high-pass filter is used to separate the high-frequency fluctuation component that characterizes environmental or physical disturbances. The average rate of change of the slow-changing trend component within a preset trend period and the standard deviation of the high-frequency fluctuation component within a preset fluctuation period are extracted as quantitative features to help identify abnormal behavior patterns in the environmental monitoring data.

6. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 1, characterized in that, The information entropy is calculated based on the following formula: Where n represents the total number of the plurality of preset potential abnormal causes, P i H represents the probability assigned to the i-th potential cause of the anomaly, and H represents the information entropy.

7. The method for analyzing denitrification data in ecological wetlands based on environmental data according to claim 1, characterized in that, It also includes the following steps: Obtain the feedback results after executing the diagnostic intervention instructions; Query the probability update relationship between the operation-result association model and the feedback result; Based on the retrieved probability update relationship, update the current probability assigned to each of the potential abnormal causes; Based on the updated probabilities, the information entropy of the current diagnostic status is recalculated; If the recalculated information entropy is higher than the preset entropy threshold, then based on the updated current probability, the expected reduction value after performing each diagnostic operation is re-predicted, and the diagnostic operation that maximizes the expected reduction value is selected as the subsequently generated diagnostic intervention instruction. Otherwise, generate the corresponding processing instructions based on the most probable potential cause of the anomaly.

8. An operation and maintenance system using the ecological wetland denitrification data analysis method based on environmental data according to claim 1, characterized in that, include: The data acquisition module is used to continuously collect environmental monitoring data from multiple different areas in the ecological wetland; wherein, the environmental monitoring data includes dissolved oxygen sensor data, water temperature data, pH value data, and nitrogen concentration data reflecting denitrification efficiency; The consistency assessment module is used to evaluate the consistency of changes in dissolved oxygen sensor data in different regions and obtain the assessment results. The correlation verification module is used to perform correlation verification between the dissolved oxygen sensor data and at least one of the water temperature data, the pH value data, and the nitrogen concentration data to obtain the verification result; The deviation analysis module is used to analyze the degree of deviation between the behavioral characteristics presented by the dissolved oxygen sensor data and historical health behavior patterns, and to obtain the analysis results; The instruction generation module is used to identify abnormal behavior patterns in the environmental monitoring data based on the evaluation results, the verification results, and the analysis results, and to generate diagnostic intervention instructions for the corresponding region and / or the corresponding project.

Images