A water quality abnormality analysis method and terminal

Abstract

The present invention discloses a water quality anomaly analysis method and terminal. When the filtered water quality data is abnormal, the water quality data is converted into a feature vector, and a target historical scene similar to the current water quality anomaly is searched from a preset historical scene library based on the feature vector. A water quality anomaly processing solution is generated based on the target historical scene, that is, the water quality data is converted into a vector form, and similar historical scenes are quickly searched. When similar historical scenes exist, the processing solution of the historical scene can be directly used without the need for powerful hardware support or a large amount of calculation, thereby effectively improving the speed and accuracy of water quality anomaly analysis and improving the efficiency and effectiveness of water quality management.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality monitoring and analysis, and in particular to a method and terminal for analyzing abnormal water quality. Background Art

[0002] Existing water quality anomaly analysis technologies usually use a combination of multiple technologies such as mechanism models, machine learning, knowledge graphs, big data statistical analysis, large models, and spectral measurements to form water quality fingerprints for matching.

[0003] This technology combination primarily uses pollutant transmission relationships to identify possible pollutant transmission pathways, inferring the most likely transmission path based on indicators at monitoring points along the path. Spectroscopic water quality fingerprinting, on the other hand, uses the spectral characteristics of water bodies to match pollutant types, such as industrial wastewater or agricultural wastewater, thereby locating the pollution source based on the facilities surrounding the water quality anomaly.

[0004] These solutions are highly versatile and effective, but some require powerful hardware support, and some increase the amount of calculation due to the need to traverse various possible paths for analysis. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a water quality abnormality analysis method and terminal, which can effectively improve the speed and accuracy of water quality abnormality analysis.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for analyzing abnormal water quality, comprising the steps of:

[0008] Collecting water quality data and screening the water quality data to obtain screened water quality data;

[0009] Determining whether the filtered water quality data is abnormal, and if so, converting the filtered water quality data into a feature vector;

[0010] Searching for a target historical scene similar to the current water quality anomaly from a preset historical scene library based on the feature vector;

[0011] generating a water quality anomaly treatment plan based on the target historical scenario;

[0012] The step of converting the filtered water quality data into a feature vector comprises:

[0013] Extracting all water quality index values ​​from the filtered water quality data;

[0014] Normalizing each of the water quality index values ​​to generate a numerical feature matrix;

[0015] Determining the ratio of each water quality index value to the water quality standard value, and normalizing the ratio to generate an abnormal feature matrix;

[0016] Determining the degree of dispersion of each water quality index value, and constructing a numerical fluctuation characteristic matrix based on the degree of dispersion;

[0017] Generate a single indicator feature matrix according to the numerical feature matrix, the abnormal feature matrix and the numerical fluctuation feature matrix;

[0018] All single-indicator feature matrices are stacked in sequence to obtain a comprehensive feature matrix;

[0019] A characteristic vector is obtained according to the single-index characteristic matrix and the comprehensive characteristic matrix.

[0020] In order to solve the above technical problems, another technical solution adopted by the present invention is:

[0021] A water quality anomaly analysis terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0022] Collecting water quality data and screening the water quality data to obtain screened water quality data;

[0023] Determining whether the filtered water quality data is abnormal, and if so, converting the filtered water quality data into a feature vector;

[0024] Searching for a target historical scene similar to the current water quality anomaly from a preset historical scene library based on the feature vector;

[0025] generating a water quality anomaly treatment plan based on the target historical scenario;

[0026] The step of converting the filtered water quality data into a feature vector comprises:

[0027] Extracting all water quality index values ​​from the filtered water quality data;

[0028] Normalizing each of the water quality index values ​​to generate a numerical feature matrix;

[0029] Determining the ratio of each water quality index value to the water quality standard value, and normalizing the ratio to generate an abnormal feature matrix;

[0030] Determining the degree of dispersion of each water quality index value, and constructing a numerical fluctuation characteristic matrix based on the degree of dispersion;

[0031] Generate a single indicator feature matrix according to the numerical feature matrix, the abnormal feature matrix and the numerical fluctuation feature matrix;

[0032] All single-indicator feature matrices are stacked in sequence to obtain a comprehensive feature matrix;

[0033] A characteristic vector is obtained according to the single-index characteristic matrix and the comprehensive characteristic matrix.

[0034] The beneficial effects of the present invention are: when the water quality data after screening is abnormal, the water quality data is converted into a feature vector, and according to the feature vector, a target historical scene similar to the current water quality abnormality is searched from a preset historical scene library, and a water quality abnormality processing solution is generated based on the target historical scene, that is, the water quality data is converted into a vector form, and similar historical scenes are quickly searched. When similar historical scenes exist, the processing solution of the historical scene can be directly used without the need for powerful hardware support or a large amount of calculation, thereby effectively improving the speed and accuracy of water quality abnormality analysis and improving the efficiency and effectiveness of water quality management. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flow chart of the steps of a method for analyzing abnormal water quality according to an embodiment of the present invention;

[0036] Figure 2 The figure is a schematic structural diagram of a water quality abnormality analysis terminal according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] To illustrate the technical content, achieved objectives and effects of the present invention in detail, the following description is given in conjunction with the embodiments and accompanying drawings.

[0038] Please refer to Figure 1 , a method for analyzing abnormal water quality, comprising the steps of:

[0039] Collecting water quality data and screening the water quality data to obtain screened water quality data;

[0040] Determining whether the filtered water quality data is abnormal, and if so, converting the filtered water quality data into a feature vector;

[0041] Searching for a target historical scene similar to the current water quality anomaly from a preset historical scene library based on the feature vector;

[0042] generating a water quality anomaly treatment plan based on the target historical scenario;

[0043] The step of converting the filtered water quality data into a feature vector comprises:

[0044] Extracting all water quality index values ​​from the filtered water quality data;

[0045] Normalizing each of the water quality index values ​​to generate a numerical feature matrix;

[0046] Determining the ratio of each water quality index value to the water quality standard value, and normalizing the ratio to generate an abnormal feature matrix;

[0047] Determining the degree of dispersion of each water quality index value, and constructing a numerical fluctuation characteristic matrix based on the degree of dispersion;

[0048] Generate a single indicator feature matrix according to the numerical feature matrix, the abnormal feature matrix and the numerical fluctuation feature matrix;

[0049] All single-indicator feature matrices are stacked in sequence to obtain a comprehensive feature matrix;

[0050] A characteristic vector is obtained according to the single-index characteristic matrix and the comprehensive characteristic matrix.

[0051] From the above description, it can be seen that the beneficial effect of the present invention is that: when the water quality data after screening is abnormal, the water quality data is converted into a feature vector, and the target historical scene similar to the current water quality abnormality is searched from the preset historical scene library according to the feature vector, and a water quality abnormality processing plan is generated based on the target historical scene, that is, the water quality data is converted into a vector form, and similar historical scenes are quickly searched. When similar historical scenes exist, the processing plan of the historical scene can be directly used without the need for powerful hardware support or a large amount of calculation, thereby effectively improving the speed and accuracy of water quality abnormality analysis and improving the efficiency and effectiveness of water quality management.

[0052] Furthermore, the water quality data is screened to obtain the screened water quality data including:

[0053] Performing data cleaning and data fusion on the water quality data to obtain processed water quality data;

[0054] Abnormal data caused by hardware equipment failure is excluded from the processed water quality data to obtain filtered water quality data.

[0055] From the above description, it can be seen that the water quality data is cleaned and fused to obtain the processed water quality data, and the abnormal data caused by hardware equipment failure in the processed water quality data is eliminated. Then, the water quality abnormality is judged, which can ensure the accuracy of the data and thus ensure the accuracy of the water quality abnormality analysis.

[0056] Furthermore, determining whether the filtered water quality data is abnormal includes:

[0057] Extracting all water quality index values ​​and change trend characteristics from the filtered water quality data;

[0058] Based on the technical regulations for ranking urban surface water environmental quality, a comprehensive analysis is performed on all the water quality index values ​​and change trend characteristics to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the target water quality index data and change area characteristics with the greatest impact are determined from all the water quality index values ​​and change trend characteristics, and the target water quality index data and change area characteristics are separately analyzed based on the surface water environmental quality standards to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the filtered water quality data is determined to be normal.

[0059] From the above description, it can be seen that it is more accurate and reliable to first conduct a comprehensive analysis of all indicators and then analyze the indicators with the greatest impact separately to determine whether there are water quality abnormalities.

[0060] Furthermore, obtaining a feature vector according to the single-index feature matrix and the comprehensive feature matrix includes:

[0061] Calculating a characteristic polynomial of the comprehensive characteristic matrix and solving the characteristic polynomial to obtain a plurality of eigenvalues;

[0062] The linear equation system is solved for each of the eigenvalues ​​to obtain an eigenvector.

[0063] From the above description, it can be seen that the final characteristic vector is obtained by converting the data of multiple water quality indicators, thereby reflecting the numerical characteristics of the water quality data.

[0064] Furthermore, the preset historical scenario library includes historical scenarios, feature vectors of the historical scenarios, and abnormal causes corresponding to the feature vectors;

[0065] The step of searching a preset historical scene library for a target historical scene similar to the current water quality anomaly based on the feature vector includes:

[0066] Calculating the distance between the feature vector and the feature vector of each of the historical scenes;

[0067] Filtering candidate historical scenes whose distance is less than a preset value from the preset historical scene library;

[0068] Eliminating repeated historical scenes from the candidate historical scenes according to the abnormal reasons to obtain eliminated candidate historical scenes;

[0069] A preset number of target historical scenes are selected from the eliminated candidate historical scenes in ascending order of the distances.

[0070] From the above description, we can see that calculating the vector distance to find the target historical scene similar to the current water quality anomaly is more efficient and accurate.

[0071] Furthermore, the preset historical scenario library also includes troubleshooting plans and rectification measures corresponding to the feature vectors;

[0072] Generating a water quality anomaly treatment plan based on the target historical scenario includes:

[0073] Generate a water quality abnormality treatment plan based on the investigation plan and the corrective measures.

[0074] From the above description, it can be seen that a water quality anomaly treatment plan is generated based on the investigation plan and rectification measures of the target historical scenario found, which quickly realizes the water quality anomaly treatment and improves the efficiency and effectiveness of water quality management.

[0075] Furthermore, it also includes:

[0076] Obtaining feedback results corresponding to the water quality abnormality treatment plan, the feedback results including the actual abnormality cause, the actual investigation plan and the actual corrective measures;

[0077] If the actual abnormal cause is inconsistent with the abnormal cause of the target historical scene, a new historical scene is generated based on the actual abnormal cause, the actual investigation plan, the actual rectification measures and the characteristic vector of this water quality abnormality, and the new historical scene is added to the preset historical scene library.

[0078] From the above description, it can be seen that when the actual abnormal cause is inconsistent with the abnormal cause of the target historical scenario, it means that the recommended solution is not very accurate. A new historical scenario is generated based on the actual abnormal cause, the actual investigation plan, the actual rectification measures, and the characteristic vector of the water quality abnormality, and added to the preset historical scenario library, so as to realize the continuous updating of the preset historical scenario library and continuously improve the accuracy of water quality abnormality analysis.

[0079] Furthermore, it also includes:

[0080] If the target historical scene similar to the current water quality anomaly cannot be found in the preset historical scene library, the knowledge graph traceability analysis method is used to analyze the current water quality anomaly and generate a water quality anomaly treatment plan.

[0081] From the above description, it can be seen that if a similar historical scenario cannot be found, the traditional knowledge graph traceability analysis method is used to analyze the water quality anomaly to ensure the effectiveness of the water quality anomaly analysis.

[0082] Please refer to Figure 2Another embodiment of the present invention provides a water quality anomaly analysis terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, each step in the above-mentioned water quality anomaly analysis method is implemented.

[0083] The above-mentioned water quality anomaly analysis method and terminal of the present invention can be applied to water quality anomaly monitoring scenarios, and are described below through specific implementation methods:

[0084] Please refer to Figure 1 , embodiment 1 of the present invention is:

[0085] A method for analyzing abnormal water quality, comprising the steps of:

[0086] S1. Collect water quality data and screen the water quality data to obtain screened water quality data, specifically including S11-S13:

[0087] S11. Collect water quality data.

[0088] In an optional implementation, the water quality data to be collected is determined from the Technical Regulations on Urban Surface Water Environmental Quality Ranking.

[0089] S12: performing data cleaning and data fusion on the water quality data to obtain processed water quality data.

[0090] Specifically, the water quality data is cleaned using the Laida criterion (3 sigm principle), and the water quality data is fused to obtain processed water quality data.

[0091] In an optional embodiment, the data fusion includes interpolation to complete missing data (underreported or excluded), unifying data frequency by interpolation algorithm, and data widening. Data widening is to integrate the data and its related basic information such as monitoring sites, geographical locations, sampling methods, etc. into a record for statistical analysis according to different dimensions.

[0092] S13, eliminating abnormal data caused by hardware equipment failure in the processed water quality data to obtain filtered water quality data. S2, judging whether the filtered water quality data is abnormal, and if so, converting the filtered water quality data into a feature vector, specifically including S21-S29:

[0093] S21. Extracting all water quality index values ​​and change trend characteristics from the filtered water quality data.

[0094] S22. Based on the technical regulations for ranking urban surface water environmental quality, a comprehensive analysis is conducted on all the water quality index values ​​and change trend characteristics to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the target water quality index data and change area characteristics with the greatest impact are determined from all the water quality index values ​​and change trend characteristics, and the target water quality index data and change area characteristics are separately analyzed based on the surface water environmental quality standards to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the filtered water quality data is determined to be normal, and the process returns to S1 for a new round of data collection.

[0095] In an optional embodiment, the abnormal water quality events include a decline in water quality level (for example, from Class II to below Class II), a deteriorating trend in data over the past week, and a significant deteriorating trend in data over the past three days.

[0096] S23: If it is determined that the filtered water quality data is abnormal, extract all water quality index values ​​in the filtered water quality data.

[0097] The water quality index values ​​are recent water quality index values, such as water quality index values ​​over the past 15 days. In an optional embodiment, the water quality indexes include total phosphorus, total nitrogen, dissolved oxygen, and the like.

[0098] S24. Normalize each water quality index value to generate a numerical feature matrix.

[0099] The numerical feature matrix is ​​a matrix with 1 row and n columns. The larger the value of the element in the matrix, the larger the value of the indicator.

[0100] S25. Determine the ratio of each water quality index value to the water quality standard value, and normalize the ratio to generate an abnormal feature matrix.

[0101] The abnormal feature matrix is ​​a matrix with 1 row and n columns. The larger the value of the element in the matrix, the more serious the indicator exceeds the standard.

[0102] S26. Determine the degree of dispersion of each water quality index value, and construct a numerical fluctuation characteristic matrix based on the degree of dispersion.

[0103] The larger the value of the element in the numerical fluctuation characteristic matrix, the greater the amplitude of the data change.

[0104] S27. Generate a single-index feature matrix based on the numerical feature matrix, the abnormal feature matrix, and the numerical fluctuation feature matrix.

[0105] Specifically, the numerical feature matrix, the abnormal feature matrix, and the numerical fluctuation feature matrix are serially connected to generate a single-index feature matrix.

[0106] S28. Stack all single-indicator feature matrices in sequence to obtain a comprehensive feature matrix.

[0107] S29, obtaining a feature vector according to the single-index feature matrix and the comprehensive feature matrix, specifically including S291-S292:

[0108] S291. Calculate the characteristic polynomial of the comprehensive characteristic matrix, and solve the characteristic polynomial to obtain multiple eigenvalues.

[0109] S292. Solve the linear equation system for each of the eigenvalues ​​to obtain an eigenvector.

[0110] Wherein, the linear equations are:

[0111] (A-λI)x=0;

[0112] Where A represents the characteristic matrix, λ represents the eigenvalue, I represents the identity matrix, and x represents the eigenvector.

[0113] A single indicator feature matrix is ​​generated based on the numerical feature matrix, abnormal feature matrix and numerical fluctuation feature matrix, and a feature vector is obtained based on the single indicator feature matrix and the comprehensive feature matrix. This can extract specific numerical features from complex water quality changes so as to achieve subsequent matching of similar historical scenarios.

[0114] S3. Searching for a target historical scene similar to the current water quality anomaly from a preset historical scene library based on the feature vector, specifically including S31-S34:

[0115] The preset historical scenario library includes historical scenarios, feature vectors of the historical scenarios, abnormal causes corresponding to the feature vectors, and troubleshooting plans and corrective measures corresponding to the feature vectors;

[0116] S31. Calculate the distance between the feature vector and the feature vector of each historical scene.

[0117] In an optional implementation, the Euclidean distance between the feature vector and the feature vector of each of the historical scenes is calculated.

[0118] S32: Filter out candidate historical scenes whose distance is less than a preset value from the preset historical scene library.

[0119] S33: Eliminate repeated historical scenes from the candidate historical scenes according to the abnormal cause to obtain eliminated candidate historical scenes.

[0120] In an optional embodiment, a semantic recognition model is used to determine repeated historical scenes with similar causes of the anomaly from the candidate historical scenes, and the repeated historical scenes are eliminated until only one remains to obtain the eliminated candidate historical scenes.

[0121] S34 , selecting a preset number of target historical scenes from the eliminated candidate historical scenes in ascending order of the distances.

[0122] In an optional implementation, the preset number is 3. In an optional implementation, if the number of candidate historical scenes in the eliminated candidate historical scenes is less than 3, the eliminated candidate historical scenes are directly used as target historical scenes.

[0123] In an optional embodiment, the method further includes:

[0124] If the target historical scene similar to the current water quality anomaly cannot be found in the preset historical scene library, the knowledge graph traceability analysis method is used to analyze the current water quality anomaly and generate a water quality anomaly treatment plan. Please refer to the Chinese patent document with publication number CN116108198A.

[0125] S4. Generate a water quality anomaly treatment plan based on the target historical scenario.

[0126] Specifically, a water quality abnormality treatment plan is generated based on the investigation plan and the rectification measures.

[0127] In an optional embodiment, natural language generation or a large language model is used to process and polish the troubleshooting plan and the corrective measures, including but not limited to: eliminating the troubleshooting directions in the troubleshooting plan that do not match the cause of the problem, integrating the water quality indicator change information of this water quality anomaly, and adjusting the corrective measures according to the latest plan text to obtain a water quality anomaly handling plan.

[0128] In one optional embodiment, the water quality anomaly handling solution includes the first three possible causes and their corresponding investigation and corrective measures, as well as up to two eliminable causes for reference during manual handling. Obtaining eliminable causes is the reverse of the traditional cause analysis approach, using a threshold + condition to determine if the prerequisites for a certain possibility are not met. For example, if the target area for water quality data collection has rainfall less than a certain threshold over the past 24 hours, the impact of rainfall erosion on water quality can be ruled out.

[0129] In an optional embodiment, the method further includes:

[0130] S5. Obtain feedback results corresponding to the water quality abnormality treatment plan, wherein the feedback results include the actual abnormality cause, the actual investigation plan, and the actual corrective measures.

[0131] S6. If the actual abnormal cause is inconsistent with the abnormal cause of the target historical scene, a new historical scene is generated according to the actual abnormal cause, the actual investigation plan, the actual rectification measures and the characteristic vector of the water quality abnormality, and the new historical scene is added to the preset historical scene library.

[0132] Please refer to Figure 2 , the second embodiment of the present invention is:

[0133] A water quality anomaly analysis terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, each step of the water quality anomaly analysis method in the first embodiment is implemented.

[0134] In summary, the present invention provides a water quality anomaly analysis method and terminal. When the collected water quality data is abnormal, the water quality data is converted into a feature vector, and a target historical scene similar to the current water quality anomaly is searched from a preset historical scene library based on the feature vector. A water quality anomaly treatment plan is generated based on the target historical scene, that is, the water quality data is converted into a vector form, and similar historical scenes are quickly searched. When similar historical scenes exist, the treatment plan of the historical scene can be directly used without the need for powerful hardware support or a large amount of calculation, thereby effectively improving the speed and accuracy of water quality anomaly analysis and improving the efficiency and effect of water quality management. At the same time, according to the numerical feature matrix, the anomaly The constant characteristic matrix and the numerical fluctuation characteristic matrix generate a single indicator characteristic matrix, and the characteristic vector is obtained according to the single indicator characteristic matrix and the comprehensive characteristic matrix. It is possible to mine specific numerical characteristics from complex water quality changes so as to achieve the matching of similar historical scenarios in the future. Moreover, when the actual abnormal cause is inconsistent with the abnormal cause of the target historical scenario, it means that the recommended solution is not very accurate. A new historical scenario is generated according to the actual abnormal cause, the actual investigation plan, the actual rectification measures and the characteristic vector of the water quality abnormality, and it is added to the preset historical scenario library, so as to realize the continuous updating of the preset historical scenario library, and then continuously improve the accuracy of water quality abnormality analysis.

[0135] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's description and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for analyzing abnormal water quality, characterized in that: Including steps: Collecting water quality data and screening the water quality data to obtain screened water quality data; Determining whether the filtered water quality data is abnormal, and if so, converting the filtered water quality data into a feature vector; Searching for a target historical scene similar to the current water quality anomaly from a preset historical scene library based on the feature vector; generating a water quality anomaly treatment plan based on the target historical scenario; The step of converting the filtered water quality data into a feature vector comprises: Extracting all water quality index values ​​from the filtered water quality data; Normalizing each of the water quality index values ​​to generate a numerical feature matrix; Determining the ratio of each water quality index value to the water quality standard value, and normalizing the ratio to generate an abnormal feature matrix; Determining the degree of dispersion of each water quality index value, and constructing a numerical fluctuation characteristic matrix based on the degree of dispersion; Generate a single indicator feature matrix according to the numerical feature matrix, the abnormal feature matrix and the numerical fluctuation feature matrix; All single-indicator feature matrices are stacked in sequence to obtain a comprehensive feature matrix; A characteristic vector is obtained according to the single-index characteristic matrix and the comprehensive characteristic matrix.

2. A method for analyzing abnormal water quality according to claim 1, characterized in that: The water quality data is screened to obtain the screened water quality data including: Performing data cleaning and data fusion on the water quality data to obtain processed water quality data; Abnormal data caused by hardware equipment failure is excluded from the processed water quality data to obtain filtered water quality data.

3. A method for analyzing abnormal water quality according to claim 1, characterized in that: Determining whether the filtered water quality data is abnormal includes: Extracting all water quality index values ​​and change trend characteristics from the filtered water quality data; Based on the technical regulations for ranking urban surface water environmental quality, a comprehensive analysis is performed on all the water quality index values ​​and change trend characteristics to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the target water quality index data and change area characteristics with the greatest impact are determined from all the water quality index values ​​and change trend characteristics, and the target water quality index data and change area characteristics are separately analyzed based on the surface water environmental quality standards to determine whether there is any abnormal water quality event. If so, the filtered water quality data is determined to be abnormal. If not, the filtered water quality data is determined to be normal.

4. A method for analyzing abnormal water quality according to claim 1, characterized in that: The obtaining of the characteristic vector according to the single-index characteristic matrix and the comprehensive characteristic matrix includes: Calculating a characteristic polynomial of the comprehensive characteristic matrix and solving the characteristic polynomial to obtain a plurality of eigenvalues; The linear equation system is solved for each of the eigenvalues ​​to obtain an eigenvector.

5. A method for analyzing abnormal water quality according to claim 1, characterized in that: The preset historical scenario library includes historical scenarios, feature vectors of the historical scenarios, and abnormal causes corresponding to the feature vectors; The step of searching a preset historical scene library for a target historical scene similar to the current water quality anomaly based on the feature vector includes: Calculating the distance between the feature vector and the feature vector of each of the historical scenes; Filtering candidate historical scenes whose distance is less than a preset value from the preset historical scene library; Eliminating repeated historical scenes from the candidate historical scenes according to the abnormal reasons to obtain eliminated candidate historical scenes; A preset number of target historical scenes are selected from the eliminated candidate historical scenes in ascending order of the distances.

6. A method for analyzing abnormal water quality according to claim 5, characterized in that: The preset historical scenario library also includes troubleshooting plans and corrective measures corresponding to the feature vectors; Generating a water quality anomaly treatment plan based on the target historical scenario includes: Generate a water quality abnormality treatment plan based on the investigation plan and the corrective measures.

7. A method for analyzing abnormal water quality according to claim 1, characterized in that: Also includes: Obtaining feedback results corresponding to the water quality abnormality treatment plan, the feedback results including the actual abnormality cause, the actual investigation plan and the actual corrective measures; If the actual abnormal cause is inconsistent with the abnormal cause of the target historical scene, a new historical scene is generated based on the actual abnormal cause, the actual investigation plan, the actual rectification measures and the characteristic vector of this water quality abnormality, and the new historical scene is added to the preset historical scene library.

8. A method for analyzing abnormal water quality according to claim 1, characterized in that: Also includes: If the target historical scene similar to the current water quality anomaly cannot be found in the preset historical scene library, the knowledge graph traceability analysis method is used to analyze the current water quality anomaly and generate a water quality anomaly treatment plan.

9. A water quality abnormality analysis terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, each step of the water quality abnormality analysis method according to any one of claims 1 to 8 is implemented.

Images