Water affair data intelligent acquisition and management method based on machine learning

Abstract

The invention relates to the technical field of water affair management, in particular to a water affair data intelligent acquisition and management method based on machine learning. Comprising the following steps: S1, collecting the positions of devices and equipment for collecting water affair data, obtaining the water affair data for preprocessing, and determining the relationship between the devices and equipment for collecting the water affair data and the accuracy of the water affair data; s2, according to the relation between the device and equipment for collecting the water affair data and the accuracy of the water affair data, a dynamic adjustment position range is reserved for the device and equipment for collecting the water affair data. According to the method, the quality and the consistency are ensured through data preprocessing, the relationship between the device position and the data accuracy is analyzed through a Transform model, and the optimal position is defined; predicting an optimal position through an LSTM model and optimizing device control; constructing a dynamic adjustment sequence to improve the precision; risks are predicted and positions are adjusted by using an RNN model, closed-loop optimization management is formed, and intelligent and efficient data acquisition management is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of water management, and in particular to a method for intelligent collection and management of water data based on machine learning. Background Art

[0002] The intelligent collection and management method of water data covers multiple aspects such as data collection, transmission, storage, processing, analysis and visualization, aiming to improve data accuracy and management efficiency through intelligent means. The advantage of the existing intelligent collection and management method of water data is that it realizes efficient and accurate data collection and real-time analysis through automation and intelligent technology, improving the comprehensiveness of monitoring and management efficiency. However, in lakes, fixed-position sensors cannot capture dynamic changes in water quality (such as changes in water flow rate and dissolved oxygen levels), especially during high flow periods or emergencies (such as algae outbreaks), resulting in data distortion and decision-making lags, affecting the system's response speed and management effectiveness.

[0003] The advantage of the intelligent collection and management method of water data based on machine learning is that it can automatically identify and predict water quality change patterns through machine learning technology, thereby improving the accuracy of data collection and the scientific nature of management decisions. However, the collection method based on machine learning also has challenges in ensuring the objectivity and accuracy of water data. The model has poor adaptability and cannot capture the real changes in water quality in a timely manner when faced with complex and changeable hydrological conditions, resulting in data distortion. In addition, fixed-position sensors are difficult to reflect dynamic changes, especially during high-flow periods or emergencies, which affects the comprehensiveness and accuracy of the data, thereby reducing the system's response speed and the reliability of decision-making.

[0004] Therefore, the research of this scheme is of great necessity and practical significance for solving the above problems and further optimizing the intelligent collection and management method of water data based on machine learning. Summary of the invention

[0005] In order to overcome the shortcoming of poor adaptability of intelligent collection of water affairs data, the present invention provides a method for intelligent collection and management of water affairs data based on machine learning.

[0006] The technical implementation scheme of the present invention is: a method for intelligent collection and management of water affairs data based on machine learning, comprising the following steps: S1: Collect the locations of devices and equipment for collecting water service data and obtain water service data for pre-processing to determine the relationship between the devices and equipment for collecting water service data and the accuracy of water service data; S2: Based on the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, dynamically adjust the position range for the devices and equipment for collecting water affairs data, and perform optimization control; S3: Classify the collected water affairs data and use the classified data to make risk predictions; S4: Determine the final water data intelligent collection and management method based on the prediction results.

[0007] Preferably, the location of the device and equipment for collecting water data and obtaining water data for preprocessing, and determining the relationship between the device and equipment for collecting water data and the accuracy of water data, include: the preprocessing includes cleaning and standardizing the data; in view of the water body stratification phenomenon, collecting data changes at different depth levels; defining the initial collection position of the device and equipment for collecting water data as the first position, and defining the water data collected at the first position as the first data; the first position includes the locations of all devices and equipment for collecting water data in the water data collection area, and the first data includes all water data collected at the first position of all devices and equipment for collecting water data in the water data collection area; in view of the water body stratification phenomenon, performing water data accuracy analysis on the first data at the first position; the water data includes parameters reflecting the water body stratification phenomenon.

[0008] Preferably, the water data accuracy analysis of the first data at the first position includes: targeting the influence of water body stratification phenomenon, taking the process position of the device and equipment for collecting water data to reach the first position as position comparison data, and recording the position changes at different depth levels; and recording external environmental parameters; paying attention to the different water quality conditions experienced by the device and equipment when moving between different depth levels; taking the water data collected in the process of the device and equipment for collecting water data arriving at the first position as collection data comparison data; obtaining a position comparison data training set and a collection data comparison data training set, constructing a Transformer model, inputting the training set data into the Transformer model, and introducing feature engineering specially designed for water body stratification phenomenon to obtain a trained Transformer model, which is specially optimized to adapt to the water body stratification phenomenon and can effectively capture data features at different depth levels; using the Transformer model to analyze the position comparison data and the collection data comparison data, not only evaluating the data accuracy at different depth levels, but also targeting the influence of water body stratification on data collection and transmission process, thereby obtaining the relationship between the device and equipment for collecting water data and the accuracy of water data.

[0009] Preferably, according to the relationship between the devices and equipment for collecting water data and the accuracy of the water data, a dynamic adjustment position range is reserved for the devices and equipment for collecting water data, and optimization control is performed, including: obtaining the relationship between the devices and equipment for collecting water data and the accuracy of the water data, and sorting them according to the accuracy, taking the position with the highest accuracy as the first preferred position, and in view of the influence of water body stratification, combining the positions of the collecting devices and equipment in the process of analyzing the relationship between the devices and equipment for collecting water data and the accuracy of the water data with the accuracy of the collected data into a sequence, constructing an LSTM model, obtaining sequence training set data, inputting the training set data into the LSTM model, obtaining a trained LSTM model, and using the trained LSTM model to obtain the optimal position after the first position of the water data, and defining it as the second preferred position; taking the distance between the first preferred position and the second preferred position as the dynamic adjustment position range, and performing optimization control based on the dynamic adjustment position range.

[0010] Preferably, the optimization control based on the dynamic adjustment position range includes: obtaining the first position of all water data collection devices, obtaining the collection area of ​​the water data collection devices and equipment within the dynamic adjustment range with the first position as the center, and defining them as the effective collection area and the farthest collection area; considering the influence of different depth levels, obtaining the effective collection area and the farthest collection area of ​​all water data collection devices and equipment within the dynamic adjustment range, taking the intersection of the effective collection areas of all devices and equipment collecting water data as the shortest adjustment distance, and taking the intersection of the farthest collection areas as the longest adjustment distance, adjusting the position of the device and equipment collecting water data within the dynamic adjustment range when the device and equipment collecting water data is at the shortest adjustment distance, and shrinking the collection area of ​​the device and equipment when the device and equipment collecting water data is at the longest adjustment distance; and constructing a dynamic adjustment sequence for the device and equipment collecting water data.

[0011] Preferably, the dynamic adjustment sequence is constructed for the devices and equipment that collect water data, including: obtaining the first position in the collection area as a static sequence, obtaining the dynamically adjusted position in the collection area as a dynamic sequence, obtaining a static sequence data training set and a dynamic sequence data training set for data changes at different depth levels, constructing a virtual adjustment model, inputting the training set data into the virtual adjustment model to obtain a trained virtual adjustment model, and using the virtual adjustment model to obtain the adjustment amount of the device or equipment in the process of collecting water data at the first position, and then making adjustments.

[0012] Preferably, the use of the virtual adjustment model to obtain the adjustment amount of the device or equipment in the process of collecting water affairs data at the first position, and then adjusting, includes: constructing a plane rectangular coordinate system with the collection area as the center point, the water flow direction as the X-axis, and the horizontal and perpendicular direction to the X-axis as the Y-axis; the static sequence includes a fixed point, a first position, and data collection accuracy, and the dynamic sequence includes a fixed point, a dynamic adjustment range, and data collection accuracy, and the fixed point includes but is not limited to the data receiving point of the device or equipment for collecting water affairs data or the position of the fixed object used to tow and fix the device or equipment for collecting water affairs data Position; calibrate static points (points at the first position) and dynamic points (points within the dynamic adjustment range) in a plane rectangular coordinate system to form a first sequence (including a static sequence and a dynamic sequence), calculate the distance from the fixed point to the first position and the maximum adjustment range through the Euclidean distance formula, and update the first sequence to form a second sequence (a fusion sequence, adding the static sequence to the dynamic sequence), the second sequence includes the minimum distance value and the maximum distance value, the first position, data acquisition accuracy, and considers data changes at different depth levels; then segment the water conservancy data acquisition barrier medium to generate a distance adjustment sequence and a data acquisition sequence.

[0013] Preferably, the second sequence includes the minimum distance and the maximum distance, the first position, and the data collection accuracy, and then is segmented according to the water data collection barrier medium to generate a distance adjustment sequence and a data collection sequence, including: the barrier medium includes but is not limited to network transmission barriers and material transmission barriers, and for data changes at different depth levels, the distance adjustment sequence and the data collection sequence are traversed to generate a distance effective adjustment parameter set and a data effective collection parameter set, and the intersection of the distance effective adjustment parameter set and the data effective collection parameter set is obtained, and the distance amount that needs to be adjusted in the process of obtaining the intersection is adjusted as the adjustment amount.

[0014] Preferably, the collected water affairs data is classified and the classified data is used for risk prediction, including: obtaining the fluctuation period and stable period of the water affairs data in the collection area, and extracting the influencing factors of water affairs data collection in the fluctuation period and the stable period, evaluating the changes of the influencing factors at different depth levels, introducing a weight mechanism to dynamically adjust the weights according to the importance of the data in view of the impact of water body stratification on different depth levels, classifying the adjustment data of the water affairs data collection devices and equipment in the stable period as safe adjustment data, and classifying the water affairs data collection factors in the fluctuation period as dangerous adjustment data, constructing an RNN model, obtaining a training set of dangerous adjustment data, inputting the training set data into the RNN model, and introducing a feature selection mechanism optimized for the water body stratification phenomenon to obtain a trained RNN model, and using the RNN model to adjust the devices and equipment for collecting water affairs data again.

[0015] Preferably, the method for determining the final intelligent collection and management method for water affairs data based on the prediction results includes: obtaining the first position of the device and equipment in the water affairs data collection process and the position in the dynamic adjustment process, and simultaneously obtaining water affairs data, looping the entire process and continuously optimizing it, and outputting the optimized intelligent collection and management method for water affairs data.

[0016] Beneficial effects: The present invention significantly improves the objectivity and accuracy of water affairs data collection through preprocessing, dynamic adjustment optimization, risk prediction and closed-loop management: 1. Data quality assurance: Data cleaning and standardization are performed in the preprocessing stage to ensure data consistency and reliability and reduce noise interference, which is particularly suitable for complex and changeable lake environments.

[0017] 2. Intelligent location optimization: Based on the relationship between device location and data accuracy, the best collection location is automatically identified and the range is dynamically adjusted to ensure that each collection point provides the most representative data, enhancing stability and representativeness, especially when water flow changes frequently.

[0018] 3. Real-time and precise control: Build a dynamic adjustment sequence, the system predicts and executes the optimal adjustment amount, realizes real-time and precise control, improves flexibility and response speed, and ensures accurate and timely data under different hydrological conditions.

[0019] 4. Risk foresight and prevention: Classification and risk prediction identify potential risks (such as algae outbreaks) in advance, adjust strategies to avoid the impact of uncertain factors, and improve the system's foresight and response capabilities.

[0020] 5. Closed-loop optimization management: The entire process constitutes closed-loop management, continuously optimizing methods to ensure long-term efficiency and accuracy, and realize intelligent management of the entire process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flow chart of the method for intelligent collection and management of water affairs data based on machine learning of the present invention; Figure 2 A schematic diagram of the process of determining the dynamic adjustment range of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Intelligent collection and management methods of water affairs data based on machine learning, such as Figure 1As shown, the following steps are included: S1: Collect the locations of devices and equipment for collecting water service data and obtain water service data for pre-processing to determine the relationship between the devices and equipment for collecting water service data and the accuracy of water service data; S2: Based on the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, dynamically adjust the position range for the devices and equipment for collecting water affairs data, and perform optimization control; S3: Classify the collected water affairs data and use the classified data to make risk predictions; S4: Determine the final water data intelligent collection and management method based on the prediction results.

[0024] The locations of devices and equipment for collecting water affairs data are collected and the water affairs data are obtained for preprocessing, and the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data is determined, including: the preprocessing includes cleaning and standardizing the data; in view of the water body stratification phenomenon, data changes at different depth levels are collected; the initial collection location of the device and equipment for collecting water affairs data is defined as the first location, and the water affairs data collected at the first location is defined as the first data; the first location includes the locations of all devices and equipment for collecting water affairs data in the water affairs data collection area, and the first data includes all water affairs data collected at the first location of all devices and equipment for collecting water affairs data in the water affairs data collection area; in view of the water body stratification phenomenon, the first data at the first location is analyzed for water affairs data accuracy; the water affairs data includes parameters reflecting the water body stratification phenomenon.

[0025] For further explanation, suppose there are 10 buoys in a lake to collect water data. The positions of these 10 buoys are recorded in turn and defined as the first position. At the same time, the water data collected by the 10 buoys in the first position are obtained. It is particularly important to note that during the entire collection process, other water data (such as temperature, dissolved oxygen, and pH value) remain consistent, and only the water stratification phenomenon changes.

[0026] The accuracy of water affairs data is analyzed for the first data at the first position, including: in view of the influence of water body stratification phenomenon, taking the process position of the device and equipment for collecting water affairs data to reach the first position as the position comparison data, and recording the position changes at different depth levels; and recording the external environmental parameters; paying attention to the different water quality conditions experienced by the device and equipment when moving between different depth levels; taking the water affairs data collected in the process of the device and equipment for collecting water affairs data reaching the first position as the collection data comparison data; obtaining the position comparison data training set and the collection data comparison data training set, constructing a Transformer model, inputting the training set data into the Transformer model, and introducing the feature engineering designed specifically for the water body stratification phenomenon, to obtain the trained Transformer model, which is specially optimized to adapt to the water body stratification phenomenon and can effectively capture the data characteristics at different depth levels; using the Transformer model to analyze the position comparison data and the collection data comparison data, not only evaluate the data accuracy at different depth levels, but also aim at the influence of water body stratification on the data collection and transmission process, so as to obtain the relationship between the device and equipment for collecting water affairs data and the accuracy of water affairs data.

[0027] A further explanation is that, assuming there are 10 buoys used to collect lake water data, the N positions of the 10 buoys from the starting position of the fixed buoy to the first position are recorded as position comparison data, and the water data collected by the 10 buoys in the process of reaching the first position are used as collected data comparison data; the position comparison data is used to compare with the first position, and the collected data comparison data is used to compare with the data collected in the process from the fixed point to the first position; the external environmental parameters include temperature and water flow velocity.

[0028] According to the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, a dynamic adjustment position range is reserved for the devices and equipment for collecting water affairs data, and optimization control is performed, including: obtaining the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, and sorting them according to the accuracy, taking the position with the highest accuracy as the first preferred position, and in view of the influence of water body stratification, combining the positions of the collecting devices and equipment in the process of analyzing the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data with the accuracy of the collected data into a sequence, constructing an LSTM model, obtaining sequence training set data, inputting the training set data into the LSTM model, obtaining a trained LSTM model, and using the trained LSTM model to obtain the optimal position after the first position of the water affairs data, and defining it as the second preferred position; taking the distance between the first preferred position and the second preferred position as the dynamic adjustment position range, and performing optimization control based on the dynamic adjustment position range.

[0029] Further explanation is that, assuming that in a lake scenario, the first preferred position is the optimal position in the process from the fixed point to the first position, that is, the position with the highest accuracy in water affairs data collection, and the second preferred position is the optimal position other than the first position obtained by analyzing the process of the buoy from the fixed point to the first position through the LSTM model; for example, assuming that the buoy releases a total of 10 meters of cable from the fixed point, and the first preferred position is assumed to be 8 meters, then the second preferred position after analysis by the LSTM model is 12 meters (assumed value), and the dynamic adjustment position range is 8-10 meters, that is, the cable's contraction range is 0-2 meters and the release range is 0-2 meters.

[0030] The optimization control is performed based on the dynamic adjustment position range, including: obtaining the first position of all water affairs data collection, taking the first position as the center, obtaining the collection area of ​​the water affairs collection devices and equipment within the dynamic adjustment range, and defining them as the effective collection area and the farthest collection area; considering the influence of different depth levels, obtaining the effective collection area and the farthest collection area of ​​all water affairs collection devices and equipment within the dynamic adjustment range, taking the intersection of the effective collection areas of all devices and equipment collecting water affairs data as the shortest adjustment distance, and taking the intersection of the farthest collection areas as the longest adjustment distance, adjusting the position of the device and equipment collecting water affairs data within the dynamic adjustment range when the device and equipment collecting water affairs data is at the shortest adjustment distance, and shrinking the collection area of ​​the device and equipment when the device and equipment collecting water affairs data is at the longest adjustment distance; and constructing a dynamic adjustment sequence for the device and equipment collecting water affairs data.

[0031] A further explanation is that, also taking the lake scene as an example, the effective collection area, that is, the effectiveness of the buoy collecting water data, is assumed that the buoy collects water data within the dynamic adjustment range of 8-12 meters, but the water flow suddenly accelerates at 8 meters or 12 meters or other obstacles appear, such as floating branches or aquatic plants, causing the buoy to collect inaccurate water data. In this case, this area is the farthest collection area, otherwise it is an effective collection area. For ease of understanding, it is assumed here that 8-12 meters is the dynamic adjustment range, 8-8.5 and 11.5-12 are the farthest adjustment distances, and 8.5-11.5 is the effective adjustment distance; all the devices and equipment for collecting water data Taking the intersection of the effective collection areas as the shortest adjustment distance means, assuming that the data collected by buoy A is 8.5-11.5 within the effective collection range, and the data collected by buoy B is 8.5-11.5 within the effective collection range, if the collected data are the same, but both are valid data, then the contraction or release range of the cables of buoy A and buoy B is 0, or the farthest adjustment distance within the effective collection range. Taking the intersection of the farthest collection areas as the longest adjustment distance means, if the farthest collection distance between buoy A and buoy B is 8-12, if the collected data are the same, then the effective adjustment distance in the longest distance of the dynamic adjustment distance shall be adjusted.

[0032] A dynamic adjustment sequence is constructed for devices and equipment that collect water data, including: obtaining the first position in the collection area as a static sequence, obtaining the dynamically adjusted position in the collection area as a dynamic sequence, obtaining a static sequence data training set and a dynamic sequence data training set for data changes at different depth levels, constructing a virtual adjustment model, inputting the training set data into the virtual adjustment model to obtain a trained virtual adjustment model, and using the virtual adjustment model to obtain the adjustment amount of the device or equipment in the process of collecting water data at the first position, and then making adjustments.

[0033] The virtual adjustment model is used to obtain the adjustment amount of the device or equipment in the process of collecting water affairs data at the first position, and then the adjustment is performed, including: constructing a plane rectangular coordinate system with the collection area as the center point, the water flow direction as the X-axis, and the horizontal and perpendicular direction to the X-axis as the Y-axis; the static sequence includes a fixed point, a first position, and data collection accuracy, and the dynamic sequence includes a fixed point, a dynamic adjustment range, and data collection accuracy. The fixed point includes but is not limited to the data receiving point of the device or equipment for collecting water affairs data or the position of the fixed object used to tow and fix the device or equipment for collecting water affairs data; calibrating the static point (the point at the first position) and the dynamic point (the point within the dynamic adjustment range) in the plane rectangular coordinate system to form a first sequence (including the static sequence and the dynamic sequence), calculating the distance from the fixed point to the first position and the maximum adjustment range through the Euclidean distance formula, and updating the first sequence to form a second sequence (a fusion sequence, adding the static sequence to the dynamic sequence), the second sequence includes the minimum distance value and the maximum distance value, the first position, the data collection accuracy, and considering the data changes at different depth levels; then segmenting according to the barrier medium for water affairs data collection to generate a distance adjustment sequence and a data collection sequence.

[0034] Further explanation is, taking the lake as an example, assuming that there are three buoys used to collect water data, the collection area is centered on the center point of the lake, that is, in the entire lake, with the center point of the lake as the center, the first sequence includes the static sequence, [A ((0, 0), 10, high), B ((1, 1), 10, medium), C ((2, 2), 10, medium)], the dynamic sequence [A ((0, 0), 8-12, high), B ((1, 1), 8-12, medium), C ((2, 2), 8-12, medium)], and the second sequence, [A (8, 12, 10, high), B (8, 12, 10, medium), C (8, 12, 10, medium)].

[0035] The second sequence includes the minimum distance and the maximum distance, the first position, and the data collection accuracy, and then is segmented according to the water data collection barrier medium to generate a distance adjustment sequence and a data collection sequence, including: the barrier medium includes but is not limited to network transmission barriers and material transmission barriers, and for data changes at different depth levels, the distance adjustment sequence and the data collection sequence are traversed to generate a distance effective adjustment parameter set and a data effective collection parameter set, and the intersection of the distance effective adjustment parameter set and the data effective collection parameter set is obtained, and the distance amount that needs to be adjusted in the process of obtaining the intersection is adjusted as the adjustment amount.

[0036] A further explanation is that, also taking the lake scenario as an example, assuming that there are 10 buoys in the lake, the second sequence is [A (8, 12, 10, high), B (8, 12, 10, medium), C (8, 12, 10, medium), D (8, 12, 10, medium), E (8, 12, 10, high), F (8, 12, 10, high), G (8, 12, 10, medium), H (8, 12, 10, high), I (8, 12, 10, low), J (8, 12, 10, high)], and the distance sequence and data collection accuracy sequence in the second sequence are extracted, that is, the distance adjustment sequence [A (8, 12), B (8, 12), C (8, 12), D ( 8, 12), E (8, 12), F (8, 12), G (8, 12), H (8, 12), I (8, 12), J (8, 12)], data acquisition sequence [A (high), B (middle), C (middle), D (middle), E (high), F (high), G (middle), H (high), I (low), J (high)], after generating the effective distance adjustment parameter set and the effective data acquisition parameter set, there will be positions in the distance adjustment sequence and the data acquisition sequence that do not meet the data acquisition requirements. The generated effective distance adjustment parameter set and the effective data acquisition parameter set are then intersected. In this way, the buoys that do not need to be adjusted are first excluded, and then the buoys that need to be adjusted are adjusted.

[0037] The collected water affairs data are classified and the classified data are used for risk prediction, including: obtaining the fluctuation period and stable period of the water affairs data in the collection area, and extracting the influencing factors of water affairs data collection in the fluctuation period and stable period, evaluating the changes of the influencing factors at different depth levels, introducing a weight mechanism to dynamically adjust the weights according to the importance of the data in view of the impact of water body stratification on different depth levels, classifying the adjustment data of the water affairs data collection devices and equipment in the stable period as safe adjustment data, and classifying the water affairs data collection factors in the fluctuation period as dangerous adjustment data, constructing an RNN model, obtaining a training set of dangerous adjustment data, inputting the training set data into the RNN model, and introducing a feature selection mechanism optimized for the water body stratification phenomenon to obtain a trained RNN model, and using the RNN model to adjust the devices and equipment for collecting water affairs data again.

[0038] A further explanation is that by classifying the collected water service data and using the classified data for risk prediction, the data characteristics of the fluctuating period and the stable period are effectively distinguished, the influencing factors are identified, and the data in the stable period is adjusted and classified as safe adjustment data, and the data in the fluctuating period is adjusted and classified as dangerous adjustment data. The RNN model is used to predict future dangerous adjustment data, identify potential risks in advance, and adjust the position of the device accordingly to improve the accuracy and safety of data collection. For example, in lakes, by analyzing the data of water quality fluctuations during high temperatures in summer and stable water quality data in winter, influencing factors such as algae reproduction are identified, and the buoy position is adjusted accordingly to avoid high-risk areas, thereby ensuring the reliability and efficiency of data collection.

[0039] The final water data intelligent collection and management method is determined based on the prediction results, including: obtaining the first position of the device and equipment in the water data collection process and the position in the dynamic adjustment process, and simultaneously obtaining the water data, making the whole process a loop and continuously optimizing it, and outputting the optimized water data intelligent collection and management method.

[0040] Further explanation is that by determining the final intelligent collection and management method of water affairs data according to the prediction results, obtaining the initial position of the device during the collection process and its dynamically adjusted position, and synchronously obtaining water affairs data, the entire process is constructed into a closed-loop system, which is continuously optimized and finally outputs the optimized intelligent collection and management method. For example, in a lake, by recording the data of the buoy at different positions and dynamically adjusting the buoy position according to the prediction results, a closed-loop management process is formed, the position of the buoy and the data collection strategy are continuously optimized, and the efficiency and accuracy of data collection are ensured.

[0041] The above is a detailed introduction to the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for intelligent collection and management of water affairs data based on machine learning, characterized in that: The following steps are involved: S1: Collect the locations of devices and equipment for collecting water service data and obtain water service data for pre-processing to determine the relationship between the devices and equipment for collecting water service data and the accuracy of water service data; S2: Based on the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, dynamically adjust the position range for the devices and equipment for collecting water affairs data, and perform optimization control; S3: Classify the collected water affairs data and use the classified data to make risk predictions; S4: Determine the final water data intelligent collection and management method based on the prediction results.

2. The method for intelligent collection and management of water affairs data based on machine learning according to claim 1, characterized in that: The method comprises collecting the locations of the devices and equipment for collecting water affairs data and obtaining water affairs data for preprocessing, and determining the relationship between the devices and equipment for collecting water affairs data and the accuracy of water affairs data, including: the preprocessing includes cleaning and standardizing the data; in view of the water body stratification phenomenon, collecting data changes at different depth levels; defining the initial collection location of the devices and equipment for collecting water affairs data as the first location, and defining the water affairs data collected at the first location as the first data; the first location includes the locations of all devices and equipment for collecting water affairs data in the water affairs data collection area, and the first data includes all water affairs data collected at the first location of all devices and equipment for collecting water affairs data in the water affairs data collection area; in view of the water body stratification phenomenon, performing water affairs data accuracy analysis on the first data at the first location; the water affairs data includes parameters reflecting the water body stratification phenomenon.

3. The method for intelligent collection and management of water affairs data based on machine learning as claimed in claim 2, characterized in that: The water affairs data accuracy analysis of the first data at the first position includes: aiming at the influence of water body stratification phenomenon, taking the process position of the device and equipment for collecting water affairs data to reach the first position as position comparison data, and recording the position changes at different depth levels; and recording external environmental parameters; paying attention to the different water quality conditions experienced by the device and equipment when moving between different depth levels; taking the water affairs data collected in the process of the device and equipment for collecting water affairs data reaching the first position as collection data comparison data; obtaining a position comparison data training set and a collection data comparison data training set, constructing a Transformer model, inputting the training set data into the Transformer model, and introducing feature engineering designed specifically for the water body stratification phenomenon to obtain a trained Transformer model, which is specially optimized to adapt to the water body stratification phenomenon and can effectively capture data features at different depth levels; using the Transformer model to analyze the position comparison data and the collection data comparison data, not only evaluating the data accuracy at different depth levels, but also aiming at the influence of water body stratification on the data collection and transmission process, thereby obtaining the relationship between the device and equipment for collecting water affairs data and the accuracy of water affairs data.

4. The method for intelligent collection and management of water affairs data based on machine learning as claimed in claim 3, characterized in that: According to the relationship between the device and equipment for collecting water affairs data and the accuracy of the water affairs data, a dynamic adjustment position range is reserved for the device and equipment for collecting water affairs data, and optimization control is performed, including: obtaining the relationship between the device and equipment for collecting water affairs data and the accuracy of the water affairs data, and sorting them according to the accuracy, taking the position with the highest accuracy as the first preferred position, and in view of the influence of the water body stratification phenomenon, combining the position of the collecting device and equipment in the process of analyzing the relationship between the device and equipment for collecting water affairs data and the accuracy of the water affairs data and the accuracy of the collected data into a sequence, constructing an LSTM model, obtaining sequence training set data, inputting the training set data into the LSTM model, obtaining a trained LSTM model, using the trained LSTM model to obtain the optimal position after the first position of the water affairs data, and defining it as the second preferred position; taking the distance between the first preferred position and the second preferred position as the dynamic adjustment position range, and performing optimization control based on the dynamic adjustment position range.

5. The method for intelligent collection and management of water affairs data based on machine learning as claimed in claim 4, characterized in that: The optimization control based on the dynamic adjustment position range includes: obtaining the first position of all water affairs data collection, taking the first position as the center, obtaining the collection area of ​​the water affairs collection devices and equipment within the dynamic adjustment range, and defining them as the effective collection area and the farthest collection area; considering the influence of different depth levels, obtaining the effective collection area and the farthest collection area of ​​all water affairs collection devices and equipment within the dynamic adjustment range, taking the intersection of the effective collection areas of all devices and equipment collecting water affairs data as the shortest adjustment distance, and taking the intersection of the farthest collection areas as the longest adjustment distance, adjusting the position of the device and equipment collecting water affairs data within the dynamic adjustment range when the device and equipment collecting water affairs data is at the shortest adjustment distance, and shrinking the collection area of ​​the device and equipment when the device and equipment collecting water affairs data is at the longest adjustment distance; and constructing a dynamic adjustment sequence for the device and equipment collecting water affairs data.

6. The method for intelligent collection and management of water affairs data based on machine learning as claimed in claim 5, characterized in that: The method of constructing a dynamic adjustment sequence for devices and equipment that collect water data includes: obtaining the first position in the collection area as a static sequence, obtaining the dynamically adjusted position in the collection area as a dynamic sequence, obtaining a static sequence data training set and a dynamic sequence data training set for data changes at different depth levels, constructing a virtual adjustment model, inputting the training set data into the virtual adjustment model to obtain a trained virtual adjustment model, and using the virtual adjustment model to obtain the adjustment amount of the device or equipment in the process of collecting water data at the first position, and then making adjustments.

7. The method for intelligent collection and management of water affairs data based on machine learning according to claim 6, characterized in that: The method of using a virtual adjustment model to obtain the adjustment amount of the device or equipment in the process of collecting water affairs data at the first position, and then adjusting the device or equipment, includes: constructing a plane rectangular coordinate system with the collection area as the center point, the water flow direction as the X-axis, and the horizontal direction perpendicular to the X-axis as the Y-axis; the static sequence includes a fixed point, a first position, and data collection accuracy, and the dynamic sequence includes a fixed point, a dynamic adjustment range, and data collection accuracy. The fixed point includes but is not limited to the data receiving point of the device or equipment for collecting water affairs data or the position of the fixed object used to tow and fix the device or equipment for collecting water affairs data; calibrating the static point (the point at the first position) and the dynamic point (the point within the dynamic adjustment range) in the plane rectangular coordinate system to form a first sequence (including a static sequence and a dynamic sequence), calculating the distance from the fixed point to the first position and the maximum adjustment range by the Euclidean distance formula, and updating the first sequence to form a second sequence (a fusion sequence, adding the static sequence to the dynamic sequence), the second sequence includes the minimum distance value and the maximum distance value, the first position, the data collection accuracy, and considering the data changes at different depth levels; then segmenting the water affairs data collection barrier medium to generate a distance adjustment sequence and a data collection sequence.

8. The method for intelligent collection and management of water affairs data based on machine learning according to claim 7, characterized in that: The second sequence includes the minimum distance and the maximum distance, the first position, and the data collection accuracy, and then is segmented according to the water data collection barrier medium to generate a distance adjustment sequence and a data collection sequence, including: the barrier medium includes but is not limited to network transmission barriers and material transmission barriers, and for data changes at different depth levels, the distance adjustment sequence and the data collection sequence are traversed to generate a distance effective adjustment parameter set and a data effective collection parameter set, and the intersection of the distance effective adjustment parameter set and the data effective collection parameter set is obtained, and the distance amount that needs to be adjusted in the process of obtaining the intersection is adjusted as the adjustment amount.

9. The method for intelligent collection and management of water affairs data based on machine learning according to claim 1, characterized in that: The method classifies the collected water affairs data and uses the classified data for risk prediction, including: obtaining the fluctuation period and stable period of the water affairs data in the collection area, and extracting the influencing factors of water affairs data collection in the fluctuation period and stable period, evaluating the changes of the influencing factors at different depth levels, introducing a weight mechanism to dynamically adjust the weights according to the importance of the data in view of the impact of water body stratification on different depth levels, classifying the adjustment data of the water affairs data collection devices and equipment in the stable period as safe adjustment data, and classifying the water affairs data collection factors in the fluctuation period as dangerous adjustment data, constructing an RNN model, obtaining a training set of dangerous adjustment data, inputting the training set data into the RNN model, and introducing a feature selection mechanism optimized for the water body stratification phenomenon to obtain a trained RNN model, and using the RNN model to adjust the devices and equipment for collecting water affairs data again.

10. The method for intelligent collection and management of water affairs data based on machine learning according to claim 1, characterized in that: The method for determining the final intelligent collection and management method for water affairs data based on the prediction results includes: obtaining the first position of the device and equipment in the water affairs data collection process and the position in the dynamic adjustment process, and simultaneously obtaining the water affairs data, looping the entire process and continuously optimizing it, and outputting the optimized intelligent collection and management method for water affairs data.

Images