Underground water level monitoring system and method based on artificial intelligence

Abstract

The invention discloses a groundwater level monitoring system and method based on artificial intelligence, and relates to the technical field of groundwater level monitoring. The water level monitoring management center is in communication connection with a multi-source data acquisition module, a feature sequence analysis module, an abnormal water level identification module, an abnormal fluctuation detection module, a fluctuation grading module and an early warning response module; the multi-source data acquisition module is used for integrating various sensors and external data sources and acquiring and preprocessing underground water level related data. By integrating various sensors and external data sources, the system can collect and preprocess underground water level related data, performs deep analysis on water level characteristics in combination with an artificial intelligence algorithm, effectively overcomes the problem of insufficient distribution density of traditional monitoring points, realizes comprehensive reflection of dynamic changes of regional underground water, reduces monitoring blind areas, and improves the monitoring efficiency. And the monitoring precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of groundwater level monitoring, and particularly relates to an artificial intelligence-based groundwater level monitoring system and method. Background Art

[0002] With the growth of the global population and economic development, the demand for water resources is increasing continuously, while the available fresh water resources are becoming increasingly scarce. As one of the important fresh water resources, the groundwater level change is directly related to the sustainable utilization of water resources. However, in recent years, due to overexploitation, environmental problems such as land subsidence, seawater intrusion, and water quality pollution have occurred frequently. It is urgent to establish an effective groundwater monitoring system to achieve scientific protection and reasonable utilization of groundwater resources.

[0003] In the prior art, the distribution density of monitoring points for groundwater level monitoring is insufficient, making it difficult to comprehensively reflect the dynamic changes of regional groundwater, and it is easy to form monitoring blind spots. Moreover, during the groundwater monitoring process, it is difficult to detect abnormal fluctuations in the water level, which may lead to missing the warning window and being unable to take countermeasures in time. Therefore, how to conduct a hierarchical assessment of the abnormal fluctuations in the water level during the groundwater monitoring process to ensure the accuracy of the implementation of countermeasures is the problem to be solved by the present invention. For this purpose, an artificial intelligence-based groundwater level monitoring system and method are proposed herein. Summary of the Invention

[0004] The purpose of the present invention is to provide an artificial intelligence-based groundwater level monitoring system and method to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solutions adopted by the present invention are as follows:

[0006] In the first aspect, an artificial intelligence-based groundwater level monitoring system includes a water level monitoring management center, which is communicatively connected with a multi-source data acquisition module, a feature sequence analysis module, an abnormal water level identification module, an abnormal fluctuation detection module, a fluctuation grading module, and a warning response module;

[0007] The multi-source data acquisition module is used to integrate a variety of sensors and external data sources to collect and preprocess the groundwater level-related data;

[0008] The feature sequence analysis module is used to perform feature analysis on the preprocessed groundwater level-related data, extract water level features including a fluctuation intensity index and a fluctuation feature index, determine the sub-indices of the fluctuation intensity index and the fluctuation feature index respectively, and integrate them to obtain a water level feature sequence table;

[0009] The abnormal water level recognition module is used to analyze the water level features in the water level feature sequence list by using artificial intelligence algorithms, and identify the sub-indicators of the water level features with abnormalities in combination with the preset normal fluctuation range;

[0010] The abnormal fluctuation detection module is used to combine the output results of the abnormal water level recognition module, comprehensively consider the fluctuation degree of each sub-indicator, calculate the abnormal fluctuation index, analyze the severity of the abnormal fluctuation, and timely detect the abnormal fluctuation of the water level to avoid missing the warning window;

[0011] The fluctuation grading module is used to divide different abnormal levels according to historical data and the monitoring requirements of groundwater levels, classify the abnormal water level conditions, match them with the abnormal fluctuation index, and determine the specific abnormal level of the abnormal fluctuation of the water level;

[0012] The warning response module is used to automatically trigger the warning mechanism according to the abnormal level classification result and start the preset countermeasures to achieve a rapid response.

[0013] A further improvement of the technical solution of the present invention is that the multi-source data acquisition module specifically includes:

[0014] According to the geological conditions, hydrological characteristics and monitoring objectives of the monitoring area, a variety of sensors are arranged, including groundwater level sensors and water quality sensors, to collect groundwater level and water quality parameter data, and access third-party data such as meteorological APIs, agricultural irrigation records, and industrial extraction volumes to supplement spatio-temporal background information, unify the data formats of different devices, and mark time stamps and geographical location tags to ensure the traceability of the data;

[0015] The collected groundwater level and water quality parameter data are transmitted to the water level monitoring management center, and integrated with the third-party data, and then the integrated data is subjected to data cleaning and anomaly elimination;

[0016] Perform a normalization preprocessing operation on the data after data cleaning, convert the multi-source data into a unified scale, and then integrate it to form a comprehensive data set.

[0017] A further improvement of the technical solution of the present invention is that the feature sequence analysis module specifically includes:

[0018] Extract the groundwater level-related data after preprocessing, including water level, water quality parameters and related spatio-temporal background data, and preliminarily organize the groundwater level-related data to extract the key information in the time series;

[0019] Conduct characteristic analysis on the sorted groundwater level related data, extract water level characteristics, including fluctuation intensity indicators and fluctuation characteristic indicators. Among them, the fluctuation intensity indicator reflects the severity of water level changes and is used to identify short-term abnormal fluctuations. The fluctuation characteristic indicator reflects the regularity or abnormal pattern of water level changes and is used to identify periodic or abnormal conditions;

[0020] Conduct quantitative analysis on the fluctuation intensity indicators and fluctuation characteristic indicators, and respectively extract the sub-indicators of the fluctuation intensity indicators and fluctuation characteristic indicators. Among them, the sub-indicators of the fluctuation intensity indicator are the daily water level change amplitude, the maximum instantaneous fluctuation rate, the number of times of exceeding the threshold fluctuation, and the cumulative fluctuation energy. The sub-indicators of the fluctuation characteristic indicator are the deviation degree of the fluctuation period, the number of phase mutation points, and the lag response time;

[0021] Fuse the fluctuation intensity indicators and fluctuation characteristic indicators to construct a structured characteristic sequence table. Combine the sub-indicators of the fluctuation intensity indicator and the sub-indicators of the fluctuation characteristic indicator into a single record to form a water level characteristic sequence table. Each record contains 7 sub-indicators and a timestamp.

[0022] A further improvement of the technical solution of the present invention lies in that: the abnormal water level identification module specifically includes:

[0023] Collect historical data of the target area, extract the water level characteristic sequence table from the historical data, divide it into a training set and a test set according to time, use the Gaussian mixture model (GMM) to fit the multi-modal distribution for the fluctuation intensity indicator to capture seasonal differences, use the kernel density estimation (KDE) to fit the non-parametric distribution for the fluctuation characteristic indicator to adapt to irregular fluctuation patterns, and determine the upper and lower limits of normal fluctuations based on the 3σ principle, and then determine the normal fluctuation range of each sub-indicator;

[0024] Judge whether each of the 7 sub-indicators of each record in the water level characteristic sequence table exceeds the normal fluctuation range, identify the abnormal sub-indicators and locate the abnormal types;

[0025] Conduct K-means clustering on the anomalies within the continuous time window, summarize the abnormal sub-indicators, and output the specific types of the abnormal sub-indicators and the timestamps when the anomalies occur.

[0026] A further improvement of the technical solution of the present invention lies in that: the abnormal fluctuation detection module specifically includes:

[0027] Receive the abnormal sub-indicators, their types and timestamps output by the abnormal water level identification module, and analyze the differences between the sub-indicators of the fluctuation intensity indicator and the fluctuation characteristic indicator and their normal fluctuation ranges;

[0028] According to the differences between the sub - indicators of the fluctuation intensity index and their normal fluctuation ranges, comprehensively calculate the fluctuation intensity trend value, analyze the change trend of the fluctuation intensity index. At the same time, use the differences between the sub - indicators of the fluctuation characteristic index and their normal fluctuation ranges to comprehensively calculate the fluctuation characteristic trend value and analyze the change trend of the fluctuation characteristic index;

[0029] Analyze the influence degree of the fluctuation intensity index and the fluctuation characteristic index on the water level fluctuation respectively, determine their weights, and then combine the fluctuation intensity trend value and the fluctuation characteristic trend value to calculate the abnormal fluctuation index and analyze the severity of the abnormal fluctuation.

[0030] A further improvement of the technical solution of the present invention lies in that: the calculation process of the fluctuation intensity trend value is as follows:

[0031] Receive the abnormal sub - indicators, their types and timestamps from the abnormal water level identification module, extract the sub - indicator data related to the fluctuation intensity index, including: the daily water level change amplitude, the maximum instantaneous fluctuation rate, the over - threshold fluctuation frequency, and the cumulative fluctuation energy. At the same time, obtain the historical data of the sub - indicators of the fluctuation intensity index and calculate their mean and standard deviation;

[0032] Standardize each sub - indicator of the fluctuation intensity index, convert it into a dimensionless standardized value, obtain the standardized value of the daily water level change amplitude, the standardized value of the maximum instantaneous fluctuation rate, the standardized value of the over - threshold fluctuation frequency, and the standardized value of the cumulative fluctuation energy, eliminate the influence of different indicator dimensions and magnitudes, and allocate their respective weights according to the influence degree of each sub - indicator of the fluctuation intensity index on the water level fluctuation intensity;

[0033] Multiply the standardized value of each sub - indicator of the fluctuation intensity index by its corresponding weight, and then sum to obtain the fluctuation intensity trend value, which comprehensively reflects the change trend of the water level fluctuation intensity;

[0034] The calculation process of the fluctuation characteristic trend value is as follows:

[0035] Receive the abnormal sub - indicators, their types and timestamps from the abnormal water level identification module, extract the sub - indicator data related to the fluctuation characteristic index, including: the fluctuation period deviation degree, the number of phase mutation points, and the lag response time. At the same time, obtain the historical data of the sub - indicators of the fluctuation characteristic index and calculate their mean and standard deviation;

[0036] Standardize each sub - indicator of the fluctuation characteristic index, convert each sub - indicator into a dimensionless standardized value, obtain the standardized value of the fluctuation period deviation degree, the standardized value of the number of phase mutation points, and the standardized value of the lag response time, eliminate the influence of different indicator dimensions and magnitudes, and then allocate the corresponding weights according to the influence degree of each sub - indicator of the fluctuation characteristic index on the water level fluctuation characteristic;

[0037] Multiply the standardized values of the sub - indicators of each fluctuation characteristic index by their corresponding weights, and sum them to calculate the fluctuation characteristic tendency value, which comprehensively reflects the change trend of the water level fluctuation characteristics.

[0038] A further improvement of the technical solution of the present invention lies in: the calculation process of the abnormal fluctuation index is as follows:

[0039] Obtain the calculated fluctuation intensity tendency value and the fluctuation characteristic tendency value, and determine the weights of the fluctuation intensity index and the fluctuation characteristic index according to the influence degrees of the fluctuation intensity index and the fluctuation characteristic index on the abnormal water level fluctuation;

[0040] Calculate the square of the fluctuation intensity tendency value, take the natural logarithm function of adding 1 to the fluctuation intensity tendency value, multiply the square of the fluctuation intensity tendency value by the obtained natural logarithm function, and divide by the weight of the fluctuation intensity index to obtain the fluctuation intensity index sensitivity function;

[0041] Take the natural logarithm function of adding 1 to the fluctuation characteristic tendency value, divide by the weight of the fluctuation characteristic index, and then take the square root of 0.5 to obtain the fluctuation characteristic index sensitivity function;

[0042] Multiply the calculated fluctuation intensity index sensitivity function by the fluctuation characteristic index sensitivity function to obtain the abnormal fluctuation index, and analyze the abnormal water level fluctuation situation.

[0043] A further improvement of the technical solution of the present invention lies in: the fluctuation classification module specifically includes:

[0044] Based on the historical groundwater level monitoring data of the target area collected, through statistical analysis and combined with the specific requirements of groundwater level monitoring, divide the abnormal water level fluctuation into different abnormal levels, namely, the slight abnormal level, the medium abnormal level and the severe abnormal level. Each abnormal level corresponds to a discrimination threshold related to the abnormal fluctuation index;

[0045] According to the divided abnormal levels and discrimination thresholds, match the abnormal fluctuation index with the specific abnormal levels. Among them, when the abnormal fluctuation detection module calculates the abnormal fluctuation index, the fluctuation classification module compares it with the preset discrimination threshold to determine the specific abnormal level of the abnormal water level fluctuation, and ensure that the abnormal fluctuation index is accurately classified into the corresponding abnormal level;

[0046] After determining the specific abnormal level of the abnormal water level fluctuation, the fluctuation classification module outputs the abnormal level information. The output content includes the abnormal level, the specific value of the abnormal fluctuation index and the timestamp information of the abnormal occurrence, and then transmits the abnormal level information to the early warning response module so as to take corresponding early warning measures according to the abnormal level.

[0047] A further improvement of the technical solution of the present invention lies in that: The early warning response module specifically includes:

[0048] Receive the abnormal level information transmitted by the fluctuation classification module, including the specific values of the abnormal level, the abnormal fluctuation index, and the time stamp of the occurrence of the abnormality, and automatically trigger the corresponding early warning mechanism according to the determined abnormal level;

[0049] According to the abnormal level, the early warning response module calls the preset countermeasures. After the countermeasures are executed, the early warning response module feeds back the response result to the monitoring system, records the processing process and result of the abnormal event, and the feedback content includes the early warning type, the measures taken, the response time, the executor, etc. At the same time, continuously monitor the water level change, evaluate the effect of the countermeasures. If the abnormal situation is not alleviated, re-evaluate and adjust the response measures according to the real-time data to ensure the dynamics and adaptability of the processing process.

[0050] Second, an artificial intelligence-based groundwater level monitoring method, implemented based on the above-mentioned artificial intelligence-based groundwater level monitoring system, includes the following steps:

[0051] Integrate groundwater level sensors, water quality sensors, and third-party data sources, obtain and preprocess groundwater level-related data, and form a comprehensive data set;

[0052] Conduct feature analysis on the preprocessed data, extract water level features including fluctuation intensity indicators and fluctuation feature indicators, determine the sub-indicators of the fluctuation intensity indicators and the fluctuation feature indicators respectively, and construct a structured water level feature sequence table;

[0053] Use artificial intelligence algorithms to analyze the water level features in the feature sequence table, identify abnormal sub-indicators in combination with the preset normal fluctuation range, and summarize the abnormal sub-indicators through K-means clustering to locate the abnormal types of the groundwater level;

[0054] Combined with the abnormal water level recognition result, comprehensively consider the fluctuation degree of each sub-indicator, calculate the fluctuation intensity trend value and the fluctuation feature trend value, analyze their change trends, determine the weights and calculate the abnormal fluctuation index, and evaluate the severity of the abnormal fluctuation;

[0055] According to historical data and monitoring requirements, divide the abnormal levels, match the abnormal fluctuation index with the abnormal levels, determine the specific abnormal level, trigger the corresponding early warning mechanism, start the preset countermeasures, and feedback the response result.

[0056] Due to the adoption of the above technical solution, the technical progress achieved by the present invention compared with the prior art is:

[0057] 1. The present invention provides a groundwater level monitoring system and method based on artificial intelligence. By integrating a variety of sensors and external data sources, it can collect and preprocess groundwater level-related data, deeply analyze the water level characteristics in combination with artificial intelligence algorithms, effectively overcome the problem of insufficient distribution density of traditional monitoring points, comprehensively reflect the dynamic changes of regional groundwater, reduce monitoring blind spots, and improve monitoring accuracy.

[0058] 2. The present invention provides a groundwater level monitoring system and method based on artificial intelligence. By using artificial intelligence algorithms to analyze the water level characteristic sequence table and combining with a preset normal fluctuation range, it can real-time identify sub-indicators of water level characteristics with anomalies, calculate the abnormal fluctuation index, timely detect abnormal water level fluctuations, avoid missing the warning window, and provide accurate time nodes for the implementation of countermeasures.

[0059] 3. The present invention provides a groundwater level monitoring system and method based on artificial intelligence. According to historical data and groundwater level monitoring requirements, different abnormal levels are divided, the water level anomalies are evaluated hierarchically, and they are matched with the abnormal fluctuation index to determine the specific abnormal level. According to the abnormal level, the system automatically triggers the corresponding warning mechanism and activates the preset countermeasures, achieving a precise response to abnormal water level fluctuations and improving the pertinence and effectiveness of countermeasures. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0061] Figure 1 It is a schematic diagram of the working process of the present invention;

[0062] Figure 2 It is a schematic diagram of the method process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0064] Example 1, as Figure 1 、 Figure 2As shown in the figure, the present invention provides an artificial intelligence-based groundwater level monitoring system, which includes a water level monitoring management center. The water level monitoring management center is communicatively connected with a multi-source data acquisition module, a feature sequence analysis module, an abnormal water level identification module, an abnormal fluctuation detection module, a fluctuation classification module, and a warning response module;

[0065] The multi-source data acquisition module is used to integrate a variety of sensors and external data sources, collect and preprocess groundwater level-related data. According to the geological conditions, hydrological characteristics, and monitoring objectives of the monitoring area, a variety of sensors are arranged, including groundwater level sensors and water quality sensors, to collect groundwater level and water quality parameter data, and access third-party data such as meteorological APIs, agricultural irrigation records, and industrial extraction volumes to supplement spatio-temporal background information, unify the data formats of different devices, mark time stamps and geographical location tags to ensure data traceability, transmit the collected groundwater level and water quality parameter data to the water level monitoring management center, and integrate it with third-party data. Then, data cleaning and anomaly elimination are performed on the integrated data to process the noise, missing values, and abnormal points in the original data, improve data quality. Among them, moving average filtering is applied to eliminate sensor fluctuation noise. For missing data caused by device offline or transmission interruption, linear interpolation is used to complete it. Based on the 3σ principle, data beyond the reasonable range is identified and eliminated. A normalization preprocessing operation is performed on the data after data cleaning to convert the multi-source data into a unified scale, and then an integrated dataset is formed. Among them, for data with different dimensions in the multi-source data, Min-Max standardization is used to ensure the consistency of model input, and the data is resampled according to a fixed time window to solve the problem of inconsistent sensor sampling frequencies;

[0066] A feature sequence analysis module is used to perform feature analysis on the preprocessed groundwater level-related data, extract water level features including fluctuation intensity indicators and fluctuation characteristic indicators, determine the respective sub-indicators of the fluctuation intensity indicators and the fluctuation characteristic indicators, integrate them to obtain a water level feature sequence table, extract the preprocessed groundwater level-related data, including water level, water quality parameters, and relevant spatio-temporal background data, and conduct preliminary collation on the groundwater level-related data, extract key information in the time series. At the same time, standardize the data to ensure the comparability and consistency of data from different sources. Conduct feature analysis on the collated groundwater level-related data, extract water level features, including fluctuation intensity indicators and fluctuation characteristic indicators. Among them, the fluctuation intensity indicator reflects the severity of water level changes and is used to identify short-term abnormal fluctuations. The fluctuation characteristic indicator reflects the regularity or abnormal pattern of water level changes and is used to identify periodic abnormalities. Conduct quantitative analysis on the fluctuation intensity indicators and the fluctuation characteristic indicators, and respectively extract the respective sub-indicators of the fluctuation intensity indicators and the fluctuation characteristic indicators. Among them, the sub-indicators of the fluctuation intensity indicator are the daily water level change amplitude, the maximum instantaneous fluctuation rate, the number of fluctuations exceeding the threshold, and the cumulative fluctuation energy. The daily water level change amplitude is the absolute value of the difference in water levels between two adjacent days. The maximum instantaneous fluctuation rate is the maximum value of the water level change per unit time. The number of fluctuations exceeding the threshold is the number of times the water level exceeds the historical mean ± 2 standard deviations during the statistical monitoring period. The cumulative fluctuation energy is the variance of the water level time series, reflecting the overall severity of fluctuations. The sub-indicators of the fluctuation characteristic indicator are the deviation of the fluctuation period, the number of phase mutation points, and the lag response time. The deviation of the fluctuation period is the deviation of the current water level fluctuation period compared with the historical mean period. The number of phase mutation points is the number of times the slope sign changes in the statistical water level time series, reflecting the frequent reversal of the fluctuation direction. The lag response time is the response delay of the water level to rainfall events. Integrate the fluctuation intensity indicators and the fluctuation characteristic indicators to construct a structured feature sequence table, merge the sub-indicators of the fluctuation intensity indicator and the sub-indicators of the fluctuation characteristic indicator into a single record to form a water level feature sequence table, and each record contains 7 sub-indicators and a timestamp;

[0067] An abnormal water level identification module is used to analyze the water level features in the water level feature sequence list using artificial intelligence algorithms, identify sub-indicators of abnormal water level features in combination with a preset normal fluctuation range, collect historical data of the target area, extract the water level feature sequence list from the historical data, divide it into a training set and a test set according to time, fit a multi-modal distribution for the fluctuation intensity index using a Gaussian mixture model (GMM) to capture seasonal differences, fit a non-parametric distribution for the fluctuation feature index using kernel density estimation (KDE) to adapt to irregular fluctuation patterns, determine the upper and lower limits of normal fluctuations based on the 3σ principle, and then determine the normal fluctuation range of each sub-indicator. For each of the 7 sub-indicators in each record of the water level feature sequence list, it is judged whether it exceeds the normal fluctuation range, abnormal sub-indicators are identified and the types of abnormalities are located. K-means clustering is performed on the abnormalities within a continuous time window, abnormal sub-indicators are summarized, and the specific types of abnormal sub-indicators and the timestamps of the occurrence of abnormalities are output;

[0068] An abnormal fluctuation detection module is used to combine the output results of the abnormal water level identification module, comprehensively consider the fluctuation degrees of each sub-indicator, calculate the abnormal fluctuation index, analyze the severity of the abnormal fluctuations, and timely detect abnormal water level fluctuations to avoid missing the warning window;

[0069] A fluctuation grading module is used to divide different abnormal levels according to historical data and the monitoring requirements of groundwater levels, grade the water level abnormal conditions, and match them with the abnormal fluctuation index to determine the specific abnormal level of the abnormal water level fluctuations;

[0070] A warning response module is used to automatically trigger a warning mechanism according to the results of the abnormal level division, and start preset countermeasures to achieve a quick response, reduce water resource losses and environmental risks caused by abnormal water level fluctuations, and improve the accuracy and effectiveness of the countermeasures.

[0071] Embodiment 2, as Figure 1 、 Figure 2 shown, based on Embodiment 1, the present invention provides a technical solution: Preferably, the abnormal fluctuation detection module specifically includes:

[0072] Receive the abnormal sub - indicators, their types and timestamps output by the abnormal water level recognition module, analyze the differences between the sub - indicators of the fluctuation intensity index and the fluctuation characteristic index and their normal fluctuation ranges. According to the differences between the sub - indicators of the fluctuation intensity index and its normal fluctuation range, comprehensively calculate the fluctuation intensity tendency value and analyze the change trend of the fluctuation intensity index. At the same time, use the differences between the sub - indicators of the fluctuation characteristic index and its normal fluctuation range to comprehensively calculate the fluctuation characteristic tendency value, analyze the change trend of the fluctuation characteristic index, analyze the influence degree of the fluctuation intensity index and the fluctuation characteristic index on the water level fluctuation respectively, determine their weights, and then combine the fluctuation intensity tendency value and the fluctuation characteristic tendency value to calculate the abnormal fluctuation index and analyze the severity of the abnormal fluctuation;

[0073] The calculation process of the fluctuation intensity tendency value is as follows:

[0074] Receive the abnormal sub - indicators, their types and timestamps from the abnormal water level recognition module, extract the sub - indicator data related to the fluctuation intensity index, including: the daily water level change amplitude, the maximum instantaneous fluctuation rate, the number of fluctuations exceeding the threshold, and the cumulative fluctuation energy. At the same time, obtain the historical data of the sub - indicators of the fluctuation intensity index, calculate their mean values and standard deviations, perform standardization processing on each sub - indicator of the fluctuation intensity index, convert it into a dimensionless standardized value, and obtain the standardized value of the daily water level change amplitude, the standardized value of the maximum instantaneous fluctuation rate, the standardized value of the number of fluctuations exceeding the threshold, and the standardized value of the cumulative fluctuation energy, eliminating the influence of different indicator dimensions and orders of magnitude. According to the influence degree of each sub - indicator of the fluctuation intensity index on the water level fluctuation intensity, assign their respective weights, multiply the standardized values of the sub - indicators of each fluctuation intensity index by their corresponding weights, and then sum to obtain the fluctuation intensity tendency value, comprehensively reflecting the change trend of the water level fluctuation intensity;

[0075] The calculation expression of the fluctuation intensity tendency value is;

[0076] ;

[0077] In the formula, is the fluctuation intensity tendency value, is the daily water level change amplitude, and are the mean value and standard deviation of the daily water level change amplitude respectively, is the maximum instantaneous fluctuation rate, and are the mean value and standard deviation of the maximum instantaneous fluctuation rate respectively, is the number of fluctuations exceeding the threshold, and are the mean value and standard deviation of the number of fluctuations exceeding the threshold respectively, is the cumulative fluctuation energy, and They are the mean and standard deviation of the cumulative fluctuation energy respectively. are the weights of the sub - indicators of each fluctuation intensity index, and the sum of the weights is 1, that is , the higher the fluctuation intensity tendency value, the greater the intensity of water level fluctuation and the more violent the water level change;

[0078] The calculation process of the fluctuation characteristic tendency value is as follows:

[0079] Receive the abnormal sub - indicators, their types and timestamps from the abnormal water level identification module, extract the sub - indicator data related to the fluctuation characteristic indicators, including: the deviation of the fluctuation period, the number of phase mutation points, and the lag response time. At the same time, obtain the historical data of the sub - indicators of the fluctuation characteristic indicators, calculate their mean and standard deviation, standardize each sub - indicator of the fluctuation characteristic indicator, convert each sub - indicator into a dimensionless standardized value, and obtain the standardized value of the deviation of the fluctuation period, the standardized value of the number of phase mutation points, and the standardized value of the lag response time, eliminating the influence of different index dimensions and orders of magnitude. Then, according to the influence degree of each sub - indicator of the fluctuation characteristic indicator on the water level fluctuation characteristics, assign corresponding weights, multiply the standardized value of each sub - indicator of the fluctuation characteristic indicator by its corresponding weight, and sum to calculate the fluctuation characteristic tendency value, comprehensively reflecting the change trend of the water level fluctuation characteristics;

[0080] The calculation expression of the fluctuation characteristic tendency value is;

[0081] ;

[0082] In the formula, is the fluctuation characteristic tendency value, which is used to comprehensively reflect the change trend of the fluctuation characteristic indicators, is the deviation of the fluctuation period, that is, the deviation between the current water level fluctuation period and the historical average period, is the historical mean of the deviation of the fluctuation period, is the historical standard deviation of the deviation of the fluctuation period, is the number of phase mutation points, that is, the number of times the slope sign changes in the water level time series, reflecting the frequent reversal of the fluctuation direction, is the historical mean of the number of phase mutation points, is the historical standard deviation of the number of phase mutation points, is the lag response time, is the historical mean of the lag response time, is the historical standard deviation of the lag response time, are the weights of the sub - indicators of each fluctuation characteristic indicator, and the sum of the weights is 1, that is , when is close to 0, it indicates that the water level fluctuation characteristics are close to the normal range and the fluctuation regularity is strong. When When it increases, it indicates that there are abnormalities in the water level fluctuation characteristics, and the larger the value, the higher the degree of abnormality;

[0083] The calculation process of the abnormal fluctuation index is as follows:

[0084] Obtain the calculated trend value of the fluctuation intensity and the trend value of the fluctuation characteristics, and determine the weight of the fluctuation intensity index and the weight of the fluctuation characteristics index according to the influence degree of the fluctuation intensity index and the fluctuation characteristics index on the abnormal water level fluctuation. Among them, the weight is obtained by analyzing the contribution degree of the fluctuation intensity and the fluctuation characteristics to the abnormal water level fluctuation in historical data. Calculate the square of the trend value of the fluctuation intensity, take the natural logarithm function of adding 1 to the trend value of the fluctuation intensity, multiply the square of the trend value of the fluctuation intensity by the obtained natural logarithm function, and divide by the weight of the fluctuation intensity index to obtain the sensitive function of the fluctuation intensity index. Take the natural logarithm function of adding 1 to the trend value of the fluctuation characteristics, divide by the weight of the fluctuation characteristics index, and then take the square root of 0.5 to obtain the sensitive function of the fluctuation characteristics index. Multiply the calculated sensitive function of the fluctuation intensity index by the sensitive function of the fluctuation characteristics index to obtain the abnormal fluctuation index, and analyze the abnormal water level fluctuation situation;

[0085] The calculation expression of the abnormal fluctuation index is;

[0086] ;

[0087] In the formula, is the abnormal fluctuation index, comprehensively reflecting the severity of the abnormal water level fluctuation, is the trend value of the fluctuation intensity, is the trend value of the fluctuation characteristics, is the weight of the fluctuation intensity index, reflecting the influence degree of the fluctuation intensity on the abnormal water level, is the weight of the fluctuation characteristics index, reflecting the influence degree of the fluctuation characteristics on the abnormal water level, The larger the value of, the more severe the abnormal water level fluctuation, Increase or Increase, both will cause to increase, or Increase, will correspondingly increase the influence of its corresponding index on ;

[0088] The fluctuation classification module specifically includes:

[0089] Based on the historical groundwater level monitoring data collected from the target area, through statistical analysis, combined with the specific requirements of groundwater level monitoring, the abnormal fluctuations of the water level are divided into different abnormal levels, namely the slight abnormal level, the medium abnormal level, and the severe abnormal level. Each abnormal level corresponds to a discrimination threshold related to the abnormal fluctuation index. According to the divided abnormal levels and discrimination thresholds, the abnormal fluctuation index is matched with the specific abnormal level. Among them, when the abnormal fluctuation detection module calculates the abnormal fluctuation index, the fluctuation classification module compares it with the preset discrimination threshold to determine the specific abnormal level of the water level abnormal fluctuation, ensuring that the abnormal fluctuation index is accurately classified into the corresponding abnormal level. After determining the specific abnormal level of the water level abnormal fluctuation, the fluctuation classification module outputs the abnormal level information. The output content includes the abnormal level, the specific value of the abnormal fluctuation index, and the timestamp information of the abnormal occurrence. Furthermore, the abnormal level information is transmitted to the early warning response module to take corresponding early warning measures according to the abnormal level;

[0090] Multiple discrimination thresholds are set corresponding to multiple abnormal levels one by one, and the corresponding relationships are as follows:

[0091] The discrimination threshold for the slight abnormal level is: ;

[0092] The discrimination threshold for the medium abnormal level is: ;

[0093] The discrimination threshold for the severe abnormal level is: ;

[0094] Among them, is the abnormal fluctuation index, is the upper threshold of the slight abnormal level and the lower threshold of the medium abnormal level, is the upper threshold of the medium abnormal level and the lower threshold of the severe abnormal level;

[0095] The early warning response module specifically includes:

[0096] Receive the abnormal level information transmitted by the fluctuation classification module, including the specific values of the abnormal level, abnormal fluctuation index, and the timestamp of the occurrence of the abnormality, and automatically trigger the corresponding early warning mechanism according to the determined abnormal level. Among them, for the slight abnormal level, the system records and marks the abnormal information in the background; for the medium abnormal level, the system sends text messages or emails to relevant personnel for reminder; for the severe abnormal level, the system immediately activates the audible and visual alarm or pushes an emergency notice to ensure a quick response to the abnormal water level fluctuation and avoid missing the best treatment opportunity. According to the abnormal level, the early warning response module calls the preset countermeasures. Among them, in the case of a slight abnormal level, the system only records the abnormal data for subsequent analysis and review, and reminds relevant personnel to strengthen monitoring and pay attention to the water level change trend; in the case of a medium abnormal level, it is recommended to suspend the groundwater extraction activities in the relevant area to reduce the impact on the water level, activate the auxiliary monitoring equipment to improve the monitoring frequency and accuracy, so as to more accurately grasp the water level change; in the case of a severe abnormal level, the system automatically deactivates the water-using equipment in the affected area to prevent further water resource loss, activates the groundwater recharge equipment to stabilize the water level and reduce the impact of abnormal fluctuations, and urgently allocates emergency resources, including emergency water supply equipment, monitoring equipment, etc., to cope with possible emergencies. After the countermeasures are executed, the early warning response module feeds back the response results to the monitoring system, records the processing process and results of the abnormal event, and the feedback content includes the early warning type, measures taken, response time, implementation personnel, etc. At the same time, continuously monitor the water level change, evaluate the effect of the countermeasures, and if the abnormal situation has not been alleviated, re-evaluate and adjust the response measures according to the real-time data to ensure the dynamics and adaptability of the processing process.

[0097] Embodiment 3, as Figure 1 、 Figure 2 shown, on the basis of Embodiments 1-2, the present invention further provides an artificial intelligence-based groundwater level monitoring method, which is implemented based on the above artificial intelligence-based groundwater level monitoring system, and includes the following steps:

[0098] Integrate groundwater level sensors, water quality sensors and third-party data sources, obtain and preprocess groundwater level-related data, and form a comprehensive data set;

[0099] Conduct feature analysis on the preprocessed data, extract water level features including fluctuation intensity indicators and fluctuation feature indicators, determine the sub-indicators of the fluctuation intensity indicators and fluctuation feature indicators respectively, and construct a structured water level feature sequence table;

[0100] Use artificial intelligence algorithms to analyze the water level features in the feature sequence table, identify abnormal sub-indicators in combination with the preset normal fluctuation range, and summarize the abnormal sub-indicators through K-means clustering to locate the abnormal types of the groundwater level;

[0101] Combined with the abnormal water level recognition results, comprehensively consider the fluctuation degree of each sub-index, calculate the fluctuation intensity trend value and the fluctuation characteristic trend value, analyze their change trends, determine the weights and calculate the abnormal fluctuation index to evaluate the severity of the abnormal fluctuation;

[0102] According to historical data and monitoring requirements, divide the abnormal levels, match the abnormal fluctuation index with the abnormal levels to determine the specific abnormal level, trigger the corresponding early warning mechanism, initiate the preset response measures, and feedback the response results.

[0103] The above are only specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An artificial intelligence-based groundwater level monitoring system, including a water level monitoring management center, characterized in that: The water level monitoring and management center is communicatively connected to a multi-source data acquisition module, a feature sequence analysis module, an abnormal water level identification module, an abnormal fluctuation detection module, a fluctuation classification module, and a warning response module; The multi-source data acquisition module is used to integrate a variety of sensors and external data sources, and collect and preprocess groundwater level-related data; The feature sequence analysis module is used to perform feature analysis on the preprocessed groundwater level-related data, extract water level features including a fluctuation intensity index and a fluctuation feature index, determine the sub-indices of the fluctuation intensity index and the fluctuation feature index respectively, and integrate them to obtain a water level feature sequence table; The abnormal water level identification module is used to analyze the water level features in the water level feature sequence table by using an artificial intelligence algorithm, and identify the sub-indices of the water level features with abnormalities in combination with a preset normal fluctuation range; The abnormal fluctuation detection module is used to combine the output result of the abnormal water level identification module, comprehensively consider the fluctuation degree of each sub-index, calculate an abnormal fluctuation index, and analyze the severity of the abnormal fluctuation; The fluctuation classification module is used to divide different abnormal levels according to historical data and groundwater level monitoring requirements, classify the water level abnormal conditions, and match them with the abnormal fluctuation index to determine the specific abnormal level of the water level abnormal fluctuation; The warning response module is used to automatically trigger a warning mechanism and start preset countermeasures according to the abnormal level classification result.

2. The groundwater level monitoring system based on artificial intelligence according to claim 1, characterized in that: The multi-source data acquisition module specifically includes: According to the geological conditions, hydrographic characteristics and monitoring objectives of the monitoring area, a variety of sensors are arranged, including groundwater level sensors and water quality sensors, to collect groundwater level and water quality parameter data, and access third-party data such as meteorological APIs, agricultural irrigation records, and industrial extraction volumes to supplement spatio-temporal background information, and label time stamps and geographical location tags; Transmit the collected groundwater level and water quality parameter data to the water level monitoring and management center, integrate it with the third-party data, and then perform data cleaning and anomaly elimination on the integrated data; Perform a normalization preprocessing operation on the data after data cleaning, convert the multi-source data into a unified scale, and then integrate it to form a comprehensive data set.

3. The groundwater level monitoring system based on artificial intelligence according to claim 1, characterized in that: The feature sequence analysis module specifically includes: Extract the preprocessed groundwater level-related data, including water level, water quality parameters, and related spatio-temporal background data, and preliminarily organize the groundwater level-related data to extract key information in the time series; Perform feature analysis on the organized groundwater level-related data to extract water level features, including a fluctuation intensity index and a fluctuation feature index; Perform quantitative analysis on the fluctuation intensity index and the fluctuation feature index, and extract the sub-indices of the fluctuation intensity index and the fluctuation feature index respectively. Among them, the sub-indices of the fluctuation intensity index are the daily water level change range, the maximum instantaneous fluctuation rate, the frequency of fluctuations exceeding the threshold, and the cumulative fluctuation energy, and the sub-indices of the fluctuation feature index are the deviation degree of the fluctuation period, the number of phase mutation points, and the lag response time; Integrate the fluctuation intensity index and the fluctuation characteristic index to construct a structured feature sequence table. Combine the sub-indices of the fluctuation intensity index and the sub-indices of the fluctuation characteristic index into a single record to form a water level feature sequence table. Each record contains 7 sub-indices and a timestamp.

4. The groundwater level monitoring system based on artificial intelligence according to claim 3, characterized in that: The abnormal water level identification module specifically includes: Collect historical data of the target area, extract the water level feature sequence table from the historical data, divide it into a training set and a test set according to time. Use the Gaussian mixture model to fit the multi-peak distribution for the fluctuation intensity index to capture seasonal differences, and use kernel density estimation to fit the non-parametric distribution for the fluctuation characteristic index to adapt to irregular fluctuation patterns. Determine the upper and lower limits of normal fluctuations based on the 3σ principle, and then determine the normal fluctuation range of each sub-index; Judge whether each of the 7 sub-indices in each record of the water level feature sequence table exceeds the normal fluctuation range, identify abnormal sub-indices and locate the types of abnormalities; Perform K-means clustering on the abnormalities within a continuous time window, summarize the abnormal sub-indices, and output the specific types of abnormal sub-indices and the timestamps when the abnormalities occur.

5. The groundwater level monitoring system based on artificial intelligence according to claim 3, characterized in that: The abnormal fluctuation detection module specifically includes: Receive the abnormal sub-indices, their types, and timestamps output by the abnormal water level identification module, and analyze the differences between the sub-indices of the fluctuation intensity index and the fluctuation characteristic index and their normal fluctuation ranges; According to the differences between the sub-indices of the fluctuation intensity index and their normal fluctuation ranges, comprehensively calculate the fluctuation intensity trend value to analyze the change trend of the fluctuation intensity index. At the same time, use the differences between the sub-indices of the fluctuation characteristic index and their normal fluctuation ranges to comprehensively calculate the fluctuation characteristic trend value to analyze the change trend of the fluctuation characteristic index; Analyze the influence degrees of the fluctuation intensity index and the fluctuation characteristic index on the water level fluctuation respectively, determine their weights, and then combine the fluctuation intensity trend value and the fluctuation characteristic trend value to calculate the abnormal fluctuation index and analyze the severity of the abnormal fluctuation.

6. The groundwater level monitoring system based on artificial intelligence according to claim 5, characterized in that: The calculation process of the fluctuation intensity trend value is as follows: Receive the abnormal sub-indices, their types, and timestamps from the abnormal water level identification module, extract the sub-index data related to the fluctuation intensity index, including: the daily water level change amplitude, the maximum instantaneous fluctuation rate, the frequency of fluctuations exceeding the threshold, and the cumulative fluctuation energy. At the same time, obtain the historical data of the sub-indices of the fluctuation intensity index and calculate their mean and standard deviation; Perform standardization processing on each sub-index of the fluctuation intensity index, convert it into a dimensionless standardized value, obtain the standardized value of the daily water level change amplitude, the standardized value of the maximum instantaneous fluctuation rate, the standardized value of the frequency of fluctuations exceeding the threshold, and the standardized value of the cumulative fluctuation energy. Allocate their respective weights according to the influence degrees of the sub-indices of the fluctuation intensity index on the water level fluctuation intensity; Multiply the standardized values of the sub-indices of each fluctuation intensity index by their corresponding weights, and then sum to obtain the fluctuation intensity trend value, which comprehensively reflects the change trend of the water level fluctuation intensity; The calculation process of the fluctuation characteristic trend value is as follows: Receive the abnormal sub - indicators, their types and timestamps from the abnormal water level recognition module, extract the sub - indicator data related to the fluctuation characteristic indicators, including: the deviation degree of the fluctuation period, the number of phase mutation points, and the lag response time. At the same time, obtain the historical data of the sub - indicators of the fluctuation characteristic indicators, and calculate their mean and standard deviation; Standardize each sub - indicator of the fluctuation characteristic indicators, convert each sub - indicator into a dimensionless standardized value, obtain the standardized value of the deviation degree of the fluctuation period, the standardized value of the number of phase mutation points, and the standardized value of the lag response time. Then, according to the influence degree of each sub - indicator of the fluctuation characteristic indicators on the water level fluctuation characteristics, assign corresponding weights; Multiply the standardized values of the sub - indicators of each fluctuation characteristic indicator by their corresponding weights, and sum them to calculate the fluctuation characteristic tendency value, which comprehensively reflects the change trend of the water level fluctuation characteristics.

7. The groundwater level monitoring system based on artificial intelligence according to claim 6, characterized in that: The calculation process of the abnormal fluctuation index is as follows: Obtain the calculated fluctuation intensity tendency value and the fluctuation characteristic tendency value, and determine the weights of the fluctuation intensity index and the fluctuation characteristic index according to the influence degree of the fluctuation intensity index and the fluctuation characteristic index on the abnormal water level fluctuation; Calculate the square of the fluctuation intensity tendency value, take the natural logarithm function of the sum of the fluctuation intensity tendency value plus 1, multiply the square of the fluctuation intensity tendency value by the obtained natural logarithm function, and divide by the weight of the fluctuation intensity index to obtain the fluctuation intensity index sensitivity function; Take the natural logarithm function of the sum of the fluctuation characteristic tendency value plus 1, divide by the weight of the fluctuation characteristic index, and then take the square root of 0.5 to obtain the fluctuation characteristic index sensitivity function; Multiply the calculated fluctuation intensity index sensitivity function by the fluctuation characteristic index sensitivity function to obtain the abnormal fluctuation index, and analyze the abnormal water level fluctuation situation.

8. The groundwater level monitoring system based on artificial intelligence according to claim 7, characterized in that: The fluctuation classification module specifically includes: Based on the historical groundwater level monitoring data of the target area collected, through statistical analysis, combined with the specific requirements of groundwater level monitoring, divide the abnormal water level fluctuation into different abnormal levels, namely, the slight abnormal level, the medium abnormal level, and the severe abnormal level. Each abnormal level corresponds to a discrimination threshold related to the abnormal fluctuation index; According to the divided abnormal levels and discrimination thresholds, match the abnormal fluctuation index with the specific abnormal levels. Among them, when the abnormal fluctuation detection module calculates the abnormal fluctuation index, the fluctuation classification module compares it with the preset discrimination threshold to determine the specific abnormal level of the abnormal water level fluctuation; After determining the specific abnormal level of the abnormal water level fluctuation, the fluctuation classification module outputs the abnormal level information, and the output content includes the abnormal level, the specific value of the abnormal fluctuation index, and the timestamp information of the abnormal occurrence. Then, transmit the abnormal level information to the warning response module.

9. The groundwater level monitoring system based on artificial intelligence according to claim 1, wherein: The warning response module specifically includes: Receive the abnormal level information transmitted by the fluctuation classification module, including the abnormal level, the specific value of the abnormal fluctuation index, and the timestamp of the abnormal occurrence, and automatically trigger the corresponding warning mechanism according to the determined abnormal level; According to the anomaly level, the early warning response module calls the preset countermeasures. After the countermeasures are executed, the early warning response module feeds back the response result to the monitoring system, records the processing process and result of the abnormal event. At the same time, it continuously monitors the water level change, evaluates the effect of the countermeasures. If the abnormal situation is not alleviated, it re-evaluates and adjusts the response measures according to the real-time data.

10. A groundwater level monitoring method based on artificial intelligence, which is implemented based on the groundwater level monitoring system based on artificial intelligence according to any one of the above claims 1-9, characterized in that, It includes the following steps: Integrate groundwater level sensors, water quality sensors and third-party data sources, obtain and preprocess the groundwater level-related data to form a comprehensive data set; Conduct feature analysis on the preprocessed data, extract water level features including fluctuation intensity indicators and fluctuation feature indicators, determine the sub-indicators of the fluctuation intensity indicators and the fluctuation feature indicators respectively, and construct a structured water level feature sequence table; Use artificial intelligence algorithms to analyze the water level features in the feature sequence table, identify abnormal sub-indicators in combination with the preset normal fluctuation range, and summarize the abnormal sub-indicators through K-means clustering to locate the abnormal type of the groundwater level; Combined with the abnormal water level recognition result, comprehensively consider the fluctuation degree of each sub-indicator, calculate the fluctuation intensity trend value and the fluctuation feature trend value, analyze their change trends, determine the weights and calculate the abnormal fluctuation index to evaluate the severity of the abnormal fluctuation; According to the historical data and monitoring requirements, divide the anomaly levels, match the abnormal fluctuation index with the anomaly levels, determine the specific anomaly level, trigger the corresponding early warning mechanism, start the preset countermeasures, and feedback the response result.

Images