Underground water level monitoring method and system

Abstract

The invention discloses an underground water level monitoring method and system, and belongs to the technical field of water level monitoring, and the method specifically comprises the steps: arranging a pressure sensor and a temperature sensor in to-be-monitored underground water, initializing the sensors, carrying out the calibration, collecting the pressure and temperature data in an underground water well in real time, and carrying out the real-time monitoring of the pressure and temperature data; the height of the underground water level is calculated and corrected according to the pressure and temperature data, filtering calibration processing is conducted on the corrected height of the underground water level, environmental noise and sensor errors are eliminated, modeling and analysis are conducted on stored historical data of the underground water level, and the future change trend of the underground water level is predicted; transmitting the finally estimated underground water level data and the predicted change trend of the underground water level, storing the data in real time and monitoring the data in real time; model parameters can be dynamically adjusted according to historical data, high-precision prediction is performed on short-term and long-term trends of the underground water level, and the method adapts to underground water level characteristics of different regions and different time periods, so that abnormal fluctuation of the underground water level is monitored in real time.

Description

Technical Field

[0001] The present invention belongs to the technical field of water level monitoring, and in particular relates to a method and system for monitoring underground water level. Background Art

[0002] Current groundwater level monitoring methods fall into two main categories: manual monitoring and automated monitoring. Traditional manual monitoring has a long cycle, poor real-time performance, and high cost, and cannot meet the needs of high-frequency and continuous groundwater level monitoring. Automated monitoring has low prediction accuracy when faced with complex water level fluctuations and is easily affected during data transmission.

[0003] For example, Chinese patent publication number CN106382972B discloses a single-hole groundwater level monitoring method that uses a buried inclinometer tube, a laser rangefinder on its top, and a reflective buoy to calculate the groundwater level and transmit the data in real time. The invention also provides a single-hole groundwater level real-time monitoring device, including a laser rangefinder, an inclinometer tube, a reflective buoy, a communication module, and a data processing terminal. This invention offers advantages for real-time measurement and monitoring, including mature equipment technology, accurate measurement results, a wide range of applications, clean and environmentally friendly energy use, high reliability, low maintenance, and labor-saving applications. It is highly suitable for groundwater level monitoring in landslide areas.

[0004] For example, a Chinese patent with authorization announcement number CN114413996B discloses a groundwater level dynamic monitoring device and method for hydrological exploration, including a freshwater groundwater level monitoring module, a groundwater chloride ion content monitoring module, a wireless data transmission base point and a prediction and comparison module. A freshwater groundwater level liquid sensor and a groundwater chloride ion content monitoring module are provided below the freshwater groundwater level monitoring module. A freshwater chloride ion content sensor is provided inside the groundwater chloride ion content monitoring module. A saltwater groundwater level liquid sensor and a saltwater chloride ion content sensor are provided below the prediction and comparison module. In this invention, after saltwater invades the freshwater aquifer, the changes in the saltwater groundwater level and the saltwater chloride ion content in the saltwater aquifer are monitored, compared with the changes in the groundwater level and chloride ion content in the freshwater aquifer, and the dynamic monitoring of the groundwater level of the entire aquifer is predicted to ensure the accuracy of the dynamic monitoring of the groundwater level.

[0005] The defects of the above patents are: 1) low accuracy in complex environments, including severe water level fluctuations; 2) automated measurement usually requires on-site burying of sensors, data cables, power cables, etc., which requires large initial investment and is difficult to protect. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention proposes a groundwater level monitoring method and system, which can realize remote and real-time groundwater level data monitoring with high precision and low power consumption. By introducing the filtering and calibration of sensor data, and a trend prediction model combining time series analysis and deep learning, it provides an innovative solution for the real-time monitoring and management of groundwater resources.

[0007] To achieve the above object, the present invention provides the following technical solutions: A groundwater level monitoring method comprises the following specific steps: Place pressure sensors and temperature sensors in the groundwater to be monitored, initialize the sensors and calibrate them; Collect pressure and temperature data in groundwater wells in real time, calculate the height of groundwater level based on the pressure and temperature data and make corrections; Perform filtering and calibration on the corrected groundwater level to eliminate environmental noise and sensor errors; Model and analyze the stored historical groundwater level data to predict future groundwater level change trends; The final estimated groundwater level data and predicted groundwater level change trends are transmitted, stored and monitored in real time.

[0008] Specifically, the real-time collection of pressure and temperature data in the groundwater well, calculation of the groundwater level based on the pressure and temperature data, and correction thereof include: Assume that the pressure data in the groundwater well collected at time t is P(t), and the temperature data in the groundwater well is T(t); According to the pressure data P(t) in the groundwater well at time t, the height of the groundwater level at time t is calculated. The specific formula is: , in, represents the height of the groundwater level at time t, P0 represents the standard atmospheric pressure, represents the density of water under standard conditions, and g represents the acceleration due to gravity; According to the temperature data T(t) in the groundwater well at time t, the height of the groundwater level at time t is corrected. The specific formula is: , in, represents the density of water at temperature T(t), represents the volume expansion coefficient of water, T0 represents the temperature under standard environment; According to the density of water at temperature T(t) The height of the groundwater level at time t Correction is performed to obtain the corrected groundwater level h(t) at time t.

[0009] Specifically, the filtering and calibration process is performed on the corrected groundwater level to eliminate environmental noise and sensor errors, including: Assume that the process noise in calculating the groundwater level height is w t , the measurement noise is v t , process noise w t and measurement noise v t The covariance matrices are Q t and R t ; The state equation is established based on the physical characteristics of groundwater level changes. The specific formula of the state equation is: , Among them, h t represents the predicted value of groundwater level at time t, h t-1 represents the predicted value of groundwater level at time t-1, F represents the state transfer matrix, B represents the control matrix, u t Indicates the control quantity; The prediction covariance matrix formula is: , P t represents the covariance matrix of the predicted groundwater level at time t, P t-1 represents the covariance matrix of the groundwater level predicted at time t-1, and T represents the transpose of the matrix; The calculated height value of the groundwater level is combined with the predicted height value using the measurement equation. The specific formula of the measurement equation is: , Among them, H represents the observation matrix, and the formula of the observation residual is: ,y t Represents the observed residual, and the covariance formula of the observed residual is: , S t represents the covariance of the observation residuals.

[0010] Specifically, the filtering and calibration process of the corrected groundwater level to eliminate environmental noise and sensor errors also includes: An adaptive feedback control mechanism is introduced to analyze the mean square error of the observation residual in real time, evaluate the noise level and dynamically adjust the noise covariance. The specific formula is: , , in, represents the dynamically adjusted process noise covariance matrix, represents the measurement noise covariance matrix after dynamic adjustment, and Represents the adaptive adjustment coefficient; According to the gain, the observation residual is weighted and combined with the predicted value of the groundwater level to obtain the optimal estimate of the groundwater level. The specific formula is: , in, represents the optimal estimated value of the groundwater level at time t, that is, the groundwater level value after filtering and calibration, K t Represents the gain coefficient, and the prediction covariance matrix is ​​updated while calculating the optimal estimated value of the groundwater level. The specific formula for the update is: , represents the updated prediction covariance matrix, and I represents the identity matrix.

[0011] Specifically, the modeling and analysis of the stored historical groundwater level data to predict future groundwater level change trends includes: Assuming the groundwater level historical data set is X, the short-term groundwater level prediction is performed. The specific formula is: , in, represents the predicted value of groundwater level at time t1, c represents the constant term, represents the autoregressive coefficient, represents the groundwater level at time t1-i in the past, Is and J represent the total short-term time in the past, represents random error, represents the random error term at time t1-j in the past, represents the sliding mean coefficient; The LSTM network is used to predict the long-term groundwater level. The memory gating mechanism within the LSTM network maintains the long-term dependency of the data and obtains the predicted groundwater level at time t2. ; Predicted value of groundwater level at time t1 and the predicted groundwater level at time t2 Perform adaptive weighted fusion, the specific formula is: , in, represents the final predicted value of groundwater level at time t1, represents the weight coefficient; The final predicted value of groundwater level is used for anomaly detection and trend analysis, and the trend change rate of the groundwater level prediction result is calculated. The specific formula is: , in, Indicates the trend change rate of the groundwater level prediction results, represents the final predicted value of groundwater level at time z, represents the final predicted value of groundwater level at time z-1, Represents the absolute value function.

[0012] Specifically, the final predicted value of the groundwater level at time t1 is , which also represents the final predicted value of the groundwater level at time t2.

[0013] Specifically, the pressure sensor and the temperature sensor are transmitted using Bluetooth signals.

[0014] A groundwater level monitoring system for implementing the above-mentioned groundwater level monitoring method, comprising: an initial setting module, a water level calculation and correction module, a filter calibration module, a water level trend analysis module and a real-time transmission monitoring module; The initial setting module is used to place pressure sensors and temperature sensors in the groundwater to be monitored, initialize the sensors and calibrate them; The water level calculation and correction module is used to collect pressure and temperature data in the groundwater well in real time, calculate the height of the groundwater level based on the pressure and temperature data, and perform correction; The filter calibration module is used to perform filter calibration on the corrected groundwater level to eliminate environmental noise and sensor errors; The water level trend analysis module is used to model and analyze the stored historical groundwater level data and predict future groundwater level change trends; The real-time transmission and monitoring module is used to transmit the final estimated groundwater level data and the predicted groundwater level change trend, and to store and monitor them in real time.

[0015] Specifically, the filter calibration module includes: a state equation unit, a measurement equation unit and an optimal estimation unit; The state equation unit is used to establish a state equation based on the physical characteristics of groundwater level changes to predict the groundwater level at time t; The measurement equation unit is used to fuse the calculated height value and the predicted height value of the groundwater level using the measurement equation; The optimal estimation unit is used to combine the weighted observation residual with the predicted value of the groundwater level according to the gain to obtain the optimal estimated value of the groundwater level.

[0016] Specifically, the water level trend analysis module includes: a short-term prediction unit, a long-term prediction unit, a final fusion unit and a water level trend analysis unit; The short-term prediction unit is used to predict the short-term water level of groundwater; The long-term prediction unit is used to predict the long-term groundwater level using an LSTM network; The final fusion unit is used to perform adaptive weighted fusion on the groundwater level prediction value at time t1 and the groundwater level prediction value at time t2; The water level trend analysis unit is used to use the obtained final predicted value of the groundwater level to perform anomaly detection and trend analysis, and calculate the trend change rate of the groundwater level prediction result.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention proposes a groundwater level monitoring method that uses a Bluetooth sensor, which can effectively reduce the need for manual intervention, thereby reducing operating and maintenance costs, and is not easily affected by the underground environment.

[0018] 2. The present invention proposes a groundwater level monitoring method that can dynamically adjust model parameters based on historical data, make high-precision predictions of short-term and long-term trends of groundwater levels, provide reliable trend data, and provide effective governance decisions.

[0019] 3. The present invention proposes a groundwater level monitoring method with strong adaptability, which can adapt to the groundwater level characteristics of different regions and different time periods, thereby monitoring abnormal fluctuations of the groundwater level in real time, and performing abnormal detection and alarm by setting thresholds. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A flow chart of a groundwater level monitoring method provided by the present invention; Figure 2 A flow chart of the real-time optimal estimation of water level provided by the present invention; Figure 3 A flow chart of water level trend analysis provided by the present invention; Figure 4 This is an architecture diagram of a groundwater level monitoring system provided by the present invention. DETAILED DESCRIPTION

[0021] Example 1 See also Figure 1-Figure 3 The present invention provides an embodiment of a groundwater level monitoring method, comprising the following specific steps: Step S1: Place pressure sensors and temperature sensors in the groundwater to be monitored to collect groundwater level and ambient temperature data. The sensor installation depth is D s , and ensure that the sensor's range and sensitivity meet the monitoring requirements, initialize the sensor and calibrate it, and collect initial environmental parameters, including initial water level, initial temperature and other parameters; The deployed pressure sensors and temperature sensors use Bluetooth for signal transmission. Compared with network transmission, the advantages are: low energy consumption, not easily affected by the surrounding environment, and stable signal transmission.

[0022] Step S2: collecting pressure and temperature data in the groundwater well in real time, calculating the height of the groundwater level based on the pressure and temperature data and making corrections; The specific steps of step S2 include: Step S201: Set the pressure data in the groundwater well collected at time t as P(t), and the temperature data in the groundwater well as T(t); Step S202: Calculate the height of the groundwater level at time t based on the pressure data P(t) in the groundwater well at time t. The specific formula is: , in, represents the height of the groundwater level at time t, P0 represents the standard atmospheric pressure, represents the density of water under standard conditions, and g represents the acceleration due to gravity; Step S203: Correct the height of the groundwater level at time t according to the temperature data T(t) in the groundwater well at time t. The specific formula is: , in, represents the density of water at temperature T(t), Represents the volume expansion coefficient of water, T0 represents the temperature under standard environment; according to the density of water under temperature environment T(t) The height of the groundwater level at time t Correction is performed to obtain the corrected groundwater level h(t) at time t.

[0023] Step S3: Perform filtering and calibration on the corrected groundwater level to eliminate environmental noise and sensor errors; The specific steps of step S3 are: Step S301: Set the process noise in the process of calculating the groundwater level as w t , the measurement noise is v t , process noise w t and measurement noise v t The covariance matrices are Q t and R t ; Process noise covariance matrix Q t Represents the uncertainty of state change over time, reflecting the natural random changes of groundwater level fluctuations, and the measurement noise covariance matrix R t It indicates the noise generated by the sensor during measurement, which is usually related to the sensor accuracy; Step S302: Establish a state equation based on the physical characteristics of groundwater level changes. The specific formula of the state equation is: , Among them, h t represents the predicted value of groundwater level at time t, h t-1 represents the predicted value of groundwater level at time t-1, F represents the state transfer matrix, B represents the control matrix, u t Indicates the control quantity; The prediction covariance matrix formula is: , P t represents the covariance matrix of the predicted groundwater level at time t, P t-1 represents the covariance matrix of the groundwater level predicted at time t-1, and T represents the transpose of the matrix; In this embodiment, the state transfer matrix F describes the relationship between the groundwater level and time, which can be approximately set to the unit matrix. The unit matrix is ​​a matrix with 1 on the main diagonal and 0 on the rest. The control matrix B and the control variable u t In the absence of external control variables, it can be ignored, that is, 0. The groundwater level covariance matrix P predicted at time t is t This reflects the uncertainty of the predicted groundwater level.

[0024] Step S303: Using the measurement equation to merge the calculated height value of the groundwater level with the predicted height value. The specific formula of the measurement equation is: , Among them, H represents the observation matrix, which is usually the unit matrix. The formula of the observation residual is: ,y t Represents the observed residual, and the covariance formula of the observed residual is: , S t represents the covariance of the observed residuals; Step S304: Introduce an adaptive feedback control mechanism to analyze the mean square error of the observation residual in real time, evaluate the noise level and dynamically adjust the noise covariance. The specific formula is: , , in, represents the dynamically adjusted process noise covariance matrix, represents the measurement noise covariance matrix after dynamic adjustment, and Represents the adaptive adjustment coefficient; Step S304 realizes the function of adjusting the covariance according to the noise changes in the real-time measurement environment, thereby further improving the filtering accuracy. and Obtained by personnel in this field through a large number of experiments or based on actual conditions; Step S305: Based on the gain, the observation residual is weighted and combined with the predicted value of the groundwater level to obtain the optimal estimated value of the groundwater level. The specific formula is: , in, represents the optimal estimated value of the groundwater level at time t, that is, the groundwater level value after filtering and calibration, K t Represents the gain coefficient, and the prediction covariance matrix is ​​updated while calculating the optimal estimated value of the groundwater level. The specific formula for the update is: , represents the updated prediction covariance matrix, and I represents the identity matrix.

[0025] Gain coefficient K t The specific formula is: , The inverse matrix of the observed residual covariance is the relative weight of the observation noise and the prediction noise. The principle of the gain coefficient is to dynamically adjust the weights of the predicted value and the observed value by minimizing the mean square value of the error. The larger the gain, the greater the influence of the observation value on the estimate; the smaller the gain, the greater the influence of the predicted value on the estimate.

[0026] The principle of the optimal estimate is: using the idea of ​​Bayesian updating, combining the predicted distribution and the observed distribution to obtain a new posterior distribution, thereby obtaining the best estimate at the current moment. In the process of continuous updating, the gain K will be adaptively adjusted. t , dynamically balances the weights of predicted values ​​and observed values, allowing the system to gradually approach the true state in the noise and uncertainty. This update method ensures that the filtering process has the optimal estimate of the system state;

[0027] The essence of this formula is to approach the true state of the system by gradually correcting the predicted value. Whenever new observation data is input, the predicted value is adjusted slightly using the observation residual. As time goes by, this small adjustment will make the estimated value continue to approach the true value and achieve the best estimation effect.

[0028] Step S4: Modeling and analyzing the stored historical groundwater level data to predict future groundwater level change trends; The specific steps of step S4 are: Step S401: Set the groundwater level historical data set as X and predict the short-term groundwater level. The specific formula is: , in, It represents the predicted value of groundwater level at time t1, c represents the constant term, which is the average level of water level data. If the water level value has a certain constant trend as a whole, c is used to describe the benchmark level of the trend. Represents the autoregressive coefficient, that is, the influence of the water level values ​​at the past Is moments on the current water level value. These coefficients are fitted through historical data and reflect the dependence of past water level changes on the current water level. represents the groundwater level at time t1-i in the past, Is and J represent the total short-term time in the past, Represents random error, or white noise, which reflects random fluctuations that cannot be explained by the model and is a random variable of the current moment forecast error. Represents the random error terms at time t1-j in the past. These error terms can reflect the noise components in the water level data. The model adjusts the current water level value through the sliding average of these error terms. Represents the sliding average coefficient, which describes the error term at the past t1-j moment The impact on the current water level value is also estimated from historical data, describing the contribution of past errors to the current forecast; This formula is based on the principle of time series analysis, which considers the short-term change of groundwater level as the combined effect of the autoregression of the current water level value and the sliding average of past errors. :This part expresses the current value of the water level as a linear combination of the water level values ​​at previous moments, and quantifies the impact of past water level changes on the current water level through the autoregressive coefficient. It captures the trend of water level changes. For example, if the water level has continued to rise in the past, the current water level may also continue to rise; the sliding average part : The sliding average part uses a linear combination of past error terms to correct the current water level value. By introducing past random error terms and their coefficients, this part can offset the influence of random fluctuations or noise in the data, thereby improving the accuracy of model predictions. This part is also used to describe non-persistent random fluctuations or periodic errors, and can take into account the impact of past prediction deviations on current prediction results.

[0029] Step S402: Use the LSTM network to predict the long-term groundwater level. The memory gating mechanism within the LSTM network maintains the long-term dependency of the data and obtains the predicted groundwater level at time t2. ; In this embodiment, t1=t2, but different methods are used when predicting the same time. Short-term consideration refers to the analysis within a short period of time, while long-term consideration refers to the analysis within a long period of time. Step S403: Predict the groundwater level at time t1 and the predicted groundwater level at time t2 Perform adaptive weighted fusion, the specific formula is: , in, It represents the final predicted value of groundwater level at time t1, or the final predicted value of groundwater level at time t2. represents the weight coefficient; In this embodiment, the weight coefficient Dynamic adjustment based on the historical error of the model. The specific formula is: , where e represents a natural constant, E dq represents the error of short-term groundwater level prediction, E cq represents the error in long-term groundwater level prediction, i.e., the error in LSTM network prediction; Step S404: Use the obtained final predicted value of groundwater level to perform anomaly detection and trend analysis, and calculate the trend change rate of the groundwater level prediction result. The specific formula is: , in, Indicates the trend change rate of the groundwater level prediction results, represents the final predicted value of groundwater level at time z, represents the final predicted value of groundwater level at time z-1, Represents the absolute value function.

[0030] In this embodiment, short-term prediction based on time series combines autoregressive and sliding average components. By adjusting the autoregressive and sliding average terms, it is possible to capture short-term fluctuation patterns in groundwater levels. LSTM captures long-term dependencies through memory cells and gating mechanisms, making it suitable for processing long-term trend changes. LSTM retains early trend information in its deep structure, thereby providing a better description of the long-term fluctuation trend of groundwater levels.

[0031] Weighted fusion dynamically adjusts the influence of short-term and long-term models through weights to ensure the accuracy of prediction. Logistic regression formula Error feedback is used to enable the model to adaptively adjust weights when there is a large difference in prediction accuracy, thereby improving the robustness of the prediction results. Finally, the trend change rate formula can be used to effectively detect abnormal fluctuations in groundwater levels for rapid response.

[0032] Step S5: The final estimated groundwater level data and the predicted groundwater level change trend are output to the monitoring center through the data interface, and stored and monitored in real time. A safety threshold is set. Once the water level exceeds the threshold range, the system will automatically trigger an early warning and notify relevant management personnel.

[0033] Set the groundwater level anomaly threshold. If the trend change rate of the groundwater level prediction result is When the groundwater level exceeds the abnormal threshold, an abnormal alarm is triggered.

[0034] Example 2 See also Figure 4 , another embodiment provided by the present invention: a groundwater level monitoring system, comprising: an initial setting module, a water level calculation and correction module, a filter calibration module, a water level trend analysis module and a real-time transmission monitoring module; The initial setting module is used to place pressure sensors and temperature sensors in the groundwater to be monitored, initialize the sensors and calibrate them; The water level calculation and correction module is used to collect pressure and temperature data in the groundwater well in real time, calculate the height of the groundwater level based on the pressure and temperature data, and perform correction; The filter calibration module is used to perform filter calibration on the corrected groundwater level to eliminate environmental noise and sensor errors; The water level trend analysis module is used to model and analyze the stored historical groundwater level data and predict future groundwater level change trends; The real-time transmission and monitoring module is used to transmit the final estimated groundwater level data and the predicted groundwater level change trend, and to store and monitor them in real time.

[0035] The filter calibration module includes: a state equation unit, a measurement equation unit and an optimal estimation unit; The state equation unit is used to establish a state equation based on the physical characteristics of groundwater level changes to predict the groundwater level at time t; The measurement equation unit is used to fuse the calculated height value and the predicted height value of the groundwater level using the measurement equation; The optimal estimation unit is used to combine the weighted observation residual with the predicted value of the groundwater level according to the gain to obtain the optimal estimated value of the groundwater level.

[0036] The water level trend analysis module includes: a short-term prediction unit, a long-term prediction unit, a final fusion unit and a water level trend analysis unit; The short-term prediction unit is used to predict the short-term water level of groundwater; The long-term prediction unit is used to predict the long-term groundwater level using an LSTM network; The final fusion unit is used to perform adaptive weighted fusion on the groundwater level prediction value at time t1 and the groundwater level prediction value at time t2; The water level trend analysis unit is used to use the obtained final predicted value of the groundwater level to perform anomaly detection and trend analysis, and calculate the trend change rate of the groundwater level prediction result.

[0037] In addition, the parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0038] The above-described specific embodiments further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is merely a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for monitoring groundwater level, characterized in that: include: Place pressure sensors and temperature sensors in the groundwater to be monitored, initialize the sensors and calibrate them; Collect the pressure and temperature data in the groundwater well in real time, calculate the height of the groundwater level based on the pressure and temperature data and make corrections; Perform filter calibration on the corrected groundwater level to eliminate environmental noise and sensor errors; Model and analyze the stored historical groundwater level data to predict future groundwater level change trends; The final estimated groundwater level data and predicted groundwater level change trends are transmitted, stored and monitored in real time.

2. A method for monitoring groundwater level according to claim 1, characterized in that: The real-time collection of pressure and temperature data in the groundwater well, calculation of the height of the groundwater level based on the pressure and temperature data and correction thereof include: Assume that the pressure data in the groundwater well collected at time t is P(t), and the temperature data in the groundwater well is T(t); According to the pressure data P(t) in the groundwater well at time t, the height of the groundwater level at time t is calculated. The specific formula is: ， in, represents the height of the groundwater level at time t, P0 represents the standard atmospheric pressure, It represents the density of water under standard conditions, and g represents the acceleration due to gravity; According to the temperature data T(t) in the groundwater well at time t, the height of the groundwater level at time t is corrected. The specific formula is: ， in, represents the density of water at temperature T(t), represents the volume expansion coefficient of water, T0 represents the temperature under standard environment; According to the density of water at temperature T(t) The height of the groundwater level at time t Correction is performed to obtain the corrected groundwater level h(t) at time t.

3. A method for monitoring groundwater level according to claim 2, characterized in that: The filtering and calibration process is performed on the corrected groundwater level to eliminate environmental noise and sensor errors, including: The process noise in calculating the groundwater level is set to w t , the measurement noise is v t , process noise w t and measurement noise v t The covariance matrices are Q t and R t ; The state equation is established according to the physical characteristics of groundwater level changes. The specific formula of the state equation is: ， Among them, h t represents the predicted value of groundwater level at time t, h t-1 represents the predicted value of groundwater level at time t-1, F represents the state transfer matrix, B represents the control matrix, u t Indicates the control quantity; The prediction covariance matrix formula is: , P t represents the covariance matrix of the predicted groundwater level at time t, P t-1 represents the covariance matrix of the groundwater level predicted at time t-1, and T represents the transpose of the matrix; The calculated height value of the groundwater level is combined with the predicted height value using the measurement equation. The specific formula of the measurement equation is: ， Among them, H represents the observation matrix, and the formula of the observation residual is: ,y t Represents the observed residual, and the covariance formula of the observed residual is: , S t represents the covariance of the observed residuals.

4. A method for monitoring groundwater level according to claim 3, characterized in that: The filtering and calibration process of the corrected groundwater level to eliminate environmental noise and sensor errors also includes: An adaptive feedback control mechanism is introduced to analyze the mean square error of the observation residual in real time, evaluate the noise level and dynamically adjust the noise covariance. The specific formula is: ， ， in, represents the dynamically adjusted process noise covariance matrix, represents the dynamically adjusted measurement noise covariance matrix, and Represents the adaptive adjustment coefficient; According to the gain, the weighted observation residual is combined with the predicted value of the groundwater level to obtain the optimal estimate of the groundwater level. The specific formula is: ， in, represents the optimal estimated value of the groundwater level at time t, that is, the groundwater level value after filtering and calibration, K t Represents the gain coefficient, calculates the optimal estimate of the groundwater level and updates the prediction covariance matrix at the same time. The specific formula for the update is: , represents the updated prediction covariance matrix, and I represents the identity matrix.

5. A method for monitoring groundwater level according to claim 1, characterized in that: The modeling and analysis of the stored historical groundwater level data to predict future groundwater level change trends include: Assuming the groundwater level historical data set as X, the short-term groundwater level is predicted. The specific formula is: ， in, represents the predicted value of groundwater level at time t1, c represents the constant term, represents the autoregressive coefficient, represents the groundwater level at the past time t1-i, Is and J represent the total short-term time in the past, represents random error, represents the random error term at the past time t1-j, represents the sliding mean coefficient; The LSTM network is used to predict the long-term groundwater level. The memory gating mechanism inside the LSTM network maintains the long-term dependency of the data and obtains the predicted groundwater level at time t2. ; Prediction of groundwater level at time t1 And the predicted value of groundwater level at time t2 Perform adaptive weighted fusion, the specific formula is: ， in, represents the final predicted value of groundwater level at time t1, represents the weight coefficient; The final predicted value of groundwater level is used for anomaly detection and trend analysis, and the trend change rate of groundwater level prediction result is calculated. The specific formula is: ， in, Indicates the trend change rate of the groundwater level prediction results, represents the final predicted value of groundwater level at time z, represents the final predicted value of groundwater level at time z-1, Represents the absolute value function.

6. A method for monitoring groundwater level according to claim 5, characterized in that: The final predicted value of the groundwater level at time t1 , which also represents the final predicted value of the groundwater level at time t2.

7. A method for monitoring groundwater level according to claim 1, characterized in that: The pressure sensor and the temperature sensor are transmitted using Bluetooth signals.

8. A groundwater level monitoring system, used to implement a groundwater level monitoring method according to any one of claims 1 to 7, characterized in that: include: Initial setting module, water level calculation and correction module, filter calibration module, water level trend analysis module and real-time transmission monitoring module; The initial setting module is used to place pressure sensors and temperature sensors in the groundwater to be monitored, initialize the sensors and perform calibration; The water level calculation and correction module is used to collect the pressure and temperature data in the groundwater well in real time, calculate the height of the groundwater level according to the pressure and temperature data, and perform correction; The filter calibration module is used to perform filter calibration processing on the corrected groundwater level to eliminate environmental noise and sensor errors; The water level trend analysis module is used to model and analyze the stored historical groundwater level data and predict future groundwater level change trends; The real-time transmission monitoring module is used to transmit the final estimated groundwater level data and the predicted groundwater level change trend, and store and monitor them in real time.

9. A groundwater level monitoring system as claimed in claim 8, characterized in that: The filter calibration module includes: a state equation unit, a measurement equation unit and an optimal estimation unit; The state equation unit is used to establish a state equation according to the physical characteristics of the groundwater level change to predict the groundwater level at time t; The measurement equation unit is used to merge the calculated height value of the groundwater level with the predicted height value using the measurement equation; The optimal estimation unit is used to combine the weighted observation residual with the predicted value of the groundwater level according to the gain to obtain the optimal estimation value of the groundwater level.

10. A groundwater level monitoring system as claimed in claim 9, characterized in that: The water level trend analysis module includes: a short-term prediction unit, a long-term prediction unit, a final fusion unit and a water level trend analysis unit; The short-term prediction unit is used to predict the short-term water level of groundwater; The long-term prediction unit is used to predict the long-term water level of groundwater using an LSTM network; The final fusion unit is used to perform adaptive weighted fusion on the groundwater level prediction value at time t1 and the groundwater level prediction value at time t2; The water level trend analysis unit is used to use the obtained final prediction value of the groundwater level to perform anomaly detection and trend analysis, and calculate the trend change rate of the groundwater level prediction result.

Images