A method and system for real-time monitoring of lake water quality based on sensor array

Abstract

This invention relates to a method and system for real-time monitoring of lake water quality based on a sensor array. The method includes: deploying a sensor array in a target lake and setting data acquisition frequency parameters; acquiring the operating parameters of the sensor array in real time, analyzing the operating parameters to determine the sensor health status, and adaptively adjusting the sensors based on the sensor health status; acquiring water quality monitoring data in real time based on the adjusted sensor array, preprocessing the water quality monitoring data and performing quality control verification to construct a qualified monitoring dataset; comparing each water quality indicator in the qualified monitoring dataset with preset environmental standard limits, marking events that exceed the limits and issuing early warnings, thus completing the real-time monitoring of the target lake water quality. This invention realizes intelligent real-time monitoring, anomaly identification, data quality control, and early warning response for lake water quality, providing comprehensive technical support for water environment management.

Description

Technical Field

[0001] This invention relates to the field of water quality monitoring technology, and in particular to a method and system for real-time monitoring of lake water quality based on a sensor array. Background Technology

[0002] Lakes, as vital freshwater resources and ecosystem carriers, directly impact regional ecological security, drinking water supply, and sustainable economic and social development. With accelerated industrialization and intensified environmental pollution, the deterioration of lake water quality has become increasingly prominent, making the establishment of an efficient and accurate water quality monitoring system an urgent need for environmental protection and water resource management. Traditional lake water quality monitoring relies primarily on manual sampling and laboratory analysis, which suffers from significant drawbacks such as low monitoring frequency, insufficient spatial coverage, and poor data timeliness. While existing automated monitoring equipment can provide continuous data, it generally suffers from sparse monitoring points, weak data quality control, and low system reliability, making it difficult to meet the practical needs of real-time monitoring of water quality in large-scale lakes.

[0003] The core challenge of real-time lake water quality monitoring stems from the dynamic complexity of the monitoring data. Lake water quality parameters exhibit high variability in their spatiotemporal distribution, and conventional fixed-frequency data acquisition methods cannot effectively capture abrupt water quality changes, leading to the omission of critical pollution processes. This inadequacy in data acquisition further raises technical challenges in data quality control. Sensors are prone to generating abnormal readings and noise interference in complex aquatic environments, and the lack of effective real-time data screening and quality control mechanisms makes it difficult for the monitoring system to distinguish between genuine water quality changes and equipment malfunctions or environmental disturbances. The accumulation of data quality problems directly affects the long-term stable operation of the monitoring system. When equipment anomalies cannot be identified and addressed in a timely manner, the continuity and reliability of monitoring work are severely threatened, ultimately leading to the risk of failure of the entire monitoring network. Therefore, this invention proposes a real-time lake water quality monitoring method and system based on a sensor array. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for real-time monitoring of lake water quality based on a sensor array, which enables adaptive data acquisition, real-time data quality control, and ensures continuous and stable monitoring operation.

[0005] To achieve the above objectives, in one aspect, the present invention provides a method for real-time monitoring of lake water quality based on a sensor array, comprising:

[0006] Deploy a sensor array at the target lake and set the data acquisition frequency parameters;

[0007] The operating parameters of the sensor array are collected in real time, analyzed to determine the health status of the sensors, and the sensors are adaptively adjusted based on the health status of the sensors.

[0008] Based on the adjusted sensor array, water quality monitoring data is collected in real time. The water quality monitoring data is preprocessed and quality control verification is performed to construct a monitoring dataset that meets quality control requirements.

[0009] By comparing each water quality indicator in the qualified monitoring dataset with the preset environmental standard limits, marking events that exceed the limits and issuing early warnings, the real-time water quality monitoring of the target lake is completed.

[0010] Optionally, deploying a sensor array in the target lake and setting data acquisition frequency parameters includes:

[0011] The sensor array is deployed based on the area, shape, water depth, and water flow conditions of the target lake;

[0012] A multi-parameter time series model was established based on the historical water quality monitoring data of the target lake, and the spatiotemporal variation patterns of water quality indicators were analyzed based on the parameter time series model.

[0013] Based on the aforementioned spatiotemporal variation patterns, an adaptive Kalman filter algorithm is used to calculate the optimal sampling interval and optimal sampling frequency for each monitoring point.

[0014] Optionally, after obtaining the optimal sampling interval and optimal sampling frequency, the method further includes real-time adjustment of the optimal sampling interval and optimal sampling frequency based on changes in water quality parameters at the monitoring points. This real-time adjustment includes:

[0015] If the rate of change of water quality parameters at any monitoring point exceeds a preset threshold, a high-frequency acquisition mode is triggered, and the sampling frequency of the monitoring point is adjusted.

[0016] By updating the multi-parameter time series model with the adjusted sampling frequency, the spatiotemporal variation patterns of each water quality indicator were re-analyzed, and the adaptive Kalman filter algorithm was re-applied to optimize the sampling interval and sampling frequency of each monitoring point.

[0017] Optionally, the operating parameters of the sensor array are collected in real time, analyzed to determine the sensor health status, and the sensors are adaptively adjusted based on the sensor health status, including:

[0018] The operating parameters of each sensor in the sensor array are monitored in real time. If any operating indicator deviates from the preset operating range, a fault warning signal is generated and the health status of the sensor is assessed.

[0019] The remaining lifespan and optimal calibration time of the sensor are calculated based on the sensor health status assessment results, and the failure probability is predicted by analyzing the performance degradation trend curve. If the failure probability exceeds the preset probability value, a backup sensor is switched on.

[0020] Optionally, the water quality monitoring data is preprocessed and quality control verified to construct a monitoring dataset that meets quality control standards, including:

[0021] Multi-sensor data fusion technology is used to cross-validate multiple monitoring data of the same water quality parameter and mark abnormal data.

[0022] A wavelet denoising algorithm was used to filter out noise from the abnormal data, and the water quality monitoring data after noise removal was obtained.

[0023] By establishing a correlation matrix of water quality parameters, the intrinsic relationship between different water quality indicators is analyzed. Then, a multiple linear regression model is used to verify the quality control of the noise-filtered data, and a monitoring dataset with qualified quality control is obtained.

[0024] Optionally, a wavelet denoising algorithm is used to filter out noise from the abnormal data to obtain noise-filtered water quality monitoring data, including:

[0025] The abnormal data is decomposed into low-frequency trend components and high-frequency noise components by performing a wavelet denoising algorithm.

[0026] Calculate the energy percentage of the high-frequency noise component. If the energy percentage of the high-frequency noise component exceeds the preset percentage, then use a soft threshold denoising algorithm to process the high-frequency noise component and obtain the denoised high-frequency noise component.

[0027] The high-frequency noise components and low-frequency trend components after denoising are reconstructed by inverse wavelet transform to obtain denoised water quality monitoring data.

[0028] Optionally, the data after noise filtering is subjected to quality control verification, including:

[0029] A multiple linear regression model is used to fit the denoised water quality monitoring data to obtain the sum of squares of the regression residuals. If the sum of squares of the regression residuals exceeds the preset control limit, data re-acquisition is triggered to obtain new water quality monitoring parameters and perform preprocessing and quality control verification until the sum of squares of the regression residuals meets the preset control limit, and the quality control qualified monitoring dataset is obtained.

[0030] On the other hand, the present invention also provides a real-time lake water quality monitoring system based on a sensor array, comprising:

[0031] The sensor deployment module is used to deploy a sensor array in the target lake and set the data acquisition frequency parameters.

[0032] The sensor adjustment module is used to collect the operating parameters of the sensor array in real time, analyze the operating parameters, determine the health status of the sensors, and adaptively adjust the sensors based on the health status of the sensors.

[0033] The water quality acquisition module is used to acquire water quality monitoring data in real time based on the adjusted sensor array, preprocess the water quality monitoring data and perform quality control verification to construct a monitoring dataset that meets quality control requirements.

[0034] The water quality monitoring module is used to compare the water quality indicators in the monitoring data that have passed quality control with the preset environmental standard limits, mark the events that exceed the limits and issue early warnings, and complete the real-time monitoring of the water quality of the target lake.

[0035] The beneficial effects of this invention are as follows:

[0036] This invention analyzes the spatiotemporal variation patterns of key water quality indicators and dynamically adjusts the sampling frequency using an adaptive Kalman filter algorithm. It utilizes multi-sensor data fusion technology for cross-validation and combines it with a sliding window anomaly detection algorithm to identify abnormal data. The abnormal data undergoes wavelet denoising processing and quality control verification. Simultaneously, this invention monitors the sensor's operating status in real time, predicts equipment lifespan and calibration time, and ensures stable monitoring operation. This invention achieves intelligent real-time monitoring, anomaly identification, data quality control, and early warning response for lake water quality, providing comprehensive technical support for water environment management. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart of a real-time lake water quality monitoring method based on a sensor array, according to an embodiment of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] On the one hand, this embodiment provides a method for real-time monitoring of lake water quality based on a sensor array, such as... Figure 1 As shown, it includes:

[0042] Deploy a sensor array at the target lake and set the data acquisition frequency parameters;

[0043] The operating parameters of the sensor array are collected in real time, analyzed to determine the health status of the sensors, and the sensors are adaptively adjusted based on the health status of the sensors.

[0044] Based on the adjusted sensor array, water quality monitoring data is collected in real time. The water quality monitoring data is preprocessed and quality control verification is performed to construct a monitoring dataset that meets quality control requirements.

[0045] By comparing each water quality indicator in the qualified monitoring dataset with the preset environmental standard limits, marking events that exceed the limits and issuing early warnings, the real-time water quality monitoring of the target lake is completed.

[0046] Specifically, this embodiment analyzes the spatiotemporal variation patterns of key water quality indicators and dynamically adjusts the sampling frequency using an adaptive Kalman filter algorithm. It utilizes multi-sensor data fusion technology for cross-validation and combines it with a sliding window anomaly detection algorithm to identify abnormal data. The abnormal data undergoes wavelet denoising processing and quality control verification. Simultaneously, this embodiment monitors the sensor's operating status in real time, predicts equipment lifespan and calibration time, and ensures stable operation of the monitoring work. This embodiment achieves intelligent real-time monitoring, anomaly identification, data quality control, and early warning response for lake water quality, providing comprehensive technical support for water environment management.

[0047] Furthermore, deploying a sensor array in the target lake and setting data acquisition frequency parameters includes:

[0048] The sensor array is deployed based on the area, shape, water depth, and water flow conditions of the target lake;

[0049] A multi-parameter time series model was established based on the historical water quality monitoring data of the target lake, and the spatiotemporal variation patterns of water quality indicators were analyzed based on the parameter time series model.

[0050] Based on the aforementioned spatiotemporal variation patterns, an adaptive Kalman filter algorithm is used to calculate the optimal sampling interval and optimal sampling frequency for each monitoring point.

[0051] The method further includes, after obtaining the optimal sampling interval and optimal sampling frequency, adjusting the optimal sampling interval and optimal sampling frequency in real time based on the changes in water quality parameters at the monitoring points. This real-time adjustment includes:

[0052] If the rate of change of water quality parameters at any monitoring point exceeds a preset threshold, a high-frequency acquisition mode is triggered, and the sampling frequency of the monitoring point is adjusted.

[0053] By updating the multi-parameter time series model with the adjusted sampling frequency, the spatiotemporal variation patterns of each water quality indicator were re-analyzed, and the adaptive Kalman filter algorithm was re-applied to optimize the sampling interval and sampling frequency of each monitoring point.

[0054] Specifically, in this embodiment, the sensor array is rationally arranged according to the area, shape, water depth and water flow of the lake. The sensor array includes various types of water quality sensors, such as pH sensors, dissolved oxygen sensors, conductivity sensors, turbidity sensors, ammonia nitrogen sensors, total phosphorus sensors, and total nitrogen sensors.

[0055] This embodiment assumes that five monitoring points are set up in the target lake, and dissolved oxygen (DO, mg / L), total phosphorus (TP, mg / L), total nitrogen (TN, mg / L), and turbidity (NTU) data are collected monthly at each point, spanning from January 2023 to May 2025, a total of 29 months. First, missing values ​​are filled using linear interpolation. For example, if DO is missing at a certain point in March 2023, it is interpolated to 8.6 based on data from February (8.5) and April (8.7). Next, a multi-parameter time series model is constructed, using the ARIMA model. Each parameter is fitted separately; for example, the DO sequence is stationary after the ADF test, so the ARIMA(1,0,1) model is determined. After parameter estimation, the trend for the next three months is predicted. In the spatiotemporal variation analysis, the monthly rate of change of each parameter is calculated. For example, if the TP at a certain point increases from 0.03 to 0.035, the rate of change is (0.035-0.03) / 0.03 = 16.67%, exceeding the threshold of 0.1, indicating significant water quality fluctuations. Spatial analysis uses Kriging interpolation to generate spatial distribution maps of each parameter, revealing that the TP concentration is higher in the southeastern part of the lake (0.04 mg / L). An adaptive Kalman filter algorithm is used to optimize the sampling interval. The initial state noise covariance Q = 0.01 and the observation noise covariance R = 0.1 are used to iteratively update the state vector (parameter values ​​and rates of change). If the DO rate of change at a certain point reaches 0.12 (>0.1), the optimal sampling interval is adjusted from 30 days to 7 days, triggering a high-frequency acquisition mode. For other points with rates of change below the threshold, the sampling period remains at 30 days.

[0056] Furthermore, the operating parameters of the sensor array are collected in real time, analyzed to determine the sensor health status, and adaptively adjusted based on the sensor health status, including:

[0057] The operating parameters of each sensor in the sensor array are monitored in real time. If any operating indicator deviates from the preset operating range, a fault warning signal is generated and the health status of the sensor is assessed.

[0058] The remaining lifespan and optimal calibration time of the sensor are calculated based on the sensor health status assessment results, and the failure probability is predicted by analyzing the performance degradation trend curve. If the failure probability exceeds the preset probability value, a backup sensor is switched on.

[0059] Specifically, this embodiment first acquires sensor output data at a sampling frequency of 100Hz. Then, it calculates the response time, defined as the time from receiving the excitation to outputting a stable signal. The algorithm uses a sliding window method with a window size of 50ms, detecting the time it takes for the signal to move from 0 to 90% of full scale. For example, a calculated response time of 12ms falls within the normal range of 10-15ms, indicating a normal response. Signal stability analysis calculates the standard deviation of the signal. Assuming the standard deviation of 1000 collected data points is 0.02V, the normal range is 0-0.05V, indicating signal stability. Temperature drift assessment uses data from the built-in temperature sensor. Assuming the current temperature is 25℃, historical data is fitted to a temperature-output curve. Each 1℃ drift causes a 0.01V deviation; the current deviation is 0.008V, below the threshold of 0.02V, indicating normal temperature drift. If any indicator is abnormal, such as a response time exceeding 15ms, a fault warning signal is triggered. The health status assessment uses a weighted scoring method, with response time, stability, and temperature drift weighted at 0.4, 0.3, and 0.3 respectively. The comprehensive score is calculated, for example, (12 / 15)*0.4+(0.02 / 0.05)*0.3+(0.008 / 0.02)*0.3=0.84. A threshold of 0.8 or above is set as a healthy state, and the result is "equipment normal".

[0060] This embodiment constructs a performance degradation model using multi-dimensional sensor data (such as temperature 50℃, vibration frequency 10Hz, and current fluctuation 0.5A). A Long Short-Term Memory (LSTM) neural network algorithm is used to analyze time-series data and predict the remaining lifespan of the equipment. The LSTM model takes nearly 30 days of sensor data as input, containing 1000 sampling points, and outputs a predicted remaining lifespan of 1500 hours with an error range of ±50 hours. The model is optimized using mean squared error (MSE), with the MSE value controlled below 0.02. Next, based on the degradation trend curve (fitted as a quadratic function y = 0.001x),... 2The system calculates the optimal calibration time point using a formula of +0.05x+1 (where x is the number of operating hours), sets a performance threshold of 80% initial efficiency, and predicts that the device will drop to this threshold after 1200 hours. The calibration task is automatically scheduled to start at hour 1150, sending a command to adjust the sensor sensitivity to the standard value ±0.1%. Simultaneously, the system monitors the failure probability in real time, using a Bayesian network combined with historical failure data (100 failure samples, featuring characteristics including temperature, vibration, and current). When the predicted failure probability exceeds 80% (e.g., reaching 85%), a backup sensor switching procedure is triggered, sending a switching command. The backup sensor takes over within 5 seconds, and performance is reassessed after the switch to ensure output stability above 95%.

[0061] Furthermore, the water quality monitoring data is preprocessed and quality control verified to construct a qualified monitoring dataset, including:

[0062] Multi-sensor data fusion technology is used to cross-validate multiple monitoring data of the same water quality parameter and mark abnormal data.

[0063] A wavelet denoising algorithm was used to filter out noise from the abnormal data, and the water quality monitoring data after noise removal was obtained.

[0064] By establishing a correlation matrix of water quality parameters, the intrinsic relationship between different water quality indicators is analyzed. Then, a multiple linear regression model is used to verify the quality control of the noise-filtered data, and a monitoring dataset with qualified quality control is obtained.

[0065] Among them, wavelet denoising algorithm is used to filter out noise from abnormal data, and water quality monitoring data after noise filtering is obtained, including:

[0066] The abnormal data is decomposed into low-frequency trend components and high-frequency noise components by performing a wavelet denoising algorithm.

[0067] Calculate the energy percentage of the high-frequency noise component. If the energy percentage of the high-frequency noise component exceeds the preset percentage, then use a soft threshold denoising algorithm to process the high-frequency noise component and obtain the denoised high-frequency noise component.

[0068] The high-frequency noise components and low-frequency trend components after denoising are reconstructed by inverse wavelet transform to obtain denoised water quality monitoring data.

[0069] The quality control verification of the noise-filtered data includes:

[0070] A multiple linear regression model is used to fit the denoised water quality monitoring data to obtain the sum of squares of the regression residuals. If the sum of squares of the regression residuals exceeds the preset control limit, data re-acquisition is triggered to obtain new water quality monitoring parameters and perform preprocessing and quality control verification until the sum of squares of the regression residuals meets the preset control limit, and the quality control qualified monitoring dataset is obtained.

[0071] Specifically, this embodiment employs multi-sensor data fusion technology to cross-validate multiple measurement signals for the same water quality parameter. For example, using three turbidity sensors to simultaneously measure the same water sample, the readings obtained are 48 NTU, 52 NTU, and 49 NTU, respectively. A weighted average algorithm (weights set according to the historical calibration accuracy of the sensors, such as 0.4, 0.3, and 0.3) is used to calculate the fused mean, which is 49.7 NTU. The fusion process uses Kalman filtering to optimize signal noise, reducing random error by approximately 10%. If the deviation of a single sensor reading from the fused mean exceeds 15%, it is marked as a suspected anomaly. For example, if a sensor reading is 60 NTU, the deviation is (60-49.7) / 49.7≈20.7%, exceeding 15%, and is therefore marked as an anomaly. An anomaly detection algorithm is used to further confirm the anomaly by combining the historical data mean and standard deviation (e.g., historical mean 50 NTU, standard deviation 2 NTU), generating a monitoring dataset with anomaly markers.

[0072] In this embodiment, the wavelet basis function is Daubechies(db4), and the decomposition level is set to 4. The abnormal signal is decomposed into a low-frequency trend component (A4) and high-frequency noise components (D1-D4) using Discrete Wavelet Transform (DWT). The energy of each component is calculated. It is assumed that the total energy of the high-frequency components D1-D4 is 1200, accounting for 30% of the total signal energy of 4000 (1200 / 4000 = 0.3). Since the high-frequency component energy accounts for 30%, the soft-threshold denoising condition is met. Soft-threshold processing is applied to each of the high-frequency components D1-D4, while the low-frequency component A4 remains unchanged. After reconstructing the signal, the denoised data sequence is obtained.

[0073] This embodiment calculates the correlation matrix between water quality parameters such as pH, dissolved oxygen (DO), chemical oxygen demand (COD), and total phosphorus (TP) using the Pearson correlation coefficient. Assuming the dataset contains 1000 records, each record includes pH (range 6.5-8.5), DO (2-10 mg / L), COD (10-50 mg / L), and TP (0.1-1.0 mg / L), the correlation coefficient is calculated using the `corr()` function from Python's pandas library, resulting in a 4×4 matrix.

[0074] For example, the correlation coefficient between pH and DO is 0.75, indicating a strong positive correlation; the correlation coefficient between COD and TP is 0.62, showing a moderate positive correlation. The matrix was visualized as a heatmap using the `heatmap` function from the `seaborn` library, with color intensity reflecting the strength of the correlation. Subsequently, a multiple linear regression model was constructed, with pH as the dependent variable and DO, COD, and TP as independent variables. The `LinearRegression` module of scikit-learn was used to fit the model. After splitting the training set (80%) and the test set (20%), the model's R-value was [value missing]. 2 The value is 0.88, and the mean square error is 0.15. Calculate the regression residual sum of squares (RSS). If the RSS exceeds the preset control limit (e.g., 0.20), trigger data re-acquisition, automatically collect new data (sampling frequency 1 time / hour), and repeat the above steps until the RSS is below 0.20, obtaining a monitoring dataset that meets quality control standards.

[0075] Furthermore, this embodiment compares water quality indicators in the qualified quality control dataset using a preset environmental standard database. If a water quality indicator exceeds the environmental standard limit, it is marked as a limit exceedance event. Based on the severity of the limit exceedance event, a decision tree algorithm is used to classify the warning level and obtain multi-level warning signals. Through a preset response process configuration, corresponding response instructions are generated for the multi-level warning signals, and warning and response logs are recorded to update the monitoring status in real time.

[0076] On the other hand, this embodiment also provides a real-time lake water quality monitoring system based on a sensor array, including:

[0077] The sensor deployment module is used to deploy a sensor array in the target lake and set the data acquisition frequency parameters.

[0078] The sensor adjustment module is used to collect the operating parameters of the sensor array in real time, analyze the operating parameters, determine the health status of the sensors, and adaptively adjust the sensors based on the health status of the sensors.

[0079] The water quality acquisition module is used to acquire water quality monitoring data in real time based on the adjusted sensor array, preprocess the water quality monitoring data and perform quality control verification to construct a monitoring dataset that meets quality control requirements.

[0080] The water quality monitoring module is used to compare the water quality indicators in the monitoring data that have passed quality control with the preset environmental standard limits, mark the events that exceed the limits and issue early warnings, and complete the real-time monitoring of the water quality of the target lake.

[0081] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for real-time monitoring of lake water quality based on a sensor array, characterized by, include: Deploy a sensor array at the target lake and set the data acquisition frequency parameters; The operating parameters of the sensor array are collected in real time, analyzed to determine the health status of the sensors, and the sensors are adaptively adjusted based on the health status of the sensors. Based on the adjusted sensor array, water quality monitoring data is collected in real time. This data is then preprocessed and subjected to quality control verification to construct a qualified monitoring dataset, including: Multi-sensor data fusion technology is used to cross-validate multiple monitoring data of the same water quality parameter and mark abnormal data. A wavelet denoising algorithm was used to filter out noise from the abnormal data, and the water quality monitoring data after noise removal was obtained. By establishing a correlation matrix of water quality parameters, the intrinsic relationship between different water quality indicators is analyzed, and the data after noise filtering is verified by combining a multiple linear regression model to obtain a monitoring dataset that meets the quality control requirements. The quality control verification of the noise-filtered data includes: A multiple linear regression model is used to fit the denoised water quality monitoring data to obtain the sum of squares of the regression residuals. If the sum of squares of the regression residuals exceeds the preset control limit, the data is re-acquired to obtain new water quality monitoring parameters and perform preprocessing and quality control verification until the sum of squares of the regression residuals meets the preset control limit, and the quality control qualified monitoring dataset is obtained. The water quality indicators in the monitoring data that have passed quality control are compared with the preset environmental standard limits. Events that exceed the limits are marked and warnings are issued to complete the real-time monitoring of the target lake's water quality. Deploying a sensor array in the target lake and setting data acquisition frequency parameters includes: The sensor array is deployed based on the area, shape, water depth, and water flow conditions of the target lake; A multi-parameter time series model was established based on the historical water quality monitoring data of the target lake, and the spatiotemporal variation patterns of water quality indicators were analyzed based on the parameter time series model. Based on the aforementioned spatiotemporal variation patterns, an adaptive Kalman filter algorithm is used to calculate the optimal sampling interval and optimal sampling frequency for each monitoring point. After obtaining the optimal sampling interval and optimal sampling frequency, the method further includes real-time adjustment of the optimal sampling interval and optimal sampling frequency based on changes in water quality parameters at the monitoring points. This real-time adjustment includes: If the rate of change of water quality parameters at any monitoring point exceeds a preset threshold, a high-frequency acquisition mode is triggered, and the sampling frequency of the monitoring point is adjusted. By updating the multi-parameter time series model with the adjusted sampling frequency, the spatiotemporal variation patterns of each water quality indicator are re-analyzed, and the adaptive Kalman filter algorithm is re-applied to optimize the sampling interval and sampling frequency of each monitoring point. The system collects the operating parameters of the sensor array in real time, analyzes these parameters to determine the sensor health status, and adaptively adjusts the sensors based on the sensor health status, including: The operating parameters of each sensor in the sensor array are monitored in real time. If any operating indicator deviates from the preset operating range, a fault warning signal is generated and the health status of the sensor is assessed. The remaining lifespan and optimal calibration time of the sensor are calculated based on the sensor health status assessment results, and the failure probability is predicted by analyzing the performance degradation trend curve. If the failure probability exceeds the preset probability value, a backup sensor is switched on.

2. The method for real-time monitoring of lake water quality based on sensor array according to claim 1, characterized in that, Wavelet denoising algorithm is used to filter out noise from abnormal data, and water quality monitoring data after noise filtering is obtained, including: The abnormal data is decomposed into low-frequency trend components and high-frequency noise components by performing a wavelet denoising algorithm. Calculate the energy percentage of the high-frequency noise component. If the energy percentage of the high-frequency noise component exceeds the preset percentage, then use a soft threshold denoising algorithm to process the high-frequency noise component and obtain the denoised high-frequency noise component. The high-frequency noise components and low-frequency trend components after denoising are reconstructed by inverse wavelet transform to obtain denoised water quality monitoring data.

3. A real-time lake water quality monitoring system based on a sensor array, used to implement the real-time lake water quality monitoring method based on a sensor array as described in any one of claims 1-2, characterized in that, include: The sensor deployment module is used to deploy a sensor array in the target lake and set the data acquisition frequency parameters. The sensor adjustment module is used to collect the operating parameters of the sensor array in real time, analyze the operating parameters, determine the health status of the sensors, and adaptively adjust the sensors based on the health status of the sensors. The water quality acquisition module is used to acquire water quality monitoring data in real time based on the adjusted sensor array, preprocess the water quality monitoring data and perform quality control verification to construct a monitoring dataset that meets quality control requirements. The water quality monitoring module is used to compare the water quality indicators in the monitoring data that have passed quality control with the preset environmental standard limits, mark the events that exceed the limits and issue early warnings, and complete the real-time monitoring of the water quality of the target lake.

Images