Water quality on-line monitoring system and monitoring method thereof

Abstract

The application discloses a kind of water quality on-line monitoring system and monitoring method thereof, it is related to water quality monitoring technical field, it solves the problem that existing water quality on-line monitoring system only pays attention to monitoring accuracy and system energy consumption, ignores the quantitative relationship of sampling frequency and sensor life-span.The application includes sampling unit, multiple water quality sensors, control unit, data transmission unit and power supply unit, three-dimensional dynamic balance optimization module is built-in in control unit, establishes three independent optimization targets of monitoring accuracy, system energy consumption, sensor life-span, establishes life consumption ledger for each sensor and calculates optimal independent sampling frequency, it is also built-in water quality abnormal grading judgment, life consumption dynamic compensation and algorithm self-learning module.The application can effectively reduce system energy consumption, prolong sensor service life, improve the stability and adaptability of system operation under the premise of ensuring monitoring accuracy, and is suitable for continuous real-time on-line monitoring of various water bodies.

Description

Technical Field

[0001] This invention relates to the field of water quality monitoring technology, specifically to an online water quality monitoring system and its monitoring method. Background Technology

[0002] Water environment quality monitoring is a crucial foundation for water environment protection and water pollution prevention. Online water quality monitoring systems enable continuous, real-time, and automatic monitoring of various parameters in water bodies, providing timely and accurate data support for water environment management. With the continuous deepening of water environment protection efforts in my country, the application scope of online water quality monitoring systems is becoming increasingly widespread, playing a key role in monitoring various water bodies such as rivers, lakes, reservoirs, and sewage outlets.

[0003] Currently, online water quality monitoring systems generally adopt a fixed sampling frequency operating mode. In order to capture abnormal events in a timely manner when water quality changes drastically, a higher sampling frequency is usually required. To balance the contradiction between monitoring accuracy and system energy consumption, existing technologies have proposed a variety of adaptive sampling frequency control methods. For example, Chinese invention patent CN114486355B discloses a dynamic frequency water quality sampling device and method for self-matching conductivity time-varying processes, which automatically adjusts the sampling frequency according to the time-varying behavior of water conductivity; Chinese invention patent CN112051377B discloses a water environment monitoring frequency self-adjustment system that adapts to weather changes, which adjusts the monitoring frequency in advance based on weather forecast information. The above-mentioned existing technologies can reduce system energy consumption and improve monitoring efficiency to a certain extent.

[0004] However, existing adaptive sampling frequency control technologies only focus on two optimization objectives: monitoring accuracy and system energy consumption. They completely ignore the direct quantitative relationship between sampling frequency and sensor lifespan. When water quality is abnormal, existing systems will indiscriminately increase the sampling frequency of all sensors, leading to an exponential increase in the aging rate of electrochemical sensor electrodes and a significant acceleration in sensor drift. During pollution events when accurate data is most needed, data distortion occurs instead. At the same time, the frequency of sensor replacement and calibration is greatly increased, raising the overall operation and maintenance cost of the system. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an online water quality monitoring system and its monitoring method. This invention solves the problem that existing online water quality monitoring systems only focus on monitoring accuracy and system energy consumption, while ignoring the quantitative relationship between sampling frequency and sensor lifespan, which leads to accelerated sensor aging, data distortion, and high operation and maintenance costs.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On one hand, an online water quality monitoring system includes a sampling unit, multiple water quality sensors, a control unit, a data transmission unit, and a power supply unit. The control unit is characterized by having a built-in three-dimensional dynamic balance optimization module. This module establishes three independent optimization objectives: monitoring accuracy, system energy consumption, and sensor lifespan. It establishes an independent lifespan loss ledger for each water quality sensor, accumulates the lifespan loss caused by each sampling in real time, calculates the remaining lifespan of each water quality sensor, calculates the optimal independent sampling frequency for each water quality sensor based on real-time water quality data and the remaining lifespan of each sensor, and sends control commands to the sampling unit and each water quality sensor to control them to perform sampling and data acquisition operations according to the corresponding optimal sampling frequency.

[0007] Furthermore, the three-dimensional dynamic equilibrium optimization module establishes a quantitative model for single-sampling lifetime loss of electrochemical water quality sensors. The mathematical expression of the quantitative model is as follows: ,in, This indicates the lifespan loss of a single sampling session for electrochemical water quality sensors. This represents the basic loss factor of an electrochemical water quality sensor, which is calculated based on the sensor's design life and standard operating conditions. This indicates the current sampling frequency of the electrochemical water quality sensor. This indicates the sampling frequency index of the electrochemical water quality sensor, which is determined through accelerated aging experiments. This represents the influence coefficient of water quality concentration on electrochemical water quality sensors. This represents the temperature influence coefficient of electrochemical water quality sensors.

[0008] Furthermore, the three-dimensional dynamic balance optimization module establishes a quantitative model for single-sampling lifetime loss of optical water quality sensors. The mathematical expression of the quantitative model is as follows: ,in, This indicates the lifespan loss of an optical water quality sensor during a single sampling. This represents the fundamental loss factor of an optical water quality sensor, which is calculated based on the sensor's design life and standard operating conditions. This indicates the current sampling frequency of the optical water quality sensor. The sampling frequency index represents the sampling frequency index of an optical water quality sensor, which is determined through accelerated aging experiments.

[0009] Furthermore, the three-dimensional dynamic equilibrium optimization module employs a multi-objective particle swarm optimization algorithm to calculate the optimal independent sampling frequency. The objective function of the algorithm is: ,in, This represents the overall optimization target value. The weighting coefficients representing the monitoring accuracy targets. This represents the accuracy loss value when using the current sampling frequency combination. The accuracy loss value is calculated by predicting the change range of water quality parameters within a set future time period. The weighting coefficients represent the system's energy consumption target. This represents the system energy consumption value when using the current sampling frequency combination. The weighting coefficients represent the sensor lifespan target. This represents the total lifespan loss of all water quality sensors when using the current sampling frequency combination.

[0010] Furthermore, the multi-objective particle swarm optimization algorithm is configured with constraints, namely, the accuracy loss value is less than or equal to a preset maximum allowable accuracy loss value, and the weight coefficients... , , Configure according to actual monitoring needs; the default setting is... Greater than , Greater than The three-dimensional dynamic balance optimization module performs an optimization calculation once in each preset decision cycle and outputs the optimal independent sampling frequency for each water quality sensor.

[0011] Furthermore, the control unit incorporates a water quality anomaly classification and judgment module. This module classifies water quality anomalies into Level 1, Level 2, and Level 3 anomalies based on real-time collected water quality data. When a Level 1 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased, as are the sampling frequencies of sensors highly correlated with the anomaly parameter, while keeping the sampling frequencies of other uncorrelated sensors unchanged. When a Level 2 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is further increased, as are the sampling frequencies of sensors highly correlated with the anomaly parameter, while the sampling frequencies of weakly correlated sensors are decreased, causing completely uncorrelated sensors to enter a temporary sleep mode. When a Level 3 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased to a preset maximum sampling frequency, causing all other sensors to enter a temporary sleep mode.

[0012] Furthermore, when a Level 3 anomaly is detected and the sensor corresponding to the anomaly parameter continues to operate at the preset highest sampling frequency for a preset continuous high-frequency operating time, if the water quality parameter is still in a Level 3 anomaly state, the control unit automatically reduces the sampling frequency of the sensor corresponding to the anomaly parameter to a preset safe sampling frequency. The safe sampling frequency is calculated based on the remaining service life of the sensor corresponding to the anomaly parameter and the current water quality concentration influence coefficient to avoid excessive aging of the sensor.

[0013] Furthermore, the control unit incorporates a dynamic lifespan compensation module. This module monitors the cumulative lifespan loss value of each water quality sensor in real time. When the cumulative lifespan loss value of a water quality sensor exceeds a first preset threshold of its designed lifespan, it automatically reduces the basic sampling frequency of that water quality sensor while appropriately increasing the sampling frequency of other sensors related to that water quality sensor. It also sends a sensor lifespan warning message to the management personnel. When the cumulative lifespan loss value exceeds a second preset threshold of its designed lifespan, it automatically switches the water quality sensor to standby mode, only waking it up to perform sampling operations when other related sensors detect water quality abnormalities.

[0014] Furthermore, the control unit has a built-in algorithm self-learning module. The algorithm self-learning module adaptively updates the model parameters of the three-dimensional dynamic balance optimization module based on the actual operating data of the system. It updates the life loss quantification model parameters of each water quality sensor monthly based on the actual calibration data and replacement records of the sensors. It updates the correlation matrix between water quality parameters quarterly based on historical water quality data. It readjusts the weight coefficients of the three optimization objectives annually based on the actual operating effect of the system.

[0015] On the other hand, a water quality online monitoring method, applicable to a water quality online monitoring system, includes the following steps:

[0016] Step 1: The control unit establishes an independent lifespan loss ledger for each water quality sensor, accumulates the lifespan loss caused by each sampling to the corresponding water quality sensor in real time, and calculates the remaining lifespan of each water quality sensor.

[0017] Step 2: The three-dimensional dynamic balance optimization module establishes three independent optimization objectives—monitoring accuracy, system energy consumption, and sensor lifespan—based on real-time water quality data and the remaining service life of each water quality sensor. It then uses a multi-objective particle swarm optimization algorithm to calculate the optimal independent sampling frequency for each water quality sensor.

[0018] Step 3: The control unit sends control commands to the sampling unit and each water quality sensor to control the sampling unit and each water quality sensor to perform sampling and data acquisition operations according to the corresponding optimal sampling frequency.

[0019] Step 4: The water quality anomaly classification and judgment module performs anomaly classification and judgment based on the real-time collected water quality data, and executes the corresponding differentiated sampling strategy according to different anomaly levels.

[0020] Step 5: The lifespan loss dynamic compensation module performs the corresponding lifespan compensation operation based on the cumulative lifespan loss value of each water quality sensor.

[0021] Step six: The algorithm self-learning module adaptively updates the model parameters of the three-dimensional dynamic equilibrium optimization module based on the actual system operation data.

[0022] Compared with existing technologies, this online water quality monitoring system and method have the following advantages:

[0023] I. This invention establishes a dynamic balance optimization model encompassing three dimensions: monitoring accuracy, system energy consumption, and sensor lifespan. It constructs a quantitative model for single-sampling lifespan loss for different types of water quality sensors, creating an independent lifespan loss ledger for each sensor and accumulating losses and calculating remaining lifespan in real time. A multi-objective optimization algorithm is employed to determine the optimal independent sampling frequency for each sensor. This approach accurately quantifies and controls the lifespan loss caused by sampling operations while ensuring monitoring accuracy and controlling system energy consumption. It avoids the accelerated aging and data distortion caused by indiscriminately increasing the sampling frequency of all sensors during water quality anomalies, thereby significantly reducing the frequency of sensor replacement and calibration and lowering the overall system operation and maintenance costs.

[0024] Second, this invention establishes a water quality anomaly classification judgment module, which implements differentiated sampling strategies based on the level of water quality anomaly. During water quality anomalies, it prioritizes the monitoring accuracy of key parameters while simultaneously reducing the frequency or putting irrelevant sensors into sleep mode, enabling accurate capture of critical data during sudden pollution events. Furthermore, by incorporating a dynamic lifespan loss compensation module and an algorithm self-learning module, the sampling strategy can be dynamically adjusted based on actual sensor wear and tear, and model parameters can be continuously optimized based on system operation data. This improves system stability and long-term adaptability, extending the overall effective service life of the system.

[0025] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0027] Figure 1 This is a diagram of the overall architecture of the present invention;

[0028] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0029] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0030] like Figure 1 and Figure 2 As shown, an online water quality monitoring system includes a sampling unit, multiple water quality sensors, a control unit, a data transmission unit, and a power supply unit. The sampling unit is used to extract water samples from the water body to be monitored and transport them to the detection positions of each water quality sensor. The multiple water quality sensors include electrochemical water quality sensors and optical water quality sensors. The control unit is implemented using an embedded microprocessor and is responsible for the operation control and data processing of the entire system. The data transmission unit is implemented using a wireless communication module and is used to upload monitoring data to a remote monitoring platform. The power supply unit is implemented using a solar power system combined with a battery to provide a stable power supply for the entire system.

[0031] The control unit incorporates a three-dimensional dynamic balancing optimization module. This module establishes three independent optimization objectives: monitoring accuracy, system energy consumption, and sensor lifespan. It also maintains an independent lifespan loss ledger for each water quality sensor, stored in the control unit's non-volatile memory. The ledger records parameters such as cumulative lifespan loss, remaining lifespan, basic loss coefficient, and sampling frequency index for each sensor. The module continuously accumulates the lifespan loss caused by each sampling operation to the corresponding water quality sensor. It calculates the remaining lifespan of each sensor and, based on real-time water quality data and the remaining lifespan of each sensor, calculates the optimal independent sampling frequency for each sensor. Finally, the module sends control commands to the sampling unit and each water quality sensor. These commands instruct the sampling unit and each sensor to perform sampling and data acquisition operations according to their respective optimal independent sampling frequencies.

[0032] Specifically, the three-dimensional dynamic equilibrium optimization module establishes a quantitative model for the single-sampling lifetime loss of electrochemical water quality sensors. The mathematical expression of the quantitative model is:

[0033]

[0034] in, This indicates the lifespan loss of a single sampling of an electrochemical water quality sensor. This represents the fundamental loss factor of an electrochemical water quality sensor. The fundamental loss factor is calculated based on the sensor's design life and standard operating conditions. Standard operating conditions include standard sampling frequency, standard water quality concentration range, and standard operating temperature range. This indicates the current sampling frequency of the electrochemical water quality sensor. This represents the sampling frequency index of an electrochemical water quality sensor. The sampling frequency index is determined through accelerated aging experiments, conducted in a laboratory environment. The accelerated aging experiment involves continuously running the sensor at different sampling frequencies, recording the time it takes for the sensor's performance to degrade to a preset threshold, and then calculating the sampling frequency index. This represents the water quality concentration influence coefficient of an electrochemical water quality sensor. The water quality concentration influence coefficient is determined based on the ratio of the currently detected water quality parameter concentration to the standard concentration range. When the water quality parameter concentration is within the standard concentration range, the water quality concentration influence coefficient is 1. When the water quality parameter concentration exceeds the standard concentration range, the water quality concentration influence coefficient increases with increasing concentration. This represents the temperature influence coefficient of an electrochemical water quality sensor. The temperature influence coefficient is determined based on the difference between the current sensor operating temperature and the standard operating temperature range. When the sensor operating temperature is within the standard operating temperature range, the temperature influence coefficient is 1. When the sensor operating temperature exceeds the standard operating temperature range, the temperature influence coefficient increases with the increase of the temperature deviation.

[0035] The three-dimensional dynamic equilibrium optimization module establishes a quantitative model of single-sampling lifetime loss for optical water quality sensors. The mathematical expression of the quantitative model is:

[0036]

[0037] in, This indicates the lifespan loss of an optical water quality sensor during a single sampling. This represents the fundamental loss factor of an optical water quality sensor, which is calculated based on the sensor's design life and standard operating conditions. This indicates the current sampling frequency of the optical water quality sensor. This indicates the sampling frequency index of the optical water quality sensor. The sampling frequency index is determined through accelerated aging experiments. The accelerated aging experimental method for optical water quality sensors is similar to that for electrochemical water quality sensors.

[0038] The three-dimensional dynamic equilibrium optimization module employs a multi-objective particle swarm optimization algorithm to calculate the optimal independent sampling frequency. The algorithm's objective function is:

[0039]

[0040] in, This represents the overall optimization target value. This represents the weighting coefficient of the monitoring accuracy target. This represents the accuracy loss value when using the current sampling frequency combination. The accuracy loss value is calculated by predicting the change range of water quality parameters within a future set time period. Specifically, this embodiment uses an autoregressive integral moving average model to predict the change trend of each water quality parameter within a future time period T, where T ranges from 10 minutes to 60 minutes. Let the predicted value of the i-th water quality parameter at time t be... The sampling interval at the current sampling frequency is Then, within the prediction time period T, the maximum predicted change in this parameter is: When the sampling interval When the time scale exceeds the feature time scale of parameter change, accuracy loss occurs. (Accuracy loss value) The calculation formula is:

[0041]

[0042] in, This represents the total number of water quality parameters. Let be the monitoring accuracy weight for the i-th water quality parameter, and the sum of the accuracy weights for all parameters is 1. and These are the upper and lower limits of the measurement range for the i-th water quality parameter, respectively. When the predicted change range... Larger, sampling interval The longer the time, the greater the loss of accuracy. The larger.

[0043] Predicting the magnitude of changes in water quality parameters over a given future time period is achieved using time series forecasting algorithms. These algorithms include, but are not limited to, autoregressive moving average algorithms and long short-term memory network algorithms. The weighting coefficient represents the system's energy consumption target. This represents the system energy consumption value when using the current sampling frequency combination. The system energy consumption value is calculated based on the sum of the energy consumption of the sampling unit, the energy consumption of each water quality sensor, and the energy consumption of the control unit. The energy consumption of the sampling unit is directly proportional to the sampling frequency. The energy consumption of each water quality sensor is directly proportional to its respective sampling frequency. The weighting coefficient represents the target lifespan of the sensor. This represents the total lifetime loss of all water quality sensors when using the current sampling frequency combination. The total lifetime loss is the sum of the lifetime loss of each water quality sensor during a single sampling.

[0044] The multi-objective particle swarm optimization algorithm has constraints. The constraint is that the accuracy loss value is less than or equal to a preset maximum allowable accuracy loss value. Weight coefficients... , , Configure according to actual monitoring needs; the default setting is... Greater than , Greater than The three-dimensional dynamic balance optimization module performs an optimization calculation once within each preset decision cycle. The module outputs the optimal independent sampling frequency for each water quality sensor. The length of the decision cycle is set according to actual monitoring needs. The length of the decision cycle can be set to any value between 1 hour and 24 hours.

[0045] In some optional implementations, the particle swarm size of the multi-objective particle swarm optimization algorithm is set to any value between 20 and 50. The number of iterations of the multi-objective particle swarm optimization algorithm is set to any value between 100 and 500. The inertia weight of the multi-objective particle swarm optimization algorithm is set to any value between 0.4 and 0.9. The cognitive coefficient of the multi-objective particle swarm optimization algorithm is set to any value between 1.5 and 2.5.

[0046] The social coefficient of the multi-objective particle swarm optimization algorithm is set to any value between 1.5 and 2.5. The control unit incorporates a water quality anomaly classification module. This module classifies water quality anomalies into Level 1, Level 2, and Level 3 anomalies based on real-time collected water quality data. Level 1 anomalies indicate that water quality parameters exceed the normal range but have not reached the warning threshold. Level 2 anomalies indicate that water quality parameters reach the warning threshold but have not reached the emergency threshold. Level 3 anomalies indicate that water quality parameters reach the emergency threshold. When a Level 1 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased. The sampling frequency of sensors highly correlated with the anomaly parameter is increased. The sampling frequency of other uncorrelated sensors remains unchanged. When a Level 2 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is further increased. The sampling frequency of sensors highly correlated with the anomaly parameter is increased. The sampling frequency of weakly correlated sensors is decreased. Completely uncorrelated sensors enter a temporary sleep mode.

[0047] When a Level 3 anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased to the preset maximum sampling frequency. All other sensors are then put into temporary sleep mode. Specifically, the correlation between water quality parameters is represented by a correlation matrix. The correlation matrix is ​​stored in the non-volatile memory of the control unit. Elements in the correlation matrix represent the degree of correlation between two water quality parameters. The correlation value ranges from 0 to 1. When the correlation between two water quality parameters is greater than or equal to a first preset correlation threshold, the two water quality parameters are considered highly correlated. When the correlation between two water quality parameters is greater than or equal to a second preset correlation threshold but less than the first preset correlation threshold, the two water quality parameters are considered weakly correlated. When the correlation between two water quality parameters is less than the second preset correlation threshold, the two water quality parameters are considered completely uncorrelated. The first preset correlation threshold is greater than the second preset correlation threshold.

[0048] When a Level 3 anomaly is detected and the sensor corresponding to the anomaly parameter continues to operate at the preset highest sampling frequency for a preset continuous high-frequency operating time, if the water quality parameter remains in a Level 3 anomaly state, the control unit automatically reduces the sampling frequency of the sensor corresponding to the anomaly parameter to a preset safe sampling frequency. The safe sampling frequency is calculated based on the remaining service life of the sensor corresponding to the anomaly parameter and the influence coefficient of the current water quality concentration.

[0049] Specifically, safe sampling frequency The calculation formula is:

[0050]

[0051] in, This is the baseline sampling frequency of the sensor under normal water quality conditions. This represents the remaining lifespan of the sensor. In this embodiment, the preset lifespan protection threshold is used. Set to 10% of the sensor's design life. This represents the influence coefficient of the current water quality concentration. The calculation principle of this formula is: when the remaining service life of the sensor is sufficient ( And the water concentration is within the standard range. When the remaining service life is below the life protection threshold or the water quality concentration exceeds the standard, the safe sampling frequency is automatically reduced proportionally to minimize sensor lifespan loss while ensuring continuous monitoring of abnormal water quality trends.

[0052] The principle for calculating the safe sampling frequency is to minimize sensor lifespan loss while ensuring continuous monitoring of water quality anomalies. Setting a safe sampling frequency can prevent excessive aging of the sensor.

[0053] The control unit incorporates a dynamic lifespan compensation module. This module monitors the cumulative lifespan loss of each water quality sensor in real time. When the cumulative lifespan loss of a water quality sensor exceeds a first preset threshold of its designed lifespan, the module automatically reduces the sensor's base sampling frequency. Simultaneously, it appropriately increases the sampling frequency of other related sensors. A sensor lifespan warning message is sent to management personnel. When the cumulative lifespan loss exceeds a second preset threshold of its designed lifespan, the sensor is automatically switched to standby mode. This sensor is only activated for sampling when other related sensors detect an abnormal water quality. The first preset threshold is less than the second preset threshold.

[0054] In some optional implementations, the first preset threshold is set to any value between 70% and 80% of the design life. The second preset threshold is set to any value between 90% and 95% of the design life. The base sampling frequency of the water quality sensor is reduced by any value between 20% and 50% of the original base sampling frequency. The sampling frequency of other sensors associated with the water quality sensor is increased by any value between 10% and 30% of the original sampling frequency.

[0055] The control unit incorporates a built-in algorithm self-learning module. This module adaptively updates the model parameters of the three-dimensional dynamic balance optimization module based on actual system operating data. Monthly updates are made to the lifespan quantification model parameters of each water quality sensor based on actual sensor calibration data and replacement records. Quarterly updates are made to the correlation matrix between water quality parameters based on historical water quality data. Annually, the weight coefficients of the three optimization objectives are readjusted based on the system's actual operating performance.

[0056] Specifically, the actual calibration data of the sensor includes calibration time, measurement error before calibration, and measurement error after calibration. The sensor replacement record includes replacement time, reason for replacement, and the cumulative lifetime loss value of the old sensor. Based on the actual calibration data and replacement record, the actual lifetime loss rate of the sensor can be calculated. The actual lifetime loss rate is compared with the lifetime loss rate predicted by the model. Based on the comparison results, the base loss coefficient and sampling frequency exponent in the lifetime loss quantification model are adjusted.

[0057] Specifically, let the model predict the cumulative lifespan loss of the sensor as... The actual cumulative lifespan loss, measured based on the sensor's actual calibration data and performance degradation, is... Basic loss coefficient The exponentially weighted moving average method is used for updating, and the update formula is as follows:

[0058]

[0059] in, As a smoothing coefficient, in this embodiment The value range is from 0.1 to 0.3. Sampling frequency index The gradient descent method is used for updating, minimizing the mean square error between the predicted loss and the actual loss. To optimize, the formula is updated as follows:

[0060]

[0061] in, In this embodiment, the learning rate is... The value range is from 0.01 to 0.1. This is the partial derivative of the mean square error with respect to the sampling frequency exponent. The above parameter update operation is performed monthly, and the updated parameters are stored in the non-volatile memory of the control unit, taking effect in the next decision cycle.

[0062] The correlation matrix between water quality parameters is updated based on historical water quality data using the Pearson correlation coefficient calculation method. The Pearson correlation coefficient method calculates the degree of linear correlation between the time series of two water quality parameters. The absolute value of the calculated Pearson correlation coefficient is used as the value of the corresponding element in the correlation matrix.

[0063] The weight coefficients of the three optimization objectives were readjusted based on the actual operating results of the system using the Analytic Hierarchy Process (AHP). The AHP determines the weight coefficients of the three optimization objectives through steps such as constructing a hierarchical structure model, building a judgment matrix, calculating weight vectors, and performing consistency checks. The actual operating results of the system include the monitoring accuracy compliance rate, average system energy consumption, and average sensor lifespan.

[0064] This invention also discloses an online water quality monitoring method. This method is applied to the aforementioned online water quality monitoring system. The method includes the following steps:

[0065] Step 1: The control unit establishes an independent lifespan loss ledger for each water quality sensor. It accumulates the lifespan loss caused by each sampling to the corresponding water quality sensor in real time. The remaining lifespan of each water quality sensor is calculated. Specifically, after each sampling, the three-dimensional dynamic balance optimization module, based on the sensor type, calls the corresponding single-sampling lifespan loss quantification model to calculate the lifespan loss value for this sampling. This lifespan loss value is added to the sensor's cumulative lifespan loss value. The remaining lifespan equals the sensor's design lifespan minus the cumulative lifespan loss value.

[0066] Step two: The three-dimensional dynamic balance optimization module establishes three independent optimization objectives—monitoring accuracy, system energy consumption, and sensor lifespan—based on real-time water quality data and the remaining lifespan of each water quality sensor. A multi-objective particle swarm optimization algorithm is used to calculate the optimal independent sampling frequency for each water quality sensor. Specifically, firstly, the variation range of each water quality parameter within a set future time period is predicted based on real-time water quality data. The accuracy loss value under different sampling frequency combinations is calculated based on the predicted variation range. The system energy consumption value and total lifespan loss value are calculated based on the sampling frequency of each water quality sensor. Then, with the goal of minimizing the overall optimization objective value, and under the constraint that the accuracy loss value is less than or equal to the preset maximum allowable accuracy loss value, the multi-objective particle swarm optimization algorithm is used to search for the optimal sampling frequency combination.

[0067] Step three: The control unit sends control commands to the sampling unit and each water quality sensor. The control unit and each water quality sensor perform sampling and data acquisition operations according to their respective optimal independent sampling frequencies. Specifically, the control unit generates a corresponding sampling control signal based on the optimal independent sampling frequency of each water quality sensor. This sampling control signal is sent to the sampling unit and each water quality sensor. The sampling unit controls the start / stop and running time of the water pump according to the sampling control signal. Each water quality sensor controls the operating state of its detection circuit according to the sampling control signal.

[0068] Step four: The water quality anomaly classification module classifies anomalies based on real-time collected water quality data. It then executes corresponding differentiated sampling strategies according to different anomaly levels. Specifically, the module compares the real-time collected water quality data with preset normal ranges, warning thresholds, and emergency thresholds. Based on the comparison results, it determines the water quality anomaly level. Then, it executes a corresponding sampling frequency adjustment strategy based on the anomaly level. Once the water quality returns to normal, the module restores the sampling frequency of each sensor to the optimal independent sampling frequency calculated by the three-dimensional dynamic balance optimization module.

[0069] Step 5: The lifespan loss dynamic compensation module performs corresponding lifespan compensation operations based on the cumulative lifespan loss value of each water quality sensor. Specifically, the module periodically reads the cumulative lifespan loss value of each water quality sensor. It compares the cumulative lifespan loss value with a first preset threshold and a second preset threshold. Based on the comparison result, it performs corresponding sampling frequency adjustment operations and sensor mode switching operations.

[0070] Step six: The algorithm self-learning module adaptively updates the model parameters of the three-dimensional dynamic equilibrium optimization module based on the actual system operating data. Specifically, the algorithm self-learning module performs lifetime loss quantification model parameter updates, correlation matrix updates, and weight coefficient updates according to a preset update cycle. The updated model parameters are stored in the non-volatile memory of the control unit. The updated model parameters take effect in the next decision cycle.

[0071] This invention establishes a three-dimensional dynamic equilibrium optimization model, simultaneously considering three optimization objectives: monitoring accuracy, system energy consumption, and sensor lifespan. This effectively reduces system energy consumption and extends sensor lifespan while maintaining monitoring accuracy. By establishing a quantitative model of lifespan loss for different types of sensors, this invention can accurately calculate the lifespan loss caused by each sampling. Employing a multi-objective particle swarm optimization algorithm, this invention can quickly solve for the optimal combination of sampling frequencies. By setting up a water quality anomaly classification judgment module, this invention can execute differentiated sampling strategies based on different anomaly levels, prioritizing the monitoring accuracy of key parameters when water quality is abnormal. By setting up a dynamic lifespan loss compensation module, this invention can dynamically adjust the sampling strategy based on the actual lifespan loss of the sensors, preventing premature sensor failure. By setting up an algorithm self-learning module, this invention can continuously optimize model parameters based on actual system operating data, improving the system's adaptability and accuracy.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A water quality online monitoring system, comprising a sampling unit, multiple water quality sensors, a control unit, a data transmission unit, and a power supply unit, characterized in that, The control unit incorporates a three-dimensional dynamic balance optimization module. This module establishes three independent optimization objectives: monitoring accuracy, system energy consumption, and sensor lifespan. It creates an independent lifespan loss ledger for each water quality sensor, accumulates the lifespan loss caused by each sampling in real time, calculates the remaining lifespan of each water quality sensor, and calculates the optimal independent sampling frequency for each water quality sensor based on real-time water quality data and the remaining lifespan of each water quality sensor. It then sends control commands to the sampling unit and each water quality sensor to control them to perform sampling and data acquisition operations according to the corresponding optimal sampling frequency.

2. The online water quality monitoring system according to claim 1, characterized in that, The three-dimensional dynamic equilibrium optimization module establishes a quantitative model for the single-sampling lifetime loss of electrochemical water quality sensors. The mathematical expression of the quantitative model is as follows: ,in, This indicates the lifespan loss of a single sampling session for electrochemical water quality sensors. This represents the basic loss factor of an electrochemical water quality sensor, which is calculated based on the sensor's design life and standard operating conditions. This indicates the current sampling frequency of the electrochemical water quality sensor. This indicates the sampling frequency index of the electrochemical water quality sensor, which is determined through accelerated aging experiments. This represents the influence coefficient of water quality concentration on electrochemical water quality sensors. This represents the temperature influence coefficient of electrochemical water quality sensors.

3. The online water quality monitoring system according to claim 2, characterized in that, The three-dimensional dynamic balance optimization module establishes a quantitative model for the single-sampling lifetime loss of optical water quality sensors. The mathematical expression of the quantitative model is as follows: ,in, This indicates the lifespan loss of an optical water quality sensor during a single sampling. This represents the fundamental loss factor of an optical water quality sensor, which is calculated based on the sensor's design life and standard operating conditions. This indicates the current sampling frequency of the optical water quality sensor. The sampling frequency index represents the sampling frequency index of an optical water quality sensor, which is determined through accelerated aging experiments.

4. The online water quality monitoring system according to claim 1, characterized in that, The three-dimensional dynamic equilibrium optimization module uses a multi-objective particle swarm optimization algorithm to calculate the optimal independent sampling frequency. The optimization objective function of the algorithm is: ,in, This represents the overall optimization target value. The weighting coefficients representing the monitoring accuracy targets. This represents the accuracy loss value when using the current sampling frequency combination. The accuracy loss value is calculated by predicting the change range of water quality parameters within a set future time period. The weighting coefficients represent the system's energy consumption target. This represents the system energy consumption value when using the current sampling frequency combination. The weighting coefficients represent the sensor lifespan target. This represents the total lifespan loss of all water quality sensors when using the current sampling frequency combination.

5. The online water quality monitoring system according to claim 4, characterized in that, The multi-objective particle swarm optimization algorithm is subject to constraints, namely, the accuracy loss value is less than or equal to a preset maximum allowable accuracy loss value, and the weight coefficients are... , , Configure according to actual monitoring needs; the default setting is... Greater than , Greater than The three-dimensional dynamic balance optimization module performs an optimization calculation once in each preset decision cycle and outputs the optimal independent sampling frequency for each water quality sensor.

6. The online water quality monitoring system according to claim 1, characterized in that, The control unit has a built-in water quality anomaly classification and judgment module. The water quality anomaly classification and judgment module classifies water quality anomalies into level one, level two, and level three anomalies based on real-time collected water quality data. When a level one anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased, the sampling frequency of the sensor highly correlated with the anomaly parameter is increased, and the sampling frequency of other uncorrelated sensors remains unchanged. When a level two anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is further increased, the sampling frequency of the sensor highly correlated with the anomaly parameter is increased, the sampling frequency of the weakly correlated sensor is decreased, and the sensor that is completely uncorrelated enters a temporary sleep mode. When a level three anomaly is detected, the sampling frequency of the sensor corresponding to the anomaly parameter is increased to the preset maximum sampling frequency, and all other sensors enter a temporary sleep mode.

7. The online water quality monitoring system according to claim 6, characterized in that, When a Level 3 anomaly is detected and the sensor corresponding to the abnormal parameter continues to operate at the preset highest sampling frequency for a preset continuous high-frequency operating time, if the water quality parameter is still in a Level 3 abnormal state, the control unit automatically reduces the sampling frequency of the sensor corresponding to the abnormal parameter to a preset safe sampling frequency. The safe sampling frequency is calculated based on the remaining service life of the sensor corresponding to the abnormal parameter and the current water quality concentration influence coefficient to avoid excessive aging of the sensor.

8. The online water quality monitoring system according to claim 1, characterized in that, The control unit has a built-in dynamic lifespan compensation module, which monitors the cumulative lifespan loss value of each water quality sensor in real time. When the cumulative lifespan loss value of a water quality sensor exceeds the first preset threshold of its designed lifespan, the module automatically reduces the basic sampling frequency of the water quality sensor and appropriately increases the sampling frequency of other sensors related to the water quality sensor. It also sends a sensor lifespan warning message to the management personnel. When the cumulative lifespan loss value exceeds the second preset threshold of its designed lifespan, the module automatically switches the water quality sensor to standby mode. The water quality sensor is only awakened to perform sampling operations when other related sensors detect water quality abnormalities.

9. The online water quality monitoring system according to claim 1, characterized in that, The control unit has a built-in algorithm self-learning module. The algorithm self-learning module adaptively updates the model parameters of the three-dimensional dynamic balance optimization module based on the actual operating data of the system. It updates the life loss quantification model parameters of each water quality sensor monthly based on the actual calibration data and replacement records of the sensors. It updates the correlation matrix between water quality parameters quarterly based on historical water quality data. It readjusts the weight coefficients of the three optimization objectives annually based on the actual operating effect of the system.

10. A method for online water quality monitoring, applied to the online water quality monitoring system according to any one of claims 1 to 9, characterized in that, Includes the following steps: Step 1: The control unit establishes an independent lifespan loss ledger for each water quality sensor, accumulates the lifespan loss caused by each sampling to the corresponding water quality sensor in real time, and calculates the remaining lifespan of each water quality sensor. Step 2: The three-dimensional dynamic balance optimization module establishes three independent optimization objectives—monitoring accuracy, system energy consumption, and sensor lifespan—based on real-time water quality data and the remaining service life of each water quality sensor. It then uses a multi-objective particle swarm optimization algorithm to calculate the optimal independent sampling frequency for each water quality sensor. Step 3: The control unit sends control commands to the sampling unit and each water quality sensor to control the sampling unit and each water quality sensor to perform sampling and data acquisition operations according to the corresponding optimal sampling frequency. Step 4: The water quality anomaly classification and judgment module performs anomaly classification and judgment based on the real-time collected water quality data, and executes the corresponding differentiated sampling strategy according to different anomaly levels. Step 5: The lifespan loss dynamic compensation module performs the corresponding lifespan compensation operation based on the cumulative lifespan loss value of each water quality sensor. Step six: The algorithm self-learning module adaptively updates the model parameters of the three-dimensional dynamic equilibrium optimization module based on the actual system operation data.

Images