A multi-dimensional intelligent control system for fluorine removal

Abstract

The present invention discloses a multi-dimensional intelligent control system for fluorine removal, which specifically relates to the field of sewage fluorine removal, including a water quality fluctuation monitoring module, a normal measurement judgment module, an abnormal stage judgment module, a deep learning prediction module, and a multi-task regression module; the water quality fluctuation monitoring module is used to obtain water quality fluctuation information of a water source that needs to be defluorinated in real time, establish pre-measurement condition 1 and pre-measurement condition 2 based on the water quality fluctuation information, and determine whether there are water quality influencing factors in the current water source through pre-measurement condition 1 and pre-measurement condition 2. The present invention realizes the management of fluorine concentration control by introducing multi-dimensional water quality fluctuation information and real-time monitoring mechanism, and performs real-time prediction based on water quality parameters such as water body chromaticity, conductivity, reflectivity and ammonia nitrogen concentration in combination with deep neural network, which is conducive to identifying water quality fluctuations and performing corresponding regulation, dynamically responding to water quality changes, and improving the convenience of water source control.

Description

Technical Field

[0001] The present invention relates to the technical field of wastewater defluorination, and more specifically, to a multi-dimensional intelligent control system for defluorination. Background Art

[0002] Traditional wastewater defluoridation processes usually rely on fixed parameter settings and manual control, lacking a real-time dynamic adjustment mechanism. These methods often lag behind when faced with changes in water quality, and are unable to adjust the fluoride removal process in time to cope with sudden water quality fluctuations and environmental changes. In addition, many existing water quality monitoring systems lack comprehensive analysis of water quality status, resulting in the inability to determine the state adaptation and changing trend of fluoride in the water source; these deficiencies make traditional methods incapable of dealing with complex water quality fluctuations, especially abnormal water quality fluctuations, affecting fluoride removal efficiency and water source water quality safety. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a multi-dimensional intelligent control system for fluorine removal, which solves the problems raised in the above-mentioned background technology through a water quality fluctuation monitoring module, a normal measurement and judgment module, an abnormal stage judgment module, a deep learning prediction module, and a multi-task regression module.

[0004] To achieve the above-mentioned object, the present invention provides the following technical solutions: a multi-dimensional intelligent control system for fluorine removal, comprising a water quality fluctuation monitoring module, a normal measurement determination module, an abnormal stage determination module, a deep learning prediction module, and a multi-task regression module; The water quality fluctuation monitoring module is used to obtain water quality fluctuation information of the water source that needs to be defluorinated in real time, establish pre-measurement condition 1 and pre-measurement condition 2 based on the water quality fluctuation information, and determine whether there are water quality influencing factors in the current water source through pre-measurement condition 1 and pre-measurement condition 2; The water quality impact information in the water quality fluctuation monitoring module includes water chromaticity, conductivity, water reflectivity, and ammonia nitrogen concentration in water; The normal measurement judgment module is used to establish the pre-measurement condition 1, which is the normal measurement state; the pre-measurement condition 1 includes that when the water body chromaticity is greater than the water body chromaticity threshold preset by the system, and the conductivity is greater than the conductivity threshold preset by the system, then a type of abnormal stage is determined; The abnormal stage judgment module is used to establish the second pre-measurement condition when the water source is in a type of abnormal stage; the second pre-measurement condition includes marking the water quality impact factor for the current water source when the water body reflectance is lower than the water body reflectance threshold preset by the system, or the water body ammonia nitrogen concentration is greater than the water body ammonia nitrogen concentration threshold preset by the system, otherwise it returns to the normal measurement state within the preset sliding time window; The deep learning prediction module is used to predict three target variables for water sources marked with water quality influencing factors through a deep neural network model; construct feature tensors based on the input water quality parameter data, and perform end-to-end feature extraction in the shared hidden layer through the back propagation algorithm; The multi-task regression module regresses three target variables in parallel based on the task output layer of the multi-task learning framework, and optimizes the distinction of the three target variables through independent loss functions; the three target variables include seasonal influence variables, processing response lag variables, and abnormal water quality variables.

[0005] In a preferred embodiment, the water body chromaticity is calculated based on the weighted time decay model, and the weight coefficient is introduced To reflect the impact of historical data on the current chromaticity, a time attenuation factor is introduced To describe the attenuation effect of historical observation data over time; the weighted historical data is attenuated to reflect the changing trend of water body chromaticity over time; ; in For the current moment The chromaticity value of water; For historical moments Observed values ​​of water chromaticity; Indicates the time corresponding to The weight coefficient of is the time attenuation coefficient; for the current moment; timestamp for historical moments; is the size of the time window.

[0006] In a preferred embodiment, the change of conductivity over time and dynamic factor is reflected by combining weighting factor and attenuation factor. and Respectively reflect the influence of ion concentration and flow rate in water, and use Adjust its weight on conductivity; Indicates the current time The conductivity value of ; in The conductivity-related Factors; Indicates Factors that affect conductivity: Indicates the The weight of each factor on the conductivity; is the time attenuation coefficient; Indicates that it is used to adjust and the extent of its effect; is the decay index; Timestamp for historical data; For the current moment.

[0007] In a preferred embodiment, the historical reflectivity data is integrated and combined with the attenuation function to reflect the influence of the historical data on the current reflectivity; nonlinear time adjustment is introduced, and the Simulate the effect of water surface conditions on reflectivity; For the current moment Water reflectivity ; in is the function of reflectivity decay over time; Parameters that represent the surface state of water bodies; A coefficient used to adjust the effect of water surface conditions on reflectivity; represents the adjustment coefficient used to control the change of reflectivity over time; Indicates the time difference.

[0008] In a preferred embodiment, the current ammonia nitrogen concentration is calculated by weighting the historical ammonia nitrogen concentration related factors, and the weight is determined by the time decay factor Control, with time interval Increase, the impact of historical data on the current ammonia nitrogen concentration decreases, and the current ammonia nitrogen concentration is predicted by integrating multiple historical data over a period of time; To indicate the current time Ammonia nitrogen concentration; ; in For historical moments Ammonia nitrogen related factors; is the time decay factor; is the adjustment coefficient; is the time difference term; is the time decay exponent.

[0009] In a preferred embodiment, water quality parameter data is input to construct a feature tensor, and the water quality parameter data includes , , , , the water quality parameter data is divided into time steps Construct a feature tensor , used for deep neural network training and feature extraction; ; in Indicates the current time step Corresponding to the input feature tensor, the value of each water quality parameter in the water quality parameter data is expressed in time is the index, constituting an element of the input feature; In a deep neural network, the input feature tensor After multiple hidden layers, feature extraction is performed and the predicted value is finally output. The network is trained through the back propagation algorithm, and the loss function is used to optimize the weights; Output the hidden layer through the ReLU activation function; ; The output layer predicts and regresses three target variables in parallel; ; in and They are the weight matrix and bias term of the first layer, acting on the input features , and through the activation function Perform nonlinear mapping to generate hidden layer output ; In the multi-task learning framework, the task output layer converts the output of the hidden layer As input, three target variables are predicted separately: seasonal impact variable , dealing with response lagged variables , abnormal water quality variables ; represents the function used for seasonal effects; represents the function used to handle response lag; represents a function for abnormal water quality detection; Indicates at time Historical water quality data at all times; In the multi-task learning framework, the three target variables are predicted in the task output layer through parallel regression. The regression task of each target variable includes an independent loss function. The learning of the model is improved by optimizing the loss function of each task. Include the loss function as the total loss; ; in is the total loss function, which is the sum of the loss functions of multiple target tasks; is the regression loss function of seasonal effects; Regression loss function to handle response lag; is the regression loss function of abnormal water quality; The loss function corresponding to each target variable is used to measure the error between the predicted value and the actual value. The weight of the network is optimized through back propagation. Through parallel regression, each task is independently optimized based on the shared features to achieve the prediction of the target variable.

[0010] In a preferred embodiment, a back propagation algorithm is introduced, which calculates the gradient of the loss function of each task and back propagates it to each layer of the network to update the weights and biases; performs gradient calculation; ; The total loss function is calculated by back-propagation algorithm with respect to the network weights The gradient is used to update the weights of each layer. The gradient contribution of each loss function to the weight is independent, but they are jointly optimized on the shared hidden layer parameters. The learning ability of the model to distinguish the three target variables is enhanced by time series features and external environment data; Capture the delayed effect of water quality changes through hysteresis characteristics and construct hysteresis characteristics ; ; External environment characteristics Including traffic ,temperature ,humidity ; ; Trend Analysis Feature Trend analysis based on historical data captures long-term trends in water quality.

[0011] Technical effects and advantages of the present invention: The present invention realizes the management of fluorine concentration control by introducing multi-dimensional water quality fluctuation information and real-time monitoring mechanism. It performs real-time prediction based on water quality parameters such as water chromaticity, conductivity, reflectivity and ammonia nitrogen concentration combined with deep neural network, which is conducive to identifying water quality fluctuations and making corresponding adjustments, dynamically responding to water quality changes, and improving the convenience of water source control.

[0012] Through the dynamic judgment mechanism of pre-measurement conditions 1 and 2, it is possible to judge in real time whether the water source has entered an abnormal stage, and adjust the fluorine removal strategy accordingly. The pre-conditions are based on the changes in water quality parameters, combined with time window and sliding window control to ensure that the system can respond in time to any water quality changes, and prevent the fluorine concentration from exceeding the standard due to the failure to deal with abnormal water quality in time; A deep neural network model is used to extract the features of water quality fluctuations. The three target variables of seasonal impact, processing response lag and abnormal water quality are regressed in parallel through a multi-task learning framework, which improves the convenience of the model in predicting water quality fluctuations. The present invention integrates external environmental data and time series analysis, and combines meteorological factors to enhance the model's ability to predict water quality changes. The hysteresis characteristics and the time dependence of historical data help the model capture the delayed effect of water quality changes, and flexibly adjust according to historical data, thereby improving the robustness of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0015] Refer to the instruction manual Figure 1 , a multi-dimensional intelligent control system for fluorine removal according to an embodiment of the present invention includes a water quality fluctuation monitoring module, a normal measurement judgment module, an abnormal stage judgment module, a deep learning prediction module, and a multi-task regression module; The water quality fluctuation monitoring module is used to obtain water quality fluctuation information of the water source that needs to be defluorinated in real time, establish pre-measurement condition 1 and pre-measurement condition 2 based on the water quality fluctuation information, and determine whether there are water quality influencing factors in the current water source through pre-measurement condition 1 and pre-measurement condition 2; The water quality impact information in the water quality fluctuation monitoring module includes water chromaticity, conductivity, water reflectivity, and ammonia nitrogen concentration in water; The normal measurement judgment module is used to establish the pre-measurement condition 1, which is the normal measurement state; the pre-measurement condition 1 includes that when the water body chromaticity is greater than the water body chromaticity threshold preset by the system, and the conductivity is greater than the conductivity threshold preset by the system, then a type of abnormal stage is determined; The abnormal stage judgment module is used to establish the second pre-measurement condition when the water source is in a type of abnormal stage; the second pre-measurement condition includes marking the water quality impact factor for the current water source when the water body reflectance is lower than the water body reflectance threshold preset by the system, or the water body ammonia nitrogen concentration is greater than the water body ammonia nitrogen concentration threshold preset by the system, otherwise it returns to the normal measurement state within the preset sliding time window; The deep learning prediction module is used to predict three target variables for water sources marked with water quality influencing factors through a deep neural network model; construct feature tensors based on the input water quality parameter data, and perform end-to-end feature extraction in the shared hidden layer through the back propagation algorithm; The multi-task regression module regresses three target variables in parallel based on the task output layer of the multi-task learning framework, and optimizes the distinction of the three target variables through independent loss functions; the three target variables include seasonal impact variables, treatment response lag variables, and abnormal water quality variables; What needs to be further explained about the above scheme is that in order to ensure effective water quality monitoring of water sources that need to be defluorinated, this scheme adopts a strategy of obtaining water quality fluctuation information in real time, and accurately determines whether there are water quality influencing factors in the water source by establishing pre-measurement conditions one and two. First, by monitoring water quality parameters such as water body chromaticity, conductivity, water body reflectivity and ammonia nitrogen concentration, we can build a complete water quality fluctuation monitoring system, in which pre-measurement condition one is used to determine whether the water source has entered an abnormal stage. If the water body chromaticity is greater than the preset threshold and the conductivity exceeds the standard set by the system, the system determines that the water source is in a Class I abnormal stage. This stage is a preliminary screening for potential risks that cannot be identified under conventional monitoring conditions. When the water source enters this stage, the further pre-measurement condition two is measured through the water body reflectivity. and ammonia nitrogen concentration to further screen to ensure more detailed abnormal identification; if these parameters return to normal within the preset sliding time window, the system returns to the normal measurement state, otherwise it continues to be marked as abnormal; these judgment conditions can dynamically adjust the changes in the reaction water quality fluctuations to ensure the timeliness and accuracy of water source water quality control; next, for the water sources marked as water quality influencing factors, the deep neural network model is used for analysis to construct a feature tensor composed of water quality parameter data, and feature extraction is performed through the shared hidden layer in deep learning, and end-to-end training is performed using the back propagation algorithm; in this process, the network can learn the implicit pattern of water quality fluctuations in the shared feature space, and then regress the three target variables in parallel through the multi-task learning framework - seasonal influence variables, processing response lag variables, and abnormal water quality variables; The loss function of each target variable is optimized independently, which can improve the model's ability to distinguish water quality fluctuations in different dimensions and its prediction accuracy. This design not only improves the ability to identify factors affecting water quality, but also ensures that the system can monitor and predict water quality in a refined and dynamic manner from multiple dimensions, avoiding the oversimplification and limitations of traditional methods. Through such a deep learning framework, we can respond to changes in water sources in real time and make more scientific and accurate processing decisions. This design is based on a deep understanding of different types of water quality fluctuations, and through the combination of multi-task learning and time series features, the system has stronger response capabilities in complex water quality control processes.

[0016] Calculate water chromaticity based on weighted time decay model and introduce weight coefficient To reflect the impact of historical data on the current chromaticity, a time attenuation factor is introduced To describe the attenuation effect of historical observation data over time; the weighted historical data is attenuated to reflect the changing trend of water body chromaticity over time; ; in For the current moment The water chromaticity value is used to reflect the concentration of visible substances in water. In this scheme, it is related to the concentration of organic matter, suspended matter and other components in water; For historical moments Observed value of water chromaticity, Used to predict and analyze future chromaticity values; Indicates the time corresponding to The weight coefficient of It is used to describe the influence of different historical observation data on the current water color. The weight coefficient will be adjusted over time, and the more recent historical data will have a higher weight. is the time attenuation coefficient, which indicates the attenuation rate of water chromaticity over time; for the current moment; is the timestamp of the historical moment, each Each corresponds to a specific historical moment ; is the size of the time window, Used to define the retrospective scope of data collection; Index the dataset; In the above formula, the closer to the current moment of The greater the impact on the current chromaticity, the smaller the impact of the more distant historical data; and It is used to quantify the temporal importance of historical data.

[0017] By combining the weighting factor and the attenuation factor to reflect the change of conductivity with time and dynamic factors, the weighting factor and Respectively reflect the influence of ion concentration and flow rate in water, and use Adjust its weight on conductivity; Indicates the current time The conductivity value of ; in The conductivity-related Factors; Indicates Factors that affect conductivity: Indicates the The weight of each factor on the conductivity; is the time attenuation coefficient; Indicates that it is used to adjust and The degree of effect, that is, the sensitivity of adjusting the decay rate; is the decay exponent, Determines the nonlinear degree of the impact of historical data on current conductivity. This will reduce the impact of historical data even more dramatically; It is the timestamp of historical data, which is used to indicate the time point of historical conductivity observation; is the total number of factors; for the current moment; In the above formula, the attenuation factor is It means that as time goes by, the influence of early observation data on the current conductivity gradually weakens. In addition, a nonlinear attenuation function is introduced to simulate the change of conductivity at different time points.

[0018] By integrating the historical reflectivity data and combining the attenuation function, the influence of historical data on the current reflectivity is reflected; nonlinear time adjustment is introduced, through Simulate the effect of water surface state on reflectivity; historical data at recent times have a greater impact on current reflectivity, while historical data far from the current time has a smaller impact; For the current moment Water reflectivity ; in is a function of the reflectivity decay over time, indicating that the influence of historical data on the current reflectivity gradually decreases over time; Parameters that represent the state of the water surface, Including suspended matter concentration, temperature, etc., which affect the optical properties of the water surface and thus affect the reflectivity; A coefficient used to adjust the effect of the water surface state on reflectivity, which is related to the concentration of dissolved substances in the water; represents the adjustment coefficient used to control the change of reflectivity over time; Indicates the time difference, describing the time from the historical point in time To the current time The time span; is the starting point of the time window.

[0019] The current ammonia nitrogen concentration is calculated by weighting the historical ammonia nitrogen concentration related factors, and its weight is determined by the time decay factor Control, with time interval Increase, the impact of historical data on the current ammonia nitrogen concentration decreases, and the current ammonia nitrogen concentration is predicted by integrating multiple historical data over a period of time; To indicate the current time Ammonia nitrogen concentration; ; in For historical moments The ammonia nitrogen related factors represent the factors related to the ammonia nitrogen concentration in the water body at that time point, including the concentrations of amino acids and nitrogen source substances in the water. It reflects the source or generation of the change in ammonia nitrogen concentration; is the time decay factor, which is used to describe the change of the influence of historical data on the current ammonia nitrogen concentration over time. As the value of increases, the influence of historical data gradually weakens; is the adjustment coefficient, Used to control the sensitivity of time decay, affecting The degree of effect on the calculation of ammonia nitrogen concentration is greater The value will accelerate the attenuation of the impact of historical data on the current ammonia nitrogen concentration; is the time difference term, indicating the current time and historical moments the time interval between is the time decay exponent, Used to describe the rate at which the impact of historical data on current ammonia nitrogen concentration decays over time. The value indicates that the influence of historical data on current values ​​will weaken rapidly; In the above formula The term decays nonlinearly over time, indicating that the impact of historical data on the current ammonia nitrogen concentration has different attenuation rates. The nonlinear attenuation is used to simulate the dynamic change of pollutants in water over time, avoiding the shortcomings of the simple linear attenuation model, and adjusting the coefficients and decay exponent , adjust the weight of the impact of historical data on the current ammonia nitrogen concentration to meet the needs of different water quality conditions. In addition, the weighted ammonia nitrogen factor It shows the ammonia nitrogen concentration-related factors of historical data, and their contribution to the current ammonia nitrogen concentration at each historical time point. In practical applications, reasonable attenuation can be used to achieve a reasonable impact of historical data on the current ammonia nitrogen concentration. It should be noted that the above scheme establishes pre-conditions one and two to detect water quality fluctuations in real time, and determines whether it enters the abnormal stage under different water quality thresholds. The establishment of pre-conditions is based on water quality parameters such as chromaticity, conductivity, reflectivity and ammonia nitrogen concentration. When the conditions are met, they are marked as water quality influencing factors, and the conversion between normal and abnormal states is dynamically adjusted by setting a sliding window. Next, the deep neural network model is trained in combination with these water quality fluctuation data, constructs feature tensors, and performs end-to-end feature extraction through shared hidden layers. The network model adopts a multi-task learning framework and regresses three target variables in parallel through independent loss functions: seasonality Influence, processing response lag and abnormal water quality, ensuring that each target variable can be learned and optimized independently, and through deep learning and multi-task learning framework, water quality change patterns can be extracted from the data, combined with time series analysis and external environmental data to enhance the model's predictive ability for water quality changes, and achieve accurate distinction and timely response to different water quality fluctuations. In this way, the defluorination plan can be quickly adjusted according to real-time data, thereby improving the intelligence and flexibility of water source management. The core of this method is to accurately predict and process the multi-level and dynamic changes of water quality fluctuations through the fusion analysis of multi-dimensional data, avoiding the limitations and lags of traditional monitoring methods.

[0020] Input water quality parameter data to construct feature tensor. Water quality parameter data includes , , , , the water quality parameter data is divided into time steps Construct a feature tensor , used for deep neural network training and feature extraction; ; in Indicates the current time step Corresponding to the input feature tensor, the value of each water quality parameter in the water quality parameter data is expressed in time is the index, constituting an element of the input feature; In a deep neural network, the input feature tensor After multiple hidden layers, feature extraction is performed and the predicted value is finally output. The network is trained through the back propagation algorithm to minimize the loss function to optimize the weights. Output the hidden layer through the ReLU activation function; ; The output layer predicts and regresses three target variables in parallel; ; in and They are the weight matrix and bias term of the first layer, acting on the input features , and through the activation function Perform nonlinear mapping to generate hidden layer output ; In the multi-task learning framework, the task output layer converts the output of the hidden layer As input, three target variables are predicted separately: seasonal impact variable , dealing with response lagged variables , abnormal water quality variables ; represents the function used for seasonal effects; represents the function used to handle the response lag, which depends on the historical data of water quality and the lag characteristics, including the time dependence of water quality changes; represents a function for abnormal water quality detection, which identifies sudden anomalies in water quality fluctuations through historical water quality data and anomaly detection mechanism; Indicates at time Historical water quality data at all times; In the multi-task learning framework, the three target variables are predicted in the task output layer through parallel regression. The regression task of each target variable includes an independent loss function. The learning of the model is improved by optimizing the loss function of each task. Include the loss function as the total loss; ; in is the total loss function, which is the sum of the loss functions of multiple target tasks; is the regression loss function of seasonal effects; Regression loss function to handle response lag; is the regression loss function of abnormal water quality; The loss function corresponding to each target variable is used to measure the error between the predicted value and the actual value. The weight of the network is optimized through back propagation. Through parallel regression, each task is independently optimized based on the shared features to achieve the prediction of the target variable.

[0021] Introduce the back propagation algorithm, which calculates the gradient of the loss function of each task and propagates it back to each layer of the network to update the weights and biases; perform gradient calculation; ; The total loss function is calculated by back-propagation algorithm with respect to the network weights The gradient is used to update the weights of each layer. The gradient contribution of each loss function to the weight is independent, but they are jointly optimized on the shared hidden layer parameters. The learning ability of the model to distinguish the three target variables is enhanced by time series features and external environment data; Capture the delayed effect of water quality changes through hysteresis characteristics and construct hysteresis characteristics ; ; External environment characteristics Including traffic ,temperature ,humidity ; ; Trend analysis feature captures the long-term trend of water quality based on trend analysis of historical data. Analysis modes include linear regression or wavelet transform. In the learning process of deep neural networks, combined with the added features, three target variables are distinguished: Seasonal impact dimension: driven by the periodic changes in water quality parameters, seasonal fluctuations are extracted using seasonal features, including Fourier transform; Treatment response lag dimension: The lag effect of water quality fluctuations is identified by the model through lag time characteristics and time windows due to the response lag of water treatment facilities; Abnormal water quality dimension: Sudden water quality anomalies, such as a sudden increase in pollutants, can be identified by the model through anomaly detection algorithms, including IsolationForest and One-ClassSVM.

[0022] In order to effectively distinguish the three target variables of seasonal impact, processing response lag and abnormal water quality, additional features are introduced. The lag effect needs to be combined with time series data and historical water quality trends. The time dependence of water quality fluctuations is captured through lag features and external environmental data such as temperature, precipitation, flow, etc. Seasonal impact involves periodic changes, so seasonal trend analysis and climate variables are introduced to help identify seasonal fluctuation patterns; abnormal water quality fluctuations are usually sudden, so features such as instantaneous flow and anomaly detection models are used to distinguish such fluctuations; through the combination of multi-dimensional features and four water quality parameters, deep neural networks are used to distinguish the three target variables, improving the prediction accuracy and robustness of the model; It should also be noted that this solution predicts and controls the fluoride concentration in the water source by real-time monitoring of water quality fluctuations and combines deep learning with a multi-task learning framework, aiming at the removal of fluoride in the sewage treatment process. The key to fluoride removal is to ensure that the fluoride concentration in the water source is maintained within a safe range to avoid excessive fluoride from having a negative impact on the aquatic ecosystem and human health. By introducing parameters such as chromaticity, conductivity, reflectivity, and ammonia nitrogen concentration in the water quality fluctuation analysis, and using a deep neural network model to extract features from time series data, we can predict the changing trend of fluoride concentration in real time, so as to take corresponding treatment measures in advance. The pre-measurement conditions and sliding time windows mentioned in the solution enable the model to flexibly judge whether the water quality has entered an abnormal stage based on historical data and the influence of the external environment, and to adjust the fluoride removal strategy in the treatment process in a timely manner. This method not only improves the automation and intelligence of the fluoride removal process, but also enhances the sensitivity to water quality changes, ensuring that the fluoride concentration is always within a controllable range in a complex water treatment environment, thereby effectively protecting water resources and public health and achieving the ultimate goal of sewage fluoride removal.

[0023] The above description is only a preferred 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 should be included in the protection scope of the present invention.

Claims

1. A multi-dimensional intelligent control system for fluorine removal, comprising a water quality fluctuation monitoring module, a normal measurement determination module, an abnormal stage determination module, a deep learning prediction module, and a multi-task regression module, characterized in that: The water quality fluctuation monitoring module is used to obtain water quality fluctuation information of the water source that needs to be defluorinated in real time, establish pre-measurement condition 1 and pre-measurement condition 2 based on the water quality fluctuation information, and determine whether there are water quality influencing factors in the current water source through pre-measurement condition 1 and pre-measurement condition 2; The water quality impact information in the water quality fluctuation monitoring module includes water chromaticity, conductivity, water reflectivity, and ammonia nitrogen concentration in water; The normal measurement judgment module is used to establish the pre-measurement condition 1, which is the normal measurement state; the pre-measurement condition 1 includes that when the water body chromaticity is greater than the water body chromaticity threshold preset by the system, and the conductivity is greater than the conductivity threshold preset by the system, then a type of abnormal stage is determined; The abnormal stage judgment module is used to establish the second pre-measurement condition when the water source is in a type of abnormal stage; the second pre-measurement condition includes marking the water quality impact factor for the current water source when the water body reflectance is lower than the water body reflectance threshold preset by the system, or the water body ammonia nitrogen concentration is greater than the water body ammonia nitrogen concentration threshold preset by the system, otherwise it returns to the normal measurement state within the preset sliding time window; The deep learning prediction module is used to predict three target variables for water sources marked with water quality influencing factors through a deep neural network model; construct feature tensors based on the input water quality parameter data, and perform end-to-end feature extraction in the shared hidden layer through the back propagation algorithm; The multi-task regression module regresses three target variables in parallel based on the task output layer of the multi-task learning framework, and optimizes the distinction of the three target variables through independent loss functions; the three target variables include seasonal influence variables, processing response lag variables, and abnormal water quality variables.

2. A multi-dimensional intelligent control system for fluorine removal according to claim 1, characterized in that: Calculate water chromaticity based on weighted time decay model and introduce weight coefficient To reflect the impact of historical data on the current chromaticity, a time attenuation factor is introduced To describe the attenuation effect of historical observation data over time; the weighted historical data is attenuated to reflect the changing trend of water body chromaticity over time; ； in For the current moment The chromaticity value of water; For historical moments Observed values ​​of water chromaticity; Indicates the time corresponding to The weight coefficient of is the time attenuation coefficient; for the current moment; timestamp for historical moments; is the size of the time window.

3. A multi-dimensional intelligent control system for defluorination according to claim 2, characterized in that: By combining the weighting factor and the attenuation factor to reflect the change of conductivity with time and dynamic factors, the weighting factor and Respectively reflect the influence of ion concentration and flow rate in water, and use Adjust its weight on conductivity; Indicates the current time The conductivity value of ； in The conductivity-related Factors; Indicates Factors that affect conductivity: Indicates the The weight of each factor on the conductivity; is the time attenuation coefficient; Indicates that it is used to adjust and the extent of its effect; is the decay index; Timestamp for historical data; For the current moment.

4. A multi-dimensional intelligent control system for defluorination according to claim 3, characterized in that: By integrating the historical reflectivity data and combining the attenuation function, the influence of historical data on the current reflectivity is reflected; nonlinear time adjustment is introduced, and the Simulate the effect of water surface conditions on reflectivity; For the current moment Water reflectivity ； in is the function of reflectivity decay over time; Parameters that represent the surface state of water bodies; A coefficient used to adjust the effect of water surface conditions on reflectivity; represents the adjustment coefficient used to control the change of reflectivity over time; Indicates the time difference.

5. A multi-dimensional intelligent control system for defluorination according to claim 4, characterized in that: The current ammonia nitrogen concentration is calculated by weighting the historical ammonia nitrogen concentration related factors, and its weight is determined by the time decay factor Control, with time interval Increase, the impact of historical data on the current ammonia nitrogen concentration decreases, and the current ammonia nitrogen concentration is predicted by integrating multiple historical data over a period of time; To indicate the current time Ammonia nitrogen concentration; ； in For historical moments Ammonia nitrogen related factors; is the time decay factor; is the adjustment coefficient; is the time difference term; is the time decay exponent.

6. A multi-dimensional intelligent control system for fluorine removal according to claim 5, characterized in that: Input water quality parameter data to construct feature tensor. Water quality parameter data includes , , , , the water quality parameter data is divided into time steps Construct a feature tensor , used for deep neural network training and feature extraction; ； in Indicates the current time step Corresponding to the input feature tensor, the value of each water quality parameter in the water quality parameter data is expressed in time is the index, constituting an element of the input feature; In a deep neural network, the input feature tensor After multiple hidden layers, feature extraction is performed and the predicted value is finally output. The network is trained through the back propagation algorithm, and the loss function is used to optimize the weights; Output the hidden layer through the ReLU activation function; ； The output layer predicts and regresses three target variables in parallel; ； in and They are the weight matrix and bias term of the first layer, acting on the input features , and through the activation function Perform nonlinear mapping to generate hidden layer output ; In the multi-task learning framework, the task output layer converts the output of the hidden layer As input, three target variables are predicted separately: seasonal impact variable , dealing with response lagged variables , abnormal water quality variables ; represents the function used for seasonal effects; represents the function used to handle response lag; represents a function for abnormal water quality detection; Indicates at time Historical water quality data at all times; In the multi-task learning framework, the three target variables are predicted in the task output layer through parallel regression. The regression task of each target variable includes an independent loss function. The learning of the model is improved by optimizing the loss function of each task. Include the loss function as the total loss; ; in is the total loss function, which is the sum of the loss functions of multiple target tasks; is the regression loss function of seasonal effects; Regression loss function to handle response lag; is the regression loss function of abnormal water quality; The loss function corresponding to each target variable is used to measure the error between the predicted value and the actual value. The weight of the network is optimized through back propagation. Through parallel regression, each task is independently optimized based on the shared features to achieve the prediction of the target variable.

7. A multi-dimensional intelligent control system for fluorine removal according to claim 6, characterized in that: Introduce the back propagation algorithm, which calculates the gradient of the loss function of each task and propagates it back to each layer of the network to update the weights and biases; perform gradient calculation; ； The total loss function is calculated by back-propagation algorithm with respect to the network weights The gradient is used to update the weights of each layer. The gradient contribution of each loss function to the weight is independent, but they are jointly optimized on the shared hidden layer parameters. The learning ability of the model to distinguish the three target variables is enhanced by time series features and external environment data; Capture the delayed effect of water quality changes through hysteresis characteristics and construct hysteresis characteristics ; ； External environment characteristics Including traffic ,temperature ,humidity ; ； Trend Analysis Feature Trend analysis based on historical data captures long-term trends in water quality.

Images