Water plant automation energy-saving management system and method based on dosage prediction model

Abstract

The application provides a water plant automatic energy-saving management system and method based on a dosing amount prediction model, belongs to the technical field of water plant management, and realizes rapid response and accurate control of turbidity changes by constructing a dosing amount prediction model based on adaptive weighted regression and combining recursive least square method to update parameters online; the system integrates data collection, preprocessing, prediction, control and energy management into one, forms a closed-loop automatic regulation and control, and significantly improves the intelligent level and stability of the preprocessing process; meanwhile, through clean energy power generation and energy consumption dynamic optimization, the dependence of the system on external power grids is effectively reduced, the system has the characteristics of energy saving and consumption reduction, low operation cost and strong adaptability, and is particularly suitable for efficient and stable operation of small and medium-sized water plants in the Sichuan-Chongqing region under alternating working conditions of the rainy season and the dry season.

Description

Technical Field

[0001] This invention relates to the field of water plant management technology, and in particular to an automated energy-saving management system and method for water plants based on a chemical dosage prediction model. Background Technology

[0002] As a crucial link in urban water supply, the pretreatment process of water treatment plants has a decisive impact on the subsequent water purification effect. In the Sichuan-Chongqing region, water sources mainly come from the Yangtze River, Jialing River, and tributary reservoirs, and are significantly affected by seasonal rainfall. During the rainy season, torrential rains wash surface soil and rock debris into rivers, causing a sharp increase in raw water turbidity in a short period of time; while during the dry season, rivers are mainly replenished by groundwater, and the turbidity is relatively low. This drastic fluctuation in turbidity places extremely high demands on the real-time response capability of the water plant's pretreatment system.

[0003] Currently, most water plants in the Sichuan-Chongqing region still use a semi-automatic, semi-manual control method. This means that they rely on turbidity data transmitted from the source, and manual judgment is used to adjust pretreatment measures, such as controlling the opening of the outlet gate and the dosage of flocculant. This method has several problems: First, manual judgment is affected by factors such as experience, fatigue, and shift changes, resulting in slow response speed, large errors, and difficulty in adapting to rapidly changing turbidity conditions. Second, the dosage control is inaccurate, often resulting in overdosing or underdosing, which wastes chemicals and affects the quality of the effluent. Third, the existing system lacks optimization of energy utilization, resulting in high energy consumption at the water plant.

[0004] Therefore, there is an urgent need in this field to develop a water plant management system and method that can achieve fully automated control, accurate predictive dosing, and high energy efficiency, in order to solve the problems of slow response, low control accuracy, high energy consumption and complex maintenance in the existing technology. Summary of the Invention

[0005] The purpose of this invention is to provide an automated energy-saving management system and method for water plants based on a chemical dosage prediction model, so as to solve the problems existing in the prior art.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] This invention provides an automated energy-saving management system for water plants based on a chemical dosage prediction model, comprising:

[0008] The sensor data acquisition module is used to collect influent water quality parameters in real time, including turbidity, flow rate, water temperature, and pH value.

[0009] The data preprocessing and feature extraction module is used to clean the collected data by detecting and replacing outliers using a sliding window, then normalizing the data and extracting features to output a feature vector.

[0010] The dosage prediction model module predicts the optimal dosage based on an adaptive weighted regression model, and the model parameters are updated in real time using the recursive least squares method.

[0011] The intelligent control module controls the dosing pump and water inlet valve based on the predicted dosage.

[0012] The energy management module is used to monitor the status of clean energy generation and battery storage, optimize system energy consumption, dynamically adjust the system sampling frequency and operating mode, and prioritize the use of battery power.

[0013] The system monitoring and feedback module is used to display the system status in real time, record data, and provide feedback for adjusting model parameters.

[0014] Preferably, the extracted features in the data preprocessing and feature extraction module include: turbidity change rate, flow rate change rate, turbidity moving average, and flow rate moving average.

[0015] Preferably, in the data preprocessing and feature extraction module, the formula for constructing the feature vector is:

[0016] ;

[0017] in, The turbidity value is the normalized value. The turbidity change rate The normalized flow value, For the rate of change of flow, This is the normalized water temperature value. Here is the normalized pH value, and T is the transpose operator.

[0018] Preferably, in the dosage prediction model module, the dosage prediction model formula is:

[0019] ;

[0020] in, Let be the predicted dosage at time t. Let be the weight vector at time t, and T be the transpose operator used to transpose the weight vector. Let be the eigenvector at time t. For the bias term at time t;

[0021] The formula for updating the model parameters is:

[0022] ;

[0023] ;

[0024] ;

[0025] in, Let be the extended parameter vector at time t. Let be the extended parameter vector at time t-1. Let be the extended eigenvector at time t. For prediction error, Let be the gain vector at time t. Let be the covariance matrix at time t. Let be the covariance matrix at time t-1, and λ be the forgetting factor.

[0026] Preferably, the forgetting factor ranges from 0.95 to 0.99.

[0027] Preferably, in the intelligent control module, the flow control formula for the dosing pump is:

[0028] ;

[0029] in, For real-time traffic, These are the conversion factors;

[0030] The formula for controlling the opening degree of the inlet valve is:

[0031] ;

[0032] in, It is an empirical constant. For real-time turbidity, To prevent division by zero of extremely small constants.

[0033] Preferably, in the energy management module, the power calculation formula for clean energy generation is:

[0034] ;

[0035] in, η is the power output of the water turbine, ρ is the turbine efficiency, g is the acceleration due to gravity, and H is the head. This refers to the power generated by solar energy.

[0036] Preferably, the system monitoring and feedback module is further used to: adjust model parameters based on effluent turbidity feedback, and trigger an alarm when effluent turbidity exceeds the standard.

[0037] This invention also provides a water plant automation energy-saving management method based on a chemical dosage prediction model, comprising the following steps:

[0038] S1. System initialization, loading historical data and setting model parameters;

[0039] S2. Real-time collection of influent water quality data;

[0040] S3. Perform data preprocessing and feature extraction;

[0041] S4. Use an adaptive weighted regression model to predict the dosage;

[0042] S5. Control the dosing pump and valves based on the prediction results;

[0043] S6. Monitor energy status and optimize energy consumption;

[0044] S7. Monitor system status in real time and provide feedback to adjust the model.

[0045] Preferably, it further includes:

[0046] S8. Retrain the prediction model weekly using historical data;

[0047] S9. Dynamically adjust control parameters based on effluent turbidity feedback;

[0048] S10. Generate an operation report, which includes the chemical dosing savings rate and energy self-sufficiency rate.

[0049] The present invention achieves the following beneficial technical effects compared to the prior art:

[0050] This invention provides an automated energy-saving management system and method for water plants based on a chemical dosage prediction model. By constructing a chemical dosage prediction model based on adaptive weighted regression and combining it with online parameter updates using recursive least squares, the system achieves rapid response and precise control to changes in turbidity. The system integrates data acquisition, preprocessing, prediction, control, and energy management into a closed-loop automated control system, significantly improving the intelligence and stability of the preprocessing process. Simultaneously, through clean energy power generation and dynamic energy consumption optimization, the system effectively reduces its dependence on the external power grid. It features energy saving, low operating costs, and strong adaptability, making it particularly suitable for the efficient and stable operation of small and medium-sized water plants in the Sichuan-Chongqing region under alternating rainy and dry season conditions. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a schematic diagram of the water plant automated energy-saving management system based on the dosage prediction model provided by the present invention.

[0053] Figure 2The flowchart of the water plant automation energy-saving management method based on the chemical dosage prediction model provided by the present invention is shown. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] The purpose of this invention is to provide an automated energy-saving management system and method for water plants based on a chemical dosage prediction model. This system addresses the problems of drastic changes in raw water turbidity, delayed response of manual control, and high energy consumption in water plants in the Sichuan-Chongqing region. By integrating intelligent sensing, adaptive algorithms, and clean energy technologies, it achieves precise automated control and energy-saving optimization of the pretreatment process.

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] Example 1:

[0058] Please see Figure 1 The water plant automation energy-saving management system based on the dosage prediction model provided in this embodiment consists of a sensor data acquisition module, a data preprocessing and feature extraction module, a dosage prediction model module, an intelligent control module, an energy management module, and a system monitoring and feedback module. The modules are interconnected through an industrial communication network (such as Modbus TCP / IP) to form a closed-loop control structure.

[0059] Specifically, when the system is working, the sensor data acquisition module collects in-water quality parameters in real time, including turbidity, flow rate, water temperature and pH value. The sampling frequency can be set to once every 5 seconds to ensure data real-time performance.

[0060] Furthermore, the collected raw data is sent to the data preprocessing and feature extraction module. This module first performs data cleaning, uses a sliding window method (window size N=10) to detect outliers, calculates the mean and standard deviation of each parameter, and replaces the data point with the previous valid value if the deviation of the data point from the mean exceeds 3 times the standard deviation.

[0061] Then, normalization is performed to scale each parameter to the [0,1] interval, as shown in the formula:

[0062] ;

[0063] in, and This is the historical minimum and maximum value, updated monthly.

[0064] Feature extraction includes calculating the rate of change of turbidity. Flow rate of change The turbidity moving average and flow moving average are used to construct the feature vector, and the calculation formula is as follows:

[0065] ;

[0066] in, The turbidity value is the normalized value. The turbidity change rate The normalized flow value, For the rate of change of flow, This is the normalized water temperature value. is the normalized pH value, where the superscript T represents the transpose operator, used to convert a column vector into a row vector to fit the model input.

[0067] Furthermore, the dosage prediction model module calculates the optimal dosage based on an adaptive weighted regression model, with the core formula being:

[0068] ;

[0069] in, Let be the predicted dosage at time t. Let be the weight vector at time t, and T be the transpose operator used to transpose the weight vector. Let be the eigenvector at time t. This is the bias term at time t.

[0070] The model parameters are updated online using the recursive least squares method to dynamically adapt to changes in water quality, and an extended parameter vector is defined. and extended feature vectors Prediction error ,in This represents the actual amount of medicine administered at the previous moment.

[0071] The formula for updating the model parameters is:

[0072] ;

[0073] ;

[0074] ;

[0075] in, Let be the extended parameter vector at time t. Let be the extended parameter vector at time t-1. Let be the extended eigenvector at time t. For prediction error, Let be the gain vector at time t. Let be the covariance matrix at time t. Let be the covariance matrix at time t-1, and λ be the forgetting factor (ranging from 0.95 to 0.99), used to adjust the weights of historical data.

[0076] Furthermore, the intelligent control module executes control commands based on the prediction results, and the flow control formula for the dosing pump is:

[0077] ;

[0078] in, For real-time traffic, The conversion factor is 0.001.

[0079] The formula for controlling the opening degree of the inlet valve is:

[0080] ;

[0081] in, This is an empirical constant (taken as 50). For real-time turbidity, To prevent division by zero by an extremely small constant (0.1), the PLC adjusts the dosing pump speed and valve opening based on the calculation results to achieve precise dosing and flow control.

[0082] Furthermore, the energy management module monitors clean energy generation and consumption. The power calculation formula for clean energy generation is as follows:

[0083] ;

[0084] in, η is the power output of the water turbine, η is the efficiency of the water turbine (taken as 0.7), ρ is the density of water, g is the acceleration due to gravity, and H is the head. The power generated by solar energy is directly collected by solar panels.

[0085] In this embodiment, the system power consumption includes the energy consumption of the dosing pump, PLC, and sensors. The energy dispatch strategy is as follows: when the total power generated by clean energy is greater than the system power consumption, the excess energy is stored in the battery; when the total power generated by clean energy is less than the system power consumption, the battery provides supplementary power. To optimize energy consumption, the sampling frequency is dynamically adjusted; for example, when the battery charge is below 20%, the sampling period is extended to 10 seconds.

[0086] Furthermore, the system monitoring and feedback module displays the operating status in real time, records historical data, and adjusts model parameters based on effluent turbidity feedback. If the effluent turbidity exceeds the standard threshold (e.g., 5 NTU), an alarm mechanism is triggered. The prediction model is retrained weekly using accumulated data, the normalization parameters and model weights are updated, and an operation report is generated, including indicators such as chemical dosing savings and energy self-sufficiency rate, to ensure continuous system optimization.

[0087] Example 2:

[0088] Please see Figure 2 The water plant automation energy-saving management method based on a chemical dosage prediction model of the present invention includes the following steps:

[0089] S1. System initialization, loading historical data and setting model parameters;

[0090] S2. Real-time collection of influent water quality data;

[0091] S3. Perform data preprocessing and feature extraction;

[0092] S4. Use an adaptive weighted regression model to predict the dosage;

[0093] S5. Control the dosing pump and valves based on the prediction results;

[0094] S6. Monitor energy status and optimize energy consumption;

[0095] S7. Monitor system status in real time and provide feedback for model adjustment;

[0096] S8. Retrain the prediction model weekly using historical data;

[0097] S9. Dynamically adjust control parameters based on effluent turbidity feedback;

[0098] S10. Generate an operation report, which includes the pesticide application savings rate and energy self-sufficiency rate, thus forming a complete closed-loop management process.

[0099] This invention, through the aforementioned system and method, achieves fully automated, precise, and low-carbon operation of water plant pretreatment, and is particularly suitable for small and medium-sized water plants in the Sichuan-Chongqing region to cope with seasonal turbidity fluctuations, improve water quality stability and economy.

[0100] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0101] It should be noted that the components mentioned in the above embodiments are all general standard parts or components known to those skilled in the art. Their structures and principles can be learned by those skilled in the art through technical manuals or conventional experimental methods.

[0102] This invention has illustrated its principles and implementation methods using specific examples. The descriptions of these embodiments are merely illustrative of the method and its core ideas; furthermore, those skilled in the art will recognize that modifications may be made to the specific implementation methods and application scope based on the principles of this invention. Therefore, the content of this specification should not be construed as limiting the invention.

Claims

1. A water plant automated energy-saving management system based on a chemical dosage prediction model, characterized in that, include: The sensor data acquisition module is used to collect influent water quality parameters in real time, including turbidity, flow rate, water temperature, and pH value. The data preprocessing and feature extraction module is used to clean the collected data by detecting and replacing outliers using a sliding window, then normalizing the data and extracting features to output a feature vector. The dosage prediction model module predicts the optimal dosage based on an adaptive weighted regression model, and the model parameters are updated in real time using the recursive least squares method. The intelligent control module controls the dosing pump and water inlet valve based on the predicted dosage. The energy management module is used to monitor the status of clean energy generation and battery storage, optimize system energy consumption, dynamically adjust the system sampling frequency and operating mode, and prioritize the use of battery power. The system monitoring and feedback module is used to display the system status in real time, record data, and provide feedback for adjusting model parameters. The dosage prediction model module contains the following formula: ； in, Let be the predicted dosage at time t. Let be the weight vector at time t, and T be the transpose operator used to transpose the weight vector. Let be the eigenvector at time t. For the bias term at time t; The formula for updating the model parameters is: ； ； ； in, Let be the extended parameter vector at time t. Let be the extended parameter vector at time t-1. Let be the extended eigenvector at time t. For prediction error, Let be the gain vector at time t. Let be the covariance matrix at time t. Let be the covariance matrix at time t-1, and λ be the forgetting factor; In the intelligent control module, the flow control formula for the dosing pump is: ； in, For real-time traffic, These are the conversion factors; The formula for controlling the opening degree of the inlet valve is: ； in, It is an empirical constant. For real-time turbidity, To prevent division by zero of extremely small constants; In the energy management module, the power calculation formula for clean energy generation is as follows: ； in, η is the power output of the water turbine, ρ is the turbine efficiency, g is the acceleration due to gravity, and H is the head. This refers to the power generated by solar energy.

2. The water plant automated energy-saving management system based on a chemical dosage prediction model as described in claim 1, characterized in that, The data preprocessing and feature extraction module extracts the following features: turbidity change rate, flow rate change rate, turbidity moving average, and flow rate moving average.

3. The water plant automated energy-saving management system based on a chemical dosage prediction model as described in claim 2, characterized in that, In the data preprocessing and feature extraction module, the formula for constructing the feature vector is: ； in, The turbidity value is the normalized value. The turbidity change rate The normalized flow value, For the rate of change of flow, This is the normalized water temperature value. Here is the normalized pH value, and T is the transpose operator.

4. The water plant automated energy-saving management system based on a chemical dosage prediction model as described in claim 1, characterized in that, The forgetting factor ranges from 0.95 to 0.

99.

5. The water plant automated energy-saving management system based on a chemical dosage prediction model as described in claim 1, characterized in that, The system monitoring and feedback module is also used to: adjust model parameters based on effluent turbidity feedback, and trigger an alarm when effluent turbidity exceeds the standard.

6. A water plant automation energy-saving management method based on a chemical dosage prediction model, characterized in that, The water plant automation energy-saving management system based on the dosage prediction model as described in any one of claims 1-5 includes the following steps: S1. System initialization, loading historical data and setting model parameters; S2. Real-time collection of influent water quality data; S3. Perform data preprocessing and feature extraction; S4. Use an adaptive weighted regression model to predict the dosage; S5. Control the dosing pump and valves based on the prediction results; S6. Monitor energy status and optimize energy consumption; S7. Monitor system status in real time and provide feedback to adjust the model.

7. The water plant automation energy-saving management method based on a chemical dosage prediction model as described in claim 6, characterized in that, Also includes: S8. Retrain the prediction model weekly using historical data; S9. Dynamically adjust control parameters based on effluent turbidity feedback; S10. Generate an operation report, which includes the chemical dosing savings rate and energy self-sufficiency rate.

Images